Up next in 10
In this episode of the Global AI October sessions we focus on what computer vision is, the technical challenge and how to create models.
Show More Show Less View Video Transcript
0:01
[Music]
0:28
[Music]
0:28
[Music] [Applause]
0:32
[Music]
0:40
[Music]
0:40
[Music] [Applause]
0:45
[Applause]
0:45
[Applause] [Music]
1:24
[Music]
1:59
[Music]
1:59
[Music] [Applause]
1:59
[Applause]
1:59
[Applause] [Music]
2:38
[Music]
2:46
[Applause]
2:47
[Applause]
2:47
[Applause] [Music]
3:16
do
3:17
do
3:17
do [Music]
3:58
so
3:59
so
3:59
so [Music]
4:12
[Music]
4:19
[Music]
5:18
[Music]
5:38
[Music]
5:45
[Music]
6:29
[Music]
6:45
[Music]
7:31
[Music]
7:56
hey everybody we're starting a little
7:59
hey everybody we're starting a little
7:59
hey everybody we're starting a little bit early today
8:00
bit early today
8:00
bit early today uh this is sort of a pre-show so welcome
8:02
uh this is sort of a pre-show so welcome
8:02
uh this is sort of a pre-show so welcome to the pre-show
8:03
to the pre-show
8:03
to the pre-show this is the unofficial bit of the global
8:05
this is the unofficial bit of the global
8:05
this is the unofficial bit of the global ai community october sessions
8:07
ai community october sessions
8:07
ai community october sessions where alicia who's to the left of me
8:11
where alicia who's to the left of me
8:11
where alicia who's to the left of me hey there you are hello uh is joining me
8:14
hey there you are hello uh is joining me
8:14
hey there you are hello uh is joining me today as a co-host and i will announce
8:17
today as a co-host and i will announce
8:17
today as a co-host and i will announce her later for
8:17
her later for
8:17
her later for for uh officially but we were just
8:21
for uh officially but we were just
8:21
for uh officially but we were just chatting about generative
8:23
chatting about generative
8:23
chatting about generative design and designing big offices using
8:25
design and designing big offices using
8:25
design and designing big offices using ai actually
8:28
ai actually
8:28
ai actually and and that came from the fact that i
8:29
and and that came from the fact that i
8:29
and and that came from the fact that i remodeled my my bathroom and shower
8:32
remodeled my my bathroom and shower
8:32
remodeled my my bathroom and shower a few weeks ago and i was like major
8:35
a few weeks ago and i was like major
8:35
a few weeks ago and i was like major work
8:37
work
8:37
work but guess i guess if you want to design
8:39
but guess i guess if you want to design
8:39
but guess i guess if you want to design a big office with 40 floors
8:41
a big office with 40 floors
8:41
a big office with 40 floors you have to optimize for either
8:42
you have to optimize for either
8:42
you have to optimize for either conferencing or working alone or working
8:44
conferencing or working alone or working
8:44
conferencing or working alone or working in a team
8:45
in a team
8:45
in a team that's that's a job that's normally
8:47
that's that's a job that's normally
8:47
that's that's a job that's normally taking months
8:48
taking months
8:48
taking months alicia do you know actually how how much
8:51
alicia do you know actually how how much
8:51
alicia do you know actually how how much time
8:51
time
8:52
time normally would take oh have you met
8:54
normally would take oh have you met
8:54
normally would take oh have you met anybody who's working in construction
8:55
anybody who's working in construction
8:56
anybody who's working in construction for that matter
8:57
for that matter
8:57
for that matter i i haven't but um
9:00
i i haven't but um
9:00
i i haven't but um i can only imagine how much time and
9:02
i can only imagine how much time and
9:02
i can only imagine how much time and manpower you save
9:04
manpower you save
9:04
manpower you save by doing that design upfront design and
9:06
by doing that design upfront design and
9:06
by doing that design upfront design and possible redesign
9:08
possible redesign
9:08
possible redesign uh before construction yeah
9:11
uh before construction yeah
9:12
uh before construction yeah exactly yeah and i guess if you could
9:14
exactly yeah and i guess if you could
9:14
exactly yeah and i guess if you could apply ai
9:15
apply ai
9:15
apply ai and generate the floor plan for a
9:17
and generate the floor plan for a
9:17
and generate the floor plan for a building
9:18
building
9:18
building and it doesn't matter it could be a
9:19
and it doesn't matter it could be a
9:19
and it doesn't matter it could be a house but especially larger buildings
9:21
house but especially larger buildings
9:21
house but especially larger buildings that make sense i guess
9:28
if you could design for for happiness so
9:30
if you could design for for happiness so
9:30
if you could design for for happiness so employee happiness
9:32
employee happiness
9:32
employee happiness right so i i wonder what the ratio is
9:35
right so i i wonder what the ratio is
9:35
right so i i wonder what the ratio is between
9:36
between
9:36
between employee happiness and i think there is
9:39
employee happiness and i think there is
9:39
employee happiness and i think there is a correlation between
9:41
a correlation between
9:41
a correlation between um that and windows right so i've
9:43
um that and windows right so i've
9:43
um that and windows right so i've noticed a lot of european buildings have
9:46
noticed a lot of european buildings have
9:46
noticed a lot of european buildings have a lot more windows space
9:50
a lot more windows space
9:50
a lot more windows space then yeah definitely so there's math
9:53
then yeah definitely so there's math
9:53
then yeah definitely so there's math behind it and if there's math behind it
9:55
behind it and if there's math behind it
9:55
behind it and if there's math behind it you can pipe that into an ai algorithm
9:57
you can pipe that into an ai algorithm
9:57
you can pipe that into an ai algorithm right
9:58
right
9:58
right yeah i guess we could could uh maybe use
10:01
yeah i guess we could could uh maybe use
10:01
yeah i guess we could could uh maybe use the
10:01
the
10:02
the amount of light coming into a particular
10:04
amount of light coming into a particular
10:04
amount of light coming into a particular floor as a part of a loss function i
10:06
floor as a part of a loss function i
10:06
floor as a part of a loss function i guess for a neural network or something
10:09
guess for a neural network or something
10:10
guess for a neural network or something if that makes sense then why build
10:12
if that makes sense then why build
10:12
if that makes sense then why build effects
10:13
effects
10:13
effects surface area right for the windows
10:16
surface area right for the windows
10:16
surface area right for the windows yeah yeah definitely a number of
10:19
yeah yeah definitely a number of
10:19
yeah yeah definitely a number of of uh working places you can put on the
10:21
of uh working places you can put on the
10:21
of uh working places you can put on the floor could also be a component in
10:23
floor could also be a component in
10:23
floor could also be a component in in the model and i guess if you
10:27
in the model and i guess if you
10:27
in the model and i guess if you so if you feed that to a computer and
10:29
so if you feed that to a computer and
10:29
so if you feed that to a computer and just try to maximize
10:31
just try to maximize
10:31
just try to maximize the the number of points the computer
10:33
the the number of points the computer
10:33
the the number of points the computer gets for the best floor plan
10:35
gets for the best floor plan
10:35
gets for the best floor plan you could actually learn how to generate
10:37
you could actually learn how to generate
10:37
you could actually learn how to generate a correct floor plan
10:38
a correct floor plan
10:38
a correct floor plan for a building yeah
10:43
so it's interesting because i'm from las
10:45
so it's interesting because i'm from las
10:45
so it's interesting because i'm from las vegas and
10:46
vegas and
10:46
vegas and uh there's not a lot of windows in las
10:49
uh there's not a lot of windows in las
10:49
uh there's not a lot of windows in las vegas
10:50
vegas
10:50
vegas so i i don't know what that says about
10:53
so i i don't know what that says about
10:53
so i i don't know what that says about you know employee job satisfaction
10:55
you know employee job satisfaction
10:55
you know employee job satisfaction but being at mit i'm now in texas
10:58
but being at mit i'm now in texas
10:58
but being at mit i'm now in texas so i i'm just saying how are the
11:01
so i i'm just saying how are the
11:02
so i i'm just saying how are the buildings in texas by the way
11:04
buildings in texas by the way
11:04
buildings in texas by the way uh you know texas is so spread out that
11:07
uh you know texas is so spread out that
11:07
uh you know texas is so spread out that there are a lot of windows and there are
11:09
there are a lot of windows and there are
11:09
there are a lot of windows and there are a lot of trees so
11:11
a lot of trees so
11:11
a lot of trees so i think if i could also add a metric
11:14
i think if i could also add a metric
11:14
i think if i could also add a metric for um proximity to nature elements
11:18
for um proximity to nature elements
11:18
for um proximity to nature elements right proximity and exposure and if you
11:21
right proximity and exposure and if you
11:21
right proximity and exposure and if you look at
11:22
look at
11:22
look at job satisfaction between you know
11:24
job satisfaction between you know
11:24
job satisfaction between you know employees on the west coast
11:25
employees on the west coast
11:25
employees on the west coast right everybody loves california and
11:27
right everybody loves california and
11:27
right everybody loves california and seattle and
11:28
seattle and
11:28
seattle and that's there's a lot of treaties there
11:31
that's there's a lot of treaties there
11:31
that's there's a lot of treaties there right
11:31
right
11:31
right yeah a lot of greenery and in san
11:33
yeah a lot of greenery and in san
11:33
yeah a lot of greenery and in san francisco well of course you've got more
11:35
francisco well of course you've got more
11:35
francisco well of course you've got more sun than in seattle but seattle has this
11:37
sun than in seattle but seattle has this
11:37
sun than in seattle but seattle has this nice environment as well with the
11:39
nice environment as well with the
11:39
nice environment as well with the mountains close by um yeah
11:42
mountains close by um yeah
11:42
mountains close by um yeah i guess those are the more favorite
11:44
i guess those are the more favorite
11:44
i guess those are the more favorite places
11:45
places
11:45
places i think however there is also a large
11:47
i think however there is also a large
11:47
i think however there is also a large correlation to the fact that
11:49
correlation to the fact that
11:49
correlation to the fact that um large i.t companies happen to come
11:52
um large i.t companies happen to come
11:52
um large i.t companies happen to come from those places so microsoft is of
11:54
from those places so microsoft is of
11:54
from those places so microsoft is of course from redmond which
11:55
course from redmond which
11:56
course from redmond which is close to seattle amazon is also based
11:59
is close to seattle amazon is also based
11:59
is close to seattle amazon is also based in seattle
12:00
in seattle
12:00
in seattle i never knew until i came there and i
12:02
i never knew until i came there and i
12:02
i never knew until i came there and i saw amazon office
12:04
saw amazon office
12:04
saw amazon office in the middle of seattle and also
12:06
in the middle of seattle and also
12:06
in the middle of seattle and also noticed that
12:08
noticed that
12:08
noticed that people were talking about the housing
12:09
people were talking about the housing
12:09
people were talking about the housing prices that were driven up by
12:11
prices that were driven up by
12:11
prices that were driven up by uh microsoft and amazon apparently um
12:14
uh microsoft and amazon apparently um
12:14
uh microsoft and amazon apparently um which is a real problem for those cities
12:16
which is a real problem for those cities
12:16
which is a real problem for those cities um the same thing is happening in san
12:18
um the same thing is happening in san
12:18
um the same thing is happening in san francisco so
12:20
francisco so
12:20
francisco so i guess tax is not a bad place to be
12:24
i guess tax is not a bad place to be
12:24
i guess tax is not a bad place to be uh in that sense so next
12:27
uh in that sense so next
12:27
uh in that sense so next i think we need some studies about the
12:29
i think we need some studies about the
12:29
i think we need some studies about the correlation between beaches
12:31
correlation between beaches
12:31
correlation between beaches because i i think i need to spend a lot
12:34
because i i think i need to spend a lot
12:34
because i i think i need to spend a lot of time
12:35
of time
12:35
of time thinking about how much better i perform
12:39
thinking about how much better i perform
12:39
thinking about how much better i perform when i'm closer to a beach and of course
12:41
when i'm closer to a beach and of course
12:41
when i'm closer to a beach and of course you have to check that manually
12:43
you have to check that manually
12:43
you have to check that manually you have to go to the beach and make
12:44
you have to go to the beach and make
12:44
you have to go to the beach and make measurements for yourself of course
12:46
measurements for yourself of course
12:46
measurements for yourself of course yeah and i i would not mind spending
12:50
yeah and i i would not mind spending
12:50
yeah and i i would not mind spending 14 days in quarantine in hawaii um while
12:52
14 days in quarantine in hawaii um while
12:52
14 days in quarantine in hawaii um while i
12:53
i
12:53
i perform law analysis and for those of
12:56
perform law analysis and for those of
12:56
perform law analysis and for those of you who don't know i am originally from
12:57
you who don't know i am originally from
12:58
you who don't know i am originally from hawaii
12:58
hawaii
12:58
hawaii so um occasionally i do get to go back
13:01
so um occasionally i do get to go back
13:01
so um occasionally i do get to go back to see my family
13:02
to see my family
13:02
to see my family and uh there's there's no place like
13:05
and uh there's there's no place like
13:05
and uh there's there's no place like home
13:06
home
13:06
home and so um i'm i'm i'm sure there's a lot
13:10
and so um i'm i'm i'm sure there's a lot
13:10
and so um i'm i'm i'm sure there's a lot of beaches that i've yet to discover in
13:12
of beaches that i've yet to discover in
13:12
of beaches that i've yet to discover in europe
13:12
europe
13:12
europe so i can recommend a few so in the
13:16
so i can recommend a few so in the
13:16
so i can recommend a few so in the netherlands we have
13:17
netherlands we have
13:17
netherlands we have sand beaches although we don't have a
13:18
sand beaches although we don't have a
13:18
sand beaches although we don't have a lot of days during the year where it's
13:20
lot of days during the year where it's
13:20
lot of days during the year where it's sunny enough to sit on the beach and
13:22
sunny enough to sit on the beach and
13:22
sunny enough to sit on the beach and sit comfortably on the beach it's pretty
13:24
sit comfortably on the beach it's pretty
13:24
sit comfortably on the beach it's pretty busy as well
13:25
busy as well
13:25
busy as well a small country a lot of people so that
13:28
a small country a lot of people so that
13:28
a small country a lot of people so that happens
13:29
happens
13:29
happens but actually there's a part of the
13:31
but actually there's a part of the
13:31
but actually there's a part of the netherlands i don't know if it's
13:32
netherlands i don't know if it's
13:32
netherlands i don't know if it's officially part of the netherlands
13:34
officially part of the netherlands
13:34
officially part of the netherlands anymore uh curacao so we've got a couple
13:36
anymore uh curacao so we've got a couple
13:36
anymore uh curacao so we've got a couple of islands
13:37
of islands
13:37
of islands uh in the in the caribbean uh that are
13:40
uh in the in the caribbean uh that are
13:40
uh in the in the caribbean uh that are pretty nice
13:41
pretty nice
13:41
pretty nice uh i've been there a couple of times and
13:42
uh i've been there a couple of times and
13:42
uh i've been there a couple of times and that's that's also part of the
13:44
that's that's also part of the
13:44
that's that's also part of the netherlands in some
13:45
netherlands in some
13:45
netherlands in some form or shape it's pretty complicated
13:47
form or shape it's pretty complicated
13:47
form or shape it's pretty complicated but yeah that's a great place to go too
13:51
but yeah that's a great place to go too
13:51
but yeah that's a great place to go too so where is your downtown located
13:54
so where is your downtown located
13:54
so where is your downtown located relative to your beach
13:57
relative to your beach
13:57
relative to your beach i have to drive one hour and 45 minutes
14:00
i have to drive one hour and 45 minutes
14:00
i have to drive one hour and 45 minutes to get to a beach uh
14:02
to get to a beach uh
14:02
to get to a beach uh from my place um well and how happy are
14:06
from my place um well and how happy are
14:06
from my place um well and how happy are you
14:08
very happy
14:12
i know you're a very happy guy yeah yeah
14:15
i know you're a very happy guy yeah yeah
14:16
i know you're a very happy guy yeah yeah i'm just excited about this episode and
14:18
i'm just excited about this episode and
14:18
i'm just excited about this episode and i'm excited about
14:19
i'm excited about
14:19
i'm excited about ai and new technology in general so i'm
14:21
ai and new technology in general so i'm
14:22
ai and new technology in general so i'm one of those people that likes to sit
14:23
one of those people that likes to sit
14:23
one of those people that likes to sit indoor and and work on
14:25
indoor and and work on
14:25
indoor and and work on on new stuff basically that's that's
14:27
on new stuff basically that's that's
14:27
on new stuff basically that's that's what i do for a living
14:29
what i do for a living
14:29
what i do for a living um
14:33
yeah that's pretty cool actually and i'm
14:35
yeah that's pretty cool actually and i'm
14:35
yeah that's pretty cool actually and i'm excited for tonight as well so
14:37
excited for tonight as well so
14:37
excited for tonight as well so we in preparation for this episode i i
14:39
we in preparation for this episode i i
14:39
we in preparation for this episode i i got a chance to talk to sacha malik the
14:41
got a chance to talk to sacha malik the
14:41
got a chance to talk to sacha malik the ceo of opencv which is
14:42
ceo of opencv which is
14:42
ceo of opencv which is a great person to have a chat with he's
14:45
a great person to have a chat with he's
14:45
a great person to have a chat with he's uh he's
14:46
uh he's
14:46
uh he's pretty uh i mean he runs this this
14:50
pretty uh i mean he runs this this
14:50
pretty uh i mean he runs this this open source company opencv for for the
14:52
open source company opencv for for the
14:52
open source company opencv for for the last 20 years
14:54
last 20 years
14:54
last 20 years it's pretty old actually that technology
14:56
it's pretty old actually that technology
14:56
it's pretty old actually that technology and still cap it's kept up to date
14:58
and still cap it's kept up to date
14:58
and still cap it's kept up to date and improved and improved over and over
14:59
and improved and improved over and over
14:59
and improved and improved over and over again and they've been doing some great
15:01
again and they've been doing some great
15:01
again and they've been doing some great stuff he's going to talk about that
15:02
stuff he's going to talk about that
15:02
stuff he's going to talk about that later in the episode as well
15:04
later in the episode as well
15:04
later in the episode as well um and just in general i feel that
15:08
um and just in general i feel that
15:08
um and just in general i feel that uh it's a privilege to to be able to
15:10
uh it's a privilege to to be able to
15:10
uh it's a privilege to to be able to talk to those kind of people they're
15:11
talk to those kind of people they're
15:11
talk to those kind of people they're pretty
15:11
pretty
15:12
pretty busy i usually only get 30 minutes to
15:14
busy i usually only get 30 minutes to
15:14
busy i usually only get 30 minutes to prepare
15:15
prepare
15:15
prepare and i got a chance to talk much longer
15:17
and i got a chance to talk much longer
15:17
and i got a chance to talk much longer to him so that that's pretty special i
15:19
to him so that that's pretty special i
15:19
to him so that that's pretty special i have to say
15:20
have to say
15:20
have to say yeah but it's really great
15:23
yeah but it's really great
15:23
yeah but it's really great everything that you and you know all the
15:25
everything that you and you know all the
15:25
everything that you and you know all the other team members for the global ai
15:27
other team members for the global ai
15:27
other team members for the global ai community do
15:28
community do
15:28
community do to to help bring you know this content
15:31
to to help bring you know this content
15:31
to to help bring you know this content to the community and that's what really
15:34
to the community and that's what really
15:34
to the community and that's what really gets me
15:35
gets me
15:35
gets me pumped up and excited about the global
15:37
pumped up and excited about the global
15:37
pumped up and excited about the global ai community is
15:40
ai community is
15:40
ai community is getting access to all of this
15:42
getting access to all of this
15:42
getting access to all of this technologies and all of these speakers
15:44
technologies and all of these speakers
15:44
technologies and all of these speakers so thank you
15:46
so thank you
15:46
so thank you yeah i'm just happy to help out um for
15:48
yeah i'm just happy to help out um for
15:48
yeah i'm just happy to help out um for me it's
15:49
me it's
15:49
me it's it's important to help people learn more
15:51
it's important to help people learn more
15:51
it's important to help people learn more about ai the whole
15:54
about ai the whole
15:54
about ai the whole ai field is pretty new actually um
15:57
ai field is pretty new actually um
15:58
ai field is pretty new actually um i know it's been been around since the
15:59
i know it's been been around since the
15:59
i know it's been been around since the 50s richard talked about this
16:01
50s richard talked about this
16:01
50s richard talked about this that last week and um but i feel as if
16:07
that last week and um but i feel as if
16:07
that last week and um but i feel as if it's only been big in the last
16:10
it's only been big in the last
16:10
it's only been big in the last 10 years maybe it's been growing
16:12
10 years maybe it's been growing
16:12
10 years maybe it's been growing tremendously and there's a lot more
16:14
tremendously and there's a lot more
16:14
tremendously and there's a lot more people
16:14
people
16:14
people working on ai technology than ever
16:16
working on ai technology than ever
16:16
working on ai technology than ever before um
16:18
before um
16:18
before um and and i feel this is the moment in
16:19
and and i feel this is the moment in
16:19
and and i feel this is the moment in time where we
16:21
time where we
16:21
time where we should actually teach people more about
16:23
should actually teach people more about
16:23
should actually teach people more about it and make it more accessible
16:25
it and make it more accessible
16:25
it and make it more accessible both on technological side where people
16:28
both on technological side where people
16:28
both on technological side where people learn people
16:29
learn people
16:29
learn people learn about what you can do with ai
16:32
learn about what you can do with ai
16:32
learn about what you can do with ai and on the other hand there's of course
16:34
and on the other hand there's of course
16:34
and on the other hand there's of course the whole responsible ai
16:35
the whole responsible ai
16:36
the whole responsible ai aspect that that's also growing and
16:38
aspect that that's also growing and
16:38
aspect that that's also growing and becoming more important actually for
16:39
becoming more important actually for
16:39
becoming more important actually for people to
16:40
people to
16:40
people to um yeah to use um
16:43
um yeah to use um
16:43
um yeah to use um because you can build anything you want
16:46
because you can build anything you want
16:46
because you can build anything you want but it has to be safe
16:48
but it has to be safe
16:48
but it has to be safe i mean uh that's where responsible ai
16:51
i mean uh that's where responsible ai
16:51
i mean uh that's where responsible ai comes in and that's that's great
16:52
comes in and that's that's great
16:52
comes in and that's that's great we play a large role in that i believe
16:54
we play a large role in that i believe
16:54
we play a large role in that i believe uh as a global ai community we can help
16:56
uh as a global ai community we can help
16:56
uh as a global ai community we can help each other
16:57
each other
16:57
each other wherever you are if you're sitting at
16:59
wherever you are if you're sitting at
16:59
wherever you are if you're sitting at home working from home or
17:01
home working from home or
17:01
home working from home or going to the office because there's no
17:03
going to the office because there's no
17:03
going to the office because there's no other way to earn money
17:05
other way to earn money
17:05
other way to earn money we're still here to help you learn more
17:07
we're still here to help you learn more
17:07
we're still here to help you learn more about ai
17:08
about ai
17:08
about ai and thankfully i don't have to do that
17:10
and thankfully i don't have to do that
17:10
and thankfully i don't have to do that all alone
17:11
all alone
17:11
all alone i can do it with a lot of people around
17:13
i can do it with a lot of people around
17:13
i can do it with a lot of people around me such as you alicia but also eve
17:16
me such as you alicia but also eve
17:16
me such as you alicia but also eve sammy devote where people have seen seen
17:20
sammy devote where people have seen seen
17:20
sammy devote where people have seen seen all of you uh at some point in time on
17:23
all of you uh at some point in time on
17:23
all of you uh at some point in time on the stream
17:24
the stream
17:24
the stream um yeah that's pretty cool actually
17:29
um yeah that's pretty cool actually
17:29
um yeah that's pretty cool actually yeah i'm so grateful for the community i
17:32
yeah i'm so grateful for the community i
17:32
yeah i'm so grateful for the community i have been working in ai for for not as
17:35
have been working in ai for for not as
17:35
have been working in ai for for not as long as you've been working
17:36
long as you've been working
17:36
long as you've been working in ai and it's i i think the community
17:40
in ai and it's i i think the community
17:40
in ai and it's i i think the community is really what has helped me
17:41
is really what has helped me
17:41
is really what has helped me to to get up to speed so quickly and i
17:44
to to get up to speed so quickly and i
17:44
to to get up to speed so quickly and i think it's great that everybody is so
17:46
think it's great that everybody is so
17:46
think it's great that everybody is so ready and willing to to share
17:48
ready and willing to to share
17:48
ready and willing to to share information and knowledge
17:50
information and knowledge
17:50
information and knowledge oh yeah absolutely yeah there's a lot a
17:53
oh yeah absolutely yeah there's a lot a
17:53
oh yeah absolutely yeah there's a lot a lot of people that have the same
17:55
lot of people that have the same
17:55
lot of people that have the same backstory as you actually
17:56
backstory as you actually
17:56
backstory as you actually they just recently started and i've been
17:59
they just recently started and i've been
17:59
they just recently started and i've been around
18:00
around
18:00
around a bit longer uh eight years now uh which
18:03
a bit longer uh eight years now uh which
18:03
a bit longer uh eight years now uh which is not an awful long time if you ask me
18:05
is not an awful long time if you ask me
18:05
is not an awful long time if you ask me there's still so many things i don't
18:07
there's still so many things i don't
18:07
there's still so many things i don't know about and and i'm learning every
18:09
know about and and i'm learning every
18:09
know about and and i'm learning every day
18:09
day
18:09
day from from from the community actually as
18:12
from from from the community actually as
18:12
from from from the community actually as well
18:14
well
18:14
well i mean just recently we had a talk on
18:17
i mean just recently we had a talk on
18:17
i mean just recently we had a talk on global ai
18:18
global ai
18:18
global ai uh talks at the show that semi runs
18:21
uh talks at the show that semi runs
18:21
uh talks at the show that semi runs from a guy who developed a um
18:25
from a guy who developed a um
18:25
from a guy who developed a um again a generation a general adversarial
18:28
again a generation a general adversarial
18:28
again a generation a general adversarial network
18:28
network
18:28
network to mix portraits of two people together
18:32
to mix portraits of two people together
18:32
to mix portraits of two people together to form a new portrait that that's kind
18:33
to form a new portrait that that's kind
18:33
to form a new portrait that that's kind of cool too
18:40
so while we're waiting for the first
18:42
so while we're waiting for the first
18:42
so while we're waiting for the first guest guest to arrive
18:43
guest guest to arrive
18:44
guest guest to arrive we uh seem to have lost him in in
18:46
we uh seem to have lost him in in
18:46
we uh seem to have lost him in in another team's call and that's
18:48
another team's call and that's
18:48
another team's call and that's funny how that goes in online meetings
18:50
funny how that goes in online meetings
18:50
funny how that goes in online meetings uh and online shows for the better i
18:52
uh and online shows for the better i
18:52
uh and online shows for the better i guess he's in the test call
18:53
guess he's in the test call
18:53
guess he's in the test call um welcome everybody to the global ai
18:57
um welcome everybody to the global ai
18:57
um welcome everybody to the global ai show
18:57
show
18:57
show uh the global ai october sessions i'm
18:59
uh the global ai october sessions i'm
18:59
uh the global ai october sessions i'm going to say um
19:00
going to say um
19:00
going to say um in this second episode already i'm
19:03
in this second episode already i'm
19:03
in this second episode already i'm really happy you're here
19:04
really happy you're here
19:04
really happy you're here watching from facebook from twitter from
19:07
watching from facebook from twitter from
19:07
watching from facebook from twitter from uh
19:07
uh
19:07
uh youtube from our website last time we
19:11
youtube from our website last time we
19:11
youtube from our website last time we had
19:11
had
19:11
had over 10 000 people watching alicia
19:15
over 10 000 people watching alicia
19:15
over 10 000 people watching alicia so we've set a bar for ourselves to
19:17
so we've set a bar for ourselves to
19:17
so we've set a bar for ourselves to reach that again
19:19
reach that again
19:19
reach that again uh i guess and i've had tremendous
19:22
uh i guess and i've had tremendous
19:22
uh i guess and i've had tremendous feedback from everybody
19:24
feedback from everybody
19:24
feedback from everybody it actually blew me away and this is not
19:27
it actually blew me away and this is not
19:27
it actually blew me away and this is not a sort of trump
19:28
a sort of trump
19:28
a sort of trump uh saying so to say he does that a lot
19:31
uh saying so to say he does that a lot
19:31
uh saying so to say he does that a lot it's amazing it's uh it's super awesome
19:35
it's amazing it's uh it's super awesome
19:35
it's amazing it's uh it's super awesome but i'm actually blown away i didn't
19:37
but i'm actually blown away i didn't
19:37
but i'm actually blown away i didn't expect that to happen
19:39
expect that to happen
19:39
expect that to happen at all i don't know about you actually
19:41
at all i don't know about you actually
19:41
at all i don't know about you actually um
19:42
um
19:42
um well i wasn't surprised i
19:45
well i wasn't surprised i
19:45
well i wasn't surprised i saw the change format and i i saw the
19:49
saw the change format and i i saw the
19:49
saw the change format and i i saw the amount of time and effort everybody put
19:50
amount of time and effort everybody put
19:50
amount of time and effort everybody put into making the show more interactive
19:52
into making the show more interactive
19:52
into making the show more interactive and i i know the topics and your
19:54
and i i know the topics and your
19:54
and i i know the topics and your speakers were
19:55
speakers were
19:55
speakers were were just amazing so um congratulations
20:00
were just amazing so um congratulations
20:00
were just amazing so um congratulations and watching the show i i had a lot of
20:03
and watching the show i i had a lot of
20:03
and watching the show i i had a lot of fun watching
20:04
fun watching
20:04
fun watching ah cool cool cool yeah a lot of people
20:07
ah cool cool cool yeah a lot of people
20:08
ah cool cool cool yeah a lot of people from the crew have been watching as well
20:09
from the crew have been watching as well
20:09
from the crew have been watching as well and and people should know that
20:11
and and people should know that
20:11
and and people should know that we have actually a couple of our friends
20:13
we have actually a couple of our friends
20:13
we have actually a couple of our friends from the community
20:14
from the community
20:14
from the community helping us out today with uh moderating
20:16
helping us out today with uh moderating
20:16
helping us out today with uh moderating the chat um so if you have questions for
20:19
the chat um so if you have questions for
20:19
the chat um so if you have questions for us
20:20
us
20:20
us feel free to ask them they will forward
20:21
feel free to ask them they will forward
20:22
feel free to ask them they will forward them to alicia and me through our back
20:23
them to alicia and me through our back
20:23
them to alicia and me through our back channel we are using slack actually as a
20:25
channel we are using slack actually as a
20:25
channel we are using slack actually as a tool for that
20:26
tool for that
20:26
tool for that um to get in so sometimes you will see
20:29
um to get in so sometimes you will see
20:29
um to get in so sometimes you will see me watching to the to the left
20:31
me watching to the to the left
20:31
me watching to the to the left uh you're right um
20:35
uh you're right um
20:35
uh you're right um checking what's going on behind the
20:37
checking what's going on behind the
20:37
checking what's going on behind the scenes for us uh but we'll definitely
20:39
scenes for us uh but we'll definitely
20:39
scenes for us uh but we'll definitely get your questions and
20:40
get your questions and
20:40
get your questions and um what's interesting actually uh we're
20:43
um what's interesting actually uh we're
20:43
um what's interesting actually uh we're going to give away another 50
20:45
going to give away another 50
20:45
going to give away another 50 gift card today this year um
20:48
gift card today this year um
20:48
gift card today this year um for the person that has the most
20:49
for the person that has the most
20:49
for the person that has the most interesting question for us
20:51
interesting question for us
20:51
interesting question for us last time we we asked people to take
20:53
last time we we asked people to take
20:53
last time we we asked people to take pictures but we've got a different plan
20:54
pictures but we've got a different plan
20:54
pictures but we've got a different plan for that
20:55
for that
20:55
for that we'll tell you about it this time if you
20:58
we'll tell you about it this time if you
20:58
we'll tell you about it this time if you have questions please feel free to ask
21:00
have questions please feel free to ask
21:00
have questions please feel free to ask us
21:00
us
21:00
us we'll make sure that the speaker knows
21:02
we'll make sure that the speaker knows
21:02
we'll make sure that the speaker knows about them or we will answer them
21:04
about them or we will answer them
21:04
about them or we will answer them if that's at all possible um
21:07
if that's at all possible um
21:07
if that's at all possible um so that's cool and we've got more stuff
21:10
so that's cool and we've got more stuff
21:10
so that's cool and we've got more stuff to give away right
21:12
to give away right
21:12
to give away right i i think we have one more thing
21:16
i i think we have one more thing
21:16
i i think we have one more thing what is it thank you to
21:19
what is it thank you to
21:19
what is it thank you to the global ai community for sponsoring
21:22
the global ai community for sponsoring
21:22
the global ai community for sponsoring an oculus
21:23
an oculus
21:23
an oculus headset um
21:26
headset um
21:26
headset um wait what we would like to do is
21:30
wait what we would like to do is
21:30
wait what we would like to do is um you know one of the reasons
21:34
um you know one of the reasons
21:34
um you know one of the reasons that i enjoy the global ai community so
21:36
that i enjoy the global ai community so
21:36
that i enjoy the global ai community so much is because we're a community
21:38
much is because we're a community
21:38
much is because we're a community and i i really miss
21:41
and i i really miss
21:41
and i i really miss seeing everybody at the events and
21:44
seeing everybody at the events and
21:44
seeing everybody at the events and making new friends and connecting with
21:46
making new friends and connecting with
21:46
making new friends and connecting with new people
21:47
new people
21:47
new people and sharing experiences and
21:50
and sharing experiences and
21:50
and sharing experiences and you know connecting and growing together
21:52
you know connecting and growing together
21:52
you know connecting and growing together so
21:53
so
21:53
so my call to action is um i'd like you
21:57
my call to action is um i'd like you
21:57
my call to action is um i'd like you all to help me with a photo challenge so
22:00
all to help me with a photo challenge so
22:00
all to help me with a photo challenge so take a selfie of yourself watching the
22:03
take a selfie of yourself watching the
22:03
take a selfie of yourself watching the october session with us today
22:06
october session with us today
22:06
october session with us today and post it on twitter and
22:09
and post it on twitter and
22:09
and post it on twitter and hashtag globalai community
22:12
hashtag globalai community
22:12
hashtag globalai community and let's let's all say hi to each other
22:15
and let's let's all say hi to each other
22:15
and let's let's all say hi to each other and get to know the rest of our
22:17
and get to know the rest of our
22:17
and get to know the rest of our community and
22:19
community and
22:19
community and um we'll go ahead and post a couple of
22:22
um we'll go ahead and post a couple of
22:22
um we'll go ahead and post a couple of reminders throughout the show
22:24
reminders throughout the show
22:24
reminders throughout the show and maybe about an hour to
22:27
and maybe about an hour to
22:27
and maybe about an hour to after the event we'll go ahead and and
22:30
after the event we'll go ahead and and
22:30
after the event we'll go ahead and and scroll through and pick a winner
22:31
scroll through and pick a winner
22:32
scroll through and pick a winner but um if you could participate it's
22:35
but um if you could participate it's
22:35
but um if you could participate it's this is going to be a lot of fun and i'm
22:38
this is going to be a lot of fun and i'm
22:38
this is going to be a lot of fun and i'm dying
22:38
dying
22:38
dying to to connect with y'all and
22:42
to to connect with y'all and
22:42
to to connect with y'all and to get to know the rest of our community
22:44
to get to know the rest of our community
22:44
to get to know the rest of our community so
22:45
so
22:45
so thank you cool cool cool
22:49
thank you cool cool cool
22:49
thank you cool cool cool that's that's actually a pretty amazing
22:51
that's that's actually a pretty amazing
22:51
that's that's actually a pretty amazing device and and
22:53
device and and
22:53
device and and um what's even more funny we've got two
22:55
um what's even more funny we've got two
22:56
um what's even more funny we've got two community demos today
22:57
community demos today
22:57
community demos today one talking about augmented reality
23:00
one talking about augmented reality
23:00
one talking about augmented reality using hololens and ai
23:02
using hololens and ai
23:02
using hololens and ai hey satya is here that's cool
23:05
hey satya is here that's cool
23:06
hey satya is here that's cool and the other demo that we have is is
23:07
and the other demo that we have is is
23:07
and the other demo that we have is is about virtual reality so if you
23:10
about virtual reality so if you
23:10
about virtual reality so if you win you can immediately immediately
23:12
win you can immediately immediately
23:12
win you can immediately immediately start to use ai on this
23:13
start to use ai on this
23:13
start to use ai on this on this apparatus so that that's kind of
23:15
on this apparatus so that that's kind of
23:15
on this apparatus so that that's kind of cool uh
23:16
cool uh
23:16
cool uh i like that you're doing this for the
23:17
i like that you're doing this for the
23:17
i like that you're doing this for the community yeah alicia
23:19
community yeah alicia
23:19
community yeah alicia and with that i would like to introduce
23:21
and with that i would like to introduce
23:21
and with that i would like to introduce uh satya
23:22
uh satya
23:22
uh satya hey sacha how are you doing i'm good uh
23:26
hey sacha how are you doing i'm good uh
23:26
hey sacha how are you doing i'm good uh somehow the link was not working i kept
23:28
somehow the link was not working i kept
23:28
somehow the link was not working i kept waiting
23:28
waiting
23:28
waiting uh sorry about that i was 15 minutes
23:30
uh sorry about that i was 15 minutes
23:30
uh sorry about that i was 15 minutes early
23:31
early
23:31
early but oh i thought that it would just let
23:34
but oh i thought that it would just let
23:34
but oh i thought that it would just let me in
23:34
me in
23:34
me in yeah then i think they sent a new link
23:36
yeah then i think they sent a new link
23:36
yeah then i think they sent a new link and that seems to work
23:38
and that seems to work
23:38
and that seems to work yeah hank is amazing like that hank is
23:40
yeah hank is amazing like that hank is
23:40
yeah hank is amazing like that hank is uh our uh
23:41
uh our uh
23:41
uh our uh our director producer sound guy
23:44
our director producer sound guy
23:44
our director producer sound guy um cable guy extraordinaire in the
23:47
um cable guy extraordinaire in the
23:47
um cable guy extraordinaire in the background
23:48
background
23:48
background uh he's actually uh clicking together
23:50
uh he's actually uh clicking together
23:50
uh he's actually uh clicking together all the video streams that we have
23:51
all the video streams that we have
23:51
all the video streams that we have tonight and it's
23:52
tonight and it's
23:52
tonight and it's and it's happening from my home actually
23:54
and it's happening from my home actually
23:54
and it's happening from my home actually i've got a sick kid at home in the
23:56
i've got a sick kid at home in the
23:56
i've got a sick kid at home in the and the netherlands is actually in a
23:58
and the netherlands is actually in a
23:58
and the netherlands is actually in a partial lockdown at the moment so i
23:59
partial lockdown at the moment so i
23:59
partial lockdown at the moment so i can't go out the door
24:00
can't go out the door
24:00
can't go out the door when my one of my kids is sick so um
24:03
when my one of my kids is sick so um
24:03
when my one of my kids is sick so um this is from my attic from my home
24:05
this is from my attic from my home
24:05
this is from my attic from my home office
24:06
office
24:06
office and where are you located at the moment
24:08
and where are you located at the moment
24:08
and where are you located at the moment sacha
24:09
sacha
24:09
sacha uh san diego oh california
24:12
uh san diego oh california
24:12
uh san diego oh california nice also working from home uh well
24:15
nice also working from home uh well
24:15
nice also working from home uh well i have a home up i have an office which
24:18
i have a home up i have an office which
24:18
i have a home up i have an office which is like five minutes from home
24:19
is like five minutes from home
24:19
is like five minutes from home so yeah it's almost like home
24:23
so yeah it's almost like home
24:23
so yeah it's almost like home it feels like home sometimes yeah cool
24:26
it feels like home sometimes yeah cool
24:26
it feels like home sometimes yeah cool cool awesome to have you so um for those
24:29
cool awesome to have you so um for those
24:29
cool awesome to have you so um for those who don't know sacha
24:32
who don't know sacha
24:32
who don't know sacha could you introduce yourself to the
24:33
could you introduce yourself to the
24:33
could you introduce yourself to the community
24:35
community
24:35
community sure so i am the ceo of
24:38
sure so i am the ceo of
24:38
sure so i am the ceo of opencv.org it is the foundation that
24:41
opencv.org it is the foundation that
24:42
opencv.org it is the foundation that actually maintains the opencv library
24:44
actually maintains the opencv library
24:44
actually maintains the opencv library and as
24:45
and as
24:45
and as most people probably know that opencv is
24:48
most people probably know that opencv is
24:48
most people probably know that opencv is the largest computer vision library
24:50
the largest computer vision library
24:50
the largest computer vision library in the world and uh for traditional
24:53
in the world and uh for traditional
24:53
in the world and uh for traditional computer vision
24:54
computer vision
24:54
computer vision algorithms and i'm also the ceo of
24:57
algorithms and i'm also the ceo of
24:57
algorithms and i'm also the ceo of opencv.ai which is the for-profit arm
24:59
opencv.ai which is the for-profit arm
24:59
opencv.ai which is the for-profit arm for opencv sounds awesome
25:03
for opencv sounds awesome
25:03
for opencv sounds awesome and how long have you actually been
25:04
and how long have you actually been
25:04
and how long have you actually been working with opencv
25:07
working with opencv
25:07
working with opencv for a little over a year uh
25:10
for a little over a year uh
25:10
for a little over a year uh i joined as the ceo of opencv a little
25:12
i joined as the ceo of opencv a little
25:12
i joined as the ceo of opencv a little over a year uh
25:14
over a year uh
25:14
over a year uh dr gary bradsky who's the founder of
25:17
dr gary bradsky who's the founder of
25:17
dr gary bradsky who's the founder of opencv he asked me to join and then
25:20
opencv he asked me to join and then
25:20
opencv he asked me to join and then you know we decided it took a few years
25:23
you know we decided it took a few years
25:23
you know we decided it took a few years the first time he asked me i was not
25:25
the first time he asked me i was not
25:25
the first time he asked me i was not ready because
25:27
ready because
25:27
ready because things were going on and uh last year i
25:30
things were going on and uh last year i
25:30
things were going on and uh last year i decided yes
25:30
decided yes
25:30
decided yes i have a clear vision now on where the
25:34
i have a clear vision now on where the
25:34
i have a clear vision now on where the foundation should go cool well that
25:37
foundation should go cool well that
25:37
foundation should go cool well that sounds really awesome
25:38
sounds really awesome
25:38
sounds really awesome so um i know for a fact that that opencv
25:42
so um i know for a fact that that opencv
25:42
so um i know for a fact that that opencv has been around for 20 years
25:44
has been around for 20 years
25:44
has been around for 20 years uh that's a that's a really long time in
25:46
uh that's a that's a really long time in
25:46
uh that's a that's a really long time in software land
25:47
software land
25:47
software land um and and i can imagine that there's a
25:49
um and and i can imagine that there's a
25:49
um and and i can imagine that there's a lot of new stuff that you that you can
25:51
lot of new stuff that you that you can
25:51
lot of new stuff that you that you can talk about
25:52
talk about
25:52
talk about so um whenever you're ready feel free to
25:54
so um whenever you're ready feel free to
25:54
so um whenever you're ready feel free to share your screen and show
25:56
share your screen and show
25:56
share your screen and show us uh your awesome keynote sure
25:59
us uh your awesome keynote sure
25:59
us uh your awesome keynote sure so uh just before i share that uh
26:03
so uh just before i share that uh
26:03
so uh just before i share that uh opencv library like in the keynote i'm
26:05
opencv library like in the keynote i'm
26:05
opencv library like in the keynote i'm going to talk about you know how opencv
26:07
going to talk about you know how opencv
26:07
going to talk about you know how opencv has grown beyond
26:08
has grown beyond
26:08
has grown beyond the opencv library and one of the things
26:11
the opencv library and one of the things
26:11
the opencv library and one of the things the main thing today
26:13
the main thing today
26:13
the main thing today would be opencv ai kit that's what i'm
26:15
would be opencv ai kit that's what i'm
26:15
would be opencv ai kit that's what i'm going to talk about and
26:17
going to talk about and
26:17
going to talk about and show what you can do with it cool so
26:22
let's
26:24
let's
26:24
let's um okay are you guys able to
26:28
um okay are you guys able to
26:28
um okay are you guys able to absolutely my screen yeah okay
26:32
absolutely my screen yeah okay
26:32
absolutely my screen yeah okay cool so um yeah so the talk is about
26:36
cool so um yeah so the talk is about
26:36
cool so um yeah so the talk is about opencv ai kit now
26:39
opencv ai kit now
26:39
opencv ai kit now just to get you guys all excited about
26:41
just to get you guys all excited about
26:41
just to get you guys all excited about this talk
26:42
this talk
26:42
this talk i want to first talk about our amazing
26:45
i want to first talk about our amazing
26:45
i want to first talk about our amazing kickstarter campaign
26:47
kickstarter campaign
26:47
kickstarter campaign so to launch this opencv ai kit which is
26:50
so to launch this opencv ai kit which is
26:50
so to launch this opencv ai kit which is essentially a smart
26:51
essentially a smart
26:52
essentially a smart camera which can do depth estimation as
26:55
camera which can do depth estimation as
26:55
camera which can do depth estimation as well
26:55
well
26:56
well we launched a kickstarter campaign a few
26:58
we launched a kickstarter campaign a few
26:58
we launched a kickstarter campaign a few months
26:59
months
26:59
months earlier and it was the biggest
27:02
earlier and it was the biggest
27:02
earlier and it was the biggest kickstarter campaign
27:03
kickstarter campaign
27:03
kickstarter campaign in the history of uh smart cameras
27:07
in the history of uh smart cameras
27:07
in the history of uh smart cameras so this uh we raised 1.35
27:11
so this uh we raised 1.35
27:11
so this uh we raised 1.35 million dollars with the help of 6500
27:15
million dollars with the help of 6500
27:15
million dollars with the help of 6500 backers so we have a very strong
27:17
backers so we have a very strong
27:17
backers so we have a very strong community
27:18
community
27:18
community that is helping us with uh this uh with
27:21
that is helping us with uh this uh with
27:21
that is helping us with uh this uh with this
27:22
this
27:22
this uh camera so with that now you would be
27:25
uh camera so with that now you would be
27:25
uh camera so with that now you would be cur
27:26
cur
27:26
cur to know what is so special about the
27:29
to know what is so special about the
27:29
to know what is so special about the smart camera
27:30
smart camera
27:30
smart camera uh why did so many people get excited
27:32
uh why did so many people get excited
27:32
uh why did so many people get excited about this camera
27:33
about this camera
27:33
about this camera and that is what this talk is about so
27:37
and that is what this talk is about so
27:37
and that is what this talk is about so until now you know we have thought about
27:39
until now you know we have thought about
27:39
until now you know we have thought about opencv
27:41
opencv
27:42
opencv as the library whenever you talk about
27:43
as the library whenever you talk about
27:44
as the library whenever you talk about opencv it is always
27:45
opencv it is always
27:45
opencv it is always about the opencv library but now
27:48
about the opencv library but now
27:48
about the opencv library but now we are growing beyond uh the library
27:52
we are growing beyond uh the library
27:52
we are growing beyond uh the library now we have an ecosystem of things we
27:56
now we have an ecosystem of things we
27:56
now we have an ecosystem of things we think of opencv as a brand and
27:59
think of opencv as a brand and
27:59
think of opencv as a brand and we want to serve the community by
28:01
we want to serve the community by
28:01
we want to serve the community by providing various
28:02
providing various
28:02
providing various services library is still the
28:05
services library is still the
28:05
services library is still the key thing we do but we also provide
28:07
key thing we do but we also provide
28:07
key thing we do but we also provide courses we provide hardware
28:09
courses we provide hardware
28:09
courses we provide hardware software we provide community support
28:12
software we provide community support
28:12
software we provide community support and
28:13
and
28:13
and very soon we will also launch a store
28:16
very soon we will also launch a store
28:16
very soon we will also launch a store where people will be able to buy
28:17
where people will be able to buy
28:17
where people will be able to buy computer vision hardware software
28:20
computer vision hardware software
28:20
computer vision hardware software and services so we want opencv brand
28:24
and services so we want opencv brand
28:24
and services so we want opencv brand to include anything you know if you
28:26
to include anything you know if you
28:26
to include anything you know if you think about any resources you need for
28:28
think about any resources you need for
28:28
think about any resources you need for computer vision
28:29
computer vision
28:29
computer vision for your project or for your company you
28:32
for your project or for your company you
28:32
for your project or for your company you should think about
28:33
should think about
28:33
should think about the opencv brand that's where we are
28:36
the opencv brand that's where we are
28:36
the opencv brand that's where we are moving things
28:38
moving things
28:38
moving things so now that brings us to opencv ai kit
28:42
so now that brings us to opencv ai kit
28:42
so now that brings us to opencv ai kit it is a family of edge computing
28:45
it is a family of edge computing
28:45
it is a family of edge computing solution
28:46
solution
28:46
solution edge ai solutions and we'll see in a
28:49
edge ai solutions and we'll see in a
28:49
edge ai solutions and we'll see in a little bit what that means
28:51
little bit what that means
28:51
little bit what that means uh edge ai is basically the opposite of
28:54
uh edge ai is basically the opposite of
28:54
uh edge ai is basically the opposite of cloud ai in the cloud we can
28:58
cloud ai in the cloud we can
28:58
cloud ai in the cloud we can have gpus and we can have all the
29:00
have gpus and we can have all the
29:00
have gpus and we can have all the processing power we need
29:02
processing power we need
29:02
processing power we need there are no power constraints usually
29:04
there are no power constraints usually
29:04
there are no power constraints usually so we can deploy
29:05
so we can deploy
29:05
so we can deploy this very heavy duty solutions which
29:08
this very heavy duty solutions which
29:08
this very heavy duty solutions which work very well
29:09
work very well
29:09
work very well but they are they require this cloud
29:12
but they are they require this cloud
29:12
but they are they require this cloud resource
29:13
resource
29:13
resource and they require connectivity so the
29:16
and they require connectivity so the
29:16
and they require connectivity so the other option is to go
29:17
other option is to go
29:17
other option is to go and do all these things on the edge on
29:20
and do all these things on the edge on
29:20
and do all these things on the edge on the device
29:21
the device
29:21
the device where the picture is being taken not
29:24
where the picture is being taken not
29:24
where the picture is being taken not load
29:24
load
29:24
load upload anything to the cloud so
29:27
upload anything to the cloud so
29:27
upload anything to the cloud so why would you do agi the first thing is
29:29
why would you do agi the first thing is
29:29
why would you do agi the first thing is speed in
29:30
speed in
29:30
speed in a lot of applications for example if
29:33
a lot of applications for example if
29:33
a lot of applications for example if you're doing sports
29:34
you're doing sports
29:34
you're doing sports analytics things are moving so fast that
29:37
analytics things are moving so fast that
29:37
analytics things are moving so fast that you have no option but to
29:38
you have no option but to
29:38
you have no option but to do the analysis on the device in real
29:41
do the analysis on the device in real
29:41
do the analysis on the device in real time
29:42
time
29:42
time the second thing is that even though you
29:43
the second thing is that even though you
29:43
the second thing is that even though you know 5g is very
29:45
know 5g is very
29:45
know 5g is very hot these days connectivity is still
29:48
hot these days connectivity is still
29:48
hot these days connectivity is still a very big problem in a lot of places
29:52
a very big problem in a lot of places
29:52
a very big problem in a lot of places especially connectivity which can help
29:55
especially connectivity which can help
29:55
especially connectivity which can help you
29:55
you
29:55
you upload videos and get them back in good
29:58
upload videos and get them back in good
29:58
upload videos and get them back in good uh real time so we need
30:02
uh real time so we need
30:02
uh real time so we need to do this processing on the edge and
30:05
to do this processing on the edge and
30:05
to do this processing on the edge and then we have privacy concerns also
30:07
then we have privacy concerns also
30:07
then we have privacy concerns also imagine that you have a roomba which is
30:09
imagine that you have a roomba which is
30:09
imagine that you have a roomba which is using a camera
30:11
using a camera
30:11
using a camera you don't want that picture to go out
30:13
you don't want that picture to go out
30:13
you don't want that picture to go out and be processed on the cloud
30:15
and be processed on the cloud
30:15
and be processed on the cloud it is sometimes okay that you know that
30:17
it is sometimes okay that you know that
30:18
it is sometimes okay that you know that camera is processing on the device and
30:20
camera is processing on the device and
30:20
camera is processing on the device and you're guaranteed that
30:21
you're guaranteed that
30:21
you're guaranteed that the images are not going out so there is
30:23
the images are not going out so there is
30:23
the images are not going out so there is a privacy concern because of which
30:24
a privacy concern because of which
30:24
a privacy concern because of which people may want
30:25
people may want
30:26
people may want to do uh the processing on the device
30:29
to do uh the processing on the device
30:29
to do uh the processing on the device itself
30:29
itself
30:29
itself and finally it's the cost it is uh
30:32
and finally it's the cost it is uh
30:32
and finally it's the cost it is uh hugely expensive
30:33
hugely expensive
30:33
hugely expensive to do the processing on the cloud people
30:35
to do the processing on the cloud people
30:36
to do the processing on the cloud people know you know
30:37
know you know
30:37
know you know what is the cost of aws it starts out
30:40
what is the cost of aws it starts out
30:40
what is the cost of aws it starts out easy and then
30:42
easy and then
30:42
easy and then it blows up very quickly so
30:45
it blows up very quickly so
30:45
it blows up very quickly so there are several challenges in
30:47
there are several challenges in
30:47
there are several challenges in deploying an edge
30:49
deploying an edge
30:49
deploying an edge solution first of all we are resource
30:51
solution first of all we are resource
30:51
solution first of all we are resource constrained uh the memory as well as the
30:53
constrained uh the memory as well as the
30:53
constrained uh the memory as well as the processing
30:54
processing
30:54
processing is often constrained the electrical
30:57
is often constrained the electrical
30:57
is often constrained the electrical power consumption is often constrained
30:59
power consumption is often constrained
30:59
power consumption is often constrained as well some of these devices could be
31:01
as well some of these devices could be
31:01
as well some of these devices could be running on batteries and we don't then
31:03
running on batteries and we don't then
31:03
running on batteries and we don't then want to drain out the batteries and
31:05
want to drain out the batteries and
31:05
want to drain out the batteries and finally the form factor
31:07
finally the form factor
31:07
finally the form factor is a crucial they have to be small
31:11
is a crucial they have to be small
31:11
is a crucial they have to be small so we started out with a mission to
31:15
so we started out with a mission to
31:15
so we started out with a mission to democratize edge ai the way raspberry pi
31:19
democratize edge ai the way raspberry pi
31:19
democratize edge ai the way raspberry pi democratized edge computing
31:22
democratized edge computing
31:22
democratized edge computing that's a very big goal given the
31:23
that's a very big goal given the
31:24
that's a very big goal given the popularity of raspberry pi
31:25
popularity of raspberry pi
31:25
popularity of raspberry pi but i think that with oak we have taken
31:29
but i think that with oak we have taken
31:29
but i think that with oak we have taken a step in the right direction we have
31:32
a step in the right direction we have
31:32
a step in the right direction we have come up
31:33
come up
31:33
come up with two versions of opencv ai kit
31:36
with two versions of opencv ai kit
31:36
with two versions of opencv ai kit the first one is called oak one it is a
31:39
the first one is called oak one it is a
31:39
the first one is called oak one it is a smart camera
31:40
smart camera
31:40
smart camera which can do neural computing on the
31:43
which can do neural computing on the
31:43
which can do neural computing on the device
31:44
device
31:44
device and the second one is called oak d d for
31:47
and the second one is called oak d d for
31:47
and the second one is called oak d d for depth
31:48
depth
31:48
depth it can do everything that oak one can do
31:51
it can do everything that oak one can do
31:51
it can do everything that oak one can do but
31:51
but
31:51
but also depth estimation
31:55
also depth estimation
31:55
also depth estimation so let's look into oak one
31:58
so let's look into oak one
31:58
so let's look into oak one what can it do you know it's tiny it's
32:01
what can it do you know it's tiny it's
32:01
what can it do you know it's tiny it's powerful if you look at you know
32:05
powerful if you look at you know
32:05
powerful if you look at you know existing solutions
32:06
existing solutions
32:06
existing solutions uh on how people come today use
32:09
uh on how people come today use
32:09
uh on how people come today use neural accelerators in their solution
32:12
neural accelerators in their solution
32:12
neural accelerators in their solution this is a standard configuration
32:14
this is a standard configuration
32:14
this is a standard configuration you have a camera that is connected to a
32:17
you have a camera that is connected to a
32:17
you have a camera that is connected to a host
32:18
host
32:18
host like a raspberry pi and you have an
32:21
like a raspberry pi and you have an
32:21
like a raspberry pi and you have an ai processor plugged in like a mobidius
32:25
ai processor plugged in like a mobidius
32:25
ai processor plugged in like a mobidius neural compute stick
32:26
neural compute stick
32:26
neural compute stick the problem with this approach is that
32:28
the problem with this approach is that
32:28
the problem with this approach is that the camera records the frame
32:31
the camera records the frame
32:31
the camera records the frame and then it moves into the raspberry pi
32:34
and then it moves into the raspberry pi
32:34
and then it moves into the raspberry pi and then that frame goes into the neural
32:37
and then that frame goes into the neural
32:37
and then that frame goes into the neural compute stick
32:38
compute stick
32:38
compute stick so the path is pretty long and when
32:41
so the path is pretty long and when
32:41
so the path is pretty long and when you're trying to do something at 60
32:43
you're trying to do something at 60
32:43
you're trying to do something at 60 frames a second
32:44
frames a second
32:44
frames a second it becomes very difficult to you know
32:47
it becomes very difficult to you know
32:47
it becomes very difficult to you know maintain sync for example with this
32:49
maintain sync for example with this
32:49
maintain sync for example with this configuration
32:50
configuration
32:50
configuration if you run tinyulo you will probably get
32:53
if you run tinyulo you will probably get
32:53
if you run tinyulo you will probably get about
32:54
about
32:54
about eight frames a second so what we did
32:57
eight frames a second so what we did
32:57
eight frames a second so what we did was we
33:01
was we
33:01
was we combined the 4k camera and the ai
33:05
combined the 4k camera and the ai
33:05
combined the 4k camera and the ai processor
33:05
processor
33:05
processor into a single unit which we call
33:09
into a single unit which we call
33:09
into a single unit which we call oak one and object detection on oak one
33:13
oak one and object detection on oak one
33:13
oak one and object detection on oak one is eight times faster than the
33:16
is eight times faster than the
33:16
is eight times faster than the configuration
33:17
configuration
33:17
configuration that i showed here even though the
33:20
that i showed here even though the
33:20
that i showed here even though the processor
33:20
processor
33:20
processor inside movidius is exactly the same
33:24
inside movidius is exactly the same
33:24
inside movidius is exactly the same as the processor inside ok one but we
33:27
as the processor inside ok one but we
33:27
as the processor inside ok one but we get
33:27
get
33:27
get 8x speed up a part of it is our
33:30
8x speed up a part of it is our
33:30
8x speed up a part of it is our you know software stack also but a big
33:33
you know software stack also but a big
33:33
you know software stack also but a big part
33:34
part
33:34
part uh is that it's not
33:37
uh is that it's not
33:37
uh is that it's not this is not the right configuration we
33:39
this is not the right configuration we
33:39
this is not the right configuration we have we should be working on
33:40
have we should be working on
33:40
have we should be working on the camera should be connected to the
33:42
the camera should be connected to the
33:42
the camera should be connected to the neural compute stick
33:43
neural compute stick
33:43
neural compute stick or to the processor uh through a very
33:47
or to the processor uh through a very
33:47
or to the processor uh through a very you know uh very good connection very
33:49
you know uh very good connection very
33:50
you know uh very good connection very fast connection
33:53
fast connection
33:53
fast connection the other thing we want i wanted to have
33:55
the other thing we want i wanted to have
33:55
the other thing we want i wanted to have with oak one
33:56
with oak one
33:56
with oak one is that it had to be it has to be tiny
33:58
is that it had to be it has to be tiny
33:58
is that it had to be it has to be tiny and you can see
33:59
and you can see
33:59
and you can see as compared to uh a quarter
34:03
as compared to uh a quarter
34:03
as compared to uh a quarter what is the size of uh this it's it's
34:06
what is the size of uh this it's it's
34:06
what is the size of uh this it's it's really tiny it is
34:07
really tiny it is
34:07
really tiny it is smaller than a double a battery
34:10
smaller than a double a battery
34:10
smaller than a double a battery so uh the height is smaller than a
34:13
so uh the height is smaller than a
34:13
so uh the height is smaller than a double a battery
34:14
double a battery
34:14
double a battery so it's some amazing product for an edge
34:17
so it's some amazing product for an edge
34:17
so it's some amazing product for an edge uh
34:18
uh
34:18
uh for an edge device now let's go on to
34:22
for an edge device now let's go on to
34:22
for an edge device now let's go on to od and ogd is inspired by human vision
34:26
od and ogd is inspired by human vision
34:26
od and ogd is inspired by human vision because if you look at how our eyes work
34:30
because if you look at how our eyes work
34:30
because if you look at how our eyes work they have two functions the first thing
34:33
they have two functions the first thing
34:33
they have two functions the first thing is
34:34
is
34:34
is they tell us what's in the scene
34:37
they tell us what's in the scene
34:37
they tell us what's in the scene you know what are we looking at and the
34:39
you know what are we looking at and the
34:39
you know what are we looking at and the second one is how far
34:41
second one is how far
34:41
second one is how far is it from it also does depth perception
34:44
is it from it also does depth perception
34:44
is it from it also does depth perception and that's why we have stereo vision and
34:46
and that's why we have stereo vision and
34:46
and that's why we have stereo vision and that was the whole objective
34:48
that was the whole objective
34:48
that was the whole objective of oak d so ogd consists of a 4k
34:53
of oak d so ogd consists of a 4k
34:53
of oak d so ogd consists of a 4k camera a high resolution camera a depth
34:56
camera a high resolution camera a depth
34:56
camera a high resolution camera a depth uh camera so there are two cameras which
34:59
uh camera so there are two cameras which
34:59
uh camera so there are two cameras which serve as the stereo pair and we have
35:02
serve as the stereo pair and we have
35:02
serve as the stereo pair and we have the ai processor if you combine all
35:05
the ai processor if you combine all
35:05
the ai processor if you combine all these
35:05
these
35:05
these uh things if you look at the cost itself
35:08
uh things if you look at the cost itself
35:08
uh things if you look at the cost itself first
35:09
first
35:09
first forget about synchronizing the frames
35:12
forget about synchronizing the frames
35:12
forget about synchronizing the frames which is in itself
35:13
which is in itself
35:13
which is in itself a big problem when you use a
35:14
a big problem when you use a
35:14
a big problem when you use a configuration like this but the cost
35:17
configuration like this but the cost
35:17
configuration like this but the cost itself
35:17
itself
35:17
itself is so big compared to oak it is
35:21
is so big compared to oak it is
35:21
is so big compared to oak it is one compact solution that we have come
35:23
one compact solution that we have come
35:23
one compact solution that we have come up with
35:25
up with
35:25
up with so when i say you know you can do neural
35:27
so when i say you know you can do neural
35:27
so when i say you know you can do neural inference on any of these devices oak 1
35:30
inference on any of these devices oak 1
35:30
inference on any of these devices oak 1 and oak
35:31
and oak
35:31
and oak d these are the kinds of things you can
35:33
d these are the kinds of things you can
35:33
d these are the kinds of things you can do
35:34
do
35:34
do you can run any object detector
35:37
you can run any object detector
35:37
you can run any object detector which is supported by openvino for
35:39
which is supported by openvino for
35:39
which is supported by openvino for example mobilenet
35:41
example mobilenet
35:41
example mobilenet ssdv2 tiny yolo etc
35:44
ssdv2 tiny yolo etc
35:44
ssdv2 tiny yolo etc and these models will come
35:47
and these models will come
35:48
and these models will come pre-packaged with the oak solution
35:51
pre-packaged with the oak solution
35:51
pre-packaged with the oak solution and i'll show a little video where
35:53
and i'll show a little video where
35:53
and i'll show a little video where you'll see how easy it is to
35:55
you'll see how easy it is to
35:55
you'll see how easy it is to get started once you have the device in
35:57
get started once you have the device in
35:57
get started once you have the device in your hand
35:59
your hand
35:59
your hand you can also do face detection our
36:02
you can also do face detection our
36:02
you can also do face detection our devices will ship
36:03
devices will ship
36:03
devices will ship with these models where you can do a lot
36:06
with these models where you can do a lot
36:06
with these models where you can do a lot of face applications
36:07
of face applications
36:07
of face applications like phase detection facial landmark
36:09
like phase detection facial landmark
36:09
like phase detection facial landmark detection
36:11
detection
36:11
detection expression estimation age estimation
36:13
expression estimation age estimation
36:13
expression estimation age estimation pose estimation
36:14
pose estimation
36:14
pose estimation etc so we not only support the hardware
36:17
etc so we not only support the hardware
36:17
etc so we not only support the hardware you know it's not just oh take this
36:19
you know it's not just oh take this
36:19
you know it's not just oh take this hardware and figure out how to do it
36:21
hardware and figure out how to do it
36:21
hardware and figure out how to do it we also help you build applications
36:23
we also help you build applications
36:23
we also help you build applications using
36:24
using
36:24
using pre-packaged models
36:27
pre-packaged models
36:27
pre-packaged models another one is vehicle detection and
36:29
another one is vehicle detection and
36:29
another one is vehicle detection and also license plate detector so all of
36:31
also license plate detector so all of
36:31
also license plate detector so all of these
36:32
these
36:32
these models will come pre-packaged you don't
36:34
models will come pre-packaged you don't
36:34
models will come pre-packaged you don't have to do anything you plug in the
36:35
have to do anything you plug in the
36:35
have to do anything you plug in the device
36:36
device
36:36
device and it is ready to go
36:39
and it is ready to go
36:39
and it is ready to go here's another example of pedestrian
36:41
here's another example of pedestrian
36:41
here's another example of pedestrian detection uh
36:43
detection uh
36:43
detection uh also comes with the device text
36:46
also comes with the device text
36:46
also comes with the device text detection
36:48
detection
36:48
detection now if you combine neural inference with
36:51
now if you combine neural inference with
36:51
now if you combine neural inference with depth
36:51
depth
36:51
depth it starts mimicking human vision now
36:55
it starts mimicking human vision now
36:55
it starts mimicking human vision now let's see
36:55
let's see
36:55
let's see an example of uh how it works
36:59
an example of uh how it works
36:59
an example of uh how it works in this example you're looking at neural
37:01
in this example you're looking at neural
37:01
in this example you're looking at neural inference
37:02
inference
37:02
inference plus depth perception the output you're
37:05
plus depth perception the output you're
37:05
plus depth perception the output you're looking at is a depth map
37:06
looking at is a depth map
37:06
looking at is a depth map blue means things are farther away and
37:10
blue means things are farther away and
37:10
blue means things are farther away and yellow or red means things are closer
37:13
yellow or red means things are closer
37:13
yellow or red means things are closer so you can see that we are not only
37:15
so you can see that we are not only
37:15
so you can see that we are not only saying what uh
37:17
saying what uh
37:17
saying what uh what this object it is i'm not sure
37:19
what this object it is i'm not sure
37:19
what this object it is i'm not sure whether it is very
37:20
whether it is very
37:20
whether it is very clearly visible but this object has been
37:23
clearly visible but this object has been
37:23
clearly visible but this object has been identified as a chair
37:24
identified as a chair
37:24
identified as a chair and this has been identified as a person
37:28
and this has been identified as a person
37:28
and this has been identified as a person but not only that we are also saying
37:30
but not only that we are also saying
37:30
but not only that we are also saying that this chair
37:31
that this chair
37:31
that this chair is three meters away um
37:35
is three meters away um
37:35
is three meters away um so uh so you know it's
37:38
so uh so you know it's
37:38
so uh so you know it's it's a pretty um pretty
37:42
it's a pretty um pretty
37:42
it's a pretty um pretty pretty much mimics a human vision
37:45
pretty much mimics a human vision
37:45
pretty much mimics a human vision now given this capability all
37:49
now given this capability all
37:49
now given this capability all neural inference uh tasks
37:52
neural inference uh tasks
37:52
neural inference uh tasks they get an extra dimension we are
37:54
they get an extra dimension we are
37:54
they get an extra dimension we are literally adding a new dimension
37:56
literally adding a new dimension
37:56
literally adding a new dimension to the solution space for example pose
37:59
to the solution space for example pose
37:59
to the solution space for example pose estimation pose estimation is
38:00
estimation pose estimation is
38:00
estimation pose estimation is interesting
38:01
interesting
38:01
interesting it tells you where some of the body
38:03
it tells you where some of the body
38:03
it tells you where some of the body parts are located
38:05
parts are located
38:05
parts are located but when you combine it with depth
38:07
but when you combine it with depth
38:07
but when you combine it with depth estimate we get
38:08
estimate we get
38:08
estimate we get 3d pose estimation we know you know
38:11
3d pose estimation we know you know
38:11
3d pose estimation we know you know where the eyes nose are located in 3d
38:15
where the eyes nose are located in 3d
38:15
where the eyes nose are located in 3d so it becomes a very powerful device
38:18
so it becomes a very powerful device
38:18
so it becomes a very powerful device similarly if you look at the problem of
38:21
similarly if you look at the problem of
38:21
similarly if you look at the problem of semantic
38:22
semantic
38:22
semantic segmentation uh semantic segmentation
38:25
segmentation uh semantic segmentation
38:25
segmentation uh semantic segmentation means that
38:26
means that
38:26
means that you are trying to find all pixels that
38:28
you are trying to find all pixels that
38:28
you are trying to find all pixels that belong
38:29
belong
38:29
belong to an object so here this whole
38:33
to an object so here this whole
38:33
to an object so here this whole all the orange pixels belong to the
38:35
all the orange pixels belong to the
38:35
all the orange pixels belong to the chair all the
38:36
chair all the
38:36
chair all the pink pixels belong to the sofa and all
38:38
pink pixels belong to the sofa and all
38:38
pink pixels belong to the sofa and all the green pixels
38:40
the green pixels
38:40
the green pixels belong to the television now
38:43
belong to the television now
38:43
belong to the television now the question is uh you know with depth
38:46
the question is uh you know with depth
38:46
the question is uh you know with depth information
38:47
information
38:47
information the quality of semantic segmentation
38:50
the quality of semantic segmentation
38:50
the quality of semantic segmentation goes
38:50
goes
38:50
goes high up because whenever there is
38:52
high up because whenever there is
38:52
high up because whenever there is ambiguity you can resolve it
38:54
ambiguity you can resolve it
38:54
ambiguity you can resolve it using the depth estimate you know you
38:56
using the depth estimate you know you
38:56
using the depth estimate you know you can know that the chair cannot be
38:58
can know that the chair cannot be
38:58
can know that the chair cannot be uh so long and so you can resolve
39:01
uh so long and so you can resolve
39:01
uh so long and so you can resolve ambiguities using depth and it can give
39:04
ambiguities using depth and it can give
39:04
ambiguities using depth and it can give a very high quality
39:05
a very high quality
39:05
a very high quality solution but
39:08
solution but
39:08
solution but you can be very creative also now i told
39:12
you can be very creative also now i told
39:12
you can be very creative also now i told you that
39:13
you that
39:13
you that you know there is a camera there there
39:15
you know there is a camera there there
39:15
you know there is a camera there there is an rgb camera
39:16
is an rgb camera
39:16
is an rgb camera on which we do neural inference and then
39:18
on which we do neural inference and then
39:18
on which we do neural inference and then we have the depth camera
39:19
we have the depth camera
39:20
we have the depth camera on which uh we do we estimate depth
39:23
on which uh we do we estimate depth
39:23
on which uh we do we estimate depth but we have configured oak so that you
39:25
but we have configured oak so that you
39:26
but we have configured oak so that you can actually run neural networks
39:27
can actually run neural networks
39:28
can actually run neural networks on the two stereo cameras as well so you
39:30
on the two stereo cameras as well so you
39:30
on the two stereo cameras as well so you can run
39:31
can run
39:31
can run stereo a neural a network on the left
39:34
stereo a neural a network on the left
39:34
stereo a neural a network on the left camera and the right camera
39:36
camera and the right camera
39:36
camera and the right camera at the same time in real time and then
39:40
at the same time in real time and then
39:40
at the same time in real time and then combine the information to do stereo
39:42
combine the information to do stereo
39:42
combine the information to do stereo neural influence
39:43
neural influence
39:43
neural influence for example you can detect face in both
39:46
for example you can detect face in both
39:46
for example you can detect face in both the
39:47
the
39:47
the uh in both the two frames coming from
39:50
uh in both the two frames coming from
39:50
uh in both the two frames coming from the left and the right camera
39:52
the left and the right camera
39:52
the left and the right camera and then do the triangulation to figure
39:54
and then do the triangulation to figure
39:54
and then do the triangulation to figure out where the face is what the depth of
39:56
out where the face is what the depth of
39:56
out where the face is what the depth of the face is
39:57
the face is
39:57
the face is now this is interesting because some of
39:59
now this is interesting because some of
39:59
now this is interesting because some of the some objects like very shiny objects
40:02
the some objects like very shiny objects
40:02
the some objects like very shiny objects they don't do very well with stereo
40:04
they don't do very well with stereo
40:04
they don't do very well with stereo algorithms right
40:05
algorithms right
40:05
algorithms right but you can still detect them using a
40:08
but you can still detect them using a
40:08
but you can still detect them using a neural detect you know and you're
40:09
neural detect you know and you're
40:09
neural detect you know and you're using a neural network and then you get
40:11
using a neural network and then you get
40:12
using a neural network and then you get approximate depth information it's not
40:13
approximate depth information it's not
40:14
approximate depth information it's not very accurate but it is approximate
40:15
very accurate but it is approximate
40:15
very accurate but it is approximate and that may be sufficient for your uh
40:18
and that may be sufficient for your uh
40:18
and that may be sufficient for your uh for your task
40:19
for your task
40:19
for your task so you can see in this example we are
40:22
so you can see in this example we are
40:22
so you can see in this example we are running a neural network
40:24
running a neural network
40:24
running a neural network on two frames one coming from the left
40:27
on two frames one coming from the left
40:27
on two frames one coming from the left camera and one coming from the right
40:28
camera and one coming from the right
40:28
camera and one coming from the right camera at the same time and it's at 30
40:31
camera at the same time and it's at 30
40:31
camera at the same time and it's at 30 frames a second
40:32
frames a second
40:32
frames a second uh if you look very carefully we are
40:35
uh if you look very carefully we are
40:35
uh if you look very carefully we are also detecting
40:36
also detecting
40:36
also detecting facial landmarks in addition to
40:39
facial landmarks in addition to
40:39
facial landmarks in addition to the face itself and all this is
40:42
the face itself and all this is
40:42
the face itself and all this is happening at 30 frames a second
40:44
happening at 30 frames a second
40:44
happening at 30 frames a second so it's a very powerful device
40:48
so it's a very powerful device
40:48
so it's a very powerful device now i've been focusing on you know the
40:51
now i've been focusing on you know the
40:51
now i've been focusing on you know the capabilities of
40:52
capabilities of
40:52
capabilities of the two capabilities of oak which is
40:54
the two capabilities of oak which is
40:54
the two capabilities of oak which is neural inference and depth perception
40:56
neural inference and depth perception
40:56
neural inference and depth perception but the power actually lies
41:00
but the power actually lies
41:00
but the power actually lies in all the other things which are not as
41:02
in all the other things which are not as
41:02
in all the other things which are not as cool to talk about
41:03
cool to talk about
41:03
cool to talk about but when you're building an application
41:05
but when you're building an application
41:05
but when you're building an application there are so many things that
41:07
there are so many things that
41:07
there are so many things that this device does on the device for
41:09
this device does on the device for
41:09
this device does on the device for example
41:10
example
41:10
example if you're working with 4k cameras you
41:12
if you're working with 4k cameras you
41:12
if you're working with 4k cameras you can encode the 4k
41:13
can encode the 4k
41:13
can encode the 4k video on the device itself so
41:17
video on the device itself so
41:18
video on the device itself so that's that's great if you're using a
41:20
that's that's great if you're using a
41:20
that's that's great if you're using a robotic application
41:21
robotic application
41:21
robotic application you can uh detect uh april tags and this
41:24
you can uh detect uh april tags and this
41:24
you can uh detect uh april tags and this is being done on the device your host
41:26
is being done on the device your host
41:26
is being done on the device your host like the raspberry pi is never being
41:28
like the raspberry pi is never being
41:28
like the raspberry pi is never being used at all
41:29
used at all
41:29
used at all and similarly you can do feature
41:31
and similarly you can do feature
41:31
and similarly you can do feature tracking and you can warp
41:33
tracking and you can warp
41:33
tracking and you can warp or de-warp images that are coming out of
41:36
or de-warp images that are coming out of
41:36
or de-warp images that are coming out of uh the
41:36
uh the
41:36
uh the the uh camera and
41:40
the uh camera and
41:40
the uh camera and so the capabilities of oak are far
41:42
so the capabilities of oak are far
41:42
so the capabilities of oak are far beyond
41:43
beyond
41:43
beyond what most smart cameras are uh able to
41:46
what most smart cameras are uh able to
41:46
what most smart cameras are uh able to do
41:47
do
41:47
do and this is possible because of
41:50
and this is possible because of
41:50
and this is possible because of the architecture we have chosen so let's
41:53
the architecture we have chosen so let's
41:53
the architecture we have chosen so let's go over
41:53
go over
41:53
go over uh you know what the architecture looks
41:55
uh you know what the architecture looks
41:55
uh you know what the architecture looks like
41:57
like
41:57
like inside oak the brain of oak is
42:00
inside oak the brain of oak is
42:00
inside oak the brain of oak is intel's media x chip and this
42:04
intel's media x chip and this
42:04
intel's media x chip and this chip is so powerful it can do 4 trillion
42:07
chip is so powerful it can do 4 trillion
42:07
chip is so powerful it can do 4 trillion operations per second and it has
42:10
operations per second and it has
42:10
operations per second and it has built-in capability
42:11
built-in capability
42:12
built-in capability for accelerating stereo as well as
42:16
for accelerating stereo as well as
42:16
for accelerating stereo as well as uh you know warping deworping april tags
42:18
uh you know warping deworping april tags
42:18
uh you know warping deworping april tags etc we have been able to implement all
42:20
etc we have been able to implement all
42:20
etc we have been able to implement all those things
42:21
those things
42:21
those things on this device so that things are
42:24
on this device so that things are
42:24
on this device so that things are hardware accelerated so
42:27
hardware accelerated so
42:27
hardware accelerated so um i'm sometimes surprised you know it
42:30
um i'm sometimes surprised you know it
42:30
um i'm sometimes surprised you know it almost feels like
42:31
almost feels like
42:31
almost feels like we are living in the future uh this is
42:33
we are living in the future uh this is
42:33
we are living in the future uh this is you know if you look at this device
42:35
you know if you look at this device
42:35
you know if you look at this device it is the site this is the tip of
42:38
it is the site this is the tip of
42:38
it is the site this is the tip of uh you know a pencil and this is such a
42:41
uh you know a pencil and this is such a
42:41
uh you know a pencil and this is such a tiny
42:42
tiny
42:42
tiny powerful device our cameras
42:45
powerful device our cameras
42:45
powerful device our cameras these are 12 megapixel cameras uh on
42:48
these are 12 megapixel cameras uh on
42:48
these are 12 megapixel cameras uh on both
42:48
both
42:48
both uh versions of oak and they can run at a
42:51
uh versions of oak and they can run at a
42:51
uh versions of oak and they can run at a maximum speed of 60 frames
42:53
maximum speed of 60 frames
42:53
maximum speed of 60 frames a second we also have
42:56
a second we also have
42:56
a second we also have a stereo camera which is a slightly
42:59
a stereo camera which is a slightly
42:59
a stereo camera which is a slightly lower resolution
42:59
lower resolution
43:00
lower resolution at 1280 by 1 800
43:03
at 1280 by 1 800
43:03
at 1280 by 1 800 but this can run at 120 frames a second
43:07
but this can run at 120 frames a second
43:07
but this can run at 120 frames a second and just to give you an idea you know we
43:09
and just to give you an idea you know we
43:09
and just to give you an idea you know we have been focusing so much
43:10
have been focusing so much
43:10
have been focusing so much on the speed how easily
43:14
on the speed how easily
43:14
on the speed how easily developers can start using this
43:17
developers can start using this
43:17
developers can start using this that we created this video how how easy
43:20
that we created this video how how easy
43:20
that we created this video how how easy it is to get started so you
43:22
it is to get started so you
43:22
it is to get started so you take the device
43:27
take the device
43:27
take the device so
43:30
so i'm not sure whether you're able to
43:31
so i'm not sure whether you're able to
43:32
so i'm not sure whether you're able to hear the sound uh probably not
43:35
hear the sound uh probably not
43:35
hear the sound uh probably not so okay so what's happening is that we
43:37
so okay so what's happening is that we
43:38
so okay so what's happening is that we just plugged in
43:38
just plugged in
43:38
just plugged in uh the device we copy-pasted some
43:42
uh the device we copy-pasted some
43:42
uh the device we copy-pasted some code and here you go within 30 seconds
43:47
code and here you go within 30 seconds
43:47
code and here you go within 30 seconds you have a demo app running so you plug
43:50
you have a demo app running so you plug
43:50
you have a demo app running so you plug this device onto your computer or a
43:52
this device onto your computer or a
43:52
this device onto your computer or a raspberry pi
43:53
raspberry pi
43:53
raspberry pi and within 30 seconds you have a demo up
43:56
and within 30 seconds you have a demo up
43:56
and within 30 seconds you have a demo up and running
43:56
and running
43:56
and running so that was our goal you know you
43:58
so that was our goal you know you
43:58
so that was our goal you know you shouldn't be spending half a day
44:00
shouldn't be spending half a day
44:00
shouldn't be spending half a day when you get a new hardware that's what
44:01
when you get a new hardware that's what
44:01
when you get a new hardware that's what happens these days
44:03
happens these days
44:03
happens these days you know they're not very user-friendly
44:07
you know they're not very user-friendly
44:07
you know they're not very user-friendly the other thing is you can obviously
44:09
the other thing is you can obviously
44:09
the other thing is you can obviously program
44:10
program
44:10
program it the way you like using python but we
44:13
it the way you like using python but we
44:13
it the way you like using python but we also built
44:14
also built
44:14
also built a visual compute pipeline so that
44:17
a visual compute pipeline so that
44:17
a visual compute pipeline so that if you want to build a solution without
44:20
if you want to build a solution without
44:20
if you want to build a solution without needing to code
44:21
needing to code
44:21
needing to code you can build a visual pipeline using
44:23
you can build a visual pipeline using
44:23
you can build a visual pipeline using different uh
44:25
different uh
44:25
different uh models so suppose you want uh the neural
44:28
models so suppose you want uh the neural
44:28
models so suppose you want uh the neural network to first
44:29
network to first
44:29
network to first do phase detection you can run the phase
44:31
do phase detection you can run the phase
44:31
do phase detection you can run the phase detection first and then you can
44:32
detection first and then you can
44:32
detection first and then you can configure this
44:33
configure this
44:33
configure this visual pipeline builder so that the next
44:36
visual pipeline builder so that the next
44:36
visual pipeline builder so that the next step is
44:37
step is
44:37
step is facial landmark detector so it loads the
44:39
facial landmark detector so it loads the
44:39
facial landmark detector so it loads the facial landmark detector and does
44:41
facial landmark detector and does
44:41
facial landmark detector and does processing and with very little almost
44:45
processing and with very little almost
44:45
processing and with very little almost no coding
44:46
no coding
44:46
no coding you will be able to use this device
44:51
you will be able to use this device
44:51
you will be able to use this device and uh like i said that we are not just
44:54
and uh like i said that we are not just
44:54
and uh like i said that we are not just selling
44:54
selling
44:54
selling a hardware we are nurturing a community
44:58
a hardware we are nurturing a community
44:58
a hardware we are nurturing a community and we want to make sure that everybody
45:01
and we want to make sure that everybody
45:01
and we want to make sure that everybody who is a beginner
45:02
who is a beginner
45:02
who is a beginner in ai they are taken care of we just
45:05
in ai they are taken care of we just
45:05
in ai they are taken care of we just don't sell you the hardware and
45:07
don't sell you the hardware and
45:07
don't sell you the hardware and just go away so for that we are giving
45:10
just go away so for that we are giving
45:10
just go away so for that we are giving away free courses
45:11
away free courses
45:11
away free courses which will help you get started even if
45:13
which will help you get started even if
45:13
which will help you get started even if you want to build
45:14
you want to build
45:14
you want to build your own uh you know let's say
45:16
your own uh you know let's say
45:16
your own uh you know let's say classifier our course will teach you how
45:18
classifier our course will teach you how
45:18
classifier our course will teach you how to train your model
45:19
to train your model
45:20
to train your model and then uh optimize it and get it on
45:22
and then uh optimize it and get it on
45:22
and then uh optimize it and get it on the device
45:24
the device
45:24
the device in addition to all the 20 you know we
45:26
in addition to all the 20 you know we
45:26
in addition to all the 20 you know we also provide 20 free models
45:28
also provide 20 free models
45:28
also provide 20 free models that you can use uh with the device we
45:31
that you can use uh with the device we
45:31
that you can use uh with the device we will also have
45:32
will also have
45:32
will also have a community forum we have a large
45:33
a community forum we have a large
45:34
a community forum we have a large community now about 6500 developers have
45:37
community now about 6500 developers have
45:37
community now about 6500 developers have signed up for
45:37
signed up for
45:37
signed up for oak so we will have a forum where our
45:41
oak so we will have a forum where our
45:41
oak so we will have a forum where our developers our developers will support
45:42
developers our developers will support
45:42
developers our developers will support the community but you can also
45:44
the community but you can also
45:44
the community but you can also do other community stuff right you can
45:47
do other community stuff right you can
45:47
do other community stuff right you can ask questions to other community members
45:50
ask questions to other community members
45:50
ask questions to other community members network etc and finally we are creating
45:53
network etc and finally we are creating
45:53
network etc and finally we are creating a model marketplace where we will put
45:56
a model marketplace where we will put
45:56
a model marketplace where we will put the 20 models which will be free for the
45:58
the 20 models which will be free for the
45:58
the 20 models which will be free for the community but
46:00
community but
46:00
community but it will also be for the community to
46:02
it will also be for the community to
46:02
it will also be for the community to other people in the community can also
46:04
other people in the community can also
46:04
other people in the community can also upload
46:05
upload
46:05
upload their own models so that other people
46:08
their own models so that other people
46:08
their own models so that other people can try it out
46:09
can try it out
46:09
can try it out and you don't have to download any code
46:12
and you don't have to download any code
46:12
and you don't have to download any code to get started
46:13
to get started
46:13
to get started with this model marketplace because we
46:16
with this model marketplace because we
46:16
with this model marketplace because we will wrap all the models
46:18
will wrap all the models
46:18
will wrap all the models uh as a web api so you go to a web page
46:21
uh as a web api so you go to a web page
46:21
uh as a web api so you go to a web page you upload a picture
46:22
you upload a picture
46:22
you upload a picture and you can see what the model does
46:24
and you can see what the model does
46:24
and you can see what the model does without having to download
46:25
without having to download
46:25
without having to download any code this is like a service to
46:29
any code this is like a service to
46:29
any code this is like a service to to the community members so that they
46:30
to the community members so that they
46:30
to the community members so that they can try out different models and see
46:32
can try out different models and see
46:32
can try out different models and see what works for them
46:33
what works for them
46:33
what works for them and then they can obviously download it
46:34
and then they can obviously download it
46:34
and then they can obviously download it to their device and
46:36
to their device and
46:36
to their device and um work with them so
46:39
um work with them so
46:39
um work with them so uh the final thing is that oak is being
46:42
uh the final thing is that oak is being
46:42
uh the final thing is that oak is being used by
46:43
used by
46:43
used by a lot of companies unfortunately we
46:45
a lot of companies unfortunately we
46:45
a lot of companies unfortunately we don't have permission to
46:46
don't have permission to
46:46
don't have permission to use their names so i'm just uh telling
46:49
use their names so i'm just uh telling
46:49
use their names so i'm just uh telling what they do
46:50
what they do
46:50
what they do we have been certified uh you know fcc
46:53
we have been certified uh you know fcc
46:53
we have been certified uh you know fcc and ce certification we have received so
46:55
and ce certification we have received so
46:55
and ce certification we have received so it is
46:56
it is
46:56
it is industrial good to go solution
47:00
industrial good to go solution
47:00
industrial good to go solution there are companies who are using us for
47:01
there are companies who are using us for
47:02
there are companies who are using us for visual assistance they are creating
47:03
visual assistance they are creating
47:03
visual assistance they are creating smart glasses
47:05
smart glasses
47:05
smart glasses for uh the visually impaired people
47:08
for uh the visually impaired people
47:08
for uh the visually impaired people it's being used in aerial drones and one
47:11
it's being used in aerial drones and one
47:11
it's being used in aerial drones and one company is using it
47:12
company is using it
47:12
company is using it it in sub sea drones uh as well so these
47:16
it in sub sea drones uh as well so these
47:16
it in sub sea drones uh as well so these drones go inside
47:17
drones go inside
47:17
drones go inside uh the ocean and then there are
47:19
uh the ocean and then there are
47:19
uh the ocean and then there are e-scooter companies
47:20
e-scooter companies
47:20
e-scooter companies cargo handling companies uh sports
47:23
cargo handling companies uh sports
47:23
cargo handling companies uh sports monitoring
47:24
monitoring
47:24
monitoring smart agriculture safety uh and security
47:27
smart agriculture safety uh and security
47:27
smart agriculture safety uh and security one company is using it for
47:29
one company is using it for
47:29
one company is using it for gun detection in schools uh thank god
47:32
gun detection in schools uh thank god
47:32
gun detection in schools uh thank god and then one satellite company is using
47:35
and then one satellite company is using
47:35
and then one satellite company is using it as
47:35
it as
47:35
it as one of its sensor as well so
47:39
one of its sensor as well so
47:39
one of its sensor as well so the community is growing we want new
47:41
the community is growing we want new
47:41
the community is growing we want new developers but i'm also
47:42
developers but i'm also
47:42
developers but i'm also so glad to see that uh companies which
47:44
so glad to see that uh companies which
47:44
so glad to see that uh companies which are building industrial products
47:46
are building industrial products
47:46
are building industrial products are also looking at oak as the solution
47:51
are also looking at oak as the solution
47:51
are also looking at oak as the solution so uh the
47:54
so uh the
47:54
so uh the you know the even though the kickstarter
47:56
you know the even though the kickstarter
47:56
you know the even though the kickstarter campaign is
47:57
campaign is
47:57
campaign is over you can actually pre-order uh i
48:00
over you can actually pre-order uh i
48:00
over you can actually pre-order uh i have given
48:01
have given
48:01
have given the link here but you can also search
48:03
the link here but you can also search
48:03
the link here but you can also search kickstarter opencv ai kit
48:05
kickstarter opencv ai kit
48:05
kickstarter opencv ai kit and you will land on this page and there
48:07
and you will land on this page and there
48:07
and you will land on this page and there is a pre-order button that you can use
48:10
is a pre-order button that you can use
48:10
is a pre-order button that you can use uh the you won't receive uh the
48:13
uh the you won't receive uh the
48:13
uh the you won't receive uh the the the cameras instantly there is a
48:16
the the cameras instantly there is a
48:16
the the cameras instantly there is a wait time
48:17
wait time
48:17
wait time and because of that we are also giving
48:20
and because of that we are also giving
48:20
and because of that we are also giving it out for a heavily discounted price
48:22
it out for a heavily discounted price
48:22
it out for a heavily discounted price these these prices prices are not the
48:24
these these prices prices are not the
48:24
these these prices prices are not the same as kickstarter prices
48:26
same as kickstarter prices
48:26
same as kickstarter prices but they are not uh anywhere close to
48:29
but they are not uh anywhere close to
48:29
but they are not uh anywhere close to the retail price
48:31
the retail price
48:31
the retail price so that that's about it uh
48:35
so that that's about it uh
48:35
so that that's about it uh i mean i was seven minutes late to the
48:36
i mean i was seven minutes late to the
48:36
i mean i was seven minutes late to the meeting but i think i just finished two
48:38
meeting but i think i just finished two
48:38
meeting but i think i just finished two minutes late
48:41
minutes late
48:41
minutes late um yeah and when we had a chat at the
48:43
um yeah and when we had a chat at the
48:43
um yeah and when we had a chat at the beginning so
48:44
beginning so
48:44
beginning so it's it's awesome i have to say
48:48
it's it's awesome i have to say
48:48
it's it's awesome i have to say i'm impressed thank you by the device
48:50
i'm impressed thank you by the device
48:50
i'm impressed thank you by the device that uh that you're working on and
48:52
that uh that you're working on and
48:52
that uh that you're working on and um we actually got a ton of questions
48:56
um we actually got a ton of questions
48:56
um we actually got a ton of questions from the community so
48:57
from the community so
48:57
from the community so oh okay i'm happy to answer
49:00
oh okay i'm happy to answer
49:00
oh okay i'm happy to answer yeah so i will group together a few so
49:03
yeah so i will group together a few so
49:03
yeah so i will group together a few so we have a topic to talk about so
49:05
we have a topic to talk about so
49:05
we have a topic to talk about so um uh actually uh stefano is also uh
49:09
um uh actually uh stefano is also uh
49:09
um uh actually uh stefano is also uh joining us i believe for uh for for a
49:11
joining us i believe for uh for for a
49:11
joining us i believe for uh for for a bit of a discussion about your session
49:13
bit of a discussion about your session
49:13
bit of a discussion about your session um so let me let me get started with the
49:15
um so let me let me get started with the
49:15
um so let me let me get started with the first question um
49:17
first question um
49:17
first question um a lot of people are wondering um how do
49:20
a lot of people are wondering um how do
49:20
a lot of people are wondering um how do i actually get
49:21
i actually get
49:21
i actually get a custom model loaded on the device do i
49:24
a custom model loaded on the device do i
49:24
a custom model loaded on the device do i have to use
49:25
have to use
49:25
have to use it ssh or how does that work
49:29
it ssh or how does that work
49:29
it ssh or how does that work uh so we have scripts uh which will just
49:32
uh so we have scripts uh which will just
49:32
uh so we have scripts uh which will just take the model from your drive from your
49:33
take the model from your drive from your
49:33
take the model from your drive from your laptop the host and put it on the
49:36
laptop the host and put it on the
49:36
laptop the host and put it on the on the device uh and we also have the
49:38
on the device uh and we also have the
49:38
on the device uh and we also have the pipeline builder which you can use to
49:41
pipeline builder which you can use to
49:41
pipeline builder which you can use to specify you know this is the location of
49:43
specify you know this is the location of
49:43
specify you know this is the location of the model uh loaded on the device
49:45
the model uh loaded on the device
49:45
the model uh loaded on the device and so you can load multiple uh multiple
49:48
and so you can load multiple uh multiple
49:48
and so you can load multiple uh multiple models on the device
49:49
models on the device
49:49
models on the device as long as the memory permits so that uh
49:53
as long as the memory permits so that uh
49:53
as long as the memory permits so that uh you know you you get the uh you can
49:56
you know you you get the uh you can
49:56
you know you you get the uh you can chain them run this model and then run
49:58
chain them run this model and then run
49:58
chain them run this model and then run this model as long as the memory permits
50:01
this model as long as the memory permits
50:01
this model as long as the memory permits we can do that it sounds really awesome
50:04
we can do that it sounds really awesome
50:04
we can do that it sounds really awesome and also the visual pipeline i guess you
50:06
and also the visual pipeline i guess you
50:06
and also the visual pipeline i guess you don't need any experience in development
50:08
don't need any experience in development
50:08
don't need any experience in development to get going with this device yeah as
50:11
to get going with this device yeah as
50:11
to get going with this device yeah as long as you're
50:11
long as you're
50:12
long as you're you know you're familiar with computers
50:13
you know you're familiar with computers
50:13
you know you're familiar with computers and can run a little bit uh
50:15
and can run a little bit uh
50:15
and can run a little bit uh some scripts etc uh you don't really
50:18
some scripts etc uh you don't really
50:18
some scripts etc uh you don't really need ai
50:18
need ai
50:18
need ai experience oh that's really cool
50:22
experience oh that's really cool
50:22
experience oh that's really cool that that's great stuff and it's good to
50:23
that that's great stuff and it's good to
50:23
that that's great stuff and it's good to know for people
50:25
know for people
50:25
know for people the other question that we have from
50:27
the other question that we have from
50:28
the other question that we have from people in the audience is
50:29
people in the audience is
50:29
people in the audience is so you mentioned uh several standard
50:31
so you mentioned uh several standard
50:31
so you mentioned uh several standard made models one of them is the license
50:33
made models one of them is the license
50:33
made models one of them is the license plate detection
50:35
plate detection
50:35
plate detection does it actually recognize what's on the
50:37
does it actually recognize what's on the
50:37
does it actually recognize what's on the license page so the numbers and
50:39
license page so the numbers and
50:39
license page so the numbers and and uh in character yes yeah so the
50:42
and uh in character yes yeah so the
50:42
and uh in character yes yeah so the the model that will we will release uh
50:45
the model that will we will release uh
50:45
the model that will we will release uh and before the end of the year
50:47
and before the end of the year
50:47
and before the end of the year it will detect not only detect the
50:49
it will detect not only detect the
50:49
it will detect not only detect the license plate it will also detect what's
50:51
license plate it will also detect what's
50:51
license plate it will also detect what's inside the license plate
50:52
inside the license plate
50:52
inside the license plate and then how many countries does that
50:54
and then how many countries does that
50:54
and then how many countries does that work actually because
50:58
yeah so it will be based on the open
50:59
yeah so it will be based on the open
50:59
yeah so it will be based on the open vino model that's already out there
51:02
vino model that's already out there
51:02
vino model that's already out there uh so i don't know exactly uh off the
51:04
uh so i don't know exactly uh off the
51:04
uh so i don't know exactly uh off the top of my head how many countries it
51:06
top of my head how many countries it
51:06
top of my head how many countries it works on
51:08
works on
51:08
works on that's actually pretty cool that that
51:09
that's actually pretty cool that that
51:09
that's actually pretty cool that that youtube open source models
51:12
youtube open source models
51:12
youtube open source models uh instead of rolling your own so if
51:15
uh instead of rolling your own so if
51:15
uh instead of rolling your own so if people are interested in what a model
51:16
people are interested in what a model
51:16
people are interested in what a model can do
51:17
can do
51:17
can do just go to the repository of the
51:18
just go to the repository of the
51:18
just go to the repository of the original model and it will actually show
51:21
original model and it will actually show
51:21
original model and it will actually show you
51:21
you
51:22
you um what what's what's capable of what
51:25
um what what's what's capable of what
51:25
um what what's what's capable of what that model is capable of i guess
51:27
that model is capable of i guess
51:27
that model is capable of i guess so my question would be
51:30
so my question would be
51:30
so my question would be what sort of models do you mean again
51:34
what sort of models do you mean again
51:34
what sort of models do you mean again sorry
51:35
sorry
51:35
sorry what sort of models can you run does it
51:36
what sort of models can you run does it
51:36
what sort of models can you run does it run for example onyx models or
51:38
run for example onyx models or
51:38
run for example onyx models or stuff like that uh so if you have an
51:41
stuff like that uh so if you have an
51:41
stuff like that uh so if you have an onyx model
51:41
onyx model
51:42
onyx model you can convert it and it will run on
51:44
you can convert it and it will run on
51:44
you can convert it and it will run on the device any open
51:45
the device any open
51:45
the device any open or any uh openvino model
51:48
or any uh openvino model
51:48
or any uh openvino model if you go to open vino model zoo any of
51:51
if you go to open vino model zoo any of
51:51
if you go to open vino model zoo any of those
51:52
those
51:52
those will work and any onyx model can be
51:54
will work and any onyx model can be
51:54
will work and any onyx model can be converted
51:55
converted
51:55
converted but you know sometimes they come up with
51:57
but you know sometimes they come up with
51:57
but you know sometimes they come up with some custom layers with other
51:59
some custom layers with other
51:59
some custom layers with other uh platforms which are new then it will
52:01
uh platforms which are new then it will
52:01
uh platforms which are new then it will take some time
52:02
take some time
52:02
take some time but as long as it is an onyx format uh
52:06
but as long as it is an onyx format uh
52:06
but as long as it is an onyx format uh it can be done that's really awesome
52:09
it can be done that's really awesome
52:09
it can be done that's really awesome so alicia i guess i guess you're into
52:12
so alicia i guess i guess you're into
52:12
so alicia i guess i guess you're into this whole
52:13
this whole
52:13
this whole vr thing what do you think of this kind
52:16
vr thing what do you think of this kind
52:16
vr thing what do you think of this kind of device um well it's
52:19
of device um well it's
52:19
of device um well it's it's kind of interesting uh to you
52:21
it's kind of interesting uh to you
52:21
it's kind of interesting uh to you talked about the size but what about
52:23
talked about the size but what about
52:23
talked about the size but what about the weight and um these have to be
52:26
the weight and um these have to be
52:26
the weight and um these have to be pretty small devices
52:29
pretty small devices
52:29
pretty small devices and i see all the use use cases
52:33
and i see all the use use cases
52:33
and i see all the use use cases for for the drone usage and i know
52:36
for for the drone usage and i know
52:36
for for the drone usage and i know that um so i've done some knee arm
52:38
that um so i've done some knee arm
52:38
that um so i've done some knee arm programming and i know you need the
52:40
programming and i know you need the
52:40
programming and i know you need the stereo you need
52:41
stereo you need
52:41
stereo you need two cameras for depth perception and for
52:44
two cameras for depth perception and for
52:44
two cameras for depth perception and for calculation
52:44
calculation
52:44
calculation when you're you're controlling robotics
52:47
when you're you're controlling robotics
52:47
when you're you're controlling robotics so um
52:48
so um
52:48
so um it's really exciting to see
52:51
it's really exciting to see
52:51
it's really exciting to see um now what's possible with respect to
52:55
um now what's possible with respect to
52:55
um now what's possible with respect to controlling uh robotics in addition to
52:57
controlling uh robotics in addition to
52:58
controlling uh robotics in addition to controlling the drones
52:59
controlling the drones
52:59
controlling the drones with these cameras so yeah the
53:03
with these cameras so yeah the
53:03
with these cameras so yeah the possibilities are so big
53:06
possibilities are so big
53:06
possibilities are so big i mean some of the applications we were
53:08
i mean some of the applications we were
53:08
i mean some of the applications we were not even thinking about
53:09
not even thinking about
53:09
not even thinking about one of them is cargo uh you know inside
53:13
one of them is cargo uh you know inside
53:13
one of them is cargo uh you know inside uh
53:14
uh
53:14
uh when you're packing trucks uh with cargo
53:17
when you're packing trucks uh with cargo
53:17
when you're packing trucks uh with cargo they want to make sure that
53:18
they want to make sure that
53:18
they want to make sure that uh there is nothing inside sometimes
53:20
uh there is nothing inside sometimes
53:20
uh there is nothing inside sometimes they want to know that state
53:22
they want to know that state
53:22
they want to know that state that oh it is completely empty which is
53:24
that oh it is completely empty which is
53:24
that oh it is completely empty which is very difficult to do
53:26
very difficult to do
53:26
very difficult to do with depth alone right but if you add
53:28
with depth alone right but if you add
53:28
with depth alone right but if you add neural compute
53:30
neural compute
53:30
neural compute then it's a trivial problem to uh so
53:33
then it's a trivial problem to uh so
53:33
then it's a trivial problem to uh so both those things you know really people
53:35
both those things you know really people
53:35
both those things you know really people don't have both those solutions together
53:37
don't have both those solutions together
53:37
don't have both those solutions together right now they have either
53:38
right now they have either
53:38
right now they have either a depth solution and it works great
53:41
a depth solution and it works great
53:41
a depth solution and it works great whenever depth is
53:42
whenever depth is
53:42
whenever depth is sufficient or they have a neural uh
53:45
sufficient or they have a neural uh
53:45
sufficient or they have a neural uh inference solution which is also great
53:47
inference solution which is also great
53:47
inference solution which is also great but
53:48
but
53:48
but really there is no solution which
53:49
really there is no solution which
53:49
really there is no solution which combines both of them together
53:51
combines both of them together
53:51
combines both of them together in one nice package and it's nice indeed
53:55
in one nice package and it's nice indeed
53:55
in one nice package and it's nice indeed it's very small also
53:57
it's very small also
53:57
it's very small also yeah um so small and the weight is not
54:00
yeah um so small and the weight is not
54:00
yeah um so small and the weight is not very much i unfortunately i don't
54:01
very much i unfortunately i don't
54:02
very much i unfortunately i don't remember it off the top of my head
54:04
remember it off the top of my head
54:04
remember it off the top of my head uh but it's uh really lightweight i mean
54:07
uh but it's uh really lightweight i mean
54:07
uh but it's uh really lightweight i mean almost like a raspberry pi
54:08
almost like a raspberry pi
54:08
almost like a raspberry pi yeah that that would be the weight uh if
54:10
yeah that that would be the weight uh if
54:10
yeah that that would be the weight uh if you hold a raspberry pi this is about
54:12
you hold a raspberry pi this is about
54:12
you hold a raspberry pi this is about the same
54:13
the same
54:13
the same and we also came up with uh an aluminium
54:16
and we also came up with uh an aluminium
54:16
and we also came up with uh an aluminium package
54:17
package
54:17
package um the new version you know as part of
54:19
um the new version you know as part of
54:19
um the new version you know as part of the kickstarter campaign we
54:21
the kickstarter campaign we
54:21
the kickstarter campaign we promised that we will produce a
54:24
promised that we will produce a
54:24
promised that we will produce a packaging around it
54:25
packaging around it
54:25
packaging around it and the aluminium actually serves as the
54:27
and the aluminium actually serves as the
54:28
and the aluminium actually serves as the heat sink
54:28
heat sink
54:28
heat sink as well so it further reduces uh the
54:31
as well so it further reduces uh the
54:31
as well so it further reduces uh the size
54:32
size
54:32
size uh because the heat sink was the big
54:35
uh because the heat sink was the big
54:35
uh because the heat sink was the big part
54:35
part
54:35
part of the solution
54:39
of the solution
54:39
of the solution you know next thing you know we're going
54:40
you know next thing you know we're going
54:40
you know next thing you know we're going to be able to teach our drones how to
54:42
to be able to teach our drones how to
54:42
to be able to teach our drones how to play catch right
54:43
play catch right
54:43
play catch right because now now we have the stereo
54:45
because now now we have the stereo
54:45
because now now we have the stereo vision and
54:47
vision and
54:47
vision and they can do those things yeah it's
54:49
they can do those things yeah it's
54:49
they can do those things yeah it's really interesting
54:51
really interesting
54:51
really interesting it's it's surprising how small these
54:52
it's it's surprising how small these
54:52
it's it's surprising how small these devices uh are becoming
54:54
devices uh are becoming
54:54
devices uh are becoming so stefano is one of the people that's
54:56
so stefano is one of the people that's
54:56
so stefano is one of the people that's actually been using the hololens and
54:58
actually been using the hololens and
54:58
actually been using the hololens and uh maybe um um i can ask you a question
55:01
uh maybe um um i can ask you a question
55:01
uh maybe um um i can ask you a question stefano
55:02
stefano
55:02
stefano um what is actually the importance of
55:05
um what is actually the importance of
55:05
um what is actually the importance of these hardware devices that
55:06
these hardware devices that
55:06
these hardware devices that we've been talking about like the oak
55:08
we've been talking about like the oak
55:08
we've been talking about like the oak device did the specialized chips
55:12
yes hi everybody and well first of all
55:15
yes hi everybody and well first of all
55:15
yes hi everybody and well first of all thanks for
55:16
thanks for
55:16
thanks for hosting me it's a great pleasure to be
55:18
hosting me it's a great pleasure to be
55:18
hosting me it's a great pleasure to be here and to hear from society about this
55:20
here and to hear from society about this
55:20
here and to hear from society about this amazing devices i'm very fond of
55:24
amazing devices i'm very fond of
55:24
amazing devices i'm very fond of combining
55:25
combining
55:25
combining hardware and the power of the cloud and
55:28
hardware and the power of the cloud and
55:28
hardware and the power of the cloud and edge
55:29
edge
55:29
edge for artificial intelligence so you asked
55:32
for artificial intelligence so you asked
55:32
for artificial intelligence so you asked about hololens and i must say
55:34
about hololens and i must say
55:34
about hololens and i must say despite being a big microsoft fan
55:37
despite being a big microsoft fan
55:37
despite being a big microsoft fan uh hololens is yes a great device but it
55:40
uh hololens is yes a great device but it
55:40
uh hololens is yes a great device but it comes without weight
55:42
comes without weight
55:42
comes without weight it's not that lightweight so you need to
55:45
it's not that lightweight so you need to
55:45
it's not that lightweight so you need to wear it for a long time
55:48
wear it for a long time
55:48
wear it for a long time not very comfortable doesn't have a very
55:51
not very comfortable doesn't have a very
55:51
not very comfortable doesn't have a very long
55:51
long
55:52
long battery autonomy i will show you
55:53
battery autonomy i will show you
55:54
battery autonomy i will show you something during my session
55:56
something during my session
55:56
something during my session it is a great device it's not just a
55:58
it is a great device it's not just a
55:58
it is a great device it's not just a camera obviously you can do a lot of
56:00
camera obviously you can do a lot of
56:00
camera obviously you can do a lot of things
56:00
things
56:00
things on augmented reality i really welcome
56:03
on augmented reality i really welcome
56:03
on augmented reality i really welcome uh the development in this field now
56:06
uh the development in this field now
56:06
uh the development in this field now these are small cameras that can be
56:09
these are small cameras that can be
56:09
these are small cameras that can be that you can wear and they can ideally
56:12
that you can wear and they can ideally
56:12
that you can wear and they can ideally have a long
56:14
have a long
56:14
have a long battery duration that you can last a
56:16
battery duration that you can last a
56:16
battery duration that you can last a full day
56:18
full day
56:18
full day yeah so so you've seen a lot of
56:20
yeah so so you've seen a lot of
56:20
yeah so so you've seen a lot of developments the last
56:21
developments the last
56:22
developments the last few years in that department we've seen
56:24
few years in that department we've seen
56:24
few years in that department we've seen neural networks on a chip
56:25
neural networks on a chip
56:25
neural networks on a chip actually like i guess okay is sort of
56:27
actually like i guess okay is sort of
56:27
actually like i guess okay is sort of similar to that right
56:29
similar to that right
56:29
similar to that right yeah yeah so it uses mediadx
56:32
yeah yeah so it uses mediadx
56:32
yeah yeah so it uses mediadx mediatex's intel's uh you know visu
56:35
mediatex's intel's uh you know visu
56:35
mediatex's intel's uh you know visu vision processing unit
56:37
vision processing unit
56:37
vision processing unit and it runs neural networks but it also
56:39
and it runs neural networks but it also
56:39
and it runs neural networks but it also does does these other things which other
56:41
does does these other things which other
56:41
does does these other things which other companies
56:42
companies
56:42
companies uh don't utilize uh it for example
56:45
uh don't utilize uh it for example
56:46
uh don't utilize uh it for example uh has stereo capability which it can
56:49
uh has stereo capability which it can
56:49
uh has stereo capability which it can run
56:49
run
56:49
run on the device and things like that which
56:52
on the device and things like that which
56:52
on the device and things like that which we utilized but normally people don't
56:54
we utilized but normally people don't
56:54
we utilized but normally people don't so uh yes it runs neural network on
56:57
so uh yes it runs neural network on
56:58
so uh yes it runs neural network on this tiny chip which is almost the size
57:00
this tiny chip which is almost the size
57:00
this tiny chip which is almost the size of uh
57:01
of uh
57:01
of uh the tip of a pencil yeah that that's
57:04
the tip of a pencil yeah that that's
57:04
the tip of a pencil yeah that that's amazing stuff
57:05
amazing stuff
57:05
amazing stuff so we've got a number of people actually
57:07
so we've got a number of people actually
57:07
so we've got a number of people actually coming up with solutions to
57:09
coming up with solutions to
57:09
coming up with solutions to to build using this opencv camera so i
57:12
to build using this opencv camera so i
57:12
to build using this opencv camera so i guess we will see a lot more
57:13
guess we will see a lot more
57:13
guess we will see a lot more about that is soon in december around
57:16
about that is soon in december around
57:16
about that is soon in december around christmas uh i guess
57:17
christmas uh i guess
57:17
christmas uh i guess um so just to give you a few ideas and
57:20
um so just to give you a few ideas and
57:20
um so just to give you a few ideas and maybe we can have a discussion about
57:22
maybe we can have a discussion about
57:22
maybe we can have a discussion about them
57:22
them
57:22
them there's someone asking um could i use
57:26
there's someone asking um could i use
57:26
there's someone asking um could i use this device
57:27
this device
57:27
this device not only for emotion detection like
57:29
not only for emotion detection like
57:29
not only for emotion detection like smiling
57:30
smiling
57:30
smiling disgust surprise but also for
57:32
disgust surprise but also for
57:32
disgust surprise but also for interaction between two people
57:34
interaction between two people
57:34
interaction between two people hugging uh affectionate responses
57:37
hugging uh affectionate responses
57:37
hugging uh affectionate responses um those kind of things making eye
57:39
um those kind of things making eye
57:39
um those kind of things making eye contact
57:40
contact
57:40
contact what do you think uh such a yeah so that
57:43
what do you think uh such a yeah so that
57:43
what do you think uh such a yeah so that that's a very good question because
57:45
that's a very good question because
57:45
that's a very good question because it is uh exactly that kind of thing and
57:48
it is uh exactly that kind of thing and
57:48
it is uh exactly that kind of thing and hugging i mean it's almost opposite of
57:50
hugging i mean it's almost opposite of
57:50
hugging i mean it's almost opposite of what we want these days
57:51
what we want these days
57:51
what we want these days uh so it's a very good application for
57:53
uh so it's a very good application for
57:53
uh so it's a very good application for social distancing
57:54
social distancing
57:54
social distancing uh so when we did a competition for uh
57:58
uh so when we did a competition for uh
57:58
uh so when we did a competition for uh uh you know spatial ai competition was
58:00
uh you know spatial ai competition was
58:00
uh you know spatial ai competition was which was sponsored by intel
58:02
which was sponsored by intel
58:02
which was sponsored by intel a lot of applicants uh had suggested
58:05
a lot of applicants uh had suggested
58:05
a lot of applicants uh had suggested solutions for social distancing
58:07
solutions for social distancing
58:07
solutions for social distancing because it has depth perception you can
58:10
because it has depth perception you can
58:10
because it has depth perception you can know
58:10
know
58:10
know you can calculate the distance between
58:12
you can calculate the distance between
58:12
you can calculate the distance between people and
58:13
people and
58:13
people and do social distancing as well and yes you
58:16
do social distancing as well and yes you
58:16
do social distancing as well and yes you can use the neural inference
58:17
can use the neural inference
58:17
can use the neural inference to figure out the location of the people
58:20
to figure out the location of the people
58:20
to figure out the location of the people in the image
58:21
in the image
58:21
in the image and then you can use depth and figure
58:23
and then you can use depth and figure
58:23
and then you can use depth and figure out the distance between them
58:25
out the distance between them
58:25
out the distance between them so uh yes those kinds of things are very
58:28
so uh yes those kinds of things are very
58:28
so uh yes those kinds of things are very much possible
58:29
much possible
58:30
much possible very cool so i guess this person will be
58:32
very cool so i guess this person will be
58:32
very cool so i guess this person will be uh working on this
58:33
uh working on this
58:33
uh working on this hobby project the next few months um
58:36
hobby project the next few months um
58:36
hobby project the next few months um trying to figure out how to put this
58:38
trying to figure out how to put this
58:38
trying to figure out how to put this together exactly and um so if you're
58:41
together exactly and um so if you're
58:41
together exactly and um so if you're if you're watching i don't know who
58:42
if you're watching i don't know who
58:42
if you're watching i don't know who asked the question they don't tell me
58:44
asked the question they don't tell me
58:44
asked the question they don't tell me but
58:44
but
58:44
but um and if you're building this actual
58:47
um and if you're building this actual
58:47
um and if you're building this actual model for us
58:48
model for us
58:48
model for us please come by the global ai community
58:50
please come by the global ai community
58:50
please come by the global ai community and show us how he did it that that's
58:52
and show us how he did it that that's
58:52
and show us how he did it that that's that that's the sort of stuff that we're
58:53
that that's the sort of stuff that we're
58:54
that that's the sort of stuff that we're looking for the fun projects in ai
58:55
looking for the fun projects in ai
58:55
looking for the fun projects in ai especially now
58:57
especially now
58:57
especially now uh given the situation around kovy that
58:59
uh given the situation around kovy that
58:59
uh given the situation around kovy that i mean
59:00
i mean
59:00
i mean you can't have enough fun i guess during
59:02
you can't have enough fun i guess during
59:02
you can't have enough fun i guess during the weekend
59:03
the weekend
59:04
the weekend so yeah i i feel a challenge coming on
59:08
so yeah i i feel a challenge coming on
59:08
so yeah i i feel a challenge coming on do we have a ai development kit
59:10
do we have a ai development kit
59:10
do we have a ai development kit challenge
59:11
challenge
59:11
challenge that we could build into to january's
59:15
that we could build into to january's
59:15
that we could build into to january's ai boot camp
59:21
i'm going to look at sacha for for a few
59:23
i'm going to look at sacha for for a few
59:23
i'm going to look at sacha for for a few seconds and
59:24
seconds and
59:24
seconds and with my great puppy eyes and ask him can
59:27
with my great puppy eyes and ask him can
59:27
with my great puppy eyes and ask him can we have a chat about that and
59:28
we have a chat about that and
59:28
we have a chat about that and maybe fix something up because that
59:30
maybe fix something up because that
59:30
maybe fix something up because that sounds like exactly the sort of
59:32
sounds like exactly the sort of
59:32
sounds like exactly the sort of challenge that the global ai community
59:33
challenge that the global ai community
59:33
challenge that the global ai community is uh is capable of
59:35
is uh is capable of
59:35
is uh is capable of completing um so yeah actually uh
59:39
completing um so yeah actually uh
59:39
completing um so yeah actually uh there is there is another challenge
59:41
there is there is another challenge
59:41
there is there is another challenge coming so our
59:42
coming so our
59:42
coming so our uh spatial ai competition is coming to
59:45
uh spatial ai competition is coming to
59:45
uh spatial ai competition is coming to an end
59:45
an end
59:46
an end the they will there were two phases
59:48
the they will there were two phases
59:48
the they will there were two phases first of all we
59:49
first of all we
59:49
first of all we selected 32 uh 32 people
59:52
selected 32 uh 32 people
59:52
selected 32 uh 32 people from about 235 people who had submitted
59:55
from about 235 people who had submitted
59:55
from about 235 people who had submitted uh the application and those 32 they
59:58
uh the application and those 32 they
59:58
uh the application and those 32 they actually got
59:59
actually got
59:59
actually got the opencv ai kit and they are building
1:00:01
the opencv ai kit and they are building
1:00:01
the opencv ai kit and they are building a solution
1:00:02
a solution
1:00:02
a solution using that and uh they will submit their
1:00:05
using that and uh they will submit their
1:00:05
using that and uh they will submit their final solutions by
1:00:07
final solutions by
1:00:07
final solutions by october 30th and then we will declare
1:00:09
october 30th and then we will declare
1:00:09
october 30th and then we will declare the winners
1:00:10
the winners
1:00:10
the winners and intel was so impressed with the
1:00:13
and intel was so impressed with the
1:00:13
and intel was so impressed with the response that we got
1:00:15
response that we got
1:00:15
response that we got that we are actually working on uh
1:00:18
that we are actually working on uh
1:00:18
that we are actually working on uh on a bigger challenge with other
1:00:20
on a bigger challenge with other
1:00:20
on a bigger challenge with other companies involved also
1:00:22
companies involved also
1:00:22
companies involved also i cannot talk a lot about this but there
1:00:24
i cannot talk a lot about this but there
1:00:24
i cannot talk a lot about this but there are more challenges coming
1:00:26
are more challenges coming
1:00:26
are more challenges coming from where opencv ai kit uh would be
1:00:29
from where opencv ai kit uh would be
1:00:29
from where opencv ai kit uh would be involved
1:00:30
involved
1:00:30
involved very very cool so if people are
1:00:32
very very cool so if people are
1:00:32
very very cool so if people are interested in that i guess they can go
1:00:34
interested in that i guess they can go
1:00:34
interested in that i guess they can go to the opencv website and sign up for
1:00:35
to the opencv website and sign up for
1:00:36
to the opencv website and sign up for that
1:00:36
that
1:00:36
that um yeah and uh one other thing i forgot
1:00:39
um yeah and uh one other thing i forgot
1:00:39
um yeah and uh one other thing i forgot to mention
1:00:40
to mention
1:00:40
to mention is uh that microsoft uh was also very
1:00:43
is uh that microsoft uh was also very
1:00:43
is uh that microsoft uh was also very generous
1:00:44
generous
1:00:44
generous in supporting us with every opencv ai
1:00:47
in supporting us with every opencv ai
1:00:47
in supporting us with every opencv ai kit
1:00:47
kit
1:00:47
kit who the people who purchased it through
1:00:49
who the people who purchased it through
1:00:50
who the people who purchased it through the kickstarter campaign
1:00:51
the kickstarter campaign
1:00:51
the kickstarter campaign they also get 50 hours of free gpu time
1:00:54
they also get 50 hours of free gpu time
1:00:54
they also get 50 hours of free gpu time so that they can train their model and
1:00:55
so that they can train their model and
1:00:56
so that they can train their model and then bring it to the device
1:00:57
then bring it to the device
1:00:57
then bring it to the device uh through uh the azure platform so i
1:01:00
uh through uh the azure platform so i
1:01:00
uh through uh the azure platform so i just wanted to
1:01:01
just wanted to
1:01:01
just wanted to you know thank them publicly here one
1:01:03
you know thank them publicly here one
1:01:03
you know thank them publicly here one more time
1:01:04
more time
1:01:04
more time that's really really cool that they do
1:01:06
that's really really cool that they do
1:01:06
that's really really cool that they do that training
1:01:07
that training
1:01:08
that training neural networks on a cpu is not going to
1:01:10
neural networks on a cpu is not going to
1:01:10
neural networks on a cpu is not going to be fun so
1:01:11
be fun so
1:01:11
be fun so and not everybody has the money to buy a
1:01:13
and not everybody has the money to buy a
1:01:13
and not everybody has the money to buy a gpu so
1:01:15
gpu so
1:01:15
gpu so i guess if you're low on budget then uh
1:01:17
i guess if you're low on budget then uh
1:01:17
i guess if you're low on budget then uh and you can and you bought this camera
1:01:19
and you can and you bought this camera
1:01:19
and you can and you bought this camera that's a great addition
1:01:20
that's a great addition
1:01:20
that's a great addition uh to the package that's i also like the
1:01:24
uh to the package that's i also like the
1:01:24
uh to the package that's i also like the fact
1:01:24
fact
1:01:24
fact i also like the fact that when you're
1:01:26
i also like the fact that when you're
1:01:26
i also like the fact that when you're doing something uh
1:01:28
doing something uh
1:01:28
doing something uh something really good you know we have
1:01:30
something really good you know we have
1:01:30
something really good you know we have always focused on ai for good
1:01:32
always focused on ai for good
1:01:32
always focused on ai for good that companies come together you know uh
1:01:36
that companies come together you know uh
1:01:36
that companies come together you know uh intel is sponsoring uh the competition
1:01:39
intel is sponsoring uh the competition
1:01:39
intel is sponsoring uh the competition uh microsoft is giving free uh
1:01:42
uh microsoft is giving free uh
1:01:42
uh microsoft is giving free uh gpu hours so all these different
1:01:46
gpu hours so all these different
1:01:46
gpu hours so all these different companies right
1:01:47
companies right
1:01:47
companies right they are coming together because we they
1:01:49
they are coming together because we they
1:01:50
they are coming together because we they know that we are serving the community
1:01:51
know that we are serving the community
1:01:51
know that we are serving the community and that's something that i
1:01:53
and that's something that i
1:01:53
and that's something that i really like people there is such a big
1:01:56
really like people there is such a big
1:01:56
really like people there is such a big developer community 6500
1:01:58
developer community 6500
1:01:58
developer community 6500 that uh it's substantial you know people
1:02:01
that uh it's substantial you know people
1:02:01
that uh it's substantial you know people will come up with new ideas
1:02:03
will come up with new ideas
1:02:03
will come up with new ideas and make this platform uh really
1:02:05
and make this platform uh really
1:02:05
and make this platform uh really successful
1:02:06
successful
1:02:06
successful so that's something i'm really excited
1:02:07
so that's something i'm really excited
1:02:08
so that's something i'm really excited about yeah
1:02:09
about yeah
1:02:09
about yeah sounds good um so talking about
1:02:12
sounds good um so talking about
1:02:12
sounds good um so talking about ai for good we've we've talked about
1:02:14
ai for good we've we've talked about
1:02:14
ai for good we've we've talked about edgy devices i guess hololens is uh
1:02:17
edgy devices i guess hololens is uh
1:02:17
edgy devices i guess hololens is uh is the same occasionally occasionally
1:02:20
is the same occasionally occasionally
1:02:20
is the same occasionally occasionally connected
1:02:21
connected
1:02:21
connected we've got somebody wondering about
1:02:22
we've got somebody wondering about
1:02:22
we've got somebody wondering about privacy so how
1:02:24
privacy so how
1:02:24
privacy so how does it help uh privacy wise to have
1:02:26
does it help uh privacy wise to have
1:02:26
does it help uh privacy wise to have offline models
1:02:28
offline models
1:02:28
offline models oh that's that's a big deal uh so
1:02:30
oh that's that's a big deal uh so
1:02:30
oh that's that's a big deal uh so privacy you can actually
1:02:33
privacy you can actually
1:02:33
privacy you can actually if you are building a solution you can
1:02:35
if you are building a solution you can
1:02:35
if you are building a solution you can come to us and we will build a camera
1:02:37
come to us and we will build a camera
1:02:37
come to us and we will build a camera for
1:02:38
for
1:02:38
for uh for you which uh the firmware
1:02:41
uh for you which uh the firmware
1:02:41
uh for you which uh the firmware locks the fact it will never let an
1:02:44
locks the fact it will never let an
1:02:44
locks the fact it will never let an image go out of the camera right that
1:02:46
image go out of the camera right that
1:02:46
image go out of the camera right that can be guaranteed
1:02:47
can be guaranteed
1:02:47
can be guaranteed by uh in the firmware so
1:02:50
by uh in the firmware so
1:02:50
by uh in the firmware so now you can do all the processing and
1:02:52
now you can do all the processing and
1:02:52
now you can do all the processing and only give out
1:02:54
only give out
1:02:54
only give out the meta information for example um
1:02:57
the meta information for example um
1:02:57
the meta information for example um for example in case uh of say
1:03:00
for example in case uh of say
1:03:00
for example in case uh of say uh like let's say a gun shooting
1:03:03
uh like let's say a gun shooting
1:03:03
uh like let's say a gun shooting application you don't want to send
1:03:05
application you don't want to send
1:03:05
application you don't want to send images again and again you just want to
1:03:08
images again and again you just want to
1:03:08
images again and again you just want to send
1:03:08
send
1:03:08
send uh the metadata that okay
1:03:11
uh the metadata that okay
1:03:11
uh the metadata that okay uh there is you know shooter in the
1:03:14
uh there is you know shooter in the
1:03:14
uh there is you know shooter in the scene and they are located
1:03:16
scene and they are located
1:03:16
scene and they are located at this uh location right
1:03:19
at this uh location right
1:03:19
at this uh location right and that's enough you don't need a very
1:03:21
and that's enough you don't need a very
1:03:21
and that's enough you don't need a very heavy network
1:03:22
heavy network
1:03:22
heavy network connection and the privacy is also
1:03:24
connection and the privacy is also
1:03:24
connection and the privacy is also maintained because most of the time
1:03:26
maintained because most of the time
1:03:26
maintained because most of the time let's say
1:03:26
let's say
1:03:26
let's say in a school situation you don't want
1:03:29
in a school situation you don't want
1:03:29
in a school situation you don't want kids pictures to be uploaded to
1:03:31
kids pictures to be uploaded to
1:03:31
kids pictures to be uploaded to a cloud or anything you just want the
1:03:33
a cloud or anything you just want the
1:03:33
a cloud or anything you just want the meta
1:03:34
meta
1:03:34
meta information to go out and this works
1:03:37
information to go out and this works
1:03:37
information to go out and this works beautifully
1:03:37
beautifully
1:03:38
beautifully because only when there is uh like a
1:03:40
because only when there is uh like a
1:03:40
because only when there is uh like a shooter situation
1:03:42
shooter situation
1:03:42
shooter situation would uh something that information be
1:03:44
would uh something that information be
1:03:44
would uh something that information be conveyed to
1:03:45
conveyed to
1:03:45
conveyed to law enforcement and only the location
1:03:48
law enforcement and only the location
1:03:48
law enforcement and only the location you know where
1:03:48
you know where
1:03:48
you know where is the shooter located in the scene
1:03:51
is the shooter located in the scene
1:03:51
is the shooter located in the scene instead of uh
1:03:53
instead of uh
1:03:53
instead of uh hurting the privacy of the teachers as
1:03:54
hurting the privacy of the teachers as
1:03:54
hurting the privacy of the teachers as well as the students
1:03:56
well as the students
1:03:56
well as the students yeah and so stefano you you've been
1:03:59
yeah and so stefano you you've been
1:03:59
yeah and so stefano you you've been working with
1:04:00
working with
1:04:00
working with uh ai a lot in your work uh do you run
1:04:03
uh ai a lot in your work uh do you run
1:04:03
uh ai a lot in your work uh do you run into these kind of privacy issues a lot
1:04:05
into these kind of privacy issues a lot
1:04:05
into these kind of privacy issues a lot in your daily work
1:04:07
in your daily work
1:04:07
in your daily work it is a concern yes i echo satya is
1:04:10
it is a concern yes i echo satya is
1:04:10
it is a concern yes i echo satya is saying um
1:04:12
saying um
1:04:12
saying um there is a concern of storing this
1:04:15
there is a concern of storing this
1:04:15
there is a concern of storing this information
1:04:16
information
1:04:16
information there is a a concern of uh detecting
1:04:19
there is a a concern of uh detecting
1:04:19
there is a a concern of uh detecting because this uh
1:04:20
because this uh
1:04:20
because this uh information we there is a this
1:04:23
information we there is a this
1:04:23
information we there is a this technology can be used
1:04:24
technology can be used
1:04:24
technology can be used to detect face of people and it can be
1:04:27
to detect face of people and it can be
1:04:27
to detect face of people and it can be used for good purposes
1:04:29
used for good purposes
1:04:29
used for good purposes obviously you don't know how it can be
1:04:30
obviously you don't know how it can be
1:04:30
obviously you don't know how it can be used now there is a
1:04:32
used now there is a
1:04:32
used now there is a very interesting evolution of
1:04:36
very interesting evolution of
1:04:36
very interesting evolution of you know face detection which is also a
1:04:39
you know face detection which is also a
1:04:39
you know face detection which is also a combination
1:04:40
combination
1:04:40
combination with a sort of
1:04:43
with a sort of
1:04:43
with a sort of a blockchain and digital ledger is a
1:04:46
a blockchain and digital ledger is a
1:04:46
a blockchain and digital ledger is a technology called
1:04:47
technology called
1:04:47
technology called photo dna that microsoft has developed
1:04:50
photo dna that microsoft has developed
1:04:50
photo dna that microsoft has developed some time ago where you actually take
1:04:52
some time ago where you actually take
1:04:52
some time ago where you actually take uh imprint of the image and
1:04:56
uh imprint of the image and
1:04:56
uh imprint of the image and create a ash value which is a unique
1:04:59
create a ash value which is a unique
1:04:59
create a ash value which is a unique signature
1:04:59
signature
1:05:00
signature but that can be used for detecting
1:05:03
but that can be used for detecting
1:05:03
but that can be used for detecting phases over and over time but without
1:05:06
phases over and over time but without
1:05:06
phases over and over time but without actually being able to reconstruct the
1:05:08
actually being able to reconstruct the
1:05:08
actually being able to reconstruct the face
1:05:09
face
1:05:09
face and therefore cannot be used if
1:05:13
and therefore cannot be used if
1:05:13
and therefore cannot be used if data leaks cannot be used to go back to
1:05:16
data leaks cannot be used to go back to
1:05:16
data leaks cannot be used to go back to the identity of the person
1:05:17
the identity of the person
1:05:17
the identity of the person so it's like you take a signature of
1:05:20
so it's like you take a signature of
1:05:20
so it's like you take a signature of your face that has
1:05:21
your face that has
1:05:21
your face that has unique landmarks and attributes and then
1:05:24
unique landmarks and attributes and then
1:05:24
unique landmarks and attributes and then you
1:05:24
you
1:05:24
you store only the the signature and that
1:05:27
store only the the signature and that
1:05:27
store only the the signature and that signature can be used
1:05:29
signature can be used
1:05:29
signature can be used to detect that phase somewhere else
1:05:33
to detect that phase somewhere else
1:05:33
to detect that phase somewhere else but it cannot be used to rebuild the
1:05:35
but it cannot be used to rebuild the
1:05:36
but it cannot be used to rebuild the identity
1:05:36
identity
1:05:36
identity so there are ways of protecting your
1:05:39
so there are ways of protecting your
1:05:39
so there are ways of protecting your privacy it's all about using technology
1:05:41
privacy it's all about using technology
1:05:41
privacy it's all about using technology in the right way
1:05:42
in the right way
1:05:42
in the right way and i love that you mentioned ai for
1:05:44
and i love that you mentioned ai for
1:05:44
and i love that you mentioned ai for good is one of the
1:05:46
good is one of the
1:05:46
good is one of the foundation of creating ethical ai
1:05:51
foundation of creating ethical ai
1:05:51
foundation of creating ethical ai so that you know there are some
1:05:52
so that you know there are some
1:05:52
so that you know there are some principles some guidelines that can be
1:05:54
principles some guidelines that can be
1:05:54
principles some guidelines that can be used
1:05:55
used
1:05:55
used to make sure that artificial
1:05:57
to make sure that artificial
1:05:57
to make sure that artificial intelligence is this big power
1:05:59
intelligence is this big power
1:05:59
intelligence is this big power but you use in a sensitive and human way
1:06:03
but you use in a sensitive and human way
1:06:03
but you use in a sensitive and human way yeah sounds really good that those
1:06:05
yeah sounds really good that those
1:06:05
yeah sounds really good that those devices
1:06:06
devices
1:06:06
devices on the one hand offer you ways of of
1:06:09
on the one hand offer you ways of of
1:06:09
on the one hand offer you ways of of losing
1:06:10
losing
1:06:10
losing using less bandwidth to project results
1:06:14
using less bandwidth to project results
1:06:14
using less bandwidth to project results back to a server or
1:06:15
back to a server or
1:06:15
back to a server or another part of your solution and at the
1:06:17
another part of your solution and at the
1:06:17
another part of your solution and at the same time protect your privacy i think
1:06:19
same time protect your privacy i think
1:06:19
same time protect your privacy i think this is one of the major steps that
1:06:21
this is one of the major steps that
1:06:21
this is one of the major steps that we're making as a community in in 2020
1:06:23
we're making as a community in in 2020
1:06:23
we're making as a community in in 2020 and beyond
1:06:24
and beyond
1:06:24
and beyond going uh thinking about how much data
1:06:27
going uh thinking about how much data
1:06:27
going uh thinking about how much data are we actually
1:06:28
are we actually
1:06:28
are we actually using um to reduce the the carbon
1:06:31
using um to reduce the the carbon
1:06:31
using um to reduce the the carbon footprint and on the other hand how do
1:06:32
footprint and on the other hand how do
1:06:32
footprint and on the other hand how do we protect
1:06:33
we protect
1:06:33
we protect users against abuse by ai
1:06:36
users against abuse by ai
1:06:36
users against abuse by ai and that's really great that that it
1:06:38
and that's really great that that it
1:06:38
and that's really great that that it works for uh
1:06:39
works for uh
1:06:40
works for uh opencv cameras and as well for the
1:06:43
opencv cameras and as well for the
1:06:43
opencv cameras and as well for the hololens
1:06:44
hololens
1:06:44
hololens so um i've got also
1:06:50
the the ai for good you know the very
1:06:53
the the ai for good you know the very
1:06:53
the the ai for good you know the very first uh
1:06:53
first uh
1:06:54
first uh when we did the competition we received
1:06:55
when we did the competition we received
1:06:55
when we did the competition we received so many uh
1:06:57
so many uh
1:06:57
so many uh applicants who built visual assistance
1:07:00
applicants who built visual assistance
1:07:00
applicants who built visual assistance devices and one of the applicant was
1:07:02
devices and one of the applicant was
1:07:02
devices and one of the applicant was marx malencio
1:07:04
marx malencio
1:07:04
marx malencio who was shot in the head back in the
1:07:05
who was shot in the head back in the
1:07:06
who was shot in the head back in the year 2003
1:07:07
year 2003
1:07:07
year 2003 and he lost his vision and he now
1:07:11
and he lost his vision and he now
1:07:11
and he lost his vision and he now runs a company with 200 people and he's
1:07:14
runs a company with 200 people and he's
1:07:14
runs a company with 200 people and he's building visual assistance device
1:07:16
building visual assistance device
1:07:16
building visual assistance device and i was so glad that he his solution
1:07:20
and i was so glad that he his solution
1:07:20
and i was so glad that he his solution required a sonar for distance
1:07:21
required a sonar for distance
1:07:22
required a sonar for distance measurement and but that was very heavy
1:07:24
measurement and but that was very heavy
1:07:24
measurement and but that was very heavy and they replaced that sonar with uh oak
1:07:28
and they replaced that sonar with uh oak
1:07:28
and they replaced that sonar with uh oak and that that made my day you know wow
1:07:31
and that that made my day you know wow
1:07:31
and that that made my day you know wow if we yeah that was such a powerful
1:07:33
if we yeah that was such a powerful
1:07:33
if we yeah that was such a powerful story
1:07:34
story
1:07:34
story uh it just blew me away
1:07:37
uh it just blew me away
1:07:37
uh it just blew me away yeah i mean if you could replace a a uh
1:07:39
yeah i mean if you could replace a a uh
1:07:39
yeah i mean if you could replace a a uh several pound weighing
1:07:40
several pound weighing
1:07:40
several pound weighing device with such a small camera and it's
1:07:43
device with such a small camera and it's
1:07:43
device with such a small camera and it's a lot cheaper too sonar is
1:07:45
a lot cheaper too sonar is
1:07:45
a lot cheaper too sonar is is very expensive to run yeah um so yeah
1:07:48
is very expensive to run yeah um so yeah
1:07:48
is very expensive to run yeah um so yeah that sounds really cool
1:07:50
that sounds really cool
1:07:50
that sounds really cool so um we've got a bunch more questions
1:07:52
so um we've got a bunch more questions
1:07:52
so um we've got a bunch more questions from the audience alicia
1:07:54
from the audience alicia
1:07:54
from the audience alicia um it involves raspberry pi's uh i have
1:07:57
um it involves raspberry pi's uh i have
1:07:57
um it involves raspberry pi's uh i have to say a lot of people are wondering
1:07:59
to say a lot of people are wondering
1:07:59
to say a lot of people are wondering um um so could i connect this device up
1:08:03
um um so could i connect this device up
1:08:03
um um so could i connect this device up to a raspberry pi and
1:08:04
to a raspberry pi and
1:08:04
to a raspberry pi and and build a larger solution around it is
1:08:06
and build a larger solution around it is
1:08:06
and build a larger solution around it is that possible
1:08:07
that possible
1:08:07
that possible that's yeah so this device is uh
1:08:10
that's yeah so this device is uh
1:08:10
that's yeah so this device is uh supposed to be connected to a host
1:08:12
supposed to be connected to a host
1:08:12
supposed to be connected to a host like a raspberry pi and raspberry pi is
1:08:15
like a raspberry pi and raspberry pi is
1:08:16
like a raspberry pi and raspberry pi is what we expect people to be connecting
1:08:18
what we expect people to be connecting
1:08:18
what we expect people to be connecting it to
1:08:19
it to
1:08:19
it to and one other thing i forgot to mention
1:08:21
and one other thing i forgot to mention
1:08:21
and one other thing i forgot to mention while uh while i was mentioning that
1:08:23
while uh while i was mentioning that
1:08:23
while uh while i was mentioning that uh you know it is eight times faster
1:08:25
uh you know it is eight times faster
1:08:25
uh you know it is eight times faster than uh if you run this thing
1:08:27
than uh if you run this thing
1:08:28
than uh if you run this thing you know if you run a neural compute
1:08:29
you know if you run a neural compute
1:08:29
you know if you run a neural compute stick with a raspberry pi
1:08:31
stick with a raspberry pi
1:08:31
stick with a raspberry pi uh in that solution the raspberry pi
1:08:33
uh in that solution the raspberry pi
1:08:33
uh in that solution the raspberry pi processor is completely engaged it goes
1:08:35
processor is completely engaged it goes
1:08:35
processor is completely engaged it goes to 100
1:08:37
to 100
1:08:37
to 100 percent in our solution even though it
1:08:39
percent in our solution even though it
1:08:40
percent in our solution even though it is eight times faster
1:08:41
is eight times faster
1:08:41
is eight times faster raspberry pi is not you know it is at
1:08:43
raspberry pi is not you know it is at
1:08:43
raspberry pi is not you know it is at one percent two percent
1:08:45
one percent two percent
1:08:45
one percent two percent so the raspberry pi itself is free uh to
1:08:48
so the raspberry pi itself is free uh to
1:08:48
so the raspberry pi itself is free uh to be used
1:08:49
be used
1:08:49
be used so you can you can run other things on
1:08:51
so you can you can run other things on
1:08:51
so you can you can run other things on there like a small
1:08:53
there like a small
1:08:53
there like a small uh http surface or an application that
1:08:57
uh http surface or an application that
1:08:57
uh http surface or an application that actually
1:08:57
actually
1:08:57
actually combines maybe several cameras together
1:09:00
combines maybe several cameras together
1:09:00
combines maybe several cameras together i mean uh the sky is the limit in this
1:09:02
i mean uh the sky is the limit in this
1:09:02
i mean uh the sky is the limit in this case
1:09:03
case
1:09:03
case you can also combine multiple oak
1:09:05
you can also combine multiple oak
1:09:05
you can also combine multiple oak devices to the same
1:09:06
devices to the same
1:09:06
devices to the same raspberry pi yeah wow
1:09:10
raspberry pi yeah wow
1:09:10
raspberry pi yeah wow yeah another cool question that people
1:09:12
yeah another cool question that people
1:09:12
yeah another cool question that people have so we've
1:09:13
have so we've
1:09:13
have so we've we're giving away an oculus quest too
1:09:16
we're giving away an oculus quest too
1:09:16
we're giving away an oculus quest too today
1:09:16
today
1:09:16
today and people are really into the whole vr
1:09:19
and people are really into the whole vr
1:09:19
and people are really into the whole vr scene and ar scene i guess that's
1:09:21
scene and ar scene i guess that's
1:09:21
scene and ar scene i guess that's because we have a lot of speakers around
1:09:23
because we have a lot of speakers around
1:09:23
because we have a lot of speakers around that today as well stefano is going to
1:09:24
that today as well stefano is going to
1:09:24
that today as well stefano is going to talk about hololens we've got
1:09:26
talk about hololens we've got
1:09:26
talk about hololens we've got um uh other people talking about
1:09:30
um uh other people talking about
1:09:30
um uh other people talking about vr um so people are wondering
1:09:34
vr um so people are wondering
1:09:34
vr um so people are wondering is the oak camera powerful enough to
1:09:37
is the oak camera powerful enough to
1:09:37
is the oak camera powerful enough to connect it to an
1:09:37
connect it to an
1:09:38
connect it to an ar vr headset actually is that at all
1:09:40
ar vr headset actually is that at all
1:09:40
ar vr headset actually is that at all possible i don't know i've never
1:09:42
possible i don't know i've never
1:09:42
possible i don't know i've never worked with one uh i don't know actually
1:09:45
worked with one uh i don't know actually
1:09:45
worked with one uh i don't know actually i don't know
1:09:45
i don't know
1:09:45
i don't know the answer to that um you know it
1:09:48
the answer to that um you know it
1:09:48
the answer to that um you know it depends on the operating system
1:09:49
depends on the operating system
1:09:50
depends on the operating system anything that supports openvino as the
1:09:51
anything that supports openvino as the
1:09:51
anything that supports openvino as the host uh
1:09:53
host uh
1:09:53
host uh any host that supports openvenus no we
1:09:56
any host that supports openvenus no we
1:09:56
any host that supports openvenus no we it will can be connected to it so
1:09:59
it will can be connected to it so
1:09:59
it will can be connected to it so i don't know what headsets actually uh
1:10:03
i don't know what headsets actually uh
1:10:03
i don't know what headsets actually uh you know support uh openvino
1:10:06
you know support uh openvino
1:10:06
you know support uh openvino but i guess we could could of course
1:10:08
but i guess we could could of course
1:10:08
but i guess we could could of course connect the oak device to raspberry pi
1:10:10
connect the oak device to raspberry pi
1:10:10
connect the oak device to raspberry pi and then use a network connection to
1:10:11
and then use a network connection to
1:10:12
and then use a network connection to talk to the vr headset that works
1:10:13
talk to the vr headset that works
1:10:14
talk to the vr headset that works yes either way yes yeah yeah i will be
1:10:16
yes either way yes yeah yeah i will be
1:10:16
yes either way yes yeah yeah i will be slower i guess
1:10:19
no because uh you know you're only
1:10:21
no because uh you know you're only
1:10:21
no because uh you know you're only getting metadata
1:10:23
getting metadata
1:10:23
getting metadata out of it you don't need to send the
1:10:26
out of it you don't need to send the
1:10:26
out of it you don't need to send the image you're using the oak camera as a
1:10:29
image you're using the oak camera as a
1:10:30
image you're using the oak camera as a sensor
1:10:30
sensor
1:10:30
sensor you're getting metadata out of it not
1:10:32
you're getting metadata out of it not
1:10:32
you're getting metadata out of it not really the image if that's
1:10:34
really the image if that's
1:10:34
really the image if that's uh yeah that's really awesome
1:10:38
uh yeah that's really awesome
1:10:38
uh yeah that's really awesome yeah even on the network it would be
1:10:39
yeah even on the network it would be
1:10:39
yeah even on the network it would be really fast
1:10:41
really fast
1:10:41
really fast yeah and so um and and i guess this will
1:10:45
yeah and so um and and i guess this will
1:10:45
yeah and so um and and i guess this will be
1:10:45
be
1:10:45
be my uh i don't know if we got more
1:10:49
my uh i don't know if we got more
1:10:49
my uh i don't know if we got more questions we've got a ton of questions
1:10:50
questions we've got a ton of questions
1:10:50
questions we've got a ton of questions coming in but
1:10:51
coming in but
1:10:51
coming in but um one of the interesting things that
1:10:53
um one of the interesting things that
1:10:53
um one of the interesting things that i've seen is that um
1:10:56
i've seen is that um
1:10:56
i've seen is that um so with all the vr people are combining
1:10:58
so with all the vr people are combining
1:10:58
so with all the vr people are combining vr with ai a lot
1:11:00
vr with ai a lot
1:11:00
vr with ai a lot what do you think um are they essential
1:11:03
what do you think um are they essential
1:11:03
what do you think um are they essential to put together
1:11:04
to put together
1:11:04
to put together or um yeah what's going on actually with
1:11:08
or um yeah what's going on actually with
1:11:08
or um yeah what's going on actually with the i i
1:11:11
the i i
1:11:11
the i i i'll give you one example where it's
1:11:13
i'll give you one example where it's
1:11:13
i'll give you one example where it's used in the opposite way
1:11:14
used in the opposite way
1:11:14
used in the opposite way where ar is using uh is helping ar and
1:11:18
where ar is using uh is helping ar and
1:11:18
where ar is using uh is helping ar and vr they are helping
1:11:20
vr they are helping
1:11:20
vr they are helping ai and that's in the training process uh
1:11:22
ai and that's in the training process uh
1:11:22
ai and that's in the training process uh in a lot of cases
1:11:24
in a lot of cases
1:11:24
in a lot of cases you cannot produce uh enough data from
1:11:27
you cannot produce uh enough data from
1:11:27
you cannot produce uh enough data from real world
1:11:28
real world
1:11:28
real world there could be privacy concerns but you
1:11:30
there could be privacy concerns but you
1:11:30
there could be privacy concerns but you can simulate
1:11:31
can simulate
1:11:31
can simulate that data right either fully uh
1:11:34
that data right either fully uh
1:11:34
that data right either fully uh you know it is fully uh virtual or
1:11:37
you know it is fully uh virtual or
1:11:37
you know it is fully uh virtual or it could be half virtual and half real
1:11:40
it could be half virtual and half real
1:11:40
it could be half virtual and half real kind of
1:11:40
kind of
1:11:40
kind of scenarios the simplest application it's
1:11:43
scenarios the simplest application it's
1:11:43
scenarios the simplest application it's not even
1:11:44
not even
1:11:44
not even vr but a simplest application would be
1:11:46
vr but a simplest application would be
1:11:46
vr but a simplest application would be uh text
1:11:47
uh text
1:11:47
uh text recognition so we did a project where we
1:11:50
recognition so we did a project where we
1:11:50
recognition so we did a project where we did text recognition on id cards
1:11:52
did text recognition on id cards
1:11:52
did text recognition on id cards and obviously we did not have as many
1:11:54
and obviously we did not have as many
1:11:54
and obviously we did not have as many passports and things like that
1:11:56
passports and things like that
1:11:56
passports and things like that and the the company that we were working
1:11:59
and the the company that we were working
1:11:59
and the the company that we were working for did not
1:12:00
for did not
1:12:00
for did not want to give us obviously so we
1:12:02
want to give us obviously so we
1:12:02
want to give us obviously so we simulated that data
1:12:04
simulated that data
1:12:04
simulated that data using uh you know different backgrounds
1:12:07
using uh you know different backgrounds
1:12:07
using uh you know different backgrounds different lighting conditions etc so we
1:12:10
different lighting conditions etc so we
1:12:10
different lighting conditions etc so we rendered those images using different
1:12:12
rendered those images using different
1:12:12
rendered those images using different fonts etc and created 10 million
1:12:15
fonts etc and created 10 million
1:12:15
fonts etc and created 10 million images of different text right and
1:12:17
images of different text right and
1:12:17
images of different text right and similarly uh
1:12:19
similarly uh
1:12:19
similarly uh i think facebook is using they are
1:12:22
i think facebook is using they are
1:12:22
i think facebook is using they are trying to
1:12:23
trying to
1:12:23
trying to do gaze detection and for that gauge
1:12:25
do gaze detection and for that gauge
1:12:25
do gaze detection and for that gauge detection application
1:12:27
detection application
1:12:27
detection application they have a simulated environment you
1:12:30
they have a simulated environment you
1:12:30
they have a simulated environment you know
1:12:30
know
1:12:30
know uh the eyes are uh they don't take
1:12:34
uh the eyes are uh they don't take
1:12:34
uh the eyes are uh they don't take actual data it's very difficult to get
1:12:36
actual data it's very difficult to get
1:12:36
actual data it's very difficult to get actual data especially in the millions
1:12:39
actual data especially in the millions
1:12:39
actual data especially in the millions so they have uh a simulator which for
1:12:41
so they have uh a simulator which for
1:12:41
so they have uh a simulator which for the eye
1:12:42
the eye
1:12:42
the eye which produces different gazes
1:12:45
which produces different gazes
1:12:45
which produces different gazes based on different eye shapes etc so it
1:12:47
based on different eye shapes etc so it
1:12:47
based on different eye shapes etc so it produces the data for ai
1:12:49
produces the data for ai
1:12:49
produces the data for ai so they are intertwined in both ways
1:12:51
so they are intertwined in both ways
1:12:51
so they are intertwined in both ways first of all you know
1:12:53
first of all you know
1:12:53
first of all you know vr and ar they are helping uh to
1:12:56
vr and ar they are helping uh to
1:12:56
vr and ar they are helping uh to generate data for ai but then you can
1:12:59
generate data for ai but then you can
1:12:59
generate data for ai but then you can also use ai
1:13:00
also use ai
1:13:00
also use ai to play a game in that scenario
1:13:04
to play a game in that scenario
1:13:04
to play a game in that scenario yeah um so stefano you've been working a
1:13:07
yeah um so stefano you've been working a
1:13:07
yeah um so stefano you've been working a lot with hololens
1:13:08
lot with hololens
1:13:08
lot with hololens uh how do the applications with ai and
1:13:10
uh how do the applications with ai and
1:13:10
uh how do the applications with ai and hololens
1:13:11
hololens
1:13:11
hololens uh what do they look like
1:13:14
uh what do they look like
1:13:14
uh what do they look like well look uh autolens is a device that
1:13:17
well look uh autolens is a device that
1:13:17
well look uh autolens is a device that is
1:13:18
is
1:13:18
is unfettered so it can completely work in
1:13:20
unfettered so it can completely work in
1:13:20
unfettered so it can completely work in isolation
1:13:21
isolation
1:13:21
isolation and you don't need to connect to to your
1:13:24
and you don't need to connect to to your
1:13:24
and you don't need to connect to to your laptop you don't need to connect to your
1:13:26
laptop you don't need to connect to your
1:13:26
laptop you don't need to connect to your mobile device you don't need to connect
1:13:27
mobile device you don't need to connect
1:13:27
mobile device you don't need to connect to the cloud
1:13:28
to the cloud
1:13:28
to the cloud so it just work on his own so
1:13:31
so it just work on his own so
1:13:31
so it just work on his own so technically you wouldn't need
1:13:33
technically you wouldn't need
1:13:33
technically you wouldn't need any ai capability attached to it it just
1:13:36
any ai capability attached to it it just
1:13:36
any ai capability attached to it it just does
1:13:36
does
1:13:36
does whatever is programmed to do and you
1:13:39
whatever is programmed to do and you
1:13:39
whatever is programmed to do and you know i i'll show you a few
1:13:41
know i i'll show you a few
1:13:41
know i i'll show you a few things in a moment uh but
1:13:44
things in a moment uh but
1:13:44
things in a moment uh but said that uh you will miss out a lot if
1:13:47
said that uh you will miss out a lot if
1:13:48
said that uh you will miss out a lot if you just
1:13:48
you just
1:13:48
you just just so to speak use it for augmented
1:13:51
just so to speak use it for augmented
1:13:51
just so to speak use it for augmented reality
1:13:52
reality
1:13:52
reality uh the power of you know face detection
1:13:55
uh the power of you know face detection
1:13:55
uh the power of you know face detection or object detection in general
1:13:57
or object detection in general
1:13:57
or object detection in general a text to speech at a speech to text
1:14:00
a text to speech at a speech to text
1:14:00
a text to speech at a speech to text which is another capability
1:14:02
which is another capability
1:14:02
which is another capability uh that is available in in the azure
1:14:04
uh that is available in in the azure
1:14:04
uh that is available in in the azure cloud as a cognitive service
1:14:06
cloud as a cognitive service
1:14:06
cloud as a cognitive service will enhance the device itself the
1:14:09
will enhance the device itself the
1:14:09
will enhance the device itself the device
1:14:10
device
1:14:10
device take it as a no put simply a piece of
1:14:14
take it as a no put simply a piece of
1:14:14
take it as a no put simply a piece of hardware
1:14:14
hardware
1:14:14
hardware then as a as you said the sky's the
1:14:16
then as a as you said the sky's the
1:14:16
then as a as you said the sky's the limit try to use it in the best way
1:14:18
limit try to use it in the best way
1:14:18
limit try to use it in the best way possible
1:14:19
possible
1:14:19
possible to make it a more interactive more
1:14:23
to make it a more interactive more
1:14:23
to make it a more interactive more interesting also from an application
1:14:24
interesting also from an application
1:14:24
interesting also from an application perspective if you don't mind
1:14:27
perspective if you don't mind
1:14:27
perspective if you don't mind i just want to add the on this
1:14:28
i just want to add the on this
1:14:28
i just want to add the on this simulation topic very very briefly
1:14:30
simulation topic very very briefly
1:14:30
simulation topic very very briefly because uh
1:14:32
because uh
1:14:32
because uh i'm based in australia and uh if you
1:14:35
i'm based in australia and uh if you
1:14:35
i'm based in australia and uh if you guys
1:14:36
guys
1:14:36
guys heard the news at the beginning of this
1:14:38
heard the news at the beginning of this
1:14:38
heard the news at the beginning of this year which has been a fantastic year
1:14:40
year which has been a fantastic year
1:14:40
year which has been a fantastic year uh we had the bush fires that have been
1:14:42
uh we had the bush fires that have been
1:14:42
uh we had the bush fires that have been impacted the
1:14:43
impacted the
1:14:43
impacted the the countryside and they've devastated
1:14:46
the countryside and they've devastated
1:14:46
the countryside and they've devastated the the country massively
1:14:48
the the country massively
1:14:48
the the country massively more recently also in california so it's
1:14:50
more recently also in california so it's
1:14:50
more recently also in california so it's actually really a global problem
1:14:52
actually really a global problem
1:14:52
actually really a global problem and uh by the time no this was a
1:14:54
and uh by the time no this was a
1:14:54
and uh by the time no this was a developing
1:14:56
developing
1:14:56
developing and impacting the impact in the
1:14:59
and impacting the impact in the
1:14:59
and impacting the impact in the the country massively then some
1:15:02
the country massively then some
1:15:02
the country massively then some technologies evolved
1:15:03
technologies evolved
1:15:03
technologies evolved using drones ground drones and flying
1:15:07
using drones ground drones and flying
1:15:07
using drones ground drones and flying drones
1:15:08
drones
1:15:08
drones to uh to look and identifying using some
1:15:11
to uh to look and identifying using some
1:15:11
to uh to look and identifying using some models
1:15:11
models
1:15:11
models to pretty predict the spread of fires
1:15:15
to pretty predict the spread of fires
1:15:15
to pretty predict the spread of fires around using
1:15:19
temperature humidity pressure whatever
1:15:22
temperature humidity pressure whatever
1:15:22
temperature humidity pressure whatever contribute not to the spread of fire now
1:15:25
contribute not to the spread of fire now
1:15:25
contribute not to the spread of fire now the the the good thing is that after a
1:15:28
the the the good thing is that after a
1:15:28
the the the good thing is that after a few months that this fire
1:15:30
few months that this fire
1:15:30
few months that this fire uh extinguished and uh so we
1:15:33
uh extinguished and uh so we
1:15:33
uh extinguished and uh so we the the technology builders didn't have
1:15:37
the the technology builders didn't have
1:15:37
the the technology builders didn't have enough time and data to to train the
1:15:40
enough time and data to to train the
1:15:40
enough time and data to to train the models
1:15:41
models
1:15:41
models so they couldn't go and you know
1:15:44
so they couldn't go and you know
1:15:44
so they couldn't go and you know light another fire again just to capture
1:15:46
light another fire again just to capture
1:15:46
light another fire again just to capture the image so in the tree
1:15:48
the image so in the tree
1:15:48
the image so in the tree and train the model so need to
1:15:50
and train the model so need to
1:15:50
and train the model so need to assimilate all these
1:15:51
assimilate all these
1:15:52
assimilate all these in different scenario in different
1:15:53
in different scenario in different
1:15:53
in different scenario in different contexts using also a sort of
1:15:56
contexts using also a sort of
1:15:56
contexts using also a sort of augmented reality technology where with
1:15:59
augmented reality technology where with
1:15:59
augmented reality technology where with stock
1:16:00
stock
1:16:00
stock images of fires they overlaid over
1:16:02
images of fires they overlaid over
1:16:02
images of fires they overlaid over forest
1:16:04
forest
1:16:04
forest australian forests to simulate a
1:16:07
australian forests to simulate a
1:16:07
australian forests to simulate a uh um an augmented reality a virtual
1:16:10
uh um an augmented reality a virtual
1:16:10
uh um an augmented reality a virtual reality where
1:16:11
reality where
1:16:11
reality where this fire is actually burning in
1:16:13
this fire is actually burning in
1:16:13
this fire is actually burning in different contexts and see how they
1:16:15
different contexts and see how they
1:16:15
different contexts and see how they spread
1:16:16
spread
1:16:16
spread in their mother they could in that mode
1:16:18
in their mother they could in that mode
1:16:18
in their mother they could in that mode they could trade the model
1:16:20
they could trade the model
1:16:20
they could trade the model to to to have these drones able to
1:16:23
to to to have these drones able to
1:16:23
to to to have these drones able to capture
1:16:23
capture
1:16:24
capture fire awfully we will not use it again
1:16:27
fire awfully we will not use it again
1:16:27
fire awfully we will not use it again but if it happens next year we are
1:16:29
but if it happens next year we are
1:16:29
but if it happens next year we are technology ready
1:16:30
technology ready
1:16:30
technology ready to prevent and react properly
1:16:33
to prevent and react properly
1:16:33
to prevent and react properly wow yeah that's that's a really
1:16:36
wow yeah that's that's a really
1:16:36
wow yeah that's that's a really i mean that's one of the more important
1:16:39
i mean that's one of the more important
1:16:39
i mean that's one of the more important applications of ai
1:16:41
applications of ai
1:16:41
applications of ai and it's good to hear about it so um
1:16:44
and it's good to hear about it so um
1:16:44
and it's good to hear about it so um i've got one more question for um such a
1:16:48
i've got one more question for um such a
1:16:48
i've got one more question for um such a before we move on to the next section so
1:16:49
before we move on to the next section so
1:16:49
before we move on to the next section so such a
1:16:50
such a
1:16:50
such a we've talked about these new devices
1:16:51
we've talked about these new devices
1:16:52
we've talked about these new devices where can people buy it and how can they
1:16:53
where can people buy it and how can they
1:16:53
where can people buy it and how can they get started
1:16:55
get started
1:16:55
get started so uh if you search for opencv ai kit
1:16:59
so uh if you search for opencv ai kit
1:16:59
so uh if you search for opencv ai kit kickstarter then you will land on the
1:17:01
kickstarter then you will land on the
1:17:01
kickstarter then you will land on the kickstarter page
1:17:02
kickstarter page
1:17:02
kickstarter page on the kickstarter page there is a
1:17:03
on the kickstarter page there is a
1:17:03
on the kickstarter page there is a button that says pre-order
1:17:05
button that says pre-order
1:17:05
button that says pre-order and you can pre-order now and i don't
1:17:08
and you can pre-order now and i don't
1:17:08
and you can pre-order now and i don't remember
1:17:09
remember
1:17:09
remember exactly what is uh the time we are
1:17:11
exactly what is uh the time we are
1:17:11
exactly what is uh the time we are promising right now but i think it will
1:17:13
promising right now but i think it will
1:17:13
promising right now but i think it will be in the january
1:17:14
be in the january
1:17:14
be in the january february framework you know time frame
1:17:17
february framework you know time frame
1:17:18
february framework you know time frame uh so people will have to wait a little
1:17:20
uh so people will have to wait a little
1:17:20
uh so people will have to wait a little bit but we are also giving
1:17:21
bit but we are also giving
1:17:21
bit but we are also giving a pretty good discount over
1:17:25
a pretty good discount over
1:17:25
a pretty good discount over uh the retail price so so they should
1:17:27
uh the retail price so so they should
1:17:27
uh the retail price so so they should ask for money this christmas instead of
1:17:29
ask for money this christmas instead of
1:17:29
ask for money this christmas instead of presents and then get one in january
1:17:31
presents and then get one in january
1:17:31
presents and then get one in january and start hacking i guess that's right
1:17:34
and start hacking i guess that's right
1:17:34
and start hacking i guess that's right and
1:17:34
and
1:17:34
and um so you're offering free courses i i
1:17:36
um so you're offering free courses i i
1:17:36
um so you're offering free courses i i guess those are available on the website
1:17:38
guess those are available on the website
1:17:38
guess those are available on the website as well
1:17:39
as well
1:17:39
as well so people can so the free course is not
1:17:41
so people can so the free course is not
1:17:41
so people can so the free course is not yet available on the website it will be
1:17:43
yet available on the website it will be
1:17:43
yet available on the website it will be available by the end of the year uh that
1:17:45
available by the end of the year uh that
1:17:45
available by the end of the year uh that is you know that's the first deliverable
1:17:47
is you know that's the first deliverable
1:17:47
is you know that's the first deliverable we have
1:17:48
we have
1:17:48
we have as part of the kickstarter campaign
1:17:50
as part of the kickstarter campaign
1:17:50
as part of the kickstarter campaign pledges
1:17:51
pledges
1:17:51
pledges and uh by the end of december people
1:17:53
and uh by the end of december people
1:17:53
and uh by the end of december people will start receiving their uh devices
1:17:55
will start receiving their uh devices
1:17:55
will start receiving their uh devices and we will release courses
1:17:58
and we will release courses
1:17:58
and we will release courses as well as free models ah cool
1:18:01
as well as free models ah cool
1:18:01
as well as free models ah cool cool awesome so that's all we have time
1:18:05
cool awesome so that's all we have time
1:18:05
cool awesome so that's all we have time for today for
1:18:06
for today for
1:18:06
for today for uh as far as opencv is concerned so if
1:18:09
uh as far as opencv is concerned so if
1:18:09
uh as far as opencv is concerned so if people have any further questions
1:18:11
people have any further questions
1:18:11
people have any further questions i guess they can go to your website and
1:18:13
i guess they can go to your website and
1:18:13
i guess they can go to your website and and ask or
1:18:14
and ask or
1:18:14
and ask or ask the questions there or are there any
1:18:16
ask the questions there or are there any
1:18:16
ask the questions there or are there any other places they can go for
1:18:18
other places they can go for
1:18:18
other places they can go for questions they can simply email to kids
1:18:22
questions they can simply email to kids
1:18:22
questions they can simply email to kids sorry uh they can simply email to
1:18:23
sorry uh they can simply email to
1:18:23
sorry uh they can simply email to kickstarter at opencv.org
1:18:25
kickstarter at opencv.org
1:18:26
kickstarter at opencv.org and uh somebody will respond sorry
1:18:29
and uh somebody will respond sorry
1:18:29
and uh somebody will respond sorry stefano now a quick question for you
1:18:31
stefano now a quick question for you
1:18:31
stefano now a quick question for you actually
1:18:32
actually
1:18:32
actually is that the opencv ai kit available
1:18:34
is that the opencv ai kit available
1:18:34
is that the opencv ai kit available globally
1:18:35
globally
1:18:35
globally i'm asking because hololens is not for
1:18:37
i'm asking because hololens is not for
1:18:37
i'm asking because hololens is not for some reason microsoft decided to
1:18:40
some reason microsoft decided to
1:18:40
some reason microsoft decided to open to deliver to ship hololens
1:18:43
open to deliver to ship hololens
1:18:43
open to deliver to ship hololens to some countries first us first and
1:18:45
to some countries first us first and
1:18:45
to some countries first us first and then others
1:18:46
then others
1:18:46
then others and then going global yes
1:18:50
and then going global yes
1:18:50
and then going global yes it is available there are obviously some
1:18:52
it is available there are obviously some
1:18:52
it is available there are obviously some countries which are restricted like
1:18:54
countries which are restricted like
1:18:54
countries which are restricted like north korea etc you cannot ship there
1:18:57
north korea etc you cannot ship there
1:18:57
north korea etc you cannot ship there and i think iran is also in that
1:18:59
and i think iran is also in that
1:18:59
and i think iran is also in that category but other than that
1:19:01
category but other than that
1:19:01
category but other than that uh we ship everywhere there is only one
1:19:03
uh we ship everywhere there is only one
1:19:03
uh we ship everywhere there is only one catch that some
1:19:05
catch that some
1:19:05
catch that some in some countries the duties are pretty
1:19:07
in some countries the duties are pretty
1:19:07
in some countries the duties are pretty high something like 20 30 percent
1:19:09
high something like 20 30 percent
1:19:09
high something like 20 30 percent so that people have to be uh you know
1:19:12
so that people have to be uh you know
1:19:12
so that people have to be uh you know cognizant about
1:19:13
cognizant about
1:19:13
cognizant about that the price could be much higher than
1:19:16
that the price could be much higher than
1:19:16
that the price could be much higher than uh what they anticipate
1:19:18
uh what they anticipate
1:19:18
uh what they anticipate sure good to know yeah thank you very
1:19:20
sure good to know yeah thank you very
1:19:20
sure good to know yeah thank you very much um
1:19:22
much um
1:19:22
much um um yeah and as mentioned uh if you have
1:19:24
um yeah and as mentioned uh if you have
1:19:24
um yeah and as mentioned uh if you have any questions for us
1:19:25
any questions for us
1:19:25
any questions for us uh as a global ai community feel free to
1:19:27
uh as a global ai community feel free to
1:19:27
uh as a global ai community feel free to ask them on twitter
1:19:28
ask them on twitter
1:19:28
ask them on twitter on our live chat uh on youtube or other
1:19:31
on our live chat uh on youtube or other
1:19:31
on our live chat uh on youtube or other places
1:19:32
places
1:19:32
places and to remember for the best question
1:19:34
and to remember for the best question
1:19:34
and to remember for the best question question we're giving away a 50
1:19:36
question we're giving away a 50
1:19:36
question we're giving away a 50 gift card from amazon uh c sharp corner
1:19:39
gift card from amazon uh c sharp corner
1:19:39
gift card from amazon uh c sharp corner has been uh
1:19:41
has been uh
1:19:41
has been uh friendly to uh to us and and they
1:19:43
friendly to uh to us and and they
1:19:43
friendly to uh to us and and they promised us
1:19:44
promised us
1:19:44
promised us to give away a 50 gift card every
1:19:46
to give away a 50 gift card every
1:19:46
to give away a 50 gift card every episode so that's uh good to know
1:19:48
episode so that's uh good to know
1:19:48
episode so that's uh good to know uh and um i haven't checked
1:19:51
uh and um i haven't checked
1:19:51
uh and um i haven't checked actually alicia have you seen any
1:19:53
actually alicia have you seen any
1:19:53
actually alicia have you seen any pictures yet on uh on twitter for the
1:19:56
pictures yet on uh on twitter for the
1:19:56
pictures yet on uh on twitter for the oculus quest uh device
1:19:59
oculus quest uh device
1:19:59
oculus quest uh device um i haven't been checking because i
1:20:02
um i haven't been checking because i
1:20:02
um i haven't been checking because i have
1:20:03
have
1:20:03
have been enraptured by this conversation
1:20:06
been enraptured by this conversation
1:20:06
been enraptured by this conversation about
1:20:07
about
1:20:07
about the ai toolkit and you know i i see all
1:20:10
the ai toolkit and you know i i see all
1:20:10
the ai toolkit and you know i i see all the interests
1:20:11
the interests
1:20:11
the interests and i'm sure everybody who saw that
1:20:13
and i'm sure everybody who saw that
1:20:13
and i'm sure everybody who saw that episode is gonna go out and get one for
1:20:15
episode is gonna go out and get one for
1:20:15
episode is gonna go out and get one for christmas
1:20:16
christmas
1:20:16
christmas um
1:20:28
[Laughter]
1:20:31
[Laughter]
1:20:31
[Laughter] so who's the next speaker alicia
1:20:34
so who's the next speaker alicia
1:20:34
so who's the next speaker alicia yeah so next up is stefano tempesta and
1:20:37
yeah so next up is stefano tempesta and
1:20:37
yeah so next up is stefano tempesta and he's here to talk to us about the
1:20:39
he's here to talk to us about the
1:20:39
he's here to talk to us about the hololens
1:20:40
hololens
1:20:40
hololens so virtual eye version i've heard
1:20:44
so virtual eye version i've heard
1:20:44
so virtual eye version i've heard so that that's that connects nicely
1:20:46
so that that's that connects nicely
1:20:46
so that that's that connects nicely together with with the stuff that satya
1:20:47
together with with the stuff that satya
1:20:48
together with with the stuff that satya talked before
1:20:49
talked before
1:20:49
talked before in the previous section um uh giving
1:20:51
in the previous section um uh giving
1:20:51
in the previous section um uh giving actually people
1:20:52
actually people
1:20:52
actually people who are blind or sort of an eye vision
1:20:54
who are blind or sort of an eye vision
1:20:54
who are blind or sort of an eye vision i'm i'm curious
1:20:55
i'm i'm curious
1:20:55
i'm i'm curious so um yeah sure show us what you've got
1:20:59
so um yeah sure show us what you've got
1:20:59
so um yeah sure show us what you've got to share
1:21:00
to share
1:21:00
to share yeah no yes thank you and look it really
1:21:03
yeah no yes thank you and look it really
1:21:03
yeah no yes thank you and look it really does connect really well with what satya
1:21:05
does connect really well with what satya
1:21:05
does connect really well with what satya said there i i'll give a
1:21:07
said there i i'll give a
1:21:07
said there i i'll give a another angle is that you know the the
1:21:09
another angle is that you know the the
1:21:09
another angle is that you know the the the cloud angle
1:21:11
the cloud angle
1:21:11
the cloud angle satya said the very important thing the
1:21:14
satya said the very important thing the
1:21:14
satya said the very important thing the cost of cloud
1:21:15
cost of cloud
1:21:15
cost of cloud can skyrocket if it's not under control
1:21:18
can skyrocket if it's not under control
1:21:18
can skyrocket if it's not under control and
1:21:19
and
1:21:19
and especially if you use it for training or
1:21:22
especially if you use it for training or
1:21:22
especially if you use it for training or using gpus any sort of things
1:21:24
using gpus any sort of things
1:21:24
using gpus any sort of things obviously there are ways around it to
1:21:27
obviously there are ways around it to
1:21:27
obviously there are ways around it to optimize
1:21:28
optimize
1:21:28
optimize a cost and performance at the same time
1:21:31
a cost and performance at the same time
1:21:31
a cost and performance at the same time but what i want to show you now is a
1:21:33
but what i want to show you now is a
1:21:33
but what i want to show you now is a combination of hardware or
1:21:35
combination of hardware or
1:21:35
combination of hardware or lens but not only as we will see towards
1:21:38
lens but not only as we will see towards
1:21:38
lens but not only as we will see towards the end
1:21:39
the end
1:21:39
the end spoiler and and also the possibility to
1:21:43
spoiler and and also the possibility to
1:21:43
spoiler and and also the possibility to access a cloud services that are hosted
1:21:47
access a cloud services that are hosted
1:21:47
access a cloud services that are hosted in azure that are
1:21:48
in azure that are
1:21:48
in azure that are completely on a server-less platform
1:21:50
completely on a server-less platform
1:21:50
completely on a server-less platform that
1:21:51
that
1:21:52
that run is opposite to the edge so
1:21:55
run is opposite to the edge so
1:21:55
run is opposite to the edge so for the audience here today is the great
1:21:57
for the audience here today is the great
1:21:57
for the audience here today is the great opportunity to hear
1:21:59
opportunity to hear
1:21:59
opportunity to hear both opportunities you build at the edge
1:22:02
both opportunities you build at the edge
1:22:02
both opportunities you build at the edge you build in the cloud and then it's up
1:22:04
you build in the cloud and then it's up
1:22:04
you build in the cloud and then it's up to you you really choose
1:22:06
to you you really choose
1:22:06
to you you really choose what is more convenient according to
1:22:08
what is more convenient according to
1:22:08
what is more convenient according to your specific needs and your business
1:22:10
your specific needs and your business
1:22:10
your specific needs and your business model
1:22:11
model
1:22:11
model so let me go and share my screen i have
1:22:14
so let me go and share my screen i have
1:22:14
so let me go and share my screen i have a few slides and then i want to show you
1:22:16
a few slides and then i want to show you
1:22:16
a few slides and then i want to show you also something
1:22:17
also something
1:22:17
also something in action uh this is a technology not
1:22:21
in action uh this is a technology not
1:22:21
in action uh this is a technology not that
1:22:21
that
1:22:21
that um is meant to
1:22:25
um is meant to
1:22:25
um is meant to describe world and people around
1:22:29
describe world and people around
1:22:29
describe world and people around the visually impaired
1:22:32
the visually impaired
1:22:32
the visually impaired i invite you to watch this video for a
1:22:34
i invite you to watch this video for a
1:22:34
i invite you to watch this video for a little moment
1:22:35
little moment
1:22:35
little moment all the technology of virtual reality
1:22:37
all the technology of virtual reality
1:22:37
all the technology of virtual reality and augmented reality is about
1:22:39
and augmented reality is about
1:22:39
and augmented reality is about acquiring the information about what's
1:22:41
acquiring the information about what's
1:22:41
acquiring the information about what's out there in the scene and then
1:22:42
out there in the scene and then
1:22:42
out there in the scene and then converting it to
1:22:43
converting it to
1:22:44
converting it to other uses all these things are coming
1:22:46
other uses all these things are coming
1:22:46
other uses all these things are coming together largely as a result of consumer
1:22:48
together largely as a result of consumer
1:22:48
together largely as a result of consumer technology not because someone wanted to
1:22:50
technology not because someone wanted to
1:22:50
technology not because someone wanted to build a device for the blind
1:22:52
build a device for the blind
1:22:52
build a device for the blind and we're just riding that wave and
1:22:54
and we're just riding that wave and
1:22:54
and we're just riding that wave and combining these different
1:22:55
combining these different
1:22:55
combining these different devices and concepts to make a new thing
1:22:58
devices and concepts to make a new thing
1:22:58
devices and concepts to make a new thing that never existed before
1:23:00
that never existed before
1:23:00
that never existed before the hololens creates a 3d mesh of the
1:23:02
the hololens creates a 3d mesh of the
1:23:02
the hololens creates a 3d mesh of the physical space
1:23:03
physical space
1:23:03
physical space to overlay virtual information on top of
1:23:06
to overlay virtual information on top of
1:23:06
to overlay virtual information on top of physical space what we did is to
1:23:10
physical space what we did is to
1:23:10
physical space what we did is to try to take advantage of that but
1:23:13
try to take advantage of that but
1:23:13
try to take advantage of that but overlay
1:23:14
overlay
1:23:14
overlay sound or auditory objects on top of
1:23:17
sound or auditory objects on top of
1:23:17
sound or auditory objects on top of physical scenes
1:23:18
physical scenes
1:23:18
physical scenes imagine you are in a world where all the
1:23:21
imagine you are in a world where all the
1:23:21
imagine you are in a world where all the objects around you have voices
1:23:23
objects around you have voices
1:23:23
objects around you have voices when you first walk into a new space you
1:23:25
when you first walk into a new space you
1:23:25
when you first walk into a new space you can simply have the objects
1:23:27
can simply have the objects
1:23:27
can simply have the objects call themselves out one at a time from
1:23:29
call themselves out one at a time from
1:23:29
call themselves out one at a time from left to right
1:23:30
left to right
1:23:30
left to right floor lamp laptop picture and you can
1:23:33
floor lamp laptop picture and you can
1:23:33
floor lamp laptop picture and you can switch to spotlight mode for example and
1:23:35
switch to spotlight mode for example and
1:23:35
switch to spotlight mode for example and point your head towards different
1:23:37
point your head towards different
1:23:37
point your head towards different objects in the scene
1:23:39
objects in the scene
1:23:39
objects in the scene they get activated and speak their names
1:23:41
they get activated and speak their names
1:23:41
they get activated and speak their names to you
1:23:42
to you
1:23:42
to you it tries to communicate the visual world
1:23:44
it tries to communicate the visual world
1:23:44
it tries to communicate the visual world at the cognitive level at the level of
1:23:46
at the cognitive level at the level of
1:23:46
at the cognitive level at the level of language
1:23:47
language
1:23:47
language at the level of understanding navigation
1:23:49
at the level of understanding navigation
1:23:49
at the level of understanding navigation started
1:23:50
started
1:23:50
started follow me follow me essentially we
1:23:53
follow me follow me essentially we
1:23:53
follow me follow me essentially we invented this virtual guide
1:23:55
invented this virtual guide
1:23:55
invented this virtual guide it always stays a few steps ahead of the
1:23:57
it always stays a few steps ahead of the
1:23:57
it always stays a few steps ahead of the user it calls out
1:23:59
user it calls out
1:23:59
user it calls out follow me up two flights of stairs
1:24:01
follow me up two flights of stairs
1:24:01
follow me up two flights of stairs around a few corners a long stretch of
1:24:03
around a few corners a long stretch of
1:24:03
around a few corners a long stretch of corridor and then a final turn into my
1:24:06
corridor and then a final turn into my
1:24:06
corridor and then a final turn into my office
1:24:07
office
1:24:07
office that was pretty cool we have seven
1:24:11
that was pretty cool we have seven
1:24:11
that was pretty cool we have seven blind subjects and they all managed to
1:24:13
blind subjects and they all managed to
1:24:13
blind subjects and they all managed to do the task to my second floor office
1:24:15
do the task to my second floor office
1:24:15
do the task to my second floor office the first trial
1:24:16
the first trial
1:24:16
the first trial it was very gratifying to see that a
1:24:18
it was very gratifying to see that a
1:24:18
it was very gratifying to see that a blind person could come to the lab put
1:24:20
blind person could come to the lab put
1:24:20
blind person could come to the lab put on our device
1:24:21
on our device
1:24:21
on our device and make their way through the building
1:24:23
and make their way through the building
1:24:23
and make their way through the building without further assistance
1:24:26
without further assistance
1:24:26
without further assistance the sky's the limit for what kind of
1:24:28
the sky's the limit for what kind of
1:24:28
the sky's the limit for what kind of functionalities you want to build
1:24:29
functionalities you want to build
1:24:30
functionalities you want to build into a device like that because it's
1:24:32
into a device like that because it's
1:24:32
into a device like that because it's essentially a software problem
1:24:34
essentially a software problem
1:24:34
essentially a software problem we'd like to see this device used and
1:24:36
we'd like to see this device used and
1:24:36
we'd like to see this device used and offered at the entrance of large spaces
1:24:39
offered at the entrance of large spaces
1:24:39
offered at the entrance of large spaces more or less the way you would adopt an
1:24:41
more or less the way you would adopt an
1:24:41
more or less the way you would adopt an audio guide when you walk into a museum
1:24:43
audio guide when you walk into a museum
1:24:43
audio guide when you walk into a museum plug the headphones in and off you go
1:24:55
well yes this is a great use of
1:24:56
well yes this is a great use of
1:24:56
well yes this is a great use of technology uh
1:24:58
technology uh
1:24:58
technology uh the combination of the hololens device
1:25:00
the combination of the hololens device
1:25:00
the combination of the hololens device the combination of
1:25:01
the combination of
1:25:01
the combination of uh some services that i will show you in
1:25:04
uh some services that i will show you in
1:25:04
uh some services that i will show you in a moment so
1:25:06
a moment so
1:25:06
a moment so how about we dig inside the technology
1:25:09
how about we dig inside the technology
1:25:09
how about we dig inside the technology and see
1:25:09
and see
1:25:09
and see how something like these can be built
1:25:12
how something like these can be built
1:25:12
how something like these can be built in the azure platform now if you were
1:25:16
in the azure platform now if you were
1:25:16
in the azure platform now if you were wondering this is a an experience done
1:25:19
wondering this is a an experience done
1:25:19
wondering this is a an experience done by caltech
1:25:20
by caltech
1:25:20
by caltech yes that's exactly the university of
1:25:22
yes that's exactly the university of
1:25:22
yes that's exactly the university of sheldon cooper
1:25:25
sheldon cooper
1:25:25
sheldon cooper all right hololens hololens is an
1:25:28
all right hololens hololens is an
1:25:28
all right hololens hololens is an unfettered device
1:25:30
unfettered device
1:25:30
unfettered device it means that it can run completely on
1:25:32
it means that it can run completely on
1:25:32
it means that it can run completely on his own
1:25:33
his own
1:25:33
his own uh without requiring connectivity
1:25:36
uh without requiring connectivity
1:25:36
uh without requiring connectivity without requiring a
1:25:37
without requiring a
1:25:37
without requiring a been plugged and without being connected
1:25:40
been plugged and without being connected
1:25:40
been plugged and without being connected wired or wireless to any other device
1:25:43
wired or wireless to any other device
1:25:43
wired or wireless to any other device like a laptop or a mobile phone
1:25:45
like a laptop or a mobile phone
1:25:46
like a laptop or a mobile phone yes cameras yes speakers yes sensors
1:25:49
yes cameras yes speakers yes sensors
1:25:49
yes cameras yes speakers yes sensors and yes indeed it is
1:25:52
and yes indeed it is
1:25:52
and yes indeed it is need a little computer in his way he has
1:25:55
need a little computer in his way he has
1:25:55
need a little computer in his way he has a gpu
1:25:56
a gpu
1:25:56
a gpu as a holographic processing unit storage
1:26:00
as a holographic processing unit storage
1:26:00
as a holographic processing unit storage and different ways of connecting the
1:26:03
and different ways of connecting the
1:26:03
and different ways of connecting the outside world
1:26:05
outside world
1:26:05
outside world as i briefly mentioned before uh
1:26:08
as i briefly mentioned before uh
1:26:08
as i briefly mentioned before uh is not that lightweight so it
1:26:11
is not that lightweight so it
1:26:11
is not that lightweight so it you know wearing it for a long time is
1:26:14
you know wearing it for a long time is
1:26:14
you know wearing it for a long time is probably not the best experience you can
1:26:16
probably not the best experience you can
1:26:16
probably not the best experience you can have
1:26:16
have
1:26:16
have and there is development in this space
1:26:18
and there is development in this space
1:26:18
and there is development in this space which i will mention towards the end
1:26:20
which i will mention towards the end
1:26:20
which i will mention towards the end and also the duration of the battery
1:26:22
and also the duration of the battery
1:26:22
and also the duration of the battery probably isn't fantastic so you can
1:26:25
probably isn't fantastic so you can
1:26:25
probably isn't fantastic so you can really use it comfortably for a couple
1:26:27
really use it comfortably for a couple
1:26:27
really use it comfortably for a couple of hours after that
1:26:28
of hours after that
1:26:28
of hours after that you will probably want to take a break
1:26:30
you will probably want to take a break
1:26:30
you will probably want to take a break and
1:26:31
and
1:26:31
and as being a a computer in his own
1:26:36
as being a a computer in his own
1:26:36
as being a a computer in his own iran's a special version of a windows
1:26:39
iran's a special version of a windows
1:26:39
iran's a special version of a windows operating system
1:26:40
operating system
1:26:40
operating system for holographic functionality age as a
1:26:43
for holographic functionality age as a
1:26:43
for holographic functionality age as a browser
1:26:44
browser
1:26:44
browser and a few other apps in the dynamics 365
1:26:47
and a few other apps in the dynamics 365
1:26:47
and a few other apps in the dynamics 365 for remote assistance to enable
1:26:51
for remote assistance to enable
1:26:51
for remote assistance to enable uh technicians on the field to operate
1:26:53
uh technicians on the field to operate
1:26:54
uh technicians on the field to operate with a subject matter expert
1:26:55
with a subject matter expert
1:26:55
with a subject matter expert back in the office to assist over
1:26:59
back in the office to assist over
1:26:59
back in the office to assist over machineries and so on s cortana for
1:27:03
machineries and so on s cortana for
1:27:03
machineries and so on s cortana for voice recognition and a few other things
1:27:06
voice recognition and a few other things
1:27:06
voice recognition and a few other things so
1:27:06
so
1:27:06
so let's say okay i want to build an app
1:27:09
let's say okay i want to build an app
1:27:09
let's say okay i want to build an app for hololens
1:27:10
for hololens
1:27:10
for hololens well at a very high level the way an app
1:27:13
well at a very high level the way an app
1:27:13
well at a very high level the way an app for all lens works is you have to
1:27:15
for all lens works is you have to
1:27:15
for all lens works is you have to consider it as
1:27:18
consider it as
1:27:18
consider it as you will consider any other a computer
1:27:21
you will consider any other a computer
1:27:21
you will consider any other a computer to connect to as any other device
1:27:23
to connect to as any other device
1:27:23
to connect to as any other device so you build a you know
1:27:26
so you build a you know
1:27:26
so you build a you know the holographic app that connect
1:27:29
the holographic app that connect
1:27:29
the holographic app that connect the device to some services that will
1:27:33
the device to some services that will
1:27:33
the device to some services that will process all this flowing information
1:27:35
process all this flowing information
1:27:35
process all this flowing information that you retrieve of this data this
1:27:37
that you retrieve of this data this
1:27:37
that you retrieve of this data this telemetry data you've seen
1:27:39
telemetry data you've seen
1:27:39
telemetry data you've seen in the video before that there is a
1:27:42
in the video before that there is a
1:27:42
in the video before that there is a scanning
1:27:43
scanning
1:27:43
scanning of the environment around and as we will
1:27:46
of the environment around and as we will
1:27:46
of the environment around and as we will see in a moment
1:27:47
see in a moment
1:27:47
see in a moment there are capabilities around the
1:27:49
there are capabilities around the
1:27:49
there are capabilities around the detecting faces the letting people
1:27:52
detecting faces the letting people
1:27:52
detecting faces the letting people identifying these people as well so
1:27:55
identifying these people as well so
1:27:55
identifying these people as well so working
1:27:55
working
1:27:55
working against a database of celebrities and
1:27:59
against a database of celebrities and
1:27:59
against a database of celebrities and recognize
1:28:01
recognize
1:28:01
recognize these celebrities by their face or
1:28:04
these celebrities by their face or
1:28:04
these celebrities by their face or pre-training a model of your friends and
1:28:07
pre-training a model of your friends and
1:28:07
pre-training a model of your friends and family
1:28:07
family
1:28:08
family and colleagues and using face detection
1:28:11
and colleagues and using face detection
1:28:11
and colleagues and using face detection and identification
1:28:12
and identification
1:28:12
and identification as well now all these
1:28:16
as well now all these
1:28:16
as well now all these is possible because of some services
1:28:19
is possible because of some services
1:28:19
is possible because of some services that we use and today i want to speak to
1:28:21
that we use and today i want to speak to
1:28:21
that we use and today i want to speak to a couple of services they are called
1:28:23
a couple of services they are called
1:28:23
a couple of services they are called azure cognitive services and they are
1:28:26
azure cognitive services and they are
1:28:26
azure cognitive services and they are comp is a serverless
1:28:27
comp is a serverless
1:28:28
comp is a serverless platform completely hosted in the azure
1:28:30
platform completely hosted in the azure
1:28:30
platform completely hosted in the azure cloud
1:28:31
cloud
1:28:31
cloud with the potential also to run at the
1:28:33
with the potential also to run at the
1:28:33
with the potential also to run at the edge but the power of it
1:28:35
edge but the power of it
1:28:36
edge but the power of it is that all these services are all
1:28:38
is that all these services are all
1:28:38
is that all these services are all machine learning power
1:28:40
machine learning power
1:28:40
machine learning power power they are
1:28:43
power they are
1:28:43
power they are exposed as a rest api in the cloud in
1:28:46
exposed as a rest api in the cloud in
1:28:46
exposed as a rest api in the cloud in the azure cloud
1:28:47
the azure cloud
1:28:47
the azure cloud so you don't have to provision any
1:28:49
so you don't have to provision any
1:28:49
so you don't have to provision any infrastructure for it
1:28:51
infrastructure for it
1:28:51
infrastructure for it you just have to subscribe and start
1:28:53
you just have to subscribe and start
1:28:53
you just have to subscribe and start using it
1:28:54
using it
1:28:54
using it and the good news is that you can start
1:28:56
and the good news is that you can start
1:28:56
and the good news is that you can start using for free
1:28:57
using for free
1:28:57
using for free so there is a free tier that allows you
1:29:00
so there is a free tier that allows you
1:29:00
so there is a free tier that allows you to do
1:29:01
to do
1:29:01
to do a number of calls to the service without
1:29:03
a number of calls to the service without
1:29:03
a number of calls to the service without paying a single
1:29:05
paying a single
1:29:05
paying a single dollar and still
1:29:08
dollar and still
1:29:08
dollar and still but despite being strongly powered by
1:29:11
but despite being strongly powered by
1:29:12
but despite being strongly powered by machine learning you're not
1:29:13
machine learning you're not
1:29:13
machine learning you're not required to be a data scientist other
1:29:15
required to be a data scientist other
1:29:15
required to be a data scientist other engineer at all
1:29:17
engineer at all
1:29:17
engineer at all these are rest api they com they hide
1:29:20
these are rest api they com they hide
1:29:20
these are rest api they com they hide the complexity of it
1:29:22
the complexity of it
1:29:22
the complexity of it and then you have the possibility to uh
1:29:25
and then you have the possibility to uh
1:29:25
and then you have the possibility to uh use the consumer from any device from
1:29:28
use the consumer from any device from
1:29:28
use the consumer from any device from any application i'll show you
1:29:29
any application i'll show you
1:29:29
any application i'll show you a few in a moment so for example
1:29:32
a few in a moment so for example
1:29:32
a few in a moment so for example computer vision
1:29:33
computer vision
1:29:33
computer vision is probably one of the largest of these
1:29:36
is probably one of the largest of these
1:29:36
is probably one of the largest of these apis can do
1:29:38
apis can do
1:29:38
apis can do a detection of objects you can do
1:29:41
a detection of objects you can do
1:29:41
a detection of objects you can do detection of faces
1:29:42
detection of faces
1:29:42
detection of faces form ink recognition
1:29:45
form ink recognition
1:29:45
form ink recognition and as you can imagine being a sort of a
1:29:49
and as you can imagine being a sort of a
1:29:49
and as you can imagine being a sort of a serverless platform
1:29:51
serverless platform
1:29:51
serverless platform all these models you know with the
1:29:52
all these models you know with the
1:29:52
all these models you know with the images of any type
1:29:55
images of any type
1:29:55
images of any type they've been pre-trained
1:29:58
they've been pre-trained
1:29:58
they've been pre-trained by microsoft engineers who have spent
1:30:02
by microsoft engineers who have spent
1:30:02
by microsoft engineers who have spent uncountable number of hours to upload
1:30:05
uncountable number of hours to upload
1:30:05
uncountable number of hours to upload thousands and thousands of
1:30:07
thousands and thousands of
1:30:07
thousands and thousands of images and labeling all these images now
1:30:10
images and labeling all these images now
1:30:10
images and labeling all these images now this is great because you can use this
1:30:13
this is great because you can use this
1:30:13
this is great because you can use this the computer vision capability
1:30:14
the computer vision capability
1:30:14
the computer vision capability straight away without going through the
1:30:17
straight away without going through the
1:30:17
straight away without going through the training process
1:30:19
training process
1:30:19
training process so here the cloud doesn't cost you a
1:30:22
so here the cloud doesn't cost you a
1:30:22
so here the cloud doesn't cost you a cent
1:30:22
cent
1:30:22
cent or so because you don't have to
1:30:25
or so because you don't have to
1:30:25
or so because you don't have to provision gpus
1:30:27
provision gpus
1:30:27
provision gpus and do any sort of training it's already
1:30:29
and do any sort of training it's already
1:30:29
and do any sort of training it's already been done for you
1:30:31
been done for you
1:30:31
been done for you and but still you have the possibility
1:30:34
and but still you have the possibility
1:30:34
and but still you have the possibility with custom vision to customize
1:30:37
with custom vision to customize
1:30:38
with custom vision to customize the um the object detection
1:30:41
the um the object detection
1:30:41
the um the object detection using your own model so you can increase
1:30:44
using your own model so you can increase
1:30:44
using your own model so you can increase the the accuracy of your computer vision
1:30:47
the the accuracy of your computer vision
1:30:47
the the accuracy of your computer vision capability
1:30:48
capability
1:30:48
capability by bringing your own data set now
1:30:51
by bringing your own data set now
1:30:51
by bringing your own data set now face detection is the one that i want to
1:30:54
face detection is the one that i want to
1:30:54
face detection is the one that i want to focus a bit today
1:30:55
focus a bit today
1:30:55
focus a bit today you know as an api to detect and
1:30:58
you know as an api to detect and
1:30:58
you know as an api to detect and identify
1:30:59
identify
1:30:59
identify people emotions and attributes and
1:31:02
people emotions and attributes and
1:31:02
people emotions and attributes and rather than showing these in a still
1:31:06
rather than showing these in a still
1:31:06
rather than showing these in a still image how about we go in the browser
1:31:10
image how about we go in the browser
1:31:10
image how about we go in the browser and uh super simple you just have to
1:31:12
and uh super simple you just have to
1:31:12
and uh super simple you just have to search
1:31:13
search
1:31:13
search for azure cognitive services it will end
1:31:16
for azure cognitive services it will end
1:31:16
for azure cognitive services it will end up
1:31:17
up
1:31:17
up on this page here where you can see all
1:31:19
on this page here where you can see all
1:31:19
on this page here where you can see all the different services that are
1:31:21
the different services that are
1:31:21
the different services that are available
1:31:22
available
1:31:22
available decision language speech vision in the
1:31:25
decision language speech vision in the
1:31:25
decision language speech vision in the one that i'm after okay
1:31:26
one that i'm after okay
1:31:26
one that i'm after okay and all the different services are
1:31:28
and all the different services are
1:31:28
and all the different services are available in here so click on
1:31:30
available in here so click on
1:31:30
available in here so click on face and you will have the possibility
1:31:33
face and you will have the possibility
1:31:33
face and you will have the possibility to understand
1:31:34
to understand
1:31:34
to understand how that works uh let me just switch
1:31:37
how that works uh let me just switch
1:31:37
how that works uh let me just switch here so we can get
1:31:38
here so we can get
1:31:38
here so we can get more attributes and as you can see there
1:31:41
more attributes and as you can see there
1:31:41
more attributes and as you can see there is a
1:31:43
is a
1:31:43
is a an object here a json object which is
1:31:46
an object here a json object which is
1:31:46
an object here a json object which is the response of the api
1:31:47
the response of the api
1:31:47
the response of the api that identifies the rectangle of the
1:31:49
that identifies the rectangle of the
1:31:50
that identifies the rectangle of the face or where the face
1:31:51
face or where the face
1:31:51
face or where the face is inside the picture and some
1:31:53
is inside the picture and some
1:31:53
is inside the picture and some attributes
1:31:54
attributes
1:31:54
attributes in here so i i have a i'll be a
1:31:58
in here so i i have a i'll be a
1:31:58
in here so i i have a i'll be a little simulator of the app that shows
1:32:01
little simulator of the app that shows
1:32:01
little simulator of the app that shows saturday's in action which i'm going to
1:32:03
saturday's in action which i'm going to
1:32:03
saturday's in action which i'm going to show you in a moment
1:32:04
show you in a moment
1:32:04
show you in a moment so let me open it is a
1:32:08
so let me open it is a
1:32:08
so let me open it is a is an emulate a simulator of all the
1:32:10
is an emulate a simulator of all the
1:32:10
is an emulate a simulator of all the different cognitive services
1:32:12
different cognitive services
1:32:12
different cognitive services okay so we'll focus on phase for today
1:32:15
okay so we'll focus on phase for today
1:32:15
okay so we'll focus on phase for today and as you can see there are some
1:32:17
and as you can see there are some
1:32:17
and as you can see there are some celebrities in there
1:32:19
celebrities in there
1:32:19
celebrities in there that are already been pre-classified
1:32:23
that are already been pre-classified
1:32:23
that are already been pre-classified and when you know i pick one of these
1:32:27
and when you know i pick one of these
1:32:27
and when you know i pick one of these i'm doing a call to this api which is
1:32:30
i'm doing a call to this api which is
1:32:30
i'm doing a call to this api which is super simple it's just
1:32:31
super simple it's just
1:32:31
super simple it's just a http call to arrest
1:32:35
a http call to arrest
1:32:35
a http call to arrest endpoint and i get back that json object
1:32:39
endpoint and i get back that json object
1:32:39
endpoint and i get back that json object which detects both images so both
1:32:42
which detects both images so both
1:32:42
which detects both images so both sorry both face without within an image
1:32:46
sorry both face without within an image
1:32:46
sorry both face without within an image and the coordinate of image one and
1:32:48
and the coordinate of image one and
1:32:48
and the coordinate of image one and image two
1:32:49
image two
1:32:49
image two and as you can see all the different
1:32:51
and as you can see all the different
1:32:51
and as you can see all the different attributes uh gender
1:32:53
attributes uh gender
1:32:53
attributes uh gender uh estimate age whether they're smiling
1:32:56
uh estimate age whether they're smiling
1:32:56
uh estimate age whether they're smiling whether they're wearing glasses
1:32:58
whether they're wearing glasses
1:32:58
whether they're wearing glasses any other face attributes i guess for
1:33:01
any other face attributes i guess for
1:33:01
any other face attributes i guess for the men
1:33:01
the men
1:33:02
the men we should get some values in here as you
1:33:04
we should get some values in here as you
1:33:04
we should get some values in here as you can see and
1:33:05
can see and
1:33:05
can see and emotions all these values are between
1:33:08
emotions all these values are between
1:33:08
emotions all these values are between zero and one you know where
1:33:09
zero and one you know where
1:33:10
zero and one you know where zero means and not detected at all and
1:33:12
zero means and not detected at all and
1:33:12
zero means and not detected at all and one
1:33:13
one
1:33:13
one happy family happy couple that uh have
1:33:16
happy family happy couple that uh have
1:33:16
happy family happy couple that uh have been detected and this obviously works
1:33:19
been detected and this obviously works
1:33:19
been detected and this obviously works uh you know
1:33:20
uh you know
1:33:20
uh you know if you upload a picture if you the take
1:33:22
if you upload a picture if you the take
1:33:22
if you upload a picture if you the take another picture so how about that now i
1:33:24
another picture so how about that now i
1:33:24
another picture so how about that now i need to stop
1:33:26
need to stop
1:33:26
need to stop my own video otherwise there is a
1:33:28
my own video otherwise there is a
1:33:28
my own video otherwise there is a conflict of camera
1:33:30
conflict of camera
1:33:30
conflict of camera and then i can take it here it's gonna
1:33:32
and then i can take it here it's gonna
1:33:32
and then i can take it here it's gonna be me
1:33:33
be me
1:33:33
be me hello there how about we do that phase
1:33:37
hello there how about we do that phase
1:33:38
hello there how about we do that phase all right and then confirm send it to
1:33:41
all right and then confirm send it to
1:33:41
all right and then confirm send it to the api
1:33:42
the api
1:33:42
the api it will come up here face detected
1:33:45
it will come up here face detected
1:33:46
it will come up here face detected and again gender age thank you
1:33:49
and again gender age thank you
1:33:49
and again gender age thank you very generous uh reading glasses
1:33:52
very generous uh reading glasses
1:33:52
very generous uh reading glasses and a few other attributes what am i he
1:33:55
and a few other attributes what am i he
1:33:55
and a few other attributes what am i he came out as a neutral
1:33:56
came out as a neutral
1:33:56
came out as a neutral okay i didn't do my angry face too much
1:34:00
okay i didn't do my angry face too much
1:34:00
okay i didn't do my angry face too much well this morning morning for me all
1:34:03
well this morning morning for me all
1:34:03
well this morning morning for me all right then
1:34:04
right then
1:34:04
right then the other thing you can do is that you
1:34:05
the other thing you can do is that you
1:34:05
the other thing you can do is that you can also detect a face you know so you
1:34:08
can also detect a face you know so you
1:34:08
can also detect a face you know so you can actually
1:34:09
can actually
1:34:09
can actually uh see uh who is in in in the picture
1:34:12
uh see uh who is in in in the picture
1:34:12
uh see uh who is in in in the picture now it's not gonna work with me i'm not
1:34:15
now it's not gonna work with me i'm not
1:34:15
now it's not gonna work with me i'm not a celebrity so i'm not in the database
1:34:17
a celebrity so i'm not in the database
1:34:17
a celebrity so i'm not in the database of vips
1:34:18
of vips
1:34:18
of vips yet i'm working on it but we can do
1:34:22
yet i'm working on it but we can do
1:34:22
yet i'm working on it but we can do something like that so i can invite here
1:34:24
something like that so i can invite here
1:34:24
something like that so i can invite here uh
1:34:24
uh
1:34:24
uh harry which is sitting next to me to
1:34:27
harry which is sitting next to me to
1:34:27
harry which is sitting next to me to take a picture harry do you mind taking
1:34:29
take a picture harry do you mind taking
1:34:29
take a picture harry do you mind taking a picture for
1:34:37
here we go her hair is coming
1:34:43
okay so send the picture to the api
1:34:47
okay so send the picture to the api
1:34:47
okay so send the picture to the api thank you harry that was generous of you
1:34:50
thank you harry that was generous of you
1:34:50
thank you harry that was generous of you face detected is still the same guy in
1:34:52
face detected is still the same guy in
1:34:52
face detected is still the same guy in here
1:34:53
here
1:34:53
here so if we find a match now the second
1:34:55
so if we find a match now the second
1:34:56
so if we find a match now the second part
1:34:56
part
1:34:56
part of the api is to detect and boom
1:34:59
of the api is to detect and boom
1:35:00
of the api is to detect and boom it came out here in red that the picture
1:35:03
it came out here in red that the picture
1:35:03
it came out here in red that the picture where the that person has been detected
1:35:07
where the that person has been detected
1:35:07
where the that person has been detected okay so that's the way it works
1:35:09
okay so that's the way it works
1:35:09
okay so that's the way it works obviously if i had
1:35:10
obviously if i had
1:35:10
obviously if i had a custom model i can
1:35:14
a custom model i can
1:35:14
a custom model i can upload pictures of my family friends uh
1:35:18
upload pictures of my family friends uh
1:35:18
upload pictures of my family friends uh pets why not it doesn't work only with
1:35:19
pets why not it doesn't work only with
1:35:20
pets why not it doesn't work only with people it works with any kind of object
1:35:22
people it works with any kind of object
1:35:22
people it works with any kind of object as long as they are labeled and then it
1:35:24
as long as they are labeled and then it
1:35:24
as long as they are labeled and then it will detect
1:35:25
will detect
1:35:25
will detect and recognize so this is the foundation
1:35:28
and recognize so this is the foundation
1:35:28
and recognize so this is the foundation this is how
1:35:28
this is how
1:35:28
this is how what we will do not to start
1:35:32
what we will do not to start
1:35:32
what we will do not to start building this application an application
1:35:35
building this application an application
1:35:35
building this application an application that
1:35:35
that
1:35:36
that identifies the world around you and
1:35:39
identifies the world around you and
1:35:39
identifies the world around you and detect people people if
1:35:42
detect people people if
1:35:42
detect people people if possible identify these people describe
1:35:46
possible identify these people describe
1:35:46
possible identify these people describe how they look like what they're doing
1:35:48
how they look like what they're doing
1:35:48
how they look like what they're doing and then translate these
1:35:50
and then translate these
1:35:50
and then translate these into a voice command and this is the
1:35:53
into a voice command and this is the
1:35:53
into a voice command and this is the second part of it
1:35:55
second part of it
1:35:55
second part of it in voice instructions i mean this is the
1:35:57
in voice instructions i mean this is the
1:35:57
in voice instructions i mean this is the second part of it
1:35:58
second part of it
1:35:58
second part of it is using the speech api which is again
1:36:02
is using the speech api which is again
1:36:02
is using the speech api which is again different flavor
1:36:03
different flavor
1:36:03
different flavor speech to text text-to-speech speech
1:36:06
speech to text text-to-speech speech
1:36:06
speech to text text-to-speech speech translation and speaker recognition
1:36:08
translation and speaker recognition
1:36:08
translation and speaker recognition so recognize the voice of a person
1:36:11
so recognize the voice of a person
1:36:11
so recognize the voice of a person and identify that person so from the
1:36:14
and identify that person so from the
1:36:14
and identify that person so from the tone of voice
1:36:15
tone of voice
1:36:15
tone of voice it will be able to say hey this is john
1:36:17
it will be able to say hey this is john
1:36:17
it will be able to say hey this is john this is mary
1:36:19
this is mary
1:36:19
this is mary and uh and again this has to work
1:36:21
and uh and again this has to work
1:36:21
and uh and again this has to work against
1:36:22
against
1:36:22
against a pre-loaded database of voices that has
1:36:25
a pre-loaded database of voices that has
1:36:25
a pre-loaded database of voices that has been trained
1:36:26
been trained
1:36:26
been trained and you can personalize this with your
1:36:28
and you can personalize this with your
1:36:28
and you can personalize this with your own
1:36:29
own
1:36:29
own family and friends and colleagues and so
1:36:31
family and friends and colleagues and so
1:36:31
family and friends and colleagues and so on so in
1:36:33
on so in
1:36:33
on so in the specific technology that i'm after
1:36:35
the specific technology that i'm after
1:36:35
the specific technology that i'm after now is a text-to-speech
1:36:37
now is a text-to-speech
1:36:37
now is a text-to-speech because all the data that has been
1:36:39
because all the data that has been
1:36:39
because all the data that has been detected
1:36:40
detected
1:36:40
detected by the vision of the computer vision
1:36:42
by the vision of the computer vision
1:36:42
by the vision of the computer vision technology
1:36:43
technology
1:36:43
technology is translated into a text
1:36:47
is translated into a text
1:36:47
is translated into a text and this text now has to be voiced to
1:36:50
and this text now has to be voiced to
1:36:50
and this text now has to be voiced to the
1:36:51
the
1:36:51
the uh to to the visually impaired people
1:36:54
uh to to the visually impaired people
1:36:54
uh to to the visually impaired people that
1:36:54
that
1:36:54
that are using the the device and the way it
1:36:57
are using the the device and the way it
1:36:58
are using the the device and the way it works again there is
1:36:59
works again there is
1:36:59
works again there is something in there that we can test
1:37:02
something in there that we can test
1:37:02
something in there that we can test immediately so let me go back to my
1:37:05
immediately so let me go back to my
1:37:05
immediately so let me go back to my uh to my cognitive services go to speech
1:37:09
uh to my cognitive services go to speech
1:37:09
uh to my cognitive services go to speech this is going to be text to speech and
1:37:12
this is going to be text to speech and
1:37:12
this is going to be text to speech and we can test this directly in the browser
1:37:14
we can test this directly in the browser
1:37:14
we can test this directly in the browser here
1:37:15
here
1:37:15
here so you see i have a different voices
1:37:17
so you see i have a different voices
1:37:17
so you see i have a different voices that i can pick
1:37:18
that i can pick
1:37:18
that i can pick from arabic all the way to tamil
1:37:22
from arabic all the way to tamil
1:37:22
from arabic all the way to tamil and different accent as well i can pick
1:37:26
and different accent as well i can pick
1:37:26
and different accent as well i can pick australian english it defaulted to this
1:37:28
australian english it defaulted to this
1:37:28
australian english it defaulted to this because as i mentioned
1:37:29
because as i mentioned
1:37:29
because as i mentioned based in australia but there are also
1:37:31
based in australia but there are also
1:37:31
based in australia but there are also different other flavors so how about
1:37:33
different other flavors so how about
1:37:33
different other flavors so how about we take english u.s different voices
1:37:36
we take english u.s different voices
1:37:36
we take english u.s different voices female and male voice style of the voice
1:37:40
female and male voice style of the voice
1:37:40
female and male voice style of the voice you want something more cheerful
1:37:42
you want something more cheerful
1:37:42
you want something more cheerful something more professional and then you
1:37:45
something more professional and then you
1:37:45
something more professional and then you can also tune
1:37:46
can also tune
1:37:46
can also tune the speed and the pitch so we can say
1:37:49
the speed and the pitch so we can say
1:37:50
the speed and the pitch so we can say hello everybody welcome
1:37:54
hello everybody welcome
1:37:54
hello everybody welcome to the global ai
1:37:58
to the global ai
1:37:58
to the global ai community all right and
1:38:01
community all right and
1:38:01
community all right and just listen
1:38:05
hello everybody welcome to the global ai
1:38:08
hello everybody welcome to the global ai
1:38:08
hello everybody welcome to the global ai community
1:38:09
community
1:38:09
community and obviously different different accent
1:38:11
and obviously different different accent
1:38:11
and obviously different different accent will produce different voice
1:38:13
will produce different voice
1:38:13
will produce different voice different intonation and if you if you
1:38:16
different intonation and if you if you
1:38:16
different intonation and if you if you want to do the cheap monk effect
1:38:18
want to do the cheap monk effect
1:38:18
want to do the cheap monk effect just tune up on the picture
1:38:23
just tune up on the picture
1:38:23
just tune up on the picture hello everybody welcome to the global ai
1:38:26
hello everybody welcome to the global ai
1:38:26
hello everybody welcome to the global ai community
1:38:26
community
1:38:26
community and you're gonna have fun as much as you
1:38:28
and you're gonna have fun as much as you
1:38:28
and you're gonna have fun as much as you want okay so this is all available in
1:38:30
want okay so this is all available in
1:38:30
want okay so this is all available in there
1:38:31
there
1:38:31
there you can start completely for free there
1:38:34
you can start completely for free there
1:38:34
you can start completely for free there are
1:38:36
are
1:38:36
are you start paying for the services after
1:38:38
you start paying for the services after
1:38:38
you start paying for the services after a significant
1:38:39
a significant
1:38:39
a significant volume of data and this is great to
1:38:42
volume of data and this is great to
1:38:42
volume of data and this is great to enable developers
1:38:44
enable developers
1:38:44
enable developers without provisioning any infrastructure
1:38:46
without provisioning any infrastructure
1:38:46
without provisioning any infrastructure without
1:38:47
without
1:38:47
without doing uploading any models without doing
1:38:51
doing uploading any models without doing
1:38:51
doing uploading any models without doing any training
1:38:52
any training
1:38:52
any training you just get started and focus on
1:38:55
you just get started and focus on
1:38:55
you just get started and focus on building your application rather than
1:38:58
building your application rather than
1:38:58
building your application rather than the ai part of it
1:39:00
the ai part of it
1:39:00
the ai part of it which may not be accessible to everybody
1:39:04
which may not be accessible to everybody
1:39:04
which may not be accessible to everybody so oh too much
1:39:07
so oh too much
1:39:07
so oh too much in in summer in case you were wondering
1:39:10
in in summer in case you were wondering
1:39:10
in in summer in case you were wondering this text speech
1:39:11
this text speech
1:39:11
this text speech is able to synthesize a speech into
1:39:14
is able to synthesize a speech into
1:39:14
is able to synthesize a speech into live into speakers or into an audio file
1:39:17
live into speakers or into an audio file
1:39:17
live into speakers or into an audio file or into a memory stream
1:39:19
or into a memory stream
1:39:19
or into a memory stream so the memory stream it will be useful
1:39:22
so the memory stream it will be useful
1:39:22
so the memory stream it will be useful when connecting to a device
1:39:24
when connecting to a device
1:39:24
when connecting to a device so you can the the api can synthesize
1:39:27
so you can the the api can synthesize
1:39:27
so you can the the api can synthesize the speech then you store it in memory
1:39:30
the speech then you store it in memory
1:39:30
the speech then you store it in memory and then
1:39:30
and then
1:39:30
and then send this stream to to the device to
1:39:33
send this stream to to the device to
1:39:33
send this stream to to the device to whatever speakers all lenses you're
1:39:35
whatever speakers all lenses you're
1:39:35
whatever speakers all lenses you're using
1:39:36
using
1:39:36
using it can even do long asynchronous
1:39:39
it can even do long asynchronous
1:39:39
it can even do long asynchronous audios and it will work completely
1:39:44
audios and it will work completely
1:39:44
audios and it will work completely asynchronously so then you know just
1:39:46
asynchronously so then you know just
1:39:46
asynchronously so then you know just render
1:39:47
render
1:39:47
render the video over 10 minutes typically and
1:39:49
the video over 10 minutes typically and
1:39:50
the video over 10 minutes typically and then get back
1:39:50
then get back
1:39:50
then get back when it's done and saved as a support
1:39:53
when it's done and saved as a support
1:39:54
when it's done and saved as a support for standard and neural voices so voices
1:39:56
for standard and neural voices so voices
1:39:56
for standard and neural voices so voices that
1:39:57
that
1:39:57
that improve and learn over time support for
1:40:00
improve and learn over time support for
1:40:00
improve and learn over time support for over 45 languages
1:40:02
over 45 languages
1:40:02
over 45 languages different boys male and female and also
1:40:04
different boys male and female and also
1:40:04
different boys male and female and also different other flavors
1:40:06
different other flavors
1:40:06
different other flavors there is an sdk a software development
1:40:08
there is an sdk a software development
1:40:08
there is an sdk a software development kit for
1:40:09
kit for
1:40:09
kit for a few programming languages so that
1:40:12
a few programming languages so that
1:40:12
a few programming languages so that means
1:40:13
means
1:40:13
means that you don't have to code directly
1:40:19
the api so you don't have to do an http
1:40:22
the api so you don't have to do an http
1:40:22
the api so you don't have to do an http request
1:40:23
request
1:40:23
request what you can do is a a single call to
1:40:26
what you can do is a a single call to
1:40:26
what you can do is a a single call to to uh two two lines of of
1:40:30
to uh two two lines of of
1:40:30
to uh two two lines of of code actually you know what i can show
1:40:32
code actually you know what i can show
1:40:32
code actually you know what i can show you here because i have the project open
1:40:33
you here because i have the project open
1:40:33
you here because i have the project open in front of me
1:40:35
in front of me
1:40:35
in front of me which is gonna be somewhere like um
1:40:38
which is gonna be somewhere like um
1:40:38
which is gonna be somewhere like um hey this is speech this is
1:40:40
hey this is speech this is
1:40:40
hey this is speech this is text-to-speech so i'll just show you
1:40:42
text-to-speech so i'll just show you
1:40:42
text-to-speech so i'll just show you very briefly how that looks like you
1:40:45
very briefly how that looks like you
1:40:45
very briefly how that looks like you create a speech config
1:40:46
create a speech config
1:40:46
create a speech config here this is in c sharp but it's my
1:40:49
here this is in c sharp but it's my
1:40:49
here this is in c sharp but it's my profile programming language
1:40:51
profile programming language
1:40:51
profile programming language and then speak to the speaker audio
1:40:53
and then speak to the speaker audio
1:40:54
and then speak to the speaker audio config
1:40:54
config
1:40:54
config speech synthesizer speaks to text speech
1:40:57
speech synthesizer speaks to text speech
1:40:57
speech synthesizer speaks to text speech text
1:40:58
text
1:40:58
text that's it oh this is all you have to do
1:41:00
that's it oh this is all you have to do
1:41:00
that's it oh this is all you have to do to render
1:41:01
to render
1:41:01
to render your text directly to your the default
1:41:04
your text directly to your the default
1:41:04
your text directly to your the default speaker
1:41:05
speaker
1:41:05
speaker so it is super simple by using the sdk
1:41:09
so it is super simple by using the sdk
1:41:09
so it is super simple by using the sdk so it makes a very very accessible
1:41:12
so it makes a very very accessible
1:41:12
so it makes a very very accessible and it also supports ssml thus
1:41:15
and it also supports ssml thus
1:41:16
and it also supports ssml thus speech synthesis markup language which
1:41:18
speech synthesis markup language which
1:41:18
speech synthesis markup language which is a
1:41:19
is a
1:41:19
is a xml type of language where you can
1:41:22
xml type of language where you can
1:41:22
xml type of language where you can define
1:41:23
define
1:41:23
define the tone of voice some breaks
1:41:26
the tone of voice some breaks
1:41:26
the tone of voice some breaks some sequence of sentences
1:41:29
some sequence of sentences
1:41:29
some sequence of sentences so make it more and more natural it
1:41:32
so make it more and more natural it
1:41:32
so make it more and more natural it comes out actually
1:41:34
comes out actually
1:41:34
comes out actually pretty pretty uh natural you it doesn't
1:41:36
pretty pretty uh natural you it doesn't
1:41:36
pretty pretty uh natural you it doesn't sound as a robotic
1:41:39
sound as a robotic
1:41:39
sound as a robotic voice at all there is a massive
1:41:41
voice at all there is a massive
1:41:41
voice at all there is a massive development in
1:41:42
development in
1:41:42
development in in this space all right so before i take
1:41:45
in this space all right so before i take
1:41:45
in this space all right so before i take a few questions
1:41:47
a few questions
1:41:47
a few questions uh i invite you to pay attention again
1:41:49
uh i invite you to pay attention again
1:41:49
uh i invite you to pay attention again for another couple of minutes on this
1:41:51
for another couple of minutes on this
1:41:51
for another couple of minutes on this video where we have a satya nadella ceo
1:41:55
video where we have a satya nadella ceo
1:41:55
video where we have a satya nadella ceo microsoft uh showing inviting
1:41:59
microsoft uh showing inviting
1:41:59
microsoft uh showing inviting um an engineer and microsoft to showcase
1:42:03
um an engineer and microsoft to showcase
1:42:03
um an engineer and microsoft to showcase an enhancement to this technology which
1:42:06
an enhancement to this technology which
1:42:06
an enhancement to this technology which goes
1:42:06
goes
1:42:06
goes beyond the limitation of the orleans
1:42:10
beyond the limitation of the orleans
1:42:10
beyond the limitation of the orleans and use this cognitive services in more
1:42:13
and use this cognitive services in more
1:42:13
and use this cognitive services in more advanced way
1:42:15
advanced way
1:42:15
advanced way with everything that i show you so far
1:42:18
with everything that i show you so far
1:42:18
with everything that i show you so far but bundle into one real application
1:42:21
but bundle into one real application
1:42:21
but bundle into one real application that
1:42:22
that
1:42:22
that does uh face recognition people
1:42:25
does uh face recognition people
1:42:25
does uh face recognition people recognition
1:42:26
recognition
1:42:26
recognition object detection and text
1:42:29
object detection and text
1:42:29
object detection and text to speech synthesis
1:42:32
and one developer truly inspired me this
1:42:37
and one developer truly inspired me this
1:42:37
and one developer truly inspired me this last year
1:42:38
last year
1:42:38
last year very close to home for us at microsoft i
1:42:41
very close to home for us at microsoft i
1:42:41
very close to home for us at microsoft i want to
1:42:42
want to
1:42:42
want to see what he dreamt of roll the video
1:42:50
i love making things which improve
1:42:52
i love making things which improve
1:42:52
i love making things which improve people's lives
1:42:53
people's lives
1:42:53
people's lives and one of the things i've always dreamt
1:42:55
and one of the things i've always dreamt
1:42:55
and one of the things i've always dreamt of since i was at university
1:42:57
of since i was at university
1:42:57
of since i was at university was this idea of something that could
1:43:00
was this idea of something that could
1:43:00
was this idea of something that could tell you at any moment what's going on
1:43:02
tell you at any moment what's going on
1:43:02
tell you at any moment what's going on around you
1:43:05
i think it's a man jumping in the air
1:43:07
i think it's a man jumping in the air
1:43:07
i think it's a man jumping in the air doing a trick on escape
1:43:11
doing a trick on escape
1:43:11
doing a trick on escape all right i really hope you get inspired
1:43:14
all right i really hope you get inspired
1:43:14
all right i really hope you get inspired to build
1:43:15
to build
1:43:15
to build apps that can be socially useful and you
1:43:18
apps that can be socially useful and you
1:43:18
apps that can be socially useful and you know improve
1:43:19
know improve
1:43:20
know improve this world a little better if you want
1:43:21
this world a little better if you want
1:43:21
this world a little better if you want to get started cognitive services
1:43:24
to get started cognitive services
1:43:24
to get started cognitive services in azure face recognition text-to-speech
1:43:27
in azure face recognition text-to-speech
1:43:27
in azure face recognition text-to-speech are just
1:43:28
are just
1:43:28
are just some of the few that are available and
1:43:31
some of the few that are available and
1:43:31
some of the few that are available and if you are after the hololens device
1:43:34
if you are after the hololens device
1:43:34
if you are after the hololens device programming hololens
1:43:36
programming hololens
1:43:36
programming hololens as well so thank you very much uh social
1:43:39
as well so thank you very much uh social
1:43:39
as well so thank you very much uh social coordinates on screen if you want to get
1:43:41
coordinates on screen if you want to get
1:43:41
coordinates on screen if you want to get in touch
1:43:42
in touch
1:43:42
in touch any follow-up questions please feel free
1:43:44
any follow-up questions please feel free
1:43:44
any follow-up questions please feel free to do so
1:43:45
to do so
1:43:45
to do so and um any questions for me right now
1:43:49
and um any questions for me right now
1:43:49
and um any questions for me right now thank you stefano and i i think it's
1:43:52
thank you stefano and i i think it's
1:43:52
thank you stefano and i i think it's great
1:43:52
great
1:43:52
great how um when we have
1:43:56
how um when we have
1:43:56
how um when we have um optimization and enhancements and the
1:43:59
um optimization and enhancements and the
1:43:59
um optimization and enhancements and the different services
1:44:00
different services
1:44:00
different services um they kind of improve the other
1:44:03
um they kind of improve the other
1:44:03
um they kind of improve the other services
1:44:03
services
1:44:04
services so i when there's enhancements and texts
1:44:07
so i when there's enhancements and texts
1:44:07
so i when there's enhancements and texts cognitive services then vision also has
1:44:11
cognitive services then vision also has
1:44:11
cognitive services then vision also has additional capabilities
1:44:12
additional capabilities
1:44:12
additional capabilities so um i i really enjoy working in the
1:44:15
so um i i really enjoy working in the
1:44:15
so um i i really enjoy working in the ecosystem
1:44:16
ecosystem
1:44:16
ecosystem we do have a question from the viewer
1:44:18
we do have a question from the viewer
1:44:18
we do have a question from the viewer and one of
1:44:20
and one of
1:44:20
and one of the concerns that that we have as a
1:44:22
the concerns that that we have as a
1:44:22
the concerns that that we have as a society is some of these tools are
1:44:24
society is some of these tools are
1:44:24
society is some of these tools are allowing us
1:44:25
allowing us
1:44:25
allowing us to take uh objects and
1:44:29
to take uh objects and
1:44:29
to take uh objects and people in and out of images as well as
1:44:31
people in and out of images as well as
1:44:31
people in and out of images as well as recreate
1:44:32
recreate
1:44:32
recreate voices and and emulate
1:44:35
voices and and emulate
1:44:35
voices and and emulate people so um you know we
1:44:39
people so um you know we
1:44:39
people so um you know we we've seen with deep fakes that this can
1:44:41
we've seen with deep fakes that this can
1:44:42
we've seen with deep fakes that this can create
1:44:42
create
1:44:42
create a lot of disinformation so what are your
1:44:45
a lot of disinformation so what are your
1:44:45
a lot of disinformation so what are your thoughts
1:44:45
thoughts
1:44:46
thoughts on on how microsoft is going about
1:44:49
on on how microsoft is going about
1:44:49
on on how microsoft is going about addressing some of those issues
1:44:54
look there is a big problem
1:44:57
look there is a big problem
1:44:57
look there is a big problem in ai that data scientists will uh
1:45:01
in ai that data scientists will uh
1:45:01
in ai that data scientists will uh agree with is uh on
1:45:05
agree with is uh on
1:45:05
agree with is uh on removing bias from models uh
1:45:08
removing bias from models uh
1:45:08
removing bias from models uh unfortunately there is no simple
1:45:10
unfortunately there is no simple
1:45:10
unfortunately there is no simple solution except
1:45:11
solution except
1:45:11
solution except try to be more as ethical and fair
1:45:14
try to be more as ethical and fair
1:45:14
try to be more as ethical and fair as possible in creating in training
1:45:18
as possible in creating in training
1:45:18
as possible in creating in training models
1:45:19
models
1:45:19
models that are as inclusive as possible
1:45:22
that are as inclusive as possible
1:45:22
that are as inclusive as possible that for example you can think
1:45:25
that for example you can think
1:45:25
that for example you can think phase recognition can be
1:45:29
phase recognition can be
1:45:29
phase recognition can be when detecting emotions when detecting
1:45:31
when detecting emotions when detecting
1:45:31
when detecting emotions when detecting age or detecting gender
1:45:33
age or detecting gender
1:45:33
age or detecting gender is obviously biased towards the
1:45:35
is obviously biased towards the
1:45:35
is obviously biased towards the stereotype
1:45:36
stereotype
1:45:36
stereotype of a male and a female and we know the
1:45:39
of a male and a female and we know the
1:45:39
of a male and a female and we know the word is not just binary
1:45:41
word is not just binary
1:45:41
word is not just binary so there is obviously uh an exercise an
1:45:45
so there is obviously uh an exercise an
1:45:45
so there is obviously uh an exercise an effort that has to be made
1:45:47
effort that has to be made
1:45:47
effort that has to be made in trained models to remove this bias
1:45:50
in trained models to remove this bias
1:45:50
in trained models to remove this bias and improve the accuracy the there is no
1:45:53
and improve the accuracy the there is no
1:45:53
and improve the accuracy the there is no solution
1:45:54
solution
1:45:54
solution except that doing a hard work or
1:45:56
except that doing a hard work or
1:45:56
except that doing a hard work or classifying classifying
1:45:57
classifying classifying
1:45:58
classifying classifying different images and improve this level
1:46:01
different images and improve this level
1:46:01
different images and improve this level of detection
1:46:02
of detection
1:46:02
of detection as much as possible it's probably not
1:46:04
as much as possible it's probably not
1:46:04
as much as possible it's probably not possible to remove
1:46:05
possible to remove
1:46:05
possible to remove bias at all simply because we as people
1:46:08
bias at all simply because we as people
1:46:08
bias at all simply because we as people when we classify the images we are also
1:46:12
when we classify the images we are also
1:46:12
when we classify the images we are also sort of driven by buyers ourselves
1:46:16
sort of driven by buyers ourselves
1:46:16
sort of driven by buyers ourselves but uh leveraging
1:46:19
but uh leveraging
1:46:19
but uh leveraging the the power of uh open communities
1:46:23
the the power of uh open communities
1:46:23
the the power of uh open communities the the the power of a
1:46:27
the the the power of a
1:46:27
the the the power of a mass crowd intelligence it is possible
1:46:30
mass crowd intelligence it is possible
1:46:30
mass crowd intelligence it is possible to remove this
1:46:31
to remove this
1:46:31
to remove this over and over time very briefly i have
1:46:34
over and over time very briefly i have
1:46:34
over and over time very briefly i have made an experience some time ago in skin
1:46:38
made an experience some time ago in skin
1:46:38
made an experience some time ago in skin condition detection so recognizing some
1:46:41
condition detection so recognizing some
1:46:41
condition detection so recognizing some conditions on your face
1:46:43
conditions on your face
1:46:43
conditions on your face whether it's wrinkles uh acne crown feet
1:46:47
whether it's wrinkles uh acne crown feet
1:46:47
whether it's wrinkles uh acne crown feet and then there is a commercial
1:46:49
and then there is a commercial
1:46:49
and then there is a commercial application that i'm not gonna go into
1:46:50
application that i'm not gonna go into
1:46:50
application that i'm not gonna go into it
1:46:51
it
1:46:51
it uh over recommending some product to
1:46:54
uh over recommending some product to
1:46:54
uh over recommending some product to know to address this concern
1:46:56
know to address this concern
1:46:56
know to address this concern and obviously they will work really well
1:46:58
and obviously they will work really well
1:46:58
and obviously they will work really well on
1:46:59
on
1:46:59
on white middle-aged women because the
1:47:01
white middle-aged women because the
1:47:01
white middle-aged women because the model was trained
1:47:03
model was trained
1:47:03
model was trained to the target market anybody else
1:47:06
to the target market anybody else
1:47:06
to the target market anybody else that doesn't fit into this category then
1:47:10
that doesn't fit into this category then
1:47:10
that doesn't fit into this category then it wouldn't be recognized that easily
1:47:12
it wouldn't be recognized that easily
1:47:12
it wouldn't be recognized that easily and that is a problem because
1:47:14
and that is a problem because
1:47:14
and that is a problem because well it was a commercial choice to
1:47:17
well it was a commercial choice to
1:47:17
well it was a commercial choice to target that
1:47:18
target that
1:47:18
target that audience but in general is an audience
1:47:20
audience but in general is an audience
1:47:20
audience but in general is an audience when you offer public
1:47:22
when you offer public
1:47:22
when you offer public services like not the azure cognitive
1:47:24
services like not the azure cognitive
1:47:24
services like not the azure cognitive services you want to remove bias as much
1:47:26
services you want to remove bias as much
1:47:26
services you want to remove bias as much as possible
1:47:27
as possible
1:47:27
as possible to make it fair and more inclusive
1:47:34
there's also the use of domains and
1:47:36
there's also the use of domains and
1:47:36
there's also the use of domains and computer vision so
1:47:38
computer vision so
1:47:38
computer vision so do you feel that um you know as
1:47:41
do you feel that um you know as
1:47:41
do you feel that um you know as as more people use these services and
1:47:43
as more people use these services and
1:47:43
as more people use these services and you know you mentioned
1:47:44
you know you mentioned
1:47:44
you know you mentioned open source that um
1:47:48
open source that um
1:47:48
open source that um there will be kind of a proliferation of
1:47:51
there will be kind of a proliferation of
1:47:51
there will be kind of a proliferation of a
1:47:52
a
1:47:52
a domain type data sunset that people
1:47:55
domain type data sunset that people
1:47:55
domain type data sunset that people start
1:47:56
start
1:47:56
start sharing in the open source community
1:47:59
sharing in the open source community
1:47:59
sharing in the open source community yes very good question and very big hope
1:48:02
yes very good question and very big hope
1:48:02
yes very good question and very big hope in this
1:48:02
in this
1:48:02
in this there is already a sort of concept of
1:48:05
there is already a sort of concept of
1:48:05
there is already a sort of concept of domain
1:48:05
domain
1:48:06
domain skills in the conversational ai
1:48:09
skills in the conversational ai
1:48:09
skills in the conversational ai which you know addresses specifically
1:48:11
which you know addresses specifically
1:48:11
which you know addresses specifically chat bots any sort especially in the
1:48:14
chat bots any sort especially in the
1:48:14
chat bots any sort especially in the customer service
1:48:15
customer service
1:48:15
customer service you don't want to build you don't want
1:48:17
you don't want to build you don't want
1:48:17
you don't want to build you don't want any
1:48:18
any
1:48:18
any probably you cannot build a chatbot that
1:48:21
probably you cannot build a chatbot that
1:48:21
probably you cannot build a chatbot that is able to have
1:48:22
is able to have
1:48:22
is able to have any sort of conversation whatsoever from
1:48:25
any sort of conversation whatsoever from
1:48:25
any sort of conversation whatsoever from a football result to
1:48:28
a football result to
1:48:28
a football result to a you know a data scientist conversation
1:48:31
a you know a data scientist conversation
1:48:31
a you know a data scientist conversation you want to have something that is a
1:48:32
you want to have something that is a
1:48:32
you want to have something that is a specific
1:48:33
specific
1:48:33
specific tailor to the domain optimize the same
1:48:36
tailor to the domain optimize the same
1:48:36
tailor to the domain optimize the same is for computer vision
1:48:37
is for computer vision
1:48:37
is for computer vision it will very likely focus on a specific
1:48:41
it will very likely focus on a specific
1:48:41
it will very likely focus on a specific domain
1:48:42
domain
1:48:42
domain skill your computer vision ai
1:48:45
skill your computer vision ai
1:48:45
skill your computer vision ai to detect what's uh one sort of
1:48:51
object of the other one before when
1:48:53
object of the other one before when
1:48:53
object of the other one before when satya was
1:48:54
satya was
1:48:54
satya was speaking there was a question can the
1:48:57
speaking there was a question can the
1:48:57
speaking there was a question can the device identify
1:48:58
device identify
1:48:58
device identify registration numbers registration plate
1:49:01
registration numbers registration plate
1:49:01
registration numbers registration plate forecast in all the countries in the
1:49:03
forecast in all the countries in the
1:49:03
forecast in all the countries in the world
1:49:03
world
1:49:04
world you don't need that because you're not
1:49:06
you don't need that because you're not
1:49:06
you don't need that because you're not using that camera that's
1:49:08
using that camera that's
1:49:08
using that camera that's one single camera in every countries in
1:49:10
one single camera in every countries in
1:49:10
one single camera in every countries in the world unlikely
1:49:11
the world unlikely
1:49:11
the world unlikely so optimize for the us optimize for the
1:49:14
so optimize for the us optimize for the
1:49:14
so optimize for the us optimize for the uk optimize for it
1:49:15
uk optimize for it
1:49:15
uk optimize for it optimize for australia the same device
1:49:18
optimize for australia the same device
1:49:18
optimize for australia the same device but train two different models so that
1:49:19
but train two different models so that
1:49:19
but train two different models so that is optimized to detect
1:49:21
is optimized to detect
1:49:21
is optimized to detect the registration plate in this country
1:49:23
the registration plate in this country
1:49:23
the registration plate in this country improve accuracy
1:49:24
improve accuracy
1:49:24
improve accuracy reduce the training effort and i get
1:49:28
reduce the training effort and i get
1:49:28
reduce the training effort and i get better results at the end
1:49:30
better results at the end
1:49:30
better results at the end and i guess it works better on a hollow
1:49:31
and i guess it works better on a hollow
1:49:32
and i guess it works better on a hollow lens in that case because you need less
1:49:34
lens in that case because you need less
1:49:34
lens in that case because you need less space less compute power to actually
1:49:36
space less compute power to actually
1:49:36
space less compute power to actually uh predict what's on the license plate
1:49:39
uh predict what's on the license plate
1:49:39
uh predict what's on the license plate especially when running at the edge
1:49:40
especially when running at the edge
1:49:40
especially when running at the edge you're absolutely right that the edge
1:49:42
you're absolutely right that the edge
1:49:42
you're absolutely right that the edge is not like the cloud where you can
1:49:44
is not like the cloud where you can
1:49:44
is not like the cloud where you can allocate a terabyte and terabyte of
1:49:46
allocate a terabyte and terabyte of
1:49:46
allocate a terabyte and terabyte of storage
1:49:46
storage
1:49:46
storage you need to work on compressed
1:49:50
you need to work on compressed
1:49:50
you need to work on compressed compact models so anything that you can
1:49:53
compact models so anything that you can
1:49:53
compact models so anything that you can save
1:49:53
save
1:49:54
save from a space and computing perspective
1:49:56
from a space and computing perspective
1:49:56
from a space and computing perspective will improve the performance of the
1:49:57
will improve the performance of the
1:49:57
will improve the performance of the device
1:49:58
device
1:49:58
device absolutely cool well thank you very much
1:50:01
absolutely cool well thank you very much
1:50:01
absolutely cool well thank you very much we're out of time sadly
1:50:02
we're out of time sadly
1:50:02
we're out of time sadly um sure i know for a fact that the
1:50:05
um sure i know for a fact that the
1:50:05
um sure i know for a fact that the hololens is a very interesting topic on
1:50:07
hololens is a very interesting topic on
1:50:07
hololens is a very interesting topic on its own
1:50:08
its own
1:50:08
its own so um um if people are interested
1:50:11
so um um if people are interested
1:50:11
so um um if people are interested anyway so if people have questions i'll
1:50:14
anyway so if people have questions i'll
1:50:14
anyway so if people have questions i'll stay on top of the chat and answer now
1:50:16
stay on top of the chat and answer now
1:50:16
stay on top of the chat and answer now awesome thank you very much
1:50:20
alicia what what what do you think um
1:50:25
alicia what what what do you think um
1:50:25
alicia what what what do you think um should i do a bit of vr this time we've
1:50:27
should i do a bit of vr this time we've
1:50:27
should i do a bit of vr this time we've talked about ar
1:50:28
talked about ar
1:50:28
talked about ar computer version vr maybe
1:50:31
computer version vr maybe
1:50:32
computer version vr maybe um you know the day just keeps getting
1:50:35
um you know the day just keeps getting
1:50:35
um you know the day just keeps getting better and better
1:50:36
better and better
1:50:36
better and better and i i'm definitely excited to be
1:50:39
and i i'm definitely excited to be
1:50:39
and i i'm definitely excited to be um hosting our our next speaker um
1:50:42
um hosting our our next speaker um
1:50:42
um hosting our our next speaker um malicia
1:50:43
malicia
1:50:43
malicia mcgregor and she's joining us from
1:50:46
mcgregor and she's joining us from
1:50:46
mcgregor and she's joining us from oklahoma
1:50:47
oklahoma
1:50:47
oklahoma and she's here to talk to her to us
1:50:50
and she's here to talk to her to us
1:50:50
and she's here to talk to her to us about vr
1:50:53
about vr
1:50:54
about vr project with um
1:50:58
with brain.js so
1:51:01
with brain.js so
1:51:01
with brain.js so we're going to be looking at integrating
1:51:03
we're going to be looking at integrating
1:51:03
we're going to be looking at integrating javascript
1:51:05
javascript
1:51:05
javascript vr and ai and
1:51:09
vr and ai and
1:51:09
vr and ai and all these fun things um i think we are
1:51:12
all these fun things um i think we are
1:51:12
all these fun things um i think we are waiting for
1:51:13
waiting for
1:51:13
waiting for her video scene but um
1:51:17
her video scene but um
1:51:17
her video scene but um have you ever done anything with ai in
1:51:19
have you ever done anything with ai in
1:51:19
have you ever done anything with ai in javascript alicia
1:51:21
javascript alicia
1:51:21
javascript alicia and i haven't done anything in
1:51:22
and i haven't done anything in
1:51:22
and i haven't done anything in javascript in a while
1:51:24
javascript in a while
1:51:24
javascript in a while oh so javascript in general wow yeah
1:51:28
oh so javascript in general wow yeah
1:51:28
oh so javascript in general wow yeah no it's uh you know
1:51:31
no it's uh you know
1:51:31
no it's uh you know uh 80 of projects right now tend to
1:51:35
uh 80 of projects right now tend to
1:51:35
uh 80 of projects right now tend to involve ingestion and that's a lot of
1:51:39
involve ingestion and that's a lot of
1:51:39
involve ingestion and that's a lot of where my time is spent and you know
1:51:42
where my time is spent and you know
1:51:42
where my time is spent and you know i i do get to talk a lot to people about
1:51:46
i i do get to talk a lot to people about
1:51:46
i i do get to talk a lot to people about um you know iot and where they're
1:51:49
um you know iot and where they're
1:51:49
um you know iot and where they're putting their data
1:51:50
putting their data
1:51:50
putting their data and um but it's
1:51:56
so exciting how um
1:52:00
how they're enabling everyone
1:52:03
how they're enabling everyone
1:52:03
how they're enabling everyone to like we're talking about hardware
1:52:05
to like we're talking about hardware
1:52:05
to like we're talking about hardware getting smaller
1:52:06
getting smaller
1:52:06
getting smaller right so just the capabilities and then
1:52:09
right so just the capabilities and then
1:52:09
right so just the capabilities and then the
1:52:10
the
1:52:10
the the enhancement in the capabilities and
1:52:12
the enhancement in the capabilities and
1:52:12
the enhancement in the capabilities and how quickly it's happening
1:52:14
how quickly it's happening
1:52:14
how quickly it's happening you know and on top of that um the
1:52:17
you know and on top of that um the
1:52:17
you know and on top of that um the the the importance that people are
1:52:19
the the importance that people are
1:52:19
the the importance that people are placing on
1:52:21
placing on
1:52:21
placing on unbiased and ethics is just phenomenal
1:52:24
unbiased and ethics is just phenomenal
1:52:24
unbiased and ethics is just phenomenal and we were talking about this in
1:52:26
and we were talking about this in
1:52:26
and we were talking about this in another session how some of these biased
1:52:28
another session how some of these biased
1:52:28
another session how some of these biased questions
1:52:29
questions
1:52:29
questions they come up and we make improvements
1:52:31
they come up and we make improvements
1:52:31
they come up and we make improvements against them so quickly
1:52:33
against them so quickly
1:52:33
against them so quickly um that you know everybody's and you
1:52:36
um that you know everybody's and you
1:52:36
um that you know everybody's and you don't get that type of responsiveness
1:52:38
don't get that type of responsiveness
1:52:38
don't get that type of responsiveness and improvements unless everybody thinks
1:52:41
and improvements unless everybody thinks
1:52:41
and improvements unless everybody thinks it's an issue
1:52:42
it's an issue
1:52:42
it's an issue right so we have melissa here
1:52:45
right so we have melissa here
1:52:45
right so we have melissa here on the line with us now welcome thank
1:52:47
on the line with us now welcome thank
1:52:47
on the line with us now welcome thank you connection is working so
1:52:49
you connection is working so
1:52:49
you connection is working so um welcome to the show and we were just
1:52:51
um welcome to the show and we were just
1:52:51
um welcome to the show and we were just discussing that we haven't seen a lot of
1:52:53
discussing that we haven't seen a lot of
1:52:53
discussing that we haven't seen a lot of javascript stuff
1:52:54
javascript stuff
1:52:54
javascript stuff in the ai scene we're mostly building a
1:52:57
in the ai scene we're mostly building a
1:52:57
in the ai scene we're mostly building a models in c
1:52:58
models in c
1:52:58
models in c plus and python actually i do a lot of
1:53:00
plus and python actually i do a lot of
1:53:00
plus and python actually i do a lot of python and almost
1:53:01
python and almost
1:53:01
python and almost no c plus plus at all um so i guess
1:53:04
no c plus plus at all um so i guess
1:53:04
no c plus plus at all um so i guess thanks to the the moore's law we have
1:53:08
thanks to the the moore's law we have
1:53:08
thanks to the the moore's law we have enough compute power to run javascript
1:53:10
enough compute power to run javascript
1:53:10
enough compute power to run javascript in a vr device and do something
1:53:11
in a vr device and do something
1:53:11
in a vr device and do something interesting with the eye so
1:53:13
interesting with the eye so
1:53:13
interesting with the eye so i'm curious what that looks like um
1:53:16
i'm curious what that looks like um
1:53:16
i'm curious what that looks like um so but first um so
1:53:20
so but first um so
1:53:20
so but first um so uh i've talked to you before the show a
1:53:22
uh i've talked to you before the show a
1:53:22
uh i've talked to you before the show a few days ago and we had a
1:53:24
few days ago and we had a
1:53:24
few days ago and we had a lovely discussion about ai um what sort
1:53:27
lovely discussion about ai um what sort
1:53:27
lovely discussion about ai um what sort of work do you do during the day
1:53:29
of work do you do during the day
1:53:29
of work do you do during the day so i work on just different react apps
1:53:32
so i work on just different react apps
1:53:32
so i work on just different react apps and some cicd
1:53:34
and some cicd
1:53:34
and some cicd pipeline tools and i'm actually just
1:53:37
pipeline tools and i'm actually just
1:53:37
pipeline tools and i'm actually just kind of all over the place
1:53:40
kind of all over the place
1:53:40
kind of all over the place so you're sort of the the profile a
1:53:42
so you're sort of the the profile a
1:53:42
so you're sort of the the profile a regular developer that sort of ran
1:53:44
regular developer that sort of ran
1:53:44
regular developer that sort of ran into uh the whole ai thing is that that
1:53:47
into uh the whole ai thing is that that
1:53:47
into uh the whole ai thing is that that correct
1:53:48
correct
1:53:48
correct actually i have my masters in mechanical
1:53:51
actually i have my masters in mechanical
1:53:51
actually i have my masters in mechanical and aerospace engineering
1:53:52
and aerospace engineering
1:53:52
and aerospace engineering so i did some research with like
1:53:55
so i did some research with like
1:53:55
so i did some research with like robotics and
1:53:56
robotics and
1:53:56
robotics and ai i was able to build an autonomous
1:53:59
ai i was able to build an autonomous
1:53:59
ai i was able to build an autonomous golf cart so
1:54:00
golf cart so
1:54:00
golf cart so that's where i got into ai the first
1:54:03
that's where i got into ai the first
1:54:03
that's where i got into ai the first time
1:54:05
time
1:54:05
time oh alicia we made a big mistake we
1:54:07
oh alicia we made a big mistake we
1:54:07
oh alicia we made a big mistake we should have asked her about
1:54:08
should have asked her about
1:54:08
should have asked her about an automated golf cart i mean that's so
1:54:10
an automated golf cart i mean that's so
1:54:10
an automated golf cart i mean that's so much more
1:54:11
much more
1:54:11
much more uh interesting but we're i guess we're
1:54:14
uh interesting but we're i guess we're
1:54:14
uh interesting but we're i guess we're lucky enough today
1:54:16
lucky enough today
1:54:16
lucky enough today for christmas if you could help me out
1:54:18
for christmas if you could help me out
1:54:18
for christmas if you could help me out with a puppy
1:54:19
with a puppy
1:54:19
with a puppy with a puppy yeah we have a global ai
1:54:22
with a puppy yeah we have a global ai
1:54:22
with a puppy yeah we have a global ai puppy
1:54:23
puppy
1:54:23
puppy these days uh that's i adopted it
1:54:25
these days uh that's i adopted it
1:54:25
these days uh that's i adopted it somehow remotely over the internet i
1:54:27
somehow remotely over the internet i
1:54:27
somehow remotely over the internet i don't know what happened
1:54:28
don't know what happened
1:54:28
don't know what happened i made a mistake somewhere
1:54:32
so so we're going to talk
1:54:36
so so we're going to talk
1:54:36
so so we're going to talk a little bit about about vr so um
1:54:39
a little bit about about vr so um
1:54:39
a little bit about about vr so um you mentioned react vr can you explain a
1:54:42
you mentioned react vr can you explain a
1:54:42
you mentioned react vr can you explain a little bit of
1:54:43
little bit of
1:54:43
little bit of how that works actually how would i
1:54:45
how that works actually how would i
1:54:45
how that works actually how would i actually build a vr application in react
1:54:48
actually build a vr application in react
1:54:48
actually build a vr application in react so the good thing is you don't actually
1:54:50
so the good thing is you don't actually
1:54:50
so the good thing is you don't actually need react to build a vr app
1:54:52
need react to build a vr app
1:54:52
need react to build a vr app so there's a few different javascript
1:54:55
so there's a few different javascript
1:54:55
so there's a few different javascript libraries like
1:54:56
libraries like
1:54:56
libraries like my favorite is a-frame and it makes it
1:54:59
my favorite is a-frame and it makes it
1:55:00
my favorite is a-frame and it makes it super easy to make a vr app with just
1:55:02
super easy to make a vr app with just
1:55:02
super easy to make a vr app with just some
1:55:02
some
1:55:02
some html and assets and a little bit of
1:55:05
html and assets and a little bit of
1:55:05
html and assets and a little bit of javascript cool
1:55:09
javascript cool
1:55:09
javascript cool and how did you get started
1:55:12
and how did you get started
1:55:12
and how did you get started um honestly with the vr app
1:55:15
um honestly with the vr app
1:55:15
um honestly with the vr app i just kind of wanted to make a vr app
1:55:19
i just kind of wanted to make a vr app
1:55:19
i just kind of wanted to make a vr app so i started looking around and reading
1:55:21
so i started looking around and reading
1:55:21
so i started looking around and reading different articles and
1:55:23
different articles and
1:55:23
different articles and trying to find tools and of course there
1:55:25
trying to find tools and of course there
1:55:26
trying to find tools and of course there were a lot of them in python and
1:55:28
were a lot of them in python and
1:55:28
were a lot of them in python and c-sharp and stuff like that but i really
1:55:30
c-sharp and stuff like that but i really
1:55:30
c-sharp and stuff like that but i really really like
1:55:31
really like
1:55:31
really like javascript so i wanted to find a way
1:55:34
javascript so i wanted to find a way
1:55:34
javascript so i wanted to find a way to bring vr to javascript and
1:55:37
to bring vr to javascript and
1:55:38
to bring vr to javascript and somebody already did it for me so using
1:55:40
somebody already did it for me so using
1:55:40
somebody already did it for me so using a frame made it really easy to just get
1:55:42
a frame made it really easy to just get
1:55:42
a frame made it really easy to just get started and
1:55:44
started and
1:55:44
started and up and running just in a browser
1:55:47
up and running just in a browser
1:55:47
up and running just in a browser wow vr applications in the browser how
1:55:51
wow vr applications in the browser how
1:55:51
wow vr applications in the browser how does that combine
1:55:53
does that combine
1:55:53
does that combine uh with uh the vr glasses that we just
1:55:56
uh with uh the vr glasses that we just
1:55:56
uh with uh the vr glasses that we just showed at the beginning of the episode
1:55:57
showed at the beginning of the episode
1:55:57
showed at the beginning of the episode the oculus quest can it run on there or
1:56:00
the oculus quest can it run on there or
1:56:00
the oculus quest can it run on there or how does that work actually the whole
1:56:01
how does that work actually the whole
1:56:01
how does that work actually the whole process so they can run on
1:56:04
process so they can run on
1:56:04
process so they can run on like a hololens or an oculus rift or
1:56:06
like a hololens or an oculus rift or
1:56:06
like a hololens or an oculus rift or something like that
1:56:07
something like that
1:56:08
something like that it's just you have to export them in
1:56:10
it's just you have to export them in
1:56:10
it's just you have to export them in that specific format
1:56:12
that specific format
1:56:12
that specific format okay so you're you're basically building
1:56:15
okay so you're you're basically building
1:56:15
okay so you're you're basically building this vr application in your browser on
1:56:17
this vr application in your browser on
1:56:17
this vr application in your browser on your computer and then you can package
1:56:19
your computer and then you can package
1:56:19
your computer and then you can package it up
1:56:20
it up
1:56:20
it up and export it to the vr device of your
1:56:22
and export it to the vr device of your
1:56:22
and export it to the vr device of your choice and i guess most popular ones are
1:56:24
choice and i guess most popular ones are
1:56:24
choice and i guess most popular ones are supported uh by the product i know for
1:56:28
supported uh by the product i know for
1:56:28
supported uh by the product i know for sure
1:56:28
sure
1:56:28
sure the oculus rift is supported um
1:56:32
the oculus rift is supported um
1:56:32
the oculus rift is supported um i think the thing that samsung makes is
1:56:34
i think the thing that samsung makes is
1:56:34
i think the thing that samsung makes is supported i forget what it's called but
1:56:37
supported i forget what it's called but
1:56:37
supported i forget what it's called but the main popular devices are supported
1:56:40
the main popular devices are supported
1:56:40
the main popular devices are supported by a-frame oh cool
1:56:43
by a-frame oh cool
1:56:43
by a-frame oh cool so um and then there's there's the ai
1:56:47
so um and then there's there's the ai
1:56:47
so um and then there's there's the ai component
1:56:49
component
1:56:49
component we've seen from from such at the
1:56:51
we've seen from from such at the
1:56:51
we've seen from from such at the beginning of this episode that there's a
1:56:52
beginning of this episode that there's a
1:56:52
beginning of this episode that there's a lot of stuff you can do with computer
1:56:54
lot of stuff you can do with computer
1:56:54
lot of stuff you can do with computer vision these days
1:56:55
vision these days
1:56:55
vision these days actually they've been working on it for
1:56:57
actually they've been working on it for
1:56:57
actually they've been working on it for 20 years um
1:56:59
20 years um
1:56:59
20 years um how does that combine with vr in this
1:57:01
how does that combine with vr in this
1:57:01
how does that combine with vr in this case what sort of thing have you built
1:57:02
case what sort of thing have you built
1:57:02
case what sort of thing have you built using ai in vr so in this case
1:57:06
using ai in vr so in this case
1:57:06
using ai in vr so in this case it's more of like how a user interacts
1:57:10
it's more of like how a user interacts
1:57:10
it's more of like how a user interacts with the vr world so things move around
1:57:13
with the vr world so things move around
1:57:13
with the vr world so things move around according to their interactions
1:57:16
according to their interactions
1:57:16
according to their interactions basically
1:57:17
basically
1:57:18
basically let's say we're trying to find
1:57:21
let's say we're trying to find
1:57:21
let's say we're trying to find objects in a game or just in a world or
1:57:24
objects in a game or just in a world or
1:57:24
objects in a game or just in a world or whatever
1:57:25
whatever
1:57:25
whatever depending on how fast or how well people
1:57:28
depending on how fast or how well people
1:57:28
depending on how fast or how well people are finding things we can move the
1:57:31
are finding things we can move the
1:57:31
are finding things we can move the objects around so that they'll stay
1:57:33
objects around so that they'll stay
1:57:33
objects around so that they'll stay interested in the world and it'll
1:57:37
interested in the world and it'll
1:57:37
interested in the world and it'll increase and decrease the difficulty as
1:57:39
increase and decrease the difficulty as
1:57:39
increase and decrease the difficulty as they need it so they don't just
1:57:41
they need it so they don't just
1:57:41
they need it so they don't just quit and leave that's uninteresting
1:57:46
quit and leave that's uninteresting
1:57:46
quit and leave that's uninteresting cool and what sort of tool are you using
1:57:48
cool and what sort of tool are you using
1:57:48
cool and what sort of tool are you using to build this
1:57:49
to build this
1:57:49
to build this in your application so this is all
1:57:53
in your application so this is all
1:57:53
in your application so this is all like all of the code is written in just
1:57:55
like all of the code is written in just
1:57:55
like all of the code is written in just vs code
1:57:57
vs code
1:57:57
vs code and yes all just javascript
1:58:00
and yes all just javascript
1:58:00
and yes all just javascript it's like building a really big
1:58:02
it's like building a really big
1:58:02
it's like building a really big complicated web app
1:58:04
complicated web app
1:58:04
complicated web app okay any special libraries you use for
1:58:07
okay any special libraries you use for
1:58:07
okay any special libraries you use for ai applications in vr
1:58:10
ai applications in vr
1:58:10
ai applications in vr brain js is my personal favorite just
1:58:13
brain js is my personal favorite just
1:58:13
brain js is my personal favorite just because
1:58:13
because
1:58:14
because it's super easy to get started with like
1:58:16
it's super easy to get started with like
1:58:16
it's super easy to get started with like if you can work with arrays
1:58:18
if you can work with arrays
1:58:18
if you can work with arrays objects and arrays of objects you can do
1:58:22
objects and arrays of objects you can do
1:58:22
objects and arrays of objects you can do ai in javascript that's that sounds
1:58:25
ai in javascript that's that sounds
1:58:25
ai in javascript that's that sounds exactly like my kind of ai
1:58:27
exactly like my kind of ai
1:58:27
exactly like my kind of ai large arrays of numbers that's that now
1:58:29
large arrays of numbers that's that now
1:58:29
large arrays of numbers that's that now we're talking now i get it
1:58:31
we're talking now i get it
1:58:31
we're talking now i get it i don't i mean i've done javascript in
1:58:33
i don't i mean i've done javascript in
1:58:33
i don't i mean i've done javascript in the past for a few years
1:58:34
the past for a few years
1:58:34
the past for a few years and it's it's the languages is so weird
1:58:38
and it's it's the languages is so weird
1:58:38
and it's it's the languages is so weird actually the training that i followed
1:58:40
actually the training that i followed
1:58:40
actually the training that i followed around javascript
1:58:42
around javascript
1:58:42
around javascript they teach me all these conversions
1:58:44
they teach me all these conversions
1:58:44
they teach me all these conversions rather than a type system because it has
1:58:47
rather than a type system because it has
1:58:47
rather than a type system because it has it should have one um i don't know about
1:58:49
it should have one um i don't know about
1:58:49
it should have one um i don't know about it but it sounds interesting
1:58:51
it but it sounds interesting
1:58:51
it but it sounds interesting a race of objects uh basically numbers
1:58:54
a race of objects uh basically numbers
1:58:54
a race of objects uh basically numbers and can you uh can you show us what a
1:58:56
and can you uh can you show us what a
1:58:56
and can you uh can you show us what a brain.js
1:58:57
brain.js
1:58:57
brain.js uh how it works what it looks like sure
1:59:01
uh how it works what it looks like sure
1:59:01
uh how it works what it looks like sure um let me share my screen real quick
1:59:03
um let me share my screen real quick
1:59:04
um let me share my screen real quick hopefully that won't be too
1:59:06
hopefully that won't be too
1:59:06
hopefully that won't be too too hard of a thing to do
1:59:14
otherwise our ai application can find it
1:59:16
otherwise our ai application can find it
1:59:16
otherwise our ai application can find it yeah there we go
1:59:19
yeah there we go
1:59:19
yeah there we go okay so if you're familiar with just
1:59:22
okay so if you're familiar with just
1:59:22
okay so if you're familiar with just any web apps in javascript this
1:59:25
any web apps in javascript this
1:59:25
any web apps in javascript this is a node.js app so
1:59:28
is a node.js app so
1:59:28
is a node.js app so we have some basic stuff in here like
1:59:30
we have some basic stuff in here like
1:59:30
we have some basic stuff in here like we're importing
1:59:31
we're importing
1:59:32
we're importing our packages and setting up a little
1:59:34
our packages and setting up a little
1:59:34
our packages and setting up a little express
1:59:35
express
1:59:35
express app and just some you know regular
1:59:39
app and just some you know regular
1:59:39
app and just some you know regular web app stuff there but this is where it
1:59:41
web app stuff there but this is where it
1:59:41
web app stuff there but this is where it actually gets
1:59:42
actually gets
1:59:42
actually gets interesting so this is just a sample of
1:59:46
interesting so this is just a sample of
1:59:46
interesting so this is just a sample of the training data we're going to use
1:59:48
the training data we're going to use
1:59:48
the training data we're going to use for the neural network because of course
1:59:51
for the neural network because of course
1:59:51
for the neural network because of course you can use neural networks in
1:59:52
you can use neural networks in
1:59:52
you can use neural networks in javascript now
1:59:54
javascript now
1:59:54
javascript now but um basically you can see it's just
1:59:57
but um basically you can see it's just
1:59:57
but um basically you can see it's just this array
1:59:58
this array
1:59:58
this array with these different objects in them and
2:00:00
with these different objects in them and
2:00:00
with these different objects in them and for
2:00:01
for
2:00:01
for this particular application we look at
2:00:04
this particular application we look at
2:00:04
this particular application we look at the user's
2:00:05
the user's
2:00:05
the user's position in the world so they're x y and
2:00:08
position in the world so they're x y and
2:00:08
position in the world so they're x y and z
2:00:09
z
2:00:09
z positions they're all in meters which
2:00:11
positions they're all in meters which
2:00:11
positions they're all in meters which i'm super happy about
2:00:13
i'm super happy about
2:00:13
i'm super happy about but um so we do have a request
2:00:16
but um so we do have a request
2:00:16
but um so we do have a request if we could um possibly zoom
2:00:20
if we could um possibly zoom
2:00:20
if we could um possibly zoom a little bit into the code and
2:00:23
a little bit into the code and
2:00:23
a little bit into the code and um yeah much better okay
2:00:27
um yeah much better okay
2:00:27
um yeah much better okay no problem but um yeah so we get the
2:00:30
no problem but um yeah so we get the
2:00:30
no problem but um yeah so we get the user's location
2:00:32
user's location
2:00:32
user's location we figure out if they found an item or
2:00:35
we figure out if they found an item or
2:00:35
we figure out if they found an item or not
2:00:35
not
2:00:35
not and we look at how many steps they've
2:00:38
and we look at how many steps they've
2:00:38
and we look at how many steps they've taken so
2:00:39
taken so
2:00:39
taken so those are just like some of the features
2:00:41
those are just like some of the features
2:00:41
those are just like some of the features we're looking at
2:00:42
we're looking at
2:00:42
we're looking at to start with and then just some of
2:00:46
to start with and then just some of
2:00:46
to start with and then just some of our you know expected output data
2:00:49
our you know expected output data
2:00:49
our you know expected output data we want to know based on their location
2:00:52
we want to know based on their location
2:00:52
we want to know based on their location if they found something and how many
2:00:54
if they found something and how many
2:00:54
if they found something and how many steps they've taken in the world
2:00:56
steps they've taken in the world
2:00:56
steps they've taken in the world if they'll actually finish the task that
2:00:58
if they'll actually finish the task that
2:00:58
if they'll actually finish the task that we've given them
2:01:01
we've given them
2:01:01
we've given them and pretty much based on that
2:01:04
and pretty much based on that
2:01:04
and pretty much based on that we have our training input data and our
2:01:06
we have our training input data and our
2:01:06
we have our training input data and our training
2:01:07
training
2:01:07
training output data and we go ahead and combine
2:01:09
output data and we go ahead and combine
2:01:09
output data and we go ahead and combine that in just one
2:01:10
that in just one
2:01:10
that in just one giant array then the fun part is here
2:01:14
giant array then the fun part is here
2:01:14
giant array then the fun part is here so this is a neural network just this
2:01:18
so this is a neural network just this
2:01:18
so this is a neural network just this one line of code
2:01:19
one line of code
2:01:19
one line of code it has three hidden layers in it and
2:01:23
it has three hidden layers in it and
2:01:23
it has three hidden layers in it and that is the entire neural network right
2:01:26
that is the entire neural network right
2:01:26
that is the entire neural network right there
2:01:27
there
2:01:27
there that's why i like javascript so much
2:01:30
that's why i like javascript so much
2:01:30
that's why i like javascript so much because they made it
2:01:31
because they made it
2:01:31
because they made it super super simple super easy but
2:01:34
super super simple super easy but
2:01:34
super super simple super easy but after we have our neural net set up
2:01:37
after we have our neural net set up
2:01:37
after we have our neural net set up then we go ahead and train it with that
2:01:40
then we go ahead and train it with that
2:01:40
then we go ahead and train it with that um
2:01:41
um
2:01:41
um data set that we went through earlier
2:01:43
data set that we went through earlier
2:01:43
data set that we went through earlier and
2:01:44
and
2:01:44
and it gives us our model so you can look at
2:01:47
it gives us our model so you can look at
2:01:47
it gives us our model so you can look at some of the
2:01:48
some of the
2:01:48
some of the stats on it let me see if that's still
2:01:51
stats on it let me see if that's still
2:01:51
stats on it let me see if that's still here
2:01:51
here
2:01:51
here from when i started it yeah so you can
2:01:54
from when i started it yeah so you can
2:01:54
from when i started it yeah so you can look at some of the stats on the model
2:01:56
look at some of the stats on the model
2:01:56
look at some of the stats on the model after it's been trained
2:01:58
after it's been trained
2:01:58
after it's been trained eh that that errors
2:02:01
eh that that errors
2:02:01
eh that that errors i don't know how i feel about that error
2:02:03
i don't know how i feel about that error
2:02:03
i don't know how i feel about that error but that's what it is that's how many
2:02:06
but that's what it is that's how many
2:02:06
but that's what it is that's how many iterations it took to get to
2:02:08
iterations it took to get to
2:02:08
iterations it took to get to a decent model and that
2:02:12
a decent model and that
2:02:12
a decent model and that last one the will finish that's just a
2:02:15
last one the will finish that's just a
2:02:15
last one the will finish that's just a prediction
2:02:16
prediction
2:02:16
prediction on some new data we have that tells us
2:02:19
on some new data we have that tells us
2:02:20
on some new data we have that tells us if the user or the player will actually
2:02:23
if the user or the player will actually
2:02:23
if the user or the player will actually finish the task we've given them
2:02:25
finish the task we've given them
2:02:25
finish the task we've given them but really all it comes down to is
2:02:28
but really all it comes down to is
2:02:28
but really all it comes down to is this line of code is our trained bottle
2:02:32
this line of code is our trained bottle
2:02:32
this line of code is our trained bottle so we've trained it we know all the
2:02:35
so we've trained it we know all the
2:02:35
so we've trained it we know all the things about it
2:02:36
things about it
2:02:36
things about it and this line was just to give an
2:02:39
and this line was just to give an
2:02:39
and this line was just to give an example of what it looks like when you
2:02:41
example of what it looks like when you
2:02:41
example of what it looks like when you run the model with
2:02:43
run the model with
2:02:43
run the model with this new data it hasn't seen before
2:02:46
this new data it hasn't seen before
2:02:46
this new data it hasn't seen before and then the rest of it is really just
2:02:48
and then the rest of it is really just
2:02:48
and then the rest of it is really just like how that
2:02:50
like how that
2:02:50
like how that data gets used in the application itself
2:02:55
data gets used in the application itself
2:02:55
data gets used in the application itself and if you were curious about what the
2:02:57
and if you were curious about what the
2:02:57
and if you were curious about what the vr app looks like
2:02:59
vr app looks like
2:02:59
vr app looks like this is it so this is all just in the
2:03:02
this is it so this is all just in the
2:03:02
this is it so this is all just in the browser
2:03:03
browser
2:03:03
browser you can wander around
2:03:07
you can wander around
2:03:07
you can wander around i wonder if i can still find one of the
2:03:10
i wonder if i can still find one of the
2:03:10
i wonder if i can still find one of the objects in here
2:03:12
objects in here
2:03:12
objects in here so basically you can just walk around
2:03:16
so basically you can just walk around
2:03:16
so basically you can just walk around look for the stuff that was on the list
2:03:18
look for the stuff that was on the list
2:03:18
look for the stuff that was on the list and based on how long it takes you to
2:03:21
and based on how long it takes you to
2:03:21
and based on how long it takes you to find things
2:03:22
find things
2:03:22
find things the different objects in the world will
2:03:24
the different objects in the world will
2:03:24
the different objects in the world will move around to
2:03:25
move around to
2:03:25
move around to try to accommodate like the user's skill
2:03:28
try to accommodate like the user's skill
2:03:28
try to accommodate like the user's skill level
2:03:29
level
2:03:29
level maybe it'll help them like find things
2:03:33
maybe it'll help them like find things
2:03:33
maybe it'll help them like find things faster or it'll slow them down if
2:03:35
faster or it'll slow them down if
2:03:35
faster or it'll slow them down if they're getting through stuff too fast
2:03:38
they're getting through stuff too fast
2:03:38
they're getting through stuff too fast but you find an object you click on it
2:03:41
but you find an object you click on it
2:03:41
but you find an object you click on it and it goes back to like the home base
2:03:44
and it goes back to like the home base
2:03:44
and it goes back to like the home base position so this one in particular is
2:03:48
position so this one in particular is
2:03:48
position so this one in particular is just
2:03:49
just
2:03:49
just a really big search and find because i
2:03:52
a really big search and find because i
2:03:52
a really big search and find because i like those kind of
2:03:53
like those kind of
2:03:53
like those kind of games and stuff but
2:03:56
games and stuff but
2:03:56
games and stuff but the point of adding the ai in here was
2:03:59
the point of adding the ai in here was
2:03:59
the point of adding the ai in here was to really make the game
2:04:01
to really make the game
2:04:01
to really make the game custom to the user so that it matches
2:04:03
custom to the user so that it matches
2:04:03
custom to the user so that it matches their skill level
2:04:04
their skill level
2:04:04
their skill level if they're going through the game really
2:04:06
if they're going through the game really
2:04:06
if they're going through the game really really fast we want to slow them down
2:04:08
really fast we want to slow them down
2:04:08
really fast we want to slow them down so they actually get to experience the
2:04:11
so they actually get to experience the
2:04:11
so they actually get to experience the virtual reality world and if they're
2:04:14
virtual reality world and if they're
2:04:14
virtual reality world and if they're having some problems finding things we
2:04:16
having some problems finding things we
2:04:16
having some problems finding things we want to
2:04:17
want to
2:04:17
want to encourage them to keep going so that
2:04:19
encourage them to keep going so that
2:04:19
encourage them to keep going so that they will
2:04:20
they will
2:04:20
they will eventually finish and they won't just
2:04:22
eventually finish and they won't just
2:04:22
eventually finish and they won't just abandon the world forever
2:04:26
abandon the world forever
2:04:26
abandon the world forever sounds really cool and so um i saw some
2:04:29
sounds really cool and so um i saw some
2:04:29
sounds really cool and so um i saw some some interesting pieces of code in there
2:04:31
some interesting pieces of code in there
2:04:31
some interesting pieces of code in there uh maybe if we can go back to your
2:04:33
uh maybe if we can go back to your
2:04:33
uh maybe if we can go back to your editor um so
2:04:35
editor um so
2:04:35
editor um so in here there's a one-liner to actually
2:04:37
in here there's a one-liner to actually
2:04:37
in here there's a one-liner to actually construct a neural network and i'm
2:04:39
construct a neural network and i'm
2:04:39
construct a neural network and i'm interested in
2:04:41
interested in
2:04:41
interested in so you specify hidden layers what else
2:04:44
so you specify hidden layers what else
2:04:44
so you specify hidden layers what else can i specify
2:04:45
can i specify
2:04:45
can i specify actually in this function in this
2:04:46
actually in this function in this
2:04:46
actually in this function in this constructor for the neural network
2:04:49
constructor for the neural network
2:04:49
constructor for the neural network so with the neural network in particular
2:04:52
so with the neural network in particular
2:04:52
so with the neural network in particular it's really
2:04:52
it's really
2:04:52
it's really just the hidden layers like i think
2:04:56
just the hidden layers like i think
2:04:56
just the hidden layers like i think there are a
2:04:57
there are a
2:04:57
there are a few other options let me see if
2:05:00
few other options let me see if
2:05:00
few other options let me see if i can just get the magic code button
2:05:03
i can just get the magic code button
2:05:03
i can just get the magic code button so yeah we can add an activation node
2:05:07
so yeah we can add an activation node
2:05:07
so yeah we can add an activation node some threshold stuff and if we want to
2:05:10
some threshold stuff and if we want to
2:05:10
some threshold stuff and if we want to make this a leaky
2:05:12
make this a leaky
2:05:12
make this a leaky relu alpha but these are the only other
2:05:15
relu alpha but these are the only other
2:05:15
relu alpha but these are the only other things you can add to that neural
2:05:16
things you can add to that neural
2:05:16
things you can add to that neural network
2:05:17
network
2:05:18
network at least for now i know i know that
2:05:20
at least for now i know i know that
2:05:20
at least for now i know i know that they're working on adding a lot more
2:05:22
they're working on adding a lot more
2:05:22
they're working on adding a lot more features
2:05:23
features
2:05:23
features like one thing brainjs is working on
2:05:25
like one thing brainjs is working on
2:05:25
like one thing brainjs is working on adding
2:05:26
adding
2:05:26
adding are convolutional neural networks which
2:05:28
are convolutional neural networks which
2:05:28
are convolutional neural networks which will be
2:05:29
will be
2:05:29
will be really cool to have in javascript
2:05:33
really cool to have in javascript
2:05:33
really cool to have in javascript yeah i can imagine that this works for
2:05:35
yeah i can imagine that this works for
2:05:35
yeah i can imagine that this works for the most basic scenarios this is
2:05:36
the most basic scenarios this is
2:05:36
the most basic scenarios this is basically just a feed forward network
2:05:38
basically just a feed forward network
2:05:38
basically just a feed forward network um uh so how new is brain.js actually at
2:05:42
um uh so how new is brain.js actually at
2:05:42
um uh so how new is brain.js actually at the moment
2:05:43
the moment
2:05:43
the moment um brainjs has been around for at least
2:05:46
um brainjs has been around for at least
2:05:46
um brainjs has been around for at least a few years now
2:05:48
a few years now
2:05:48
a few years now i think there was a point the original
2:05:51
i think there was a point the original
2:05:51
i think there was a point the original creator
2:05:51
creator
2:05:51
creator passed it off to the current maintainer
2:05:55
passed it off to the current maintainer
2:05:55
passed it off to the current maintainer but it's been around for at least two or
2:05:58
but it's been around for at least two or
2:05:58
but it's been around for at least two or three years
2:06:00
three years
2:06:00
three years okay so it's not completely new it's
2:06:02
okay so it's not completely new it's
2:06:02
okay so it's not completely new it's just that
2:06:03
just that
2:06:03
just that it's still it looks like it it's pretty
2:06:07
it's still it looks like it it's pretty
2:06:07
it's still it looks like it it's pretty limited in functionality right now um
2:06:10
limited in functionality right now um
2:06:10
limited in functionality right now um so what other options do i have if i
2:06:13
so what other options do i have if i
2:06:13
so what other options do i have if i want to build neural networks in
2:06:15
want to build neural networks in
2:06:15
want to build neural networks in javascript or use them for that matter
2:06:18
javascript or use them for that matter
2:06:18
javascript or use them for that matter let's see for building neural networks
2:06:20
let's see for building neural networks
2:06:20
let's see for building neural networks in javascript
2:06:22
in javascript
2:06:22
in javascript um recently i found out that there are a
2:06:25
um recently i found out that there are a
2:06:25
um recently i found out that there are a few
2:06:26
few
2:06:26
few like javascript equivalents to some of
2:06:28
like javascript equivalents to some of
2:06:28
like javascript equivalents to some of the python libraries like
2:06:30
the python libraries like
2:06:30
the python libraries like there's a num js library instead of
2:06:33
there's a num js library instead of
2:06:33
there's a num js library instead of numpy there's the keras js
2:06:37
numpy there's the keras js
2:06:37
numpy there's the keras js so that one is really powerful it's
2:06:41
so that one is really powerful it's
2:06:41
so that one is really powerful it's almost as powerful as the python version
2:06:44
almost as powerful as the python version
2:06:44
almost as powerful as the python version of it
2:06:45
of it
2:06:45
of it but it's still relatively new in
2:06:48
but it's still relatively new in
2:06:48
but it's still relatively new in javascript the whole machine learning
2:06:50
javascript the whole machine learning
2:06:50
javascript the whole machine learning ai thing so there's still a lot of
2:06:53
ai thing so there's still a lot of
2:06:53
ai thing so there's still a lot of development left
2:06:56
development left
2:06:56
development left yeah i can imagine that it's it's
2:06:57
yeah i can imagine that it's it's
2:06:57
yeah i can imagine that it's it's actually the first time that i've seen
2:06:59
actually the first time that i've seen
2:06:59
actually the first time that i've seen an implementation of machine learning
2:07:00
an implementation of machine learning
2:07:00
an implementation of machine learning algorithm
2:07:01
algorithm
2:07:01
algorithm in javascript and it looks really
2:07:03
in javascript and it looks really
2:07:03
in javascript and it looks really interesting i guess if you're
2:07:06
interesting i guess if you're
2:07:06
interesting i guess if you're coming from a background of web
2:07:07
coming from a background of web
2:07:07
coming from a background of web development and doing
2:07:09
development and doing
2:07:09
development and doing css html those kind of things this would
2:07:12
css html those kind of things this would
2:07:12
css html those kind of things this would make more sense to you
2:07:13
make more sense to you
2:07:14
make more sense to you instead of building uh a neural network
2:07:16
instead of building uh a neural network
2:07:16
instead of building uh a neural network in python
2:07:18
in python
2:07:18
in python where you have a lot more options to
2:07:19
where you have a lot more options to
2:07:19
where you have a lot more options to think about um
2:07:22
think about um
2:07:22
think about um i guess that makes it really interesting
2:07:23
i guess that makes it really interesting
2:07:23
i guess that makes it really interesting for people to get started with this
2:07:26
for people to get started with this
2:07:26
for people to get started with this yeah one thing that is really cool about
2:07:28
yeah one thing that is really cool about
2:07:28
yeah one thing that is really cool about the javascript implementations
2:07:31
the javascript implementations
2:07:31
the javascript implementations is that they do strip out a lot of the
2:07:33
is that they do strip out a lot of the
2:07:33
is that they do strip out a lot of the background knowledge you need
2:07:35
background knowledge you need
2:07:35
background knowledge you need so you don't have to know about the
2:07:38
so you don't have to know about the
2:07:38
so you don't have to know about the statistical analysis that goes behind
2:07:40
statistical analysis that goes behind
2:07:40
statistical analysis that goes behind all the calculations you don't have to
2:07:43
all the calculations you don't have to
2:07:43
all the calculations you don't have to worry as much about
2:07:44
worry as much about
2:07:44
worry as much about cleaning data or making sure it's in a
2:07:47
cleaning data or making sure it's in a
2:07:47
cleaning data or making sure it's in a certain form as long as you can put it
2:07:50
certain form as long as you can put it
2:07:50
certain form as long as you can put it in
2:07:50
in
2:07:50
in an array of objects this will handle it
2:07:55
an array of objects this will handle it
2:07:55
an array of objects this will handle it sounds cool and what about performance
2:07:57
sounds cool and what about performance
2:07:57
sounds cool and what about performance how fast uh would you say this
2:07:59
how fast uh would you say this
2:07:59
how fast uh would you say this this is that is one thing
2:08:03
this is that is one thing
2:08:03
this is that is one thing that it does lack a little bit it is
2:08:05
that it does lack a little bit it is
2:08:05
that it does lack a little bit it is kind of slow
2:08:06
kind of slow
2:08:06
kind of slow like so far i haven't used any of the
2:08:09
like so far i haven't used any of the
2:08:09
like so far i haven't used any of the javascript machine learning libraries
2:08:12
javascript machine learning libraries
2:08:12
javascript machine learning libraries in production just because the
2:08:14
in production just because the
2:08:14
in production just because the performance on them
2:08:16
performance on them
2:08:16
performance on them it isn't the greatest when you're
2:08:18
it isn't the greatest when you're
2:08:18
it isn't the greatest when you're feeding them new data at the same time
2:08:20
feeding them new data at the same time
2:08:20
feeding them new data at the same time that they're making predictions
2:08:24
yeah exactly so um
2:08:28
yeah exactly so um
2:08:28
yeah exactly so um if people want to try this out um where
2:08:31
if people want to try this out um where
2:08:31
if people want to try this out um where can they
2:08:32
can they
2:08:32
can they where can they learn more about this do
2:08:33
where can they learn more about this do
2:08:33
where can they learn more about this do you have a code repository where people
2:08:35
you have a code repository where people
2:08:35
you have a code repository where people for example can download this stuff
2:08:37
for example can download this stuff
2:08:38
for example can download this stuff oh yeah yeah i have a github for this
2:08:41
oh yeah yeah i have a github for this
2:08:41
oh yeah yeah i have a github for this in particular so i guess i can just show
2:08:44
in particular so i guess i can just show
2:08:44
in particular so i guess i can just show you the link
2:08:45
you the link
2:08:45
you the link if you go to my user at
2:08:49
if you go to my user at
2:08:49
if you go to my user at foot coder you can go in here to
2:08:52
foot coder you can go in here to
2:08:52
foot coder you can go in here to it's hot that's the name of the game
2:08:57
so yeah this is where you can go to just
2:09:00
so yeah this is where you can go to just
2:09:00
so yeah this is where you can go to just play with this vr
2:09:01
play with this vr
2:09:01
play with this vr app oh cool it looked pretty green to me
2:09:04
app oh cool it looked pretty green to me
2:09:04
app oh cool it looked pretty green to me but we were discussing earlier that
2:09:07
but we were discussing earlier that
2:09:07
but we were discussing earlier that trees make people happy
2:09:08
trees make people happy
2:09:08
trees make people happy yeah yeah exactly yeah i think this game
2:09:10
yeah yeah exactly yeah i think this game
2:09:10
yeah yeah exactly yeah i think this game will make me a happy person
2:09:12
will make me a happy person
2:09:12
will make me a happy person as long as there's enough green in there
2:09:14
as long as there's enough green in there
2:09:14
as long as there's enough green in there and i guess if we
2:09:15
and i guess if we
2:09:15
and i guess if we uh if people want to try this out they
2:09:17
uh if people want to try this out they
2:09:17
uh if people want to try this out they can certainly add more objects to find
2:09:19
can certainly add more objects to find
2:09:19
can certainly add more objects to find more tricks with neural networks i guess
2:09:21
more tricks with neural networks i guess
2:09:21
more tricks with neural networks i guess to make this even more interesting
2:09:22
to make this even more interesting
2:09:22
to make this even more interesting you'd have to add a beach yes yes
2:09:25
you'd have to add a beach yes yes
2:09:25
you'd have to add a beach yes yes definitely
2:09:25
definitely
2:09:25
definitely yeah we need a beach in there yeah
2:09:29
yeah we need a beach in there yeah
2:09:29
yeah we need a beach in there yeah if you were interested in just making
2:09:30
if you were interested in just making
2:09:30
if you were interested in just making your own vr app
2:09:32
your own vr app
2:09:32
your own vr app this is all of the code that
2:09:35
this is all of the code that
2:09:36
this is all of the code that makes up that entire world like i think
2:09:39
makes up that entire world like i think
2:09:39
makes up that entire world like i think it's not even 100 lines of code to make
2:09:41
it's not even 100 lines of code to make
2:09:42
it's not even 100 lines of code to make this whole
2:09:43
this whole
2:09:43
this whole vr world and include the ai in it
2:09:47
vr world and include the ai in it
2:09:47
vr world and include the ai in it so do you have any other plans for
2:09:50
so do you have any other plans for
2:09:50
so do you have any other plans for integrating ai in a vr project
2:09:53
integrating ai in a vr project
2:09:53
integrating ai in a vr project oh yeah so one of the things that i'm
2:09:55
oh yeah so one of the things that i'm
2:09:55
oh yeah so one of the things that i'm going to try
2:09:56
going to try
2:09:56
going to try your smile pop out and that was real
2:09:58
your smile pop out and that was real
2:09:58
your smile pop out and that was real quick it was like oh yeah
2:10:00
quick it was like oh yeah
2:10:00
quick it was like oh yeah [Laughter]
2:10:03
[Laughter]
2:10:03
[Laughter] yeah i'm super excited about this
2:10:05
yeah i'm super excited about this
2:10:05
yeah i'm super excited about this because i haven't seen a lot of
2:10:07
because i haven't seen a lot of
2:10:07
because i haven't seen a lot of ai in vr yet and i'm a little bit
2:10:10
ai in vr yet and i'm a little bit
2:10:10
ai in vr yet and i'm a little bit surprised
2:10:11
surprised
2:10:12
surprised but kind of well i'll connect you with
2:10:14
but kind of well i'll connect you with
2:10:14
but kind of well i'll connect you with some people and
2:10:15
some people and
2:10:15
some people and and we'll we'll get that accelerated
2:10:17
and we'll we'll get that accelerated
2:10:17
and we'll we'll get that accelerated right
2:10:18
right
2:10:18
right yes that would be great yeah if people
2:10:21
yes that would be great yeah if people
2:10:21
yes that would be great yeah if people are interested in this sort of stuff
2:10:23
are interested in this sort of stuff
2:10:23
are interested in this sort of stuff they should definitely check your
2:10:24
they should definitely check your
2:10:24
they should definitely check your website uh there should be a github repo
2:10:27
website uh there should be a github repo
2:10:27
website uh there should be a github repo attached to it if i'm not mistaken um
2:10:30
attached to it if i'm not mistaken um
2:10:30
attached to it if i'm not mistaken um so what we'll do is we will actually
2:10:33
so what we'll do is we will actually
2:10:33
so what we'll do is we will actually share your link as a resource to this to
2:10:36
share your link as a resource to this to
2:10:36
share your link as a resource to this to this episode
2:10:37
this episode
2:10:38
this episode um people may not know this yet but
2:10:41
um people may not know this yet but
2:10:42
um people may not know this yet but um you can watch this whole stream back
2:10:44
um you can watch this whole stream back
2:10:44
um you can watch this whole stream back after we we are done
2:10:45
after we we are done
2:10:45
after we we are done you can actually rewind i believe in in
2:10:48
you can actually rewind i believe in in
2:10:48
you can actually rewind i believe in in the stream um
2:10:50
the stream um
2:10:50
the stream um but after we are done with the show we
2:10:52
but after we are done with the show we
2:10:52
but after we are done with the show we will we will cut up
2:10:53
will we will cut up
2:10:53
will we will cut up all the videos that we've uh made all
2:10:56
all the videos that we've uh made all
2:10:56
all the videos that we've uh made all the sections of the show
2:10:58
the sections of the show
2:10:58
the sections of the show and we will have extra links and
2:10:59
and we will have extra links and
2:10:59
and we will have extra links and resources underneath that so people can
2:11:01
resources underneath that so people can
2:11:02
resources underneath that so people can find your website and give this a shot
2:11:05
find your website and give this a shot
2:11:05
find your website and give this a shot i mean it sounds really interesting
2:11:06
i mean it sounds really interesting
2:11:06
i mean it sounds really interesting trying trying to use javascript to uh
2:11:10
trying trying to use javascript to uh
2:11:10
trying trying to use javascript to uh run ai models for that matter um i i
2:11:13
run ai models for that matter um i i
2:11:13
run ai models for that matter um i i guess from your story that if people
2:11:16
guess from your story that if people
2:11:16
guess from your story that if people want to get started this is a great
2:11:17
want to get started this is a great
2:11:17
want to get started this is a great place
2:11:18
place
2:11:18
place if they want something more they should
2:11:20
if they want something more they should
2:11:20
if they want something more they should probably try the curious gis i think
2:11:23
probably try the curious gis i think
2:11:23
probably try the curious gis i think i think so that one has a little bit
2:11:25
i think so that one has a little bit
2:11:25
i think so that one has a little bit more um
2:11:27
more um
2:11:28
more um i i want to say maybe a little bit more
2:11:30
i i want to say maybe a little bit more
2:11:30
i i want to say maybe a little bit more depth to it
2:11:31
depth to it
2:11:31
depth to it so it does work a lot like the python
2:11:33
so it does work a lot like the python
2:11:34
so it does work a lot like the python version of it
2:11:35
version of it
2:11:35
version of it and that'll give you those cnns that
2:11:38
and that'll give you those cnns that
2:11:38
and that'll give you those cnns that i've been really
2:11:39
i've been really
2:11:39
i've been really really waiting for cool
2:11:42
really waiting for cool
2:11:42
really waiting for cool so once you have those you should
2:11:43
so once you have those you should
2:11:43
so once you have those you should definitely come back and show us the
2:11:44
definitely come back and show us the
2:11:44
definitely come back and show us the upgraded game with beaches
2:11:46
upgraded game with beaches
2:11:46
upgraded game with beaches more trees more grass and a cnn in there
2:11:49
more trees more grass and a cnn in there
2:11:49
more trees more grass and a cnn in there i'm excited
2:11:52
i'm excited
2:11:52
i'm excited we'll see you in february yeah
2:11:55
we'll see you in february yeah
2:11:55
we'll see you in february yeah because another project i'm working on
2:11:58
because another project i'm working on
2:11:58
because another project i'm working on is an air guitar
2:12:00
is an air guitar
2:12:00
is an air guitar so i have these like i have eeg
2:12:04
so i have these like i have eeg
2:12:04
so i have these like i have eeg and emg sensors that i wrap around my
2:12:07
and emg sensors that i wrap around my
2:12:07
and emg sensors that i wrap around my head and put on my arms
2:12:09
head and put on my arms
2:12:09
head and put on my arms and based on the notes that you think
2:12:11
and based on the notes that you think
2:12:12
and based on the notes that you think about and the way that you move your
2:12:13
about and the way that you move your
2:12:13
about and the way that you move your fingers
2:12:14
fingers
2:12:14
fingers it translates that into guitar sounds
2:12:18
it translates that into guitar sounds
2:12:18
it translates that into guitar sounds so i'm super excited about that one
2:12:22
so i'm super excited about that one
2:12:22
so i'm super excited about that one oh wow that sounds great that's going to
2:12:25
oh wow that sounds great that's going to
2:12:25
oh wow that sounds great that's going to be ready for february
2:12:26
be ready for february
2:12:26
be ready for february do you think actually yeah
2:12:29
do you think actually yeah
2:12:30
do you think actually yeah i think that one should be ready by
2:12:32
i think that one should be ready by
2:12:32
i think that one should be ready by february
2:12:34
february
2:12:34
february i i feel like putting up a
2:12:37
i i feel like putting up a
2:12:37
i i feel like putting up a a a i for music show i don't know
2:12:41
a a i for music show i don't know
2:12:41
a a i for music show i don't know how how that will work out but that
2:12:42
how how that will work out but that
2:12:42
how how that will work out but that sounds like a really awesome idea for
2:12:44
sounds like a really awesome idea for
2:12:44
sounds like a really awesome idea for the community
2:12:46
the community
2:12:46
the community if people are interested in that let us
2:12:48
if people are interested in that let us
2:12:48
if people are interested in that let us know
2:12:49
know
2:12:49
know um we could definitely work something
2:12:51
um we could definitely work something
2:12:51
um we could definitely work something out um
2:12:53
out um
2:12:53
out um so after that i would like to thank you
2:12:55
so after that i would like to thank you
2:12:55
so after that i would like to thank you very much for your time
2:12:56
very much for your time
2:12:56
very much for your time um and and for joining us for that
2:12:58
um and and for joining us for that
2:12:58
um and and for joining us for that matter so if people have more questions
2:13:00
matter so if people have more questions
2:13:00
matter so if people have more questions to you
2:13:01
to you
2:13:01
to you they can probably ask me in a live chat
2:13:03
they can probably ask me in a live chat
2:13:03
they can probably ask me in a live chat feel free to join that
2:13:05
feel free to join that
2:13:05
feel free to join that that one also if you have
2:13:08
that one also if you have
2:13:08
that one also if you have questions for us and you happen to have
2:13:11
questions for us and you happen to have
2:13:11
questions for us and you happen to have a very good and a very fun question for
2:13:13
a very good and a very fun question for
2:13:13
a very good and a very fun question for us
2:13:13
us
2:13:13
us you can get a chance at winning a oculus
2:13:16
you can get a chance at winning a oculus
2:13:16
you can get a chance at winning a oculus quest 2 device
2:13:17
quest 2 device
2:13:18
quest 2 device um we will announce a winner at the end
2:13:20
um we will announce a winner at the end
2:13:20
um we will announce a winner at the end of the show and
2:13:21
of the show and
2:13:21
of the show and uh also important to know that um if
2:13:24
uh also important to know that um if
2:13:24
uh also important to know that um if you're not much of a photo person
2:13:27
you're not much of a photo person
2:13:27
you're not much of a photo person uh sort of uh you you're a little camera
2:13:30
uh sort of uh you you're a little camera
2:13:30
uh sort of uh you you're a little camera shy
2:13:32
shy
2:13:32
shy yeah ask a question exactly that that
2:13:35
yeah ask a question exactly that that
2:13:35
yeah ask a question exactly that that was the correct order
2:13:36
was the correct order
2:13:36
was the correct order the photo is for the oculus quest
2:13:38
the photo is for the oculus quest
2:13:38
the photo is for the oculus quest because that's visual
2:13:40
because that's visual
2:13:40
because that's visual and the other thing is for writing as a
2:13:42
and the other thing is for writing as a
2:13:42
and the other thing is for writing as a great question um
2:13:44
great question um
2:13:44
great question um so feel free to ask us questions and
2:13:46
so feel free to ask us questions and
2:13:46
so feel free to ask us questions and send us your pictures of you watching
2:13:48
send us your pictures of you watching
2:13:48
send us your pictures of you watching the global ai show
2:13:49
the global ai show
2:13:49
the global ai show uh live um so thank you very much again
2:13:54
uh live um so thank you very much again
2:13:54
uh live um so thank you very much again thanks for having me with that i would
2:13:56
thanks for having me with that i would
2:13:56
thanks for having me with that i would like to invite
2:13:57
like to invite
2:13:57
like to invite elizabeth to join us
2:14:00
elizabeth to join us
2:14:00
elizabeth to join us and that was fast hi okay
2:14:05
and that was fast hi okay
2:14:05
and that was fast hi okay normally how this works alicia in his
2:14:06
normally how this works alicia in his
2:14:06
normally how this works alicia in his studio is the one guest walks away
2:14:09
studio is the one guest walks away
2:14:09
studio is the one guest walks away he has sort of a short break and the
2:14:11
he has sort of a short break and the
2:14:11
he has sort of a short break and the other guest pops in
2:14:13
other guest pops in
2:14:13
other guest pops in and i've got time to drink my glass of
2:14:16
and i've got time to drink my glass of
2:14:16
and i've got time to drink my glass of water
2:14:17
water
2:14:17
water um maybe stretch a little bit or
2:14:19
um maybe stretch a little bit or
2:14:19
um maybe stretch a little bit or something we should do that a bit so
2:14:21
something we should do that a bit so
2:14:21
something we should do that a bit so stretch ah i'm ready for the next one
2:14:26
stretch ah i'm ready for the next one
2:14:26
stretch ah i'm ready for the next one [Music]
2:14:36
um those puppies for christmas
2:14:40
um those puppies for christmas
2:14:40
um those puppies for christmas i know you're talking about human
2:14:42
i know you're talking about human
2:14:42
i know you're talking about human interactions but you know this is going
2:14:44
interactions but you know this is going
2:14:44
interactions but you know this is going to make
2:14:45
to make
2:14:45
to make my mechanical puppy more more
2:14:47
my mechanical puppy more more
2:14:47
my mechanical puppy more more interactive
2:14:48
interactive
2:14:48
interactive with my seven-year-old sorry the puppy i
2:14:50
with my seven-year-old sorry the puppy i
2:14:50
with my seven-year-old sorry the puppy i have now snores
2:14:52
have now snores
2:14:52
have now snores okay i i i would like a non-snoring
2:14:55
okay i i i would like a non-snoring
2:14:56
okay i i i would like a non-snoring mechanical puppy place
2:14:58
mechanical puppy place
2:14:58
mechanical puppy place we'll see what we can do on that by
2:14:59
we'll see what we can do on that by
2:14:59
we'll see what we can do on that by christmas yeah we'll see what we can do
2:15:01
christmas yeah we'll see what we can do
2:15:02
christmas yeah we'll see what we can do you should know elizabeth you have a
2:15:03
you should know elizabeth you have a
2:15:03
you should know elizabeth you have a global ai puppy
2:15:05
global ai puppy
2:15:05
global ai puppy since tonight in our pre-show so we
2:15:09
since tonight in our pre-show so we
2:15:09
since tonight in our pre-show so we started a little bit early to test out
2:15:10
started a little bit early to test out
2:15:10
started a little bit early to test out our sound settings and video etc
2:15:12
our sound settings and video etc
2:15:12
our sound settings and video etc and we had a short chat and alicia has a
2:15:15
and we had a short chat and alicia has a
2:15:15
and we had a short chat and alicia has a small dog
2:15:16
small dog
2:15:16
small dog and i adopted it somehow it's a global
2:15:19
and i adopted it somehow it's a global
2:15:19
and i adopted it somehow it's a global ai dog
2:15:20
ai dog
2:15:20
ai dog but i'm not sure did i adopt a
2:15:22
but i'm not sure did i adopt a
2:15:22
but i'm not sure did i adopt a mechanical puppy or the real one
2:15:24
mechanical puppy or the real one
2:15:24
mechanical puppy or the real one what what will it be he's so my training
2:15:27
what what will it be he's so my training
2:15:27
what what will it be he's so my training algorithms are awesome
2:15:29
algorithms are awesome
2:15:29
algorithms are awesome and he is just extremely well trained
2:15:32
and he is just extremely well trained
2:15:32
and he is just extremely well trained okay then i don't mind either one is
2:15:34
okay then i don't mind either one is
2:15:34
okay then i don't mind either one is fine
2:15:36
fine
2:15:36
fine so elizabeth you're you're actually uh
2:15:40
so elizabeth you're you're actually uh
2:15:40
so elizabeth you're you're actually uh working for
2:15:40
working for
2:15:40
working for cisco collaboration uh as the head of
2:15:43
cisco collaboration uh as the head of
2:15:43
cisco collaboration uh as the head of innovation
2:15:44
innovation
2:15:44
innovation that sounds like an interesting job not
2:15:47
that sounds like an interesting job not
2:15:47
that sounds like an interesting job not for all cisco for
2:15:48
for all cisco for
2:15:48
for all cisco for just specifically collaboration our cto
2:15:50
just specifically collaboration our cto
2:15:50
just specifically collaboration our cto team yes
2:15:51
team yes
2:15:51
team yes yeah you collaborate all day
2:15:55
yeah you collaborate all day
2:15:55
yeah you collaborate all day on innovation or how does this actually
2:15:56
on innovation or how does this actually
2:15:56
on innovation or how does this actually work can you tell us a little bit what
2:15:58
work can you tell us a little bit what
2:15:58
work can you tell us a little bit what your day
2:15:59
your day
2:15:59
your day looks like actually as an innovation uh
2:16:01
looks like actually as an innovation uh
2:16:01
looks like actually as an innovation uh person
2:16:03
person
2:16:03
person uh sure i actually need to try to share
2:16:05
uh sure i actually need to try to share
2:16:05
uh sure i actually need to try to share some slides here too
2:16:06
some slides here too
2:16:06
some slides here too this is giving me a little bit of an
2:16:08
this is giving me a little bit of an
2:16:08
this is giving me a little bit of an issue so i'm gonna see if i can bring
2:16:10
issue so i'm gonna see if i can bring
2:16:10
issue so i'm gonna see if i can bring this up while i'm talking
2:16:12
this up while i'm talking
2:16:12
this up while i'm talking sure i'm a picture person so let me know
2:16:15
sure i'm a picture person so let me know
2:16:15
sure i'm a picture person so let me know if that comes up and you guys can see
2:16:16
if that comes up and you guys can see
2:16:16
if that comes up and you guys can see that
2:16:17
that
2:16:17
that yes it's visible on screen you can see
2:16:20
yes it's visible on screen you can see
2:16:20
yes it's visible on screen you can see the okay perfect
2:16:22
the okay perfect
2:16:22
the okay perfect so yeah what what i do on a day-to-day
2:16:25
so yeah what what i do on a day-to-day
2:16:25
so yeah what what i do on a day-to-day basis
2:16:25
basis
2:16:25
basis changes pretty much but i head up our
2:16:28
changes pretty much but i head up our
2:16:28
changes pretty much but i head up our part of our team that's focused on
2:16:30
part of our team that's focused on
2:16:30
part of our team that's focused on disruptive innovation for cisco
2:16:32
disruptive innovation for cisco
2:16:32
disruptive innovation for cisco collaboration so we tend to look at
2:16:33
collaboration so we tend to look at
2:16:33
collaboration so we tend to look at things that are
2:16:34
things that are
2:16:34
things that are several years down the pike that could
2:16:36
several years down the pike that could
2:16:36
several years down the pike that could become products so we get to do a lot of
2:16:38
become products so we get to do a lot of
2:16:38
become products so we get to do a lot of tinkering a lot of playing a lot of
2:16:40
tinkering a lot of playing a lot of
2:16:40
tinkering a lot of playing a lot of exploring which is a lot of fun
2:16:42
exploring which is a lot of fun
2:16:42
exploring which is a lot of fun but we're also dealing with technology
2:16:43
but we're also dealing with technology
2:16:43
but we're also dealing with technology that we likely won't see
2:16:45
that we likely won't see
2:16:45
that we likely won't see in our hands and in common usage anytime
2:16:47
in our hands and in common usage anytime
2:16:47
in our hands and in common usage anytime in the very immediate future
2:16:49
in the very immediate future
2:16:49
in the very immediate future so we definitely have to have the long
2:16:51
so we definitely have to have the long
2:16:52
so we definitely have to have the long time frame in mind
2:16:54
time frame in mind
2:16:54
time frame in mind what i wanted to talk about a little bit
2:16:56
what i wanted to talk about a little bit
2:16:56
what i wanted to talk about a little bit today is
2:16:57
today is
2:16:57
today is basically taking us probably another 10
2:16:59
basically taking us probably another 10
2:16:59
basically taking us probably another 10 000 levels up from the last presentation
2:17:01
000 levels up from the last presentation
2:17:01
000 levels up from the last presentation which is awesome by the way
2:17:02
which is awesome by the way
2:17:02
which is awesome by the way and looking a little bit more at what is
2:17:05
and looking a little bit more at what is
2:17:05
and looking a little bit more at what is the often overlooked side of technology
2:17:09
the often overlooked side of technology
2:17:09
the often overlooked side of technology which is the people that are actually
2:17:10
which is the people that are actually
2:17:10
which is the people that are actually going to use it i know this is a fault
2:17:13
going to use it i know this is a fault
2:17:13
going to use it i know this is a fault of myself and my team we often get very
2:17:14
of myself and my team we often get very
2:17:14
of myself and my team we often get very excited about what the technology
2:17:16
excited about what the technology
2:17:16
excited about what the technology can do the capabilities of the
2:17:18
can do the capabilities of the
2:17:18
can do the capabilities of the technology which is great and amazing
2:17:20
technology which is great and amazing
2:17:20
technology which is great and amazing but we can very easily get bogged down
2:17:23
but we can very easily get bogged down
2:17:23
but we can very easily get bogged down into
2:17:24
into
2:17:24
into this is so cool we should do this and
2:17:26
this is so cool we should do this and
2:17:26
this is so cool we should do this and not stop to think
2:17:28
not stop to think
2:17:28
not stop to think should we actually do this do people
2:17:29
should we actually do this do people
2:17:29
should we actually do this do people want this will people use this does it
2:17:30
want this will people use this does it
2:17:30
want this will people use this does it make their life better
2:17:32
make their life better
2:17:32
make their life better so that's primarily what i want to talk
2:17:33
so that's primarily what i want to talk
2:17:34
so that's primarily what i want to talk about today and then relate that back to
2:17:36
about today and then relate that back to
2:17:36
about today and then relate that back to human and ai interaction but start by
2:17:39
human and ai interaction but start by
2:17:39
human and ai interaction but start by talking about
2:17:40
talking about
2:17:40
talking about human interaction what makes us uniquely
2:17:42
human interaction what makes us uniquely
2:17:42
human interaction what makes us uniquely human
2:17:43
human
2:17:43
human and i swear i didn't know you guys were
2:17:45
and i swear i didn't know you guys were
2:17:45
and i swear i didn't know you guys were going to be talking about trees earlier
2:17:47
going to be talking about trees earlier
2:17:47
going to be talking about trees earlier but i do have a slide with trees
2:17:49
but i do have a slide with trees
2:17:49
but i do have a slide with trees here but i wanted to start with
2:17:53
here but i wanted to start with
2:17:53
here but i wanted to start with why people do the things that we do
2:17:56
why people do the things that we do
2:17:56
why people do the things that we do so bear with me a little bit i told you
2:17:57
so bear with me a little bit i told you
2:17:57
so bear with me a little bit i told you we're going to go several layers up here
2:18:00
we're going to go several layers up here
2:18:00
we're going to go several layers up here but if you've ever walked down a
2:18:02
but if you've ever walked down a
2:18:02
but if you've ever walked down a tree-lined street in the middle of
2:18:03
tree-lined street in the middle of
2:18:03
tree-lined street in the middle of summer
2:18:04
summer
2:18:04
summer and just heard that hum of lawn mowers
2:18:06
and just heard that hum of lawn mowers
2:18:06
and just heard that hum of lawn mowers and smelled the fresh-cut grass and
2:18:08
and smelled the fresh-cut grass and
2:18:08
and smelled the fresh-cut grass and found yourself
2:18:09
found yourself
2:18:09
found yourself instantly transported back to summers as
2:18:12
instantly transported back to summers as
2:18:12
instantly transported back to summers as a kid
2:18:13
a kid
2:18:13
a kid feeling the sun on your face the warm
2:18:14
feeling the sun on your face the warm
2:18:14
feeling the sun on your face the warm breeze in your hair
2:18:16
breeze in your hair
2:18:16
breeze in your hair you know and understand how powerful
2:18:18
you know and understand how powerful
2:18:18
you know and understand how powerful certain sight smells senses can evoke
2:18:21
certain sight smells senses can evoke
2:18:21
certain sight smells senses can evoke these vivid vivid memories
2:18:23
these vivid vivid memories
2:18:23
these vivid vivid memories that's part of what we as people that's
2:18:26
that's part of what we as people that's
2:18:26
that's part of what we as people that's what we do this is how we interact
2:18:28
what we do this is how we interact
2:18:28
what we do this is how we interact so when it comes to things like the
2:18:30
so when it comes to things like the
2:18:30
so when it comes to things like the smell of
2:18:31
smell of
2:18:31
smell of cookies baking at the holiday time if
2:18:33
cookies baking at the holiday time if
2:18:33
cookies baking at the holiday time if you've ever gone through
2:18:34
you've ever gone through
2:18:34
you've ever gone through hours and hours of painful painful
2:18:36
hours and hours of painful painful
2:18:36
hours and hours of painful painful travel going for airports
2:18:37
travel going for airports
2:18:38
travel going for airports and horrendous traffic and even worse
2:18:40
and horrendous traffic and even worse
2:18:40
and horrendous traffic and even worse weather just to get home to feel
2:18:42
weather just to get home to feel
2:18:42
weather just to get home to feel hear your family argue at the holidays
2:18:44
hear your family argue at the holidays
2:18:44
hear your family argue at the holidays to smell this familiar smells
2:18:46
to smell this familiar smells
2:18:46
to smell this familiar smells you understand how important it is to
2:18:48
you understand how important it is to
2:18:48
you understand how important it is to get that sense and feeling of home that
2:18:50
get that sense and feeling of home that
2:18:50
get that sense and feeling of home that sense and feeling of community this is
2:18:52
sense and feeling of community this is
2:18:52
sense and feeling of community this is partially
2:18:53
partially
2:18:53
partially what we do this is how we understand
2:18:55
what we do this is how we understand
2:18:55
what we do this is how we understand what it is to be human these things are
2:18:57
what it is to be human these things are
2:18:57
what it is to be human these things are not necessarily
2:18:59
not necessarily
2:18:59
not necessarily logical nor programmable so when we're
2:19:01
logical nor programmable so when we're
2:19:01
logical nor programmable so when we're talking about what it is to be uniquely
2:19:03
talking about what it is to be uniquely
2:19:03
talking about what it is to be uniquely human
2:19:04
human
2:19:04
human we do we as a species do things that
2:19:07
we do we as a species do things that
2:19:07
we do we as a species do things that only humans do
2:19:08
only humans do
2:19:08
only humans do we do things that are things that are
2:19:10
we do things that are things that are
2:19:10
we do things that are things that are emotional things that are irrational
2:19:12
emotional things that are irrational
2:19:12
emotional things that are irrational things that are
2:19:13
things that are
2:19:13
things that are oftentimes illogical and it is what
2:19:16
oftentimes illogical and it is what
2:19:16
oftentimes illogical and it is what makes
2:19:16
makes
2:19:16
makes us unique we have traditions we have
2:19:19
us unique we have traditions we have
2:19:19
us unique we have traditions we have things that many of them that don't make
2:19:21
things that many of them that don't make
2:19:21
things that many of them that don't make any sense
2:19:22
any sense
2:19:22
any sense we repeat them every year we do crazy
2:19:24
we repeat them every year we do crazy
2:19:24
we repeat them every year we do crazy things when our sports teams are playing
2:19:26
things when our sports teams are playing
2:19:26
things when our sports teams are playing we do them just because it makes us
2:19:28
we do them just because it makes us
2:19:28
we do them just because it makes us happy and we also the flip side of it we
2:19:31
happy and we also the flip side of it we
2:19:31
happy and we also the flip side of it we avoid things because we have these
2:19:32
avoid things because we have these
2:19:32
avoid things because we have these completely irrational fears
2:19:34
completely irrational fears
2:19:34
completely irrational fears and any time we try to apply that to
2:19:37
and any time we try to apply that to
2:19:37
and any time we try to apply that to computers and technology and ai it's
2:19:38
computers and technology and ai it's
2:19:38
computers and technology and ai it's very hard to program
2:19:40
very hard to program
2:19:40
very hard to program that humanity into it that's what we
2:19:42
that humanity into it that's what we
2:19:42
that humanity into it that's what we want to get into a little bit
2:19:44
want to get into a little bit
2:19:44
want to get into a little bit we also as a society we do some things
2:19:46
we also as a society we do some things
2:19:46
we also as a society we do some things that are
2:19:48
that are
2:19:48
that are decidedly unpredictable as individuals
2:19:51
decidedly unpredictable as individuals
2:19:52
decidedly unpredictable as individuals but can be
2:19:52
but can be
2:19:52
but can be very predictable in a group when we do
2:19:54
very predictable in a group when we do
2:19:54
very predictable in a group when we do things in a group i like this quote from
2:19:56
things in a group i like this quote from
2:19:56
things in a group i like this quote from the french poet journalist uh anatol
2:19:58
the french poet journalist uh anatol
2:19:58
the french poet journalist uh anatol friends
2:19:59
friends
2:19:59
friends he said human nature is to think wisely
2:20:02
he said human nature is to think wisely
2:20:02
he said human nature is to think wisely and act in an absurd fashion
2:20:04
and act in an absurd fashion
2:20:04
and act in an absurd fashion today i want to talk about basically
2:20:05
today i want to talk about basically
2:20:05
today i want to talk about basically what makes us human what drives our
2:20:07
what makes us human what drives our
2:20:07
what makes us human what drives our interactions
2:20:08
interactions
2:20:08
interactions and how we can bring those human to
2:20:10
and how we can bring those human to
2:20:10
and how we can bring those human to human relational tenants
2:20:12
human relational tenants
2:20:12
human relational tenants into what we expect of our technology
2:20:15
into what we expect of our technology
2:20:15
into what we expect of our technology how we think our
2:20:16
how we think our
2:20:16
how we think our technology will be used and adopted i
2:20:19
technology will be used and adopted i
2:20:19
technology will be used and adopted i also want to talk a little bit about my
2:20:21
also want to talk a little bit about my
2:20:21
also want to talk a little bit about my personal predictions and theories
2:20:23
personal predictions and theories
2:20:23
personal predictions and theories on what i think technology adoption
2:20:24
on what i think technology adoption
2:20:24
on what i think technology adoption behavioral shifts we can expect
2:20:26
behavioral shifts we can expect
2:20:26
behavioral shifts we can expect in the next five to ten years please
2:20:28
in the next five to ten years please
2:20:28
in the next five to ten years please take them with a grain of salt these are
2:20:29
take them with a grain of salt these are
2:20:30
take them with a grain of salt these are my predictions
2:20:30
my predictions
2:20:30
my predictions feel free to argue them and what we
2:20:33
feel free to argue them and what we
2:20:33
feel free to argue them and what we think breakthroughs in the technology
2:20:34
think breakthroughs in the technology
2:20:34
think breakthroughs in the technology itself computer vision machine learning
2:20:36
itself computer vision machine learning
2:20:36
itself computer vision machine learning virtual reality artificial intelligence
2:20:37
virtual reality artificial intelligence
2:20:37
virtual reality artificial intelligence what we think that might translate to
2:20:39
what we think that might translate to
2:20:39
what we think that might translate to but first i'm going to spend more of the
2:20:40
but first i'm going to spend more of the
2:20:40
but first i'm going to spend more of the time talking about humans as a species
2:20:43
time talking about humans as a species
2:20:43
time talking about humans as a species and
2:20:43
and
2:20:43
and who we are and why we do things we do we
2:20:46
who we are and why we do things we do we
2:20:46
who we are and why we do things we do we are relational
2:20:47
are relational
2:20:47
are relational we are social creatures so much of what
2:20:49
we are social creatures so much of what
2:20:49
we are social creatures so much of what we do
2:20:50
we do
2:20:50
we do evolves around groups we do things
2:20:52
evolves around groups we do things
2:20:52
evolves around groups we do things together we go to the movies together we
2:20:54
together we go to the movies together we
2:20:54
together we go to the movies together we watch sporting events together we attend
2:20:56
watch sporting events together we attend
2:20:56
watch sporting events together we attend music concerts together
2:20:57
music concerts together
2:20:57
music concerts together we like to sweat together in fitness
2:20:59
we like to sweat together in fitness
2:20:59
we like to sweat together in fitness classes we'll go out of our way to sit
2:21:01
classes we'll go out of our way to sit
2:21:01
classes we'll go out of our way to sit shoulder and shoulder to watch a
2:21:03
shoulder and shoulder to watch a
2:21:03
shoulder and shoulder to watch a theatrical performance to watch a
2:21:05
theatrical performance to watch a
2:21:05
theatrical performance to watch a symphony we'll spend more money
2:21:07
symphony we'll spend more money
2:21:07
symphony we'll spend more money more time more time in traffic and lines
2:21:09
more time more time in traffic and lines
2:21:09
more time more time in traffic and lines and uncomfortable seats just to do
2:21:11
and uncomfortable seats just to do
2:21:11
and uncomfortable seats just to do things together
2:21:12
things together
2:21:12
things together as a group and this is something in the
2:21:14
as a group and this is something in the
2:21:14
as a group and this is something in the current world situation
2:21:16
current world situation
2:21:16
current world situation many people are missing this element of
2:21:18
many people are missing this element of
2:21:18
many people are missing this element of it because we like
2:21:19
it because we like
2:21:19
it because we like being together with other people we like
2:21:21
being together with other people we like
2:21:21
being together with other people we like that social element
2:21:22
that social element
2:21:22
that social element we work in groups live in groups play in
2:21:25
we work in groups live in groups play in
2:21:25
we work in groups live in groups play in groups it's
2:21:25
groups it's
2:21:26
groups it's who we are there's also uh as a
2:21:29
who we are there's also uh as a
2:21:29
who we are there's also uh as a group we can be very unpredictable and
2:21:31
group we can be very unpredictable and
2:21:31
group we can be very unpredictable and we can change our minds collectively
2:21:33
we can change our minds collectively
2:21:33
we can change our minds collectively i have to bring up
2:21:36
i have to bring up
2:21:36
i have to bring up open floor plans this is always
2:21:38
open floor plans this is always
2:21:38
open floor plans this is always something that hits a nerve with people
2:21:40
something that hits a nerve with people
2:21:40
something that hits a nerve with people they either love them or hate them
2:21:41
they either love them or hate them
2:21:41
they either love them or hate them but if you think about as a society as a
2:21:44
but if you think about as a society as a
2:21:44
but if you think about as a society as a group as a global society you don't have
2:21:46
group as a global society you don't have
2:21:46
group as a global society you don't have to go that many years far back to think
2:21:48
to go that many years far back to think
2:21:48
to go that many years far back to think when all the companies were moving to
2:21:50
when all the companies were moving to
2:21:50
when all the companies were moving to these open floor plans
2:21:51
these open floor plans
2:21:51
these open floor plans and the internet was flooded with these
2:21:54
and the internet was flooded with these
2:21:54
and the internet was flooded with these who's moving open floor plans why
2:21:55
who's moving open floor plans why
2:21:55
who's moving open floor plans why they're great what we love about them
2:21:57
they're great what we love about them
2:21:57
they're great what we love about them how you work better in them
2:21:58
how you work better in them
2:21:58
how you work better in them why to do it now how you're going to
2:21:59
why to do it now how you're going to
2:21:59
why to do it now how you're going to save money and then it's only
2:22:02
save money and then it's only
2:22:02
save money and then it's only a short period of time after that we
2:22:04
a short period of time after that we
2:22:04
a short period of time after that we collectively as a society
2:22:06
collectively as a society
2:22:06
collectively as a society changed our mind and we basically said
2:22:08
changed our mind and we basically said
2:22:08
changed our mind and we basically said open floor plans are bad
2:22:09
open floor plans are bad
2:22:09
open floor plans are bad they don't work this is why it's failed
2:22:11
they don't work this is why it's failed
2:22:11
they don't work this is why it's failed this is why you should ditch it even
2:22:12
this is why you should ditch it even
2:22:12
this is why you should ditch it even though you just invested in it this is
2:22:13
though you just invested in it this is
2:22:13
though you just invested in it this is how we're going to fix them going
2:22:14
how we're going to fix them going
2:22:14
how we're going to fix them going forward
2:22:15
forward
2:22:15
forward we as a society are very changeable and
2:22:18
we as a society are very changeable and
2:22:18
we as a society are very changeable and if we believe the
2:22:20
if we believe the
2:22:20
if we believe the albert einstein quote here that
2:22:22
albert einstein quote here that
2:22:22
albert einstein quote here that intelligence is the ability
2:22:23
intelligence is the ability
2:22:24
intelligence is the ability measure of intelligence the ability to
2:22:25
measure of intelligence the ability to
2:22:25
measure of intelligence the ability to change then humans are
2:22:27
change then humans are
2:22:27
change then humans are truly superior i'll bite quite fickle in
2:22:29
truly superior i'll bite quite fickle in
2:22:29
truly superior i'll bite quite fickle in that regard
2:22:30
that regard
2:22:30
that regard so when this comes to human interaction
2:22:33
so when this comes to human interaction
2:22:33
so when this comes to human interaction we're very complex in the way that we
2:22:34
we're very complex in the way that we
2:22:34
we're very complex in the way that we interact
2:22:35
interact
2:22:35
interact again this is why it is so challenging
2:22:37
again this is why it is so challenging
2:22:37
again this is why it is so challenging to make realistic ai
2:22:39
to make realistic ai
2:22:39
to make realistic ai we have social cues and norms that are
2:22:42
we have social cues and norms that are
2:22:42
we have social cues and norms that are just
2:22:43
just
2:22:43
just phenomenally interrelated and and
2:22:46
phenomenally interrelated and and
2:22:46
phenomenally interrelated and and very complex there was a a
2:22:50
very complex there was a a
2:22:50
very complex there was a a a a macarthur genius grant by the name
2:22:54
a a macarthur genius grant by the name
2:22:54
a a macarthur genius grant by the name of amos taversky
2:22:56
of amos taversky
2:22:56
of amos taversky him and his colleague daniel kahneman
2:22:58
him and his colleague daniel kahneman
2:22:58
him and his colleague daniel kahneman who was
2:22:59
who was
2:22:59
who was he won the nobel peace prize award in
2:23:02
he won the nobel peace prize award in
2:23:02
he won the nobel peace prize award in economics
2:23:03
economics
2:23:03
economics there were two psychologists back in the
2:23:04
there were two psychologists back in the
2:23:04
there were two psychologists back in the 70s and they went through this
2:23:06
70s and they went through this
2:23:06
70s and they went through this experiment to actually
2:23:08
experiment to actually
2:23:08
experiment to actually prove scientifically prove that humans
2:23:11
prove scientifically prove that humans
2:23:11
prove scientifically prove that humans as a species are not rational creatures
2:23:14
as a species are not rational creatures
2:23:14
as a species are not rational creatures and we just systematically make
2:23:17
and we just systematically make
2:23:17
and we just systematically make decisions that
2:23:18
decisions that
2:23:18
decisions that really defy logic but that that is
2:23:21
really defy logic but that that is
2:23:21
really defy logic but that that is actually a
2:23:22
actually a
2:23:22
actually a good thing because we don't make these
2:23:24
good thing because we don't make these
2:23:24
good thing because we don't make these decisions solely by
2:23:26
decisions solely by
2:23:26
decisions solely by weighing facts it's not just a factual
2:23:28
weighing facts it's not just a factual
2:23:28
weighing facts it's not just a factual based decision we often add
2:23:29
based decision we often add
2:23:30
based decision we often add all these other things that are very
2:23:31
all these other things that are very
2:23:31
all these other things that are very illogical and irrational but the outcome
2:23:33
illogical and irrational but the outcome
2:23:33
illogical and irrational but the outcome we actually often make better decisions
2:23:35
we actually often make better decisions
2:23:35
we actually often make better decisions because of it
2:23:36
because of it
2:23:36
because of it so they were really focused on the fact
2:23:38
so they were really focused on the fact
2:23:38
so they were really focused on the fact of the the relationships
2:23:40
of the the relationships
2:23:40
of the the relationships people are fairly are not that
2:23:41
people are fairly are not that
2:23:41
people are fairly are not that complicated themselves but when you
2:23:42
complicated themselves but when you
2:23:42
complicated themselves but when you start getting the relationships between
2:23:44
start getting the relationships between
2:23:44
start getting the relationships between people the dynamics it gets phenomenally
2:23:46
people the dynamics it gets phenomenally
2:23:46
people the dynamics it gets phenomenally more complicated
2:23:48
more complicated
2:23:48
more complicated and it was very hard to pinpoint that
2:23:50
and it was very hard to pinpoint that
2:23:50
and it was very hard to pinpoint that but it creates these really
2:23:52
but it creates these really
2:23:52
but it creates these really rich and multi-level level decisions
2:23:55
rich and multi-level level decisions
2:23:55
rich and multi-level level decisions that really
2:23:56
that really
2:23:56
that really are the fabric of our society one
2:23:59
are the fabric of our society one
2:23:59
are the fabric of our society one example i want to use
2:24:00
example i want to use
2:24:00
example i want to use about this is expression so this is
2:24:03
about this is expression so this is
2:24:03
about this is expression so this is actually pictures of my eldest daughter
2:24:05
actually pictures of my eldest daughter
2:24:05
actually pictures of my eldest daughter and aside from just wanting to put big
2:24:07
and aside from just wanting to put big
2:24:07
and aside from just wanting to put big pictures of my cute kid in here
2:24:09
pictures of my cute kid in here
2:24:09
pictures of my cute kid in here and have black male photos for when
2:24:10
and have black male photos for when
2:24:10
and have black male photos for when she's later older
2:24:12
she's later older
2:24:12
she's later older she was a very very expressive baby
2:24:15
she was a very very expressive baby
2:24:16
she was a very very expressive baby more so than my other kids quite frankly
2:24:18
more so than my other kids quite frankly
2:24:18
more so than my other kids quite frankly it is very easy to tell just by looking
2:24:20
it is very easy to tell just by looking
2:24:20
it is very easy to tell just by looking at some of these pictures
2:24:21
at some of these pictures
2:24:21
at some of these pictures exactly what she is thinking what she is
2:24:23
exactly what she is thinking what she is
2:24:23
exactly what she is thinking what she is feeling and experiencing
2:24:25
feeling and experiencing
2:24:25
feeling and experiencing when people are babies they don't have a
2:24:27
when people are babies they don't have a
2:24:27
when people are babies they don't have a filter you haven't yet
2:24:28
filter you haven't yet
2:24:28
filter you haven't yet learned how to have that poker face how
2:24:30
learned how to have that poker face how
2:24:30
learned how to have that poker face how to have that filter
2:24:31
to have that filter
2:24:31
to have that filter so if you've ever given a young child a
2:24:33
so if you've ever given a young child a
2:24:33
so if you've ever given a young child a food they dislike for instance
2:24:35
food they dislike for instance
2:24:35
food they dislike for instance they are not hesitant to show you
2:24:37
they are not hesitant to show you
2:24:37
they are not hesitant to show you through a variety of
2:24:39
through a variety of
2:24:39
through a variety of nonverbal reactions exactly their
2:24:41
nonverbal reactions exactly their
2:24:41
nonverbal reactions exactly their opinion and how much they dislike that
2:24:43
opinion and how much they dislike that
2:24:43
opinion and how much they dislike that this is something that over time we
2:24:46
this is something that over time we
2:24:46
this is something that over time we learn this concept of
2:24:47
learn this concept of
2:24:47
learn this concept of a poker face we learn to not wear a
2:24:50
a poker face we learn to not wear a
2:24:50
a poker face we learn to not wear a heart on our sleeve
2:24:51
heart on our sleeve
2:24:51
heart on our sleeve but there are still micro expressions
2:24:54
but there are still micro expressions
2:24:54
but there are still micro expressions that are constantly
2:24:55
that are constantly
2:24:55
that are constantly going on in day-to-day communications
2:24:57
going on in day-to-day communications
2:24:57
going on in day-to-day communications these
2:24:58
these
2:24:58
these reveal our underlying emotions and even
2:25:00
reveal our underlying emotions and even
2:25:00
reveal our underlying emotions and even though we learn more to control those
2:25:03
though we learn more to control those
2:25:03
though we learn more to control those fortunately in most work settings and
2:25:05
fortunately in most work settings and
2:25:05
fortunately in most work settings and meetings we can
2:25:06
meetings we can
2:25:06
meetings we can hide some of our inner thoughts but it
2:25:09
hide some of our inner thoughts but it
2:25:09
hide some of our inner thoughts but it is part of that complex web of nonverbal
2:25:11
is part of that complex web of nonverbal
2:25:11
is part of that complex web of nonverbal communications facial expressions are
2:25:13
communications facial expressions are
2:25:13
communications facial expressions are huge and that's just one piece of it
2:25:15
huge and that's just one piece of it
2:25:15
huge and that's just one piece of it things like body movements
2:25:16
things like body movements
2:25:16
things like body movements posture proximity eye contact tone this
2:25:19
posture proximity eye contact tone this
2:25:19
posture proximity eye contact tone this is all part of that
2:25:20
is all part of that
2:25:20
is all part of that complex web of non-verbal communications
2:25:24
complex web of non-verbal communications
2:25:24
complex web of non-verbal communications and those arguments been made that
2:25:25
and those arguments been made that
2:25:25
and those arguments been made that non-verbal is actually far more
2:25:27
non-verbal is actually far more
2:25:27
non-verbal is actually far more important than verbal
2:25:28
important than verbal
2:25:28
important than verbal there was a professor albert morabian
2:25:31
there was a professor albert morabian
2:25:31
there was a professor albert morabian who is a psychology professor at the
2:25:32
who is a psychology professor at the
2:25:32
who is a psychology professor at the university of california in la
2:25:34
university of california in la
2:25:34
university of california in la he did this study to explore how
2:25:37
he did this study to explore how
2:25:37
he did this study to explore how important
2:25:38
important
2:25:38
important verbal versus nonverbal communications
2:25:40
verbal versus nonverbal communications
2:25:40
verbal versus nonverbal communications are and
2:25:41
are and
2:25:41
are and he had this very well cited study you
2:25:43
he had this very well cited study you
2:25:43
he had this very well cited study you guys probably heard about it was in
2:25:45
guys probably heard about it was in
2:25:45
guys probably heard about it was in in early 1960s and he basically
2:25:48
in early 1960s and he basically
2:25:48
in early 1960s and he basically concluded that body language accounts
2:25:49
concluded that body language accounts
2:25:50
concluded that body language accounts for
2:25:50
for
2:25:50
for 55 personal communication this
2:25:54
55 personal communication this
2:25:54
55 personal communication this tone of your voice purely the tone
2:25:56
tone of your voice purely the tone
2:25:56
tone of your voice purely the tone accounts for another 38 percent when it
2:25:57
accounts for another 38 percent when it
2:25:57
accounts for another 38 percent when it gets to actually what you are saying the
2:25:59
gets to actually what you are saying the
2:25:59
gets to actually what you are saying the words you use
2:26:00
words you use
2:26:00
words you use that's only 7 it gives you just a
2:26:03
that's only 7 it gives you just a
2:26:03
that's only 7 it gives you just a glimpse at how important
2:26:04
glimpse at how important
2:26:04
glimpse at how important all those non-verbal cues and human and
2:26:06
all those non-verbal cues and human and
2:26:06
all those non-verbal cues and human and actual interaction really really are
2:26:09
actual interaction really really are
2:26:09
actual interaction really really are this also clues us into why first
2:26:11
this also clues us into why first
2:26:11
this also clues us into why first impressions are so important
2:26:13
impressions are so important
2:26:13
impressions are so important there was another very interesting study
2:26:15
there was another very interesting study
2:26:15
there was another very interesting study done by two psychologists at princeton
2:26:17
done by two psychologists at princeton
2:26:17
done by two psychologists at princeton janine willis and alexander todorov they
2:26:20
janine willis and alexander todorov they
2:26:20
janine willis and alexander todorov they did this series of experiments
2:26:22
did this series of experiments
2:26:22
did this series of experiments that revealed that it only takes a tenth
2:26:24
that revealed that it only takes a tenth
2:26:24
that revealed that it only takes a tenth of a second to form an impression
2:26:26
of a second to form an impression
2:26:26
of a second to form an impression of a stranger from their face so this
2:26:29
of a stranger from their face so this
2:26:29
of a stranger from their face so this really gives you an idea about how
2:26:31
really gives you an idea about how
2:26:31
really gives you an idea about how important those snap judgments are
2:26:33
important those snap judgments are
2:26:33
important those snap judgments are and what was interesting is they did
2:26:35
and what was interesting is they did
2:26:35
and what was interesting is they did these studies for longer and realized
2:26:36
these studies for longer and realized
2:26:36
these studies for longer and realized that
2:26:37
that
2:26:37
that even after you were exposed to somebody
2:26:39
even after you were exposed to somebody
2:26:39
even after you were exposed to somebody for a longer period of time
2:26:40
for a longer period of time
2:26:40
for a longer period of time those impressions usually didn't
2:26:42
those impressions usually didn't
2:26:42
those impressions usually didn't drastically change you would get more
2:26:44
drastically change you would get more
2:26:44
drastically change you would get more confidence in your initial judgment but
2:26:46
confidence in your initial judgment but
2:26:46
confidence in your initial judgment but often time that first impressions was
2:26:48
often time that first impressions was
2:26:48
often time that first impressions was what's
2:26:49
what's
2:26:49
what's uh stood with you this is also some of
2:26:51
uh stood with you this is also some of
2:26:51
uh stood with you this is also some of the science behind
2:26:53
the science behind
2:26:53
the science behind why it matters so much when we were
2:26:55
why it matters so much when we were
2:26:55
why it matters so much when we were making that first
2:26:56
making that first
2:26:56
making that first indirect interaction with someone that
2:26:58
indirect interaction with someone that
2:26:58
indirect interaction with someone that first presentation
2:27:00
first presentation
2:27:00
first presentation it also explains the entire concept of
2:27:02
it also explains the entire concept of
2:27:02
it also explains the entire concept of speed dating which is basically these
2:27:04
speed dating which is basically these
2:27:04
speed dating which is basically these short timed interactions between people
2:27:06
short timed interactions between people
2:27:06
short timed interactions between people usually it's only
2:27:07
usually it's only
2:27:07
usually it's only three minutes sometimes longer and it's
2:27:09
three minutes sometimes longer and it's
2:27:09
three minutes sometimes longer and it's based on the logic that
2:27:10
based on the logic that
2:27:10
based on the logic that it takes people such a short time to
2:27:12
it takes people such a short time to
2:27:12
it takes people such a short time to make up their mind about someone
2:27:13
make up their mind about someone
2:27:13
make up their mind about someone there was another study that was done in
2:27:15
there was another study that was done in
2:27:15
there was another study that was done in ohio ohio state university
2:27:17
ohio ohio state university
2:27:17
ohio ohio state university where scientists found that people can
2:27:19
where scientists found that people can
2:27:19
where scientists found that people can usually tell in the first one to two
2:27:20
usually tell in the first one to two
2:27:20
usually tell in the first one to two minutes
2:27:21
minutes
2:27:21
minutes whether they're interested in a
2:27:22
whether they're interested in a
2:27:22
whether they're interested in a relationship with another person
2:27:24
relationship with another person
2:27:24
relationship with another person and they continued that study to see
2:27:26
and they continued that study to see
2:27:26
and they continued that study to see that after nine weeks
2:27:27
that after nine weeks
2:27:27
that after nine weeks that initial impression made in the
2:27:29
that initial impression made in the
2:27:29
that initial impression made in the first one to two minutes hardly
2:27:31
first one to two minutes hardly
2:27:31
first one to two minutes hardly ever changed so you realize how how
2:27:34
ever changed so you realize how how
2:27:34
ever changed so you realize how how strong that is
2:27:35
strong that is
2:27:35
strong that is the whole book blink by malcolm gladwell
2:27:37
the whole book blink by malcolm gladwell
2:27:37
the whole book blink by malcolm gladwell i'm sure you've
2:27:38
i'm sure you've
2:27:38
i'm sure you've come across this as well this focuses on
2:27:41
come across this as well this focuses on
2:27:41
come across this as well this focuses on the power of thinking without thinking
2:27:42
the power of thinking without thinking
2:27:42
the power of thinking without thinking it's all about intuition and instinct
2:27:44
it's all about intuition and instinct
2:27:44
it's all about intuition and instinct and his phraseology is thin slicing but
2:27:47
and his phraseology is thin slicing but
2:27:48
and his phraseology is thin slicing but it's basically the same concept of how
2:27:49
it's basically the same concept of how
2:27:49
it's basically the same concept of how we use
2:27:50
we use
2:27:50
we use limited information for a very small
2:27:52
limited information for a very small
2:27:52
limited information for a very small period of time and draw a conclusion
2:27:54
period of time and draw a conclusion
2:27:54
period of time and draw a conclusion and these that can actually be quite
2:27:56
and these that can actually be quite
2:27:56
and these that can actually be quite accurate now again apply this to
2:27:58
accurate now again apply this to
2:27:58
accurate now again apply this to human interactions overall and what is
2:28:00
human interactions overall and what is
2:28:00
human interactions overall and what is so important to us
2:28:02
so important to us
2:28:02
so important to us because we're so relational and because
2:28:04
because we're so relational and because
2:28:04
because we're so relational and because our relationships are so complex
2:28:05
our relationships are so complex
2:28:06
our relationships are so complex and they're driven by so many things
2:28:07
and they're driven by so many things
2:28:07
and they're driven by so many things like non-verbal and intangible qualities
2:28:10
like non-verbal and intangible qualities
2:28:10
like non-verbal and intangible qualities a big underlying driver of this is trust
2:28:13
a big underlying driver of this is trust
2:28:13
a big underlying driver of this is trust this is why
2:28:14
this is why
2:28:14
this is why people stay in long-term committed
2:28:15
people stay in long-term committed
2:28:15
people stay in long-term committed relationships this is why people
2:28:17
relationships this is why people
2:28:17
relationships this is why people work with the same people for a long
2:28:19
work with the same people for a long
2:28:19
work with the same people for a long period of time and pull known quantities
2:28:21
period of time and pull known quantities
2:28:21
period of time and pull known quantities onto a new team when they're starting a
2:28:22
onto a new team when they're starting a
2:28:22
onto a new team when they're starting a new venture
2:28:23
new venture
2:28:23
new venture it's this idea can you trust another
2:28:25
it's this idea can you trust another
2:28:25
it's this idea can you trust another person is that person considered
2:28:27
person is that person considered
2:28:27
person is that person considered trustworthy
2:28:28
trustworthy
2:28:28
trustworthy and trust is repeatedly cited time and
2:28:31
and trust is repeatedly cited time and
2:28:31
and trust is repeatedly cited time and time again as
2:28:32
time again as
2:28:32
time again as the number one quality that's valued in
2:28:33
the number one quality that's valued in
2:28:34
the number one quality that's valued in a relationship that's whether it's
2:28:35
a relationship that's whether it's
2:28:35
a relationship that's whether it's romantic social
2:28:36
romantic social
2:28:36
romantic social business relationship it's also
2:28:40
business relationship it's also
2:28:40
business relationship it's also credibility has been cited as one of the
2:28:42
credibility has been cited as one of the
2:28:42
credibility has been cited as one of the top qualities
2:28:43
top qualities
2:28:43
top qualities in a leader that makes make a leader a
2:28:45
in a leader that makes make a leader a
2:28:45
in a leader that makes make a leader a desirable leader
2:28:46
desirable leader
2:28:46
desirable leader we put so much value on trust
2:28:49
we put so much value on trust
2:28:49
we put so much value on trust i like the this uh little visual here
2:28:52
i like the this uh little visual here
2:28:52
i like the this uh little visual here too
2:28:53
too
2:28:53
too one of the things that brings us all
2:28:55
one of the things that brings us all
2:28:55
one of the things that brings us all back to cookies that i had in the
2:28:56
back to cookies that i had in the
2:28:56
back to cookies that i had in the beginning i
2:28:57
beginning i
2:28:57
beginning i must be hungry when i made this
2:28:58
must be hungry when i made this
2:28:58
must be hungry when i made this presentation raising cookies that look
2:29:00
presentation raising cookies that look
2:29:00
presentation raising cookies that look like chocolate chip cookies are the main
2:29:01
like chocolate chip cookies are the main
2:29:01
like chocolate chip cookies are the main reason i have trust issues
2:29:03
reason i have trust issues
2:29:03
reason i have trust issues the fact that trust is so detrimental to
2:29:06
the fact that trust is so detrimental to
2:29:06
the fact that trust is so detrimental to the
2:29:06
the
2:29:06
the longevity of our relationships the depth
2:29:08
longevity of our relationships the depth
2:29:08
longevity of our relationships the depth of relationships it's such a deep
2:29:10
of relationships it's such a deep
2:29:10
of relationships it's such a deep subject that we will
2:29:12
subject that we will
2:29:12
subject that we will often make light of it by adding some
2:29:14
often make light of it by adding some
2:29:14
often make light of it by adding some levity to the gravity of this quality of
2:29:16
levity to the gravity of this quality of
2:29:16
levity to the gravity of this quality of trust
2:29:16
trust
2:29:16
trust by means and things like this
2:29:20
by means and things like this
2:29:20
by means and things like this the concept of breaking trust is also
2:29:22
the concept of breaking trust is also
2:29:22
the concept of breaking trust is also something that's been around
2:29:24
something that's been around
2:29:24
something that's been around since basically forever uh since the
2:29:26
since basically forever uh since the
2:29:26
since basically forever uh since the dawn of time it's been referenced in
2:29:28
dawn of time it's been referenced in
2:29:28
dawn of time it's been referenced in art and literature the concept of
2:29:30
art and literature the concept of
2:29:30
art and literature the concept of breaking bread has roots in
2:29:32
breaking bread has roots in
2:29:32
breaking bread has roots in biblical references where it basically
2:29:34
biblical references where it basically
2:29:34
biblical references where it basically became synonymous with
2:29:35
became synonymous with
2:29:36
became synonymous with the mean a meaningful connection made
2:29:37
the mean a meaningful connection made
2:29:38
the mean a meaningful connection made over a meal it was the meaningful
2:29:39
over a meal it was the meaningful
2:29:39
over a meal it was the meaningful connection with
2:29:40
connection with
2:29:40
connection with another person it basically meant the
2:29:42
another person it basically meant the
2:29:42
another person it basically meant the sense of familial
2:29:44
sense of familial
2:29:44
sense of familial familiarity with someone or a group of
2:29:46
familiarity with someone or a group of
2:29:46
familiarity with someone or a group of someone's and it allowed you to build
2:29:48
someone's and it allowed you to build
2:29:48
someone's and it allowed you to build rapport
2:29:48
rapport
2:29:48
rapport establish trust and that was crucial to
2:29:51
establish trust and that was crucial to
2:29:51
establish trust and that was crucial to really founding relationships
2:29:54
really founding relationships
2:29:54
really founding relationships so if we now take all this and apply
2:29:56
so if we now take all this and apply
2:29:56
so if we now take all this and apply this to technology
2:29:57
this to technology
2:29:57
this to technology what does this mean this was so far all
2:30:00
what does this mean this was so far all
2:30:00
what does this mean this was so far all been about
2:30:01
been about
2:30:01
been about human to human interaction so when we
2:30:04
human to human interaction so when we
2:30:04
human to human interaction so when we start talking about human to
2:30:06
start talking about human to
2:30:06
start talking about human to technology interaction we tend as people
2:30:08
technology interaction we tend as people
2:30:08
technology interaction we tend as people to want to apply
2:30:10
to want to apply
2:30:10
to want to apply humanizing capabilities to our
2:30:12
humanizing capabilities to our
2:30:12
humanizing capabilities to our technology we tend to humanize it in a
2:30:13
technology we tend to humanize it in a
2:30:13
technology we tend to humanize it in a couple different ways
2:30:15
couple different ways
2:30:15
couple different ways before i get into this let me first
2:30:17
before i get into this let me first
2:30:17
before i get into this let me first explain a little bit about who i am and
2:30:19
explain a little bit about who i am and
2:30:19
explain a little bit about who i am and what i do
2:30:19
what i do
2:30:20
what i do so i mentioned before i'm director of
2:30:21
so i mentioned before i'm director of
2:30:21
so i mentioned before i'm director of innovation for the cisco collaboration
2:30:23
innovation for the cisco collaboration
2:30:23
innovation for the cisco collaboration group
2:30:24
group
2:30:24
group our charter for our team is focused on
2:30:26
our charter for our team is focused on
2:30:26
our charter for our team is focused on disruptive innovation
2:30:27
disruptive innovation
2:30:28
disruptive innovation we are tasked with building a bridge
2:30:30
we are tasked with building a bridge
2:30:30
we are tasked with building a bridge between the state of enterprise
2:30:32
between the state of enterprise
2:30:32
between the state of enterprise collaboration today
2:30:33
collaboration today
2:30:33
collaboration today and the experiences that we think will
2:30:35
and the experiences that we think will
2:30:35
and the experiences that we think will be possible as technology advances
2:30:38
be possible as technology advances
2:30:38
be possible as technology advances in the future so we explore discover
2:30:40
in the future so we explore discover
2:30:40
in the future so we explore discover experiment with
2:30:41
experiment with
2:30:41
experiment with play with tinker with at times invent
2:30:44
play with tinker with at times invent
2:30:44
play with tinker with at times invent new technologies that ultimately shape
2:30:46
new technologies that ultimately shape
2:30:46
new technologies that ultimately shape the future of work
2:30:48
the future of work
2:30:48
the future of work i work in technology i'm in a future
2:30:50
i work in technology i'm in a future
2:30:50
i work in technology i'm in a future gazing branch of technology i work for
2:30:52
gazing branch of technology i work for
2:30:52
gazing branch of technology i work for the office of the cto
2:30:53
the office of the cto
2:30:54
the office of the cto i'm inside of an engineering technology
2:30:55
i'm inside of an engineering technology
2:30:55
i'm inside of an engineering technology group at a high-tech company
2:30:56
group at a high-tech company
2:30:56
group at a high-tech company headquartered in the heart of silicon
2:30:58
headquartered in the heart of silicon
2:30:58
headquartered in the heart of silicon valley
2:30:59
valley
2:30:59
valley it would be very easy to call myself a
2:31:01
it would be very easy to call myself a
2:31:01
it would be very easy to call myself a technologist
2:31:02
technologist
2:31:02
technologist however when people ask me what i do i
2:31:04
however when people ask me what i do i
2:31:04
however when people ask me what i do i try to be very clear that i'm not in the
2:31:06
try to be very clear that i'm not in the
2:31:06
try to be very clear that i'm not in the technology business
2:31:07
technology business
2:31:07
technology business i am in the people business if i'm truly
2:31:09
i am in the people business if i'm truly
2:31:09
i am in the people business if i'm truly doing my job and my team is truly doing
2:31:11
doing my job and my team is truly doing
2:31:11
doing my job and my team is truly doing our job
2:31:12
our job
2:31:12
our job the solutions that we build should fade
2:31:15
the solutions that we build should fade
2:31:15
the solutions that we build should fade into the background
2:31:16
into the background
2:31:16
into the background and what comes out is a more powerful
2:31:19
and what comes out is a more powerful
2:31:19
and what comes out is a more powerful way for humans to connect
2:31:20
way for humans to connect
2:31:20
way for humans to connect and transcend these traditional barriers
2:31:22
and transcend these traditional barriers
2:31:22
and transcend these traditional barriers of time and space that's what we're
2:31:24
of time and space that's what we're
2:31:24
of time and space that's what we're trying to do
2:31:25
trying to do
2:31:25
trying to do so even though i work for a technology
2:31:27
so even though i work for a technology
2:31:27
so even though i work for a technology company i am
2:31:28
company i am
2:31:28
company i am in the people business so when we talk
2:31:31
in the people business so when we talk
2:31:31
in the people business so when we talk about humanizing technology
2:31:33
about humanizing technology
2:31:33
about humanizing technology what do i mean by that i look at
2:31:35
what do i mean by that i look at
2:31:35
what do i mean by that i look at humanizing technology as
2:31:37
humanizing technology as
2:31:37
humanizing technology as two different facets firstly we try to
2:31:40
two different facets firstly we try to
2:31:40
two different facets firstly we try to use
2:31:41
use
2:31:41
use technology as a way to unlock
2:31:43
technology as a way to unlock
2:31:43
technology as a way to unlock relationships
2:31:44
relationships
2:31:44
relationships it's important to understand that when
2:31:46
it's important to understand that when
2:31:46
it's important to understand that when you introduce technology to a human
2:31:48
you introduce technology to a human
2:31:48
you introduce technology to a human human interaction it doesn't replace the
2:31:50
human interaction it doesn't replace the
2:31:50
human interaction it doesn't replace the human
2:31:50
human
2:31:50
human interaction it doesn't downplay the
2:31:53
interaction it doesn't downplay the
2:31:53
interaction it doesn't downplay the importance of it either
2:31:54
importance of it either
2:31:54
importance of it either if anything it should facilitate it
2:31:57
if anything it should facilitate it
2:31:57
if anything it should facilitate it basically
2:31:57
basically
2:31:57
basically we humanize this technology in in it by
2:32:01
we humanize this technology in in it by
2:32:01
we humanize this technology in in it by way of unlocking this connection that
2:32:03
way of unlocking this connection that
2:32:03
way of unlocking this connection that otherwise wouldn't be possible thinking
2:32:04
otherwise wouldn't be possible thinking
2:32:04
otherwise wouldn't be possible thinking of it as
2:32:05
of it as
2:32:05
of it as using technology as a bridge to get to
2:32:07
using technology as a bridge to get to
2:32:07
using technology as a bridge to get to another person
2:32:08
another person
2:32:08
another person a great example of this is the history
2:32:11
a great example of this is the history
2:32:11
a great example of this is the history of the very first telephone call
2:32:13
of the very first telephone call
2:32:13
of the very first telephone call in 1876 alexander graham bell conducted
2:32:16
in 1876 alexander graham bell conducted
2:32:16
in 1876 alexander graham bell conducted the first telephone call
2:32:17
the first telephone call
2:32:17
the first telephone call ever recorded his first words were mr
2:32:20
ever recorded his first words were mr
2:32:20
ever recorded his first words were mr watson
2:32:21
watson
2:32:21
watson come here i want to see you it's very
2:32:23
come here i want to see you it's very
2:32:23
come here i want to see you it's very interesting to me that the first usage
2:32:25
interesting to me that the first usage
2:32:25
interesting to me that the first usage of what ended up being a revolutionary
2:32:27
of what ended up being a revolutionary
2:32:27
of what ended up being a revolutionary technology ended up basically being a
2:32:29
technology ended up basically being a
2:32:29
technology ended up basically being a way to speed up and facilitate
2:32:31
way to speed up and facilitate
2:32:31
way to speed up and facilitate face-to-face interaction
2:32:33
face-to-face interaction
2:32:33
face-to-face interaction so even the early switchboards were an
2:32:35
so even the early switchboards were an
2:32:36
so even the early switchboards were an interesting reminder of the
2:32:37
interesting reminder of the
2:32:37
interesting reminder of the the connectivity by brought about by
2:32:40
the connectivity by brought about by
2:32:40
the connectivity by brought about by because they would basically literally
2:32:41
because they would basically literally
2:32:41
because they would basically literally plug
2:32:42
plug
2:32:42
plug cables into appropriate ports to connect
2:32:44
cables into appropriate ports to connect
2:32:44
cables into appropriate ports to connect to people on
2:32:45
to people on
2:32:45
to people on opposite ends of a call the second part
2:32:48
opposite ends of a call the second part
2:32:48
opposite ends of a call the second part of what i mean by humanizing technology
2:32:50
of what i mean by humanizing technology
2:32:50
of what i mean by humanizing technology is ascribing human tendencies to the
2:32:53
is ascribing human tendencies to the
2:32:53
is ascribing human tendencies to the technology itself
2:32:55
technology itself
2:32:55
technology itself so beyond just using technology to
2:32:56
so beyond just using technology to
2:32:56
so beyond just using technology to facilitate access to humans
2:32:59
facilitate access to humans
2:32:59
facilitate access to humans we are in some ways attempt to make our
2:33:01
we are in some ways attempt to make our
2:33:01
we are in some ways attempt to make our technology
2:33:02
technology
2:33:02
technology feel more human in a way that we want to
2:33:05
feel more human in a way that we want to
2:33:05
feel more human in a way that we want to interact with it so what i
2:33:07
interact with it so what i
2:33:07
interact with it so what i mean by this is a couple things so one
2:33:09
mean by this is a couple things so one
2:33:09
mean by this is a couple things so one being the whole concept of
2:33:10
being the whole concept of
2:33:10
being the whole concept of personification
2:33:11
personification
2:33:12
personification personification attribute of personal
2:33:15
personification attribute of personal
2:33:15
personification attribute of personal nature of human characteristics applying
2:33:17
nature of human characteristics applying
2:33:17
nature of human characteristics applying it to something that's non-human
2:33:19
it to something that's non-human
2:33:19
it to something that's non-human this is why kids paint smiley faces on
2:33:21
this is why kids paint smiley faces on
2:33:22
this is why kids paint smiley faces on rocks
2:33:22
rocks
2:33:22
rocks this is why the logo on the side of all
2:33:24
this is why the logo on the side of all
2:33:24
this is why the logo on the side of all your amazon boxes looks like a smile
2:33:26
your amazon boxes looks like a smile
2:33:26
your amazon boxes looks like a smile this is why we put these cute robot
2:33:29
this is why we put these cute robot
2:33:29
this is why we put these cute robot faces on
2:33:30
faces on
2:33:30
faces on robots that work in warehouses and
2:33:31
robots that work in warehouses and
2:33:31
robots that work in warehouses and retail environments to give them these
2:33:33
retail environments to give them these
2:33:33
retail environments to give them these googly eyes and cute names we are
2:33:35
googly eyes and cute names we are
2:33:35
googly eyes and cute names we are personifying our technology
2:33:38
personifying our technology
2:33:38
personifying our technology the other piece of this is a little bit
2:33:40
the other piece of this is a little bit
2:33:40
the other piece of this is a little bit more complicated
2:33:41
more complicated
2:33:41
more complicated and it goes back to the concept of trust
2:33:44
and it goes back to the concept of trust
2:33:44
and it goes back to the concept of trust so the idea of breaking bread that we
2:33:45
so the idea of breaking bread that we
2:33:46
so the idea of breaking bread that we brought up earlier why credibility is
2:33:48
brought up earlier why credibility is
2:33:48
brought up earlier why credibility is such a desired trait in our leadership
2:33:50
such a desired trait in our leadership
2:33:50
such a desired trait in our leadership the importance of trust is so strong
2:33:52
the importance of trust is so strong
2:33:52
the importance of trust is so strong strongly ingrained us as humans that we
2:33:55
strongly ingrained us as humans that we
2:33:55
strongly ingrained us as humans that we apply this to our technology as well
2:33:58
apply this to our technology as well
2:33:58
apply this to our technology as well we use the same verbiage and description
2:34:01
we use the same verbiage and description
2:34:01
we use the same verbiage and description when we describe technology i had
2:34:03
when we describe technology i had
2:34:03
when we describe technology i had a co-worker recently mentioned that
2:34:06
a co-worker recently mentioned that
2:34:06
a co-worker recently mentioned that she was in a conversation with a
2:34:07
she was in a conversation with a
2:34:08
she was in a conversation with a potential customer and
2:34:10
potential customer and
2:34:10
potential customer and they be mainly because of the current
2:34:11
they be mainly because of the current
2:34:11
they be mainly because of the current situation in the world they were using
2:34:13
situation in the world they were using
2:34:13
situation in the world they were using technology that they had never used
2:34:15
technology that they had never used
2:34:15
technology that they had never used before and when they were describing
2:34:17
before and when they were describing
2:34:17
before and when they were describing what they wanted or what they needed
2:34:18
what they wanted or what they needed
2:34:18
what they wanted or what they needed their main concern was around security
2:34:20
their main concern was around security
2:34:20
their main concern was around security but what the customer wanted in their
2:34:23
but what the customer wanted in their
2:34:23
but what the customer wanted in their technology was end-to-end encryption
2:34:25
technology was end-to-end encryption
2:34:25
technology was end-to-end encryption in that particular solution but the
2:34:26
in that particular solution but the
2:34:26
in that particular solution but the customer didn't say i want antenna
2:34:28
customer didn't say i want antenna
2:34:28
customer didn't say i want antenna encryption what the customer said was i
2:34:30
encryption what the customer said was i
2:34:30
encryption what the customer said was i want a tool i can trust
2:34:32
want a tool i can trust
2:34:32
want a tool i can trust because we take those terms that we are
2:34:34
because we take those terms that we are
2:34:34
because we take those terms that we are familiar with and those concepts that we
2:34:35
familiar with and those concepts that we
2:34:35
familiar with and those concepts that we are familiar with and we apply it to our
2:34:37
are familiar with and we apply it to our
2:34:37
are familiar with and we apply it to our technology we
2:34:38
technology we
2:34:38
technology we apply this human vernacular to our
2:34:40
apply this human vernacular to our
2:34:40
apply this human vernacular to our technology exchanges
2:34:42
technology exchanges
2:34:42
technology exchanges so how do we put this all together so
2:34:45
so how do we put this all together so
2:34:45
so how do we put this all together so insights from humans in the past can
2:34:48
insights from humans in the past can
2:34:48
insights from humans in the past can help us
2:34:49
help us
2:34:49
help us predict how humans will adapt into
2:34:50
predict how humans will adapt into
2:34:50
predict how humans will adapt into technology in the future
2:34:52
technology in the future
2:34:52
technology in the future how do we really realistically expect
2:34:54
how do we really realistically expect
2:34:54
how do we really realistically expect this to change over the next five
2:34:56
this to change over the next five
2:34:56
this to change over the next five ten years this is where i'm going to get
2:34:58
ten years this is where i'm going to get
2:34:58
ten years this is where i'm going to get into a few of my predictions
2:35:00
into a few of my predictions
2:35:00
into a few of my predictions take this with a grain of salt so
2:35:02
take this with a grain of salt so
2:35:02
take this with a grain of salt so firstly i think
2:35:03
firstly i think
2:35:03
firstly i think we are changing where we work this is
2:35:06
we are changing where we work this is
2:35:06
we are changing where we work this is not so much future state this is current
2:35:08
not so much future state this is current
2:35:08
not so much future state this is current state this is
2:35:09
state this is
2:35:09
state this is the world we live in right now for
2:35:10
the world we live in right now for
2:35:10
the world we live in right now for better or worse the current situation
2:35:13
better or worse the current situation
2:35:13
better or worse the current situation we have collectively as a global society
2:35:16
we have collectively as a global society
2:35:16
we have collectively as a global society experienced
2:35:17
experienced
2:35:17
experienced remote working on a grand scale and
2:35:19
remote working on a grand scale and
2:35:20
remote working on a grand scale and found that in some areas
2:35:21
found that in some areas
2:35:21
found that in some areas it doesn't work at all in other word
2:35:24
it doesn't work at all in other word
2:35:24
it doesn't work at all in other word areas it works great
2:35:26
areas it works great
2:35:26
areas it works great and it really depends on how and
2:35:29
and it really depends on how and
2:35:29
and it really depends on how and it's applied and the area that it's
2:35:31
it's applied and the area that it's
2:35:31
it's applied and the area that it's applied in but i take
2:35:33
applied in but i take
2:35:33
applied in but i take my my group personally i've worked
2:35:36
my my group personally i've worked
2:35:36
my my group personally i've worked remotely for
2:35:38
remotely for
2:35:38
remotely for last few years this is something that
2:35:40
last few years this is something that
2:35:40
last few years this is something that many folks on my team work remotely
2:35:41
many folks on my team work remotely
2:35:41
many folks on my team work remotely we're a global team
2:35:42
we're a global team
2:35:42
we're a global team across i think nine different time zones
2:35:45
across i think nine different time zones
2:35:46
across i think nine different time zones nine hours of time zones four different
2:35:48
nine hours of time zones four different
2:35:48
nine hours of time zones four different countries
2:35:50
countries
2:35:50
countries uh it's a very spread team so we're very
2:35:52
uh it's a very spread team so we're very
2:35:52
uh it's a very spread team so we're very used to working remotely
2:35:53
used to working remotely
2:35:54
used to working remotely and there are certain things that we
2:35:55
and there are certain things that we
2:35:55
and there are certain things that we know to do have dedicated home office
2:35:56
know to do have dedicated home office
2:35:56
know to do have dedicated home office space try to keep certain
2:35:58
space try to keep certain
2:35:58
space try to keep certain certain hours that are work hours versus
2:35:59
certain hours that are work hours versus
2:35:59
certain hours that are work hours versus home hours simple things like get out of
2:36:01
home hours simple things like get out of
2:36:01
home hours simple things like get out of your sweatpants get on video it makes it
2:36:03
your sweatpants get on video it makes it
2:36:03
your sweatpants get on video it makes it easier to do
2:36:04
easier to do
2:36:04
easier to do remote collaboration but the biggest
2:36:06
remote collaboration but the biggest
2:36:06
remote collaboration but the biggest lament of remote workers is usually this
2:36:09
lament of remote workers is usually this
2:36:09
lament of remote workers is usually this this feeling of lack of connection i
2:36:12
this feeling of lack of connection i
2:36:12
this feeling of lack of connection i don't feel connected to my co-workers
2:36:14
don't feel connected to my co-workers
2:36:14
don't feel connected to my co-workers the way i used to
2:36:15
the way i used to
2:36:15
the way i used to and it's interesting to see how people
2:36:17
and it's interesting to see how people
2:36:17
and it's interesting to see how people are changing their meetings and their
2:36:19
are changing their meetings and their
2:36:19
are changing their meetings and their approach now again as we're doing this
2:36:21
approach now again as we're doing this
2:36:21
approach now again as we're doing this on a grand scale
2:36:22
on a grand scale
2:36:22
on a grand scale we and my team we've had started doing
2:36:25
we and my team we've had started doing
2:36:25
we and my team we've had started doing these um
2:36:26
these um
2:36:26
these um regular no agenda water cooler type
2:36:29
regular no agenda water cooler type
2:36:29
regular no agenda water cooler type calls
2:36:29
calls
2:36:29
calls this usually involves wine coffee
2:36:31
this usually involves wine coffee
2:36:31
this usually involves wine coffee sometimes both depending on what time
2:36:33
sometimes both depending on what time
2:36:33
sometimes both depending on what time zone people are in
2:36:34
zone people are in
2:36:34
zone people are in and we're on video 100 of the time
2:36:37
and we're on video 100 of the time
2:36:37
and we're on video 100 of the time we don't do audio calls we do crazy
2:36:40
we don't do audio calls we do crazy
2:36:40
we don't do audio calls we do crazy themes like hawaiian shirt day and crazy
2:36:42
themes like hawaiian shirt day and crazy
2:36:42
themes like hawaiian shirt day and crazy hat day and favorite sports team day and
2:36:44
hat day and favorite sports team day and
2:36:44
hat day and favorite sports team day and a way to kind of get everyone doing the
2:36:46
a way to kind of get everyone doing the
2:36:46
a way to kind of get everyone doing the same thing while remote
2:36:48
same thing while remote
2:36:48
same thing while remote and that has really helped bring some
2:36:50
and that has really helped bring some
2:36:50
and that has really helped bring some camaraderie to it but we're realizing
2:36:52
camaraderie to it but we're realizing
2:36:52
camaraderie to it but we're realizing this is just different this changing
2:36:54
this is just different this changing
2:36:54
this is just different this changing where we work people talk about the new
2:36:56
where we work people talk about the new
2:36:56
where we work people talk about the new normal
2:36:56
normal
2:36:56
normal all the time regardless of how
2:36:59
all the time regardless of how
2:36:59
all the time regardless of how work looks at your workplace a year from
2:37:02
work looks at your workplace a year from
2:37:02
work looks at your workplace a year from now
2:37:03
now
2:37:03
now it will be different from what it is
2:37:05
it will be different from what it is
2:37:05
it will be different from what it is today and it will be different from what
2:37:06
today and it will be different from what
2:37:06
today and it will be different from what it was a year ago so we are absolutely
2:37:08
it was a year ago so we are absolutely
2:37:08
it was a year ago so we are absolutely changing
2:37:09
changing
2:37:09
changing where we work one facet of this is
2:37:12
where we work one facet of this is
2:37:12
where we work one facet of this is security i i
2:37:13
security i i
2:37:14
security i i mentioned the customer before that said
2:37:15
mentioned the customer before that said
2:37:15
mentioned the customer before that said trust is so important this concept of
2:37:17
trust is so important this concept of
2:37:17
trust is so important this concept of end-to-end and security
2:37:19
end-to-end and security
2:37:19
end-to-end and security as you're starting to see people working
2:37:22
as you're starting to see people working
2:37:22
as you're starting to see people working from
2:37:22
from
2:37:22
from all over these different places from
2:37:24
all over these different places from
2:37:24
all over these different places from different connections you're not going
2:37:26
different connections you're not going
2:37:26
different connections you're not going to see a lessening of security
2:37:27
to see a lessening of security
2:37:27
to see a lessening of security requirements you're going to see an
2:37:28
requirements you're going to see an
2:37:28
requirements you're going to see an increase in this and it really comes
2:37:31
increase in this and it really comes
2:37:31
increase in this and it really comes from our need to
2:37:32
from our need to
2:37:32
from our need to want to trust our machines but it gets
2:37:35
want to trust our machines but it gets
2:37:36
want to trust our machines but it gets very complicated when you start looking
2:37:37
very complicated when you start looking
2:37:37
very complicated when you start looking at people
2:37:37
at people
2:37:38
at people joining from different devices different
2:37:40
joining from different devices different
2:37:40
joining from different devices different locations different types of networks
2:37:41
locations different types of networks
2:37:41
locations different types of networks spotty wi-fi all all these types of
2:37:43
spotty wi-fi all all these types of
2:37:43
spotty wi-fi all all these types of things so security gets really really
2:37:45
things so security gets really really
2:37:45
things so security gets really really interesting as an area
2:37:46
interesting as an area
2:37:46
interesting as an area we probably do a whole segway on that
2:37:48
we probably do a whole segway on that
2:37:48
we probably do a whole segway on that but i'll leave that for now
2:37:50
but i'll leave that for now
2:37:50
but i'll leave that for now and also we're changing how we work uh
2:37:53
and also we're changing how we work uh
2:37:53
and also we're changing how we work uh we'll start looking at things beyond how
2:37:56
we'll start looking at things beyond how
2:37:56
we'll start looking at things beyond how do people connect in a more
2:37:57
do people connect in a more
2:37:57
do people connect in a more technologically advanced way but how do
2:38:00
technologically advanced way but how do
2:38:00
technologically advanced way but how do we use technology to make that
2:38:01
we use technology to make that
2:38:01
we use technology to make that interaction even better
2:38:02
interaction even better
2:38:02
interaction even better than what you could do purely in real
2:38:04
than what you could do purely in real
2:38:04
than what you could do purely in real life things in augmented virtual reality
2:38:06
life things in augmented virtual reality
2:38:06
life things in augmented virtual reality we've talked about some today
2:38:08
we've talked about some today
2:38:08
we've talked about some today have made major strides in the last few
2:38:09
have made major strides in the last few
2:38:10
have made major strides in the last few years but how do you go
2:38:11
years but how do you go
2:38:11
years but how do you go from a step further of just like hey on
2:38:13
from a step further of just like hey on
2:38:13
from a step further of just like hey on a video call and waving
2:38:15
a video call and waving
2:38:15
a video call and waving to actually feel as if you're being
2:38:16
to actually feel as if you're being
2:38:16
to actually feel as if you're being there together
2:38:18
there together
2:38:18
there together i have a short video in here that i
2:38:20
i have a short video in here that i
2:38:20
i have a short video in here that i don't know if it's going to work i'm
2:38:21
don't know if it's going to work i'm
2:38:21
don't know if it's going to work i'm going to try it if it doesn't work let
2:38:23
going to try it if it doesn't work let
2:38:23
going to try it if it doesn't work let me know i can talk through it instead
2:38:25
me know i can talk through it instead
2:38:25
me know i can talk through it instead but we'll we can certainly try we've got
2:38:27
but we'll we can certainly try we've got
2:38:27
but we'll we can certainly try we've got hank in the background and he's great
2:38:29
hank in the background and he's great
2:38:29
hank in the background and he's great with this uh sort of thing so
2:38:31
with this uh sort of thing so
2:38:31
with this uh sort of thing so just give it a shot and see what happens
2:38:35
oh there's no sound i don't think we
2:38:40
oh there's no sound i don't think we
2:38:40
oh there's no sound i don't think we all right um i did send the video if you
2:38:42
all right um i did send the video if you
2:38:42
all right um i did send the video if you guys have it on your end
2:38:44
guys have it on your end
2:38:44
guys have it on your end maybe you can play it after but yeah and
2:38:47
maybe you can play it after but yeah and
2:38:47
maybe you can play it after but yeah and in the interest of time i'll basically
2:38:49
in the interest of time i'll basically
2:38:49
in the interest of time i'll basically just move forward to say that
2:38:50
just move forward to say that
2:38:50
just move forward to say that what we could try actually is uh if you
2:38:53
what we could try actually is uh if you
2:38:53
what we could try actually is uh if you stop sharing your screen and share it
2:38:55
stop sharing your screen and share it
2:38:55
stop sharing your screen and share it again but enable sound there should be a
2:38:57
again but enable sound there should be a
2:38:57
again but enable sound there should be a check
2:38:57
check
2:38:58
check check box saying i want to share sound
2:39:00
check box saying i want to share sound
2:39:00
check box saying i want to share sound that that might actually fix the problem
2:39:01
that that might actually fix the problem
2:39:01
that that might actually fix the problem that we have
2:39:02
that we have
2:39:02
that we have let's give it a whirl see what happens
2:39:06
let's give it a whirl see what happens
2:39:06
let's give it a whirl see what happens it's computer technology so basically
2:39:07
it's computer technology so basically
2:39:07
it's computer technology so basically this is turning it off and on again
2:39:09
this is turning it off and on again
2:39:09
this is turning it off and on again but different it's kind of like kicking
2:39:11
but different it's kind of like kicking
2:39:11
but different it's kind of like kicking it right
2:39:13
it right
2:39:13
it right yeah sort of it's been down
2:39:20
all right it's not giving me an option
2:39:24
all right it's not giving me an option
2:39:24
all right it's not giving me an option oh it's not working okay let's do let's
2:39:26
oh it's not working okay let's do let's
2:39:26
oh it's not working okay let's do let's move on we'll make sure
2:39:27
move on we'll make sure
2:39:27
move on we'll make sure we share the video afterwards that uh
2:39:28
we share the video afterwards that uh
2:39:28
we share the video afterwards that uh that works too all right don't worry um
2:39:30
that works too all right don't worry um
2:39:30
that works too all right don't worry um are you guys seeing my slideshare again
2:39:32
are you guys seeing my slideshare again
2:39:32
are you guys seeing my slideshare again yes yeah that's working perfect so i'll
2:39:34
yes yeah that's working perfect so i'll
2:39:34
yes yeah that's working perfect so i'll give that the very short
2:39:35
give that the very short
2:39:35
give that the very short it's a short video that is just a teaser
2:39:37
it's a short video that is just a teaser
2:39:37
it's a short video that is just a teaser of a product that my team is building
2:39:39
of a product that my team is building
2:39:39
of a product that my team is building right now
2:39:40
right now
2:39:40
right now i should clarify prototype that my team
2:39:42
i should clarify prototype that my team
2:39:42
i should clarify prototype that my team is building right now not product we
2:39:43
is building right now not product we
2:39:43
is building right now not product we don't build products
2:39:45
don't build products
2:39:45
don't build products but it is in the ar vr space and it's
2:39:48
but it is in the ar vr space and it's
2:39:48
but it is in the ar vr space and it's really focused on that concept of how do
2:39:50
really focused on that concept of how do
2:39:50
really focused on that concept of how do you be there
2:39:51
you be there
2:39:51
you be there with another colleague this is an early
2:39:53
with another colleague this is an early
2:39:53
with another colleague this is an early concept drawing gives you
2:39:54
concept drawing gives you
2:39:54
concept drawing gives you basically an idea of what it is but i
2:39:56
basically an idea of what it is but i
2:39:56
basically an idea of what it is but i won't go into too much more depth
2:39:58
won't go into too much more depth
2:39:58
won't go into too much more depth so moving beyond how we work together
2:40:01
so moving beyond how we work together
2:40:01
so moving beyond how we work together we're changing with whom we work and
2:40:04
we're changing with whom we work and
2:40:04
we're changing with whom we work and this i think is an interesting concept
2:40:07
this i think is an interesting concept
2:40:07
this i think is an interesting concept too it's going to
2:40:11
too it's going to
2:40:11
too it's going to the whole definition of what my team
2:40:13
the whole definition of what my team
2:40:13
the whole definition of what my team does we are a collaboration group
2:40:15
does we are a collaboration group
2:40:15
does we are a collaboration group historically collaboration has meant
2:40:16
historically collaboration has meant
2:40:16
historically collaboration has meant the act of working with someone to
2:40:19
the act of working with someone to
2:40:19
the act of working with someone to produce or create something
2:40:20
produce or create something
2:40:20
produce or create something and that concept is really evolving over
2:40:23
and that concept is really evolving over
2:40:23
and that concept is really evolving over time
2:40:24
time
2:40:24
time we're no longer thinking of how do i
2:40:26
we're no longer thinking of how do i
2:40:26
we're no longer thinking of how do i work with someone
2:40:28
work with someone
2:40:28
work with someone but how do i work with something a
2:40:30
but how do i work with something a
2:40:30
but how do i work with something a machine a
2:40:31
machine a
2:40:31
machine a ba a device and that's a very different
2:40:34
ba a device and that's a very different
2:40:34
ba a device and that's a very different way to look at collaboration we're not
2:40:35
way to look at collaboration we're not
2:40:36
way to look at collaboration we're not looking
2:40:36
looking
2:40:36
looking no longer looking at as a purely human
2:40:38
no longer looking at as a purely human
2:40:38
no longer looking at as a purely human to human collaboration but
2:40:40
to human collaboration but
2:40:40
to human collaboration but human to machine human to ai human to
2:40:42
human to machine human to ai human to
2:40:42
human to machine human to ai human to technology interaction
2:40:44
technology interaction
2:40:44
technology interaction and how do you fit that all together in
2:40:46
and how do you fit that all together in
2:40:46
and how do you fit that all together in a way that doesn't feel clunky
2:40:48
a way that doesn't feel clunky
2:40:48
a way that doesn't feel clunky and that's a big thing that we need to
2:40:50
and that's a big thing that we need to
2:40:50
and that's a big thing that we need to solve
2:40:51
solve
2:40:51
solve but it's very interesting and very
2:40:52
but it's very interesting and very
2:40:52
but it's very interesting and very exciting as well
2:40:54
exciting as well
2:40:54
exciting as well with that some of my predictions of what
2:40:58
with that some of my predictions of what
2:40:58
with that some of my predictions of what i think
2:41:00
i think
2:41:00
i think evolution of ai communications will look
2:41:02
evolution of ai communications will look
2:41:02
evolution of ai communications will look like
2:41:03
like
2:41:03
like we still may be quite a ways off from
2:41:06
we still may be quite a ways off from
2:41:06
we still may be quite a ways off from having artificial intelligence with this
2:41:08
having artificial intelligence with this
2:41:08
having artificial intelligence with this high
2:41:09
high
2:41:09
high emotional intelligence strong perception
2:41:11
emotional intelligence strong perception
2:41:11
emotional intelligence strong perception however i think
2:41:13
however i think
2:41:13
however i think as we progressed already from very
2:41:15
as we progressed already from very
2:41:15
as we progressed already from very simplistic models things like chat bots
2:41:17
simplistic models things like chat bots
2:41:17
simplistic models things like chat bots to more sophisticated models that can
2:41:19
to more sophisticated models that can
2:41:19
to more sophisticated models that can encompass natural language processing
2:41:21
encompass natural language processing
2:41:21
encompass natural language processing and conversational
2:41:23
and conversational
2:41:23
and conversational my prediction is that we'll begin to
2:41:24
my prediction is that we'll begin to
2:41:24
my prediction is that we'll begin to layer on contextual awareness to that
2:41:26
layer on contextual awareness to that
2:41:26
layer on contextual awareness to that and eventually we will get to some
2:41:28
and eventually we will get to some
2:41:28
and eventually we will get to some element of social awareness
2:41:30
element of social awareness
2:41:30
element of social awareness and some concept of social intelligence
2:41:33
and some concept of social intelligence
2:41:33
and some concept of social intelligence we'll see this evolution beyond this
2:41:35
we'll see this evolution beyond this
2:41:35
we'll see this evolution beyond this conversational ai to layering in all
2:41:37
conversational ai to layering in all
2:41:37
conversational ai to layering in all this contextual awareness and
2:41:39
this contextual awareness and
2:41:39
this contextual awareness and basic social cues and that does get us a
2:41:41
basic social cues and that does get us a
2:41:41
basic social cues and that does get us a step further to ultimately getting
2:41:43
step further to ultimately getting
2:41:43
step further to ultimately getting achieving this element of social
2:41:45
achieving this element of social
2:41:45
achieving this element of social intelligence i also think we're going to
2:41:46
intelligence i also think we're going to
2:41:46
intelligence i also think we're going to see a greater depth of social dynamics
2:41:48
see a greater depth of social dynamics
2:41:48
see a greater depth of social dynamics in our software
2:41:49
in our software
2:41:50
in our software so not just these binary conversations
2:41:52
so not just these binary conversations
2:41:52
so not just these binary conversations but the ability to drift in and out of
2:41:54
but the ability to drift in and out of
2:41:54
but the ability to drift in and out of conversations online the same way as you
2:41:56
conversations online the same way as you
2:41:56
conversations online the same way as you could in person
2:41:57
could in person
2:41:57
could in person as we build out this complexity of the
2:41:59
as we build out this complexity of the
2:41:59
as we build out this complexity of the virtual interactions this will start to
2:42:01
virtual interactions this will start to
2:42:01
virtual interactions this will start to become possible
2:42:02
become possible
2:42:02
become possible and i'm hoping we could talk about the
2:42:03
and i'm hoping we could talk about the
2:42:03
and i'm hoping we could talk about the sum in the chat after because this is a
2:42:04
sum in the chat after because this is a
2:42:04
sum in the chat after because this is a really interesting
2:42:05
really interesting
2:42:05
really interesting area i think so as far as the future of
2:42:08
area i think so as far as the future of
2:42:08
area i think so as far as the future of collaboration
2:42:09
collaboration
2:42:09
collaboration i'm going to wrap up here it's a very
2:42:10
i'm going to wrap up here it's a very
2:42:10
i'm going to wrap up here it's a very exciting field to be working in
2:42:13
exciting field to be working in
2:42:13
exciting field to be working in there's so much opportunity on the
2:42:14
there's so much opportunity on the
2:42:14
there's so much opportunity on the horizon within my team we're asking
2:42:16
horizon within my team we're asking
2:42:16
horizon within my team we're asking questions like
2:42:17
questions like
2:42:18
questions like what if i could have real-life
2:42:19
what if i could have real-life
2:42:19
what if i could have real-life interactions that are better than what's
2:42:20
interactions that are better than what's
2:42:20
interactions that are better than what's currently possible in real life
2:42:23
currently possible in real life
2:42:23
currently possible in real life what if conferencing doesn't have to be
2:42:24
what if conferencing doesn't have to be
2:42:24
what if conferencing doesn't have to be limited to this two-dimensional plane
2:42:26
limited to this two-dimensional plane
2:42:26
limited to this two-dimensional plane what happens when real and virtual are
2:42:28
what happens when real and virtual are
2:42:28
what happens when real and virtual are no longer distinct ideas and they start
2:42:30
no longer distinct ideas and they start
2:42:30
no longer distinct ideas and they start blending together
2:42:31
blending together
2:42:31
blending together what happens when artificial
2:42:33
what happens when artificial
2:42:33
what happens when artificial intelligence begins to have this concept
2:42:35
intelligence begins to have this concept
2:42:35
intelligence begins to have this concept of emotional and social intelligence
2:42:37
of emotional and social intelligence
2:42:37
of emotional and social intelligence what are those communications and
2:42:38
what are those communications and
2:42:38
what are those communications and interactions like and that is really
2:42:40
interactions like and that is really
2:42:40
interactions like and that is really really really exciting to me
2:42:42
really really exciting to me
2:42:42
really really exciting to me so just to wrap up my thoughts here
2:42:44
so just to wrap up my thoughts here
2:42:44
so just to wrap up my thoughts here humans are unique
2:42:45
humans are unique
2:42:45
humans are unique we value meaning trust experience
2:42:48
we value meaning trust experience
2:42:48
we value meaning trust experience technology is built for humans
2:42:49
technology is built for humans
2:42:49
technology is built for humans and as much as we can remember that keep
2:42:51
and as much as we can remember that keep
2:42:51
and as much as we can remember that keep it to the forefront of everything we
2:42:53
it to the forefront of everything we
2:42:53
it to the forefront of everything we build
2:42:54
build
2:42:54
build we use technology to expedite or enhance
2:42:57
we use technology to expedite or enhance
2:42:57
we use technology to expedite or enhance these human connections
2:42:58
these human connections
2:42:58
these human connections it's an extension of our humanity and
2:43:01
it's an extension of our humanity and
2:43:01
it's an extension of our humanity and technology will be
2:43:02
technology will be
2:43:02
technology will be successfully adopted i think when it
2:43:04
successfully adopted i think when it
2:43:04
successfully adopted i think when it adheres to those values when it
2:43:06
adheres to those values when it
2:43:06
adheres to those values when it expedites our connection when it
2:43:07
expedites our connection when it
2:43:07
expedites our connection when it enhances our relationship
2:43:09
enhances our relationship
2:43:09
enhances our relationship that's when you ultimately see adoption
2:43:11
that's when you ultimately see adoption
2:43:11
that's when you ultimately see adoption so i hope you've enjoyed this talk i've
2:43:13
so i hope you've enjoyed this talk i've
2:43:13
so i hope you've enjoyed this talk i've definitely enjoyed sharing it
2:43:14
definitely enjoyed sharing it
2:43:14
definitely enjoyed sharing it feel free to reach out to me and that's
2:43:16
feel free to reach out to me and that's
2:43:16
feel free to reach out to me and that's all i have for today
2:43:18
all i have for today
2:43:18
all i have for today cool well thank you very much it's an
2:43:20
cool well thank you very much it's an
2:43:20
cool well thank you very much it's an interesting talk seeing
2:43:22
interesting talk seeing
2:43:22
interesting talk seeing uh this perspective in a field of ai
2:43:25
uh this perspective in a field of ai
2:43:25
uh this perspective in a field of ai where we
2:43:26
where we
2:43:26
where we spend an awful lot of time talking about
2:43:28
spend an awful lot of time talking about
2:43:28
spend an awful lot of time talking about neural networks and
2:43:30
neural networks and
2:43:30
neural networks and amazing technological advances in that
2:43:32
amazing technological advances in that
2:43:32
amazing technological advances in that field
2:43:33
field
2:43:33
field it's good to remember that ai is
2:43:35
it's good to remember that ai is
2:43:35
it's good to remember that ai is actually here to enhance humans rather
2:43:37
actually here to enhance humans rather
2:43:37
actually here to enhance humans rather than
2:43:38
than
2:43:38
than replace them i find that personally
2:43:41
replace them i find that personally
2:43:41
replace them i find that personally really
2:43:41
really
2:43:41
really important but it's good to see that more
2:43:43
important but it's good to see that more
2:43:43
important but it's good to see that more and more companies like cisco
2:43:45
and more companies like cisco
2:43:45
and more companies like cisco are working on this actually and
2:43:48
are working on this actually and
2:43:48
are working on this actually and we actually have got a question from um
2:43:53
we actually have got a question from um
2:43:53
we actually have got a question from um somebody in the audience uh in regards
2:43:55
somebody in the audience uh in regards
2:43:55
somebody in the audience uh in regards to that so let me find that real quick
2:44:02
um
2:44:08
i will be calling into that customer
2:44:10
i will be calling into that customer
2:44:10
i will be calling into that customer service line every day
2:44:12
service line every day
2:44:12
service line every day all right with me about my day
2:44:18
all right with me about my day
2:44:18
all right with me about my day about my finances or whatever is going
2:44:21
about my finances or whatever is going
2:44:21
about my finances or whatever is going on or my lack of tree is
2:44:25
yeah exactly so both is asking
2:44:29
yeah exactly so both is asking
2:44:30
yeah exactly so both is asking when using technology we currently use a
2:44:31
when using technology we currently use a
2:44:32
when using technology we currently use a lot of screens as you mentioned
2:44:33
lot of screens as you mentioned
2:44:33
lot of screens as you mentioned when talking about video conferencing
2:44:35
when talking about video conferencing
2:44:35
when talking about video conferencing we're using our screen to see each other
2:44:37
we're using our screen to see each other
2:44:37
we're using our screen to see each other we're using a webcam to connect
2:44:39
we're using a webcam to connect
2:44:39
we're using a webcam to connect um do you see
2:44:42
um do you see
2:44:42
um do you see a large role for screens and cameras in
2:44:45
a large role for screens and cameras in
2:44:45
a large role for screens and cameras in the future or do you think that we're
2:44:47
the future or do you think that we're
2:44:47
the future or do you think that we're moving
2:44:47
moving
2:44:47
moving back away from that um when we talk
2:44:50
back away from that um when we talk
2:44:50
back away from that um when we talk about
2:44:51
about
2:44:51
about collaborative software
2:44:54
collaborative software
2:44:54
collaborative software my personal opinion on this is there'll
2:44:56
my personal opinion on this is there'll
2:44:56
my personal opinion on this is there'll still be a large role for cameras and
2:44:57
still be a large role for cameras and
2:44:57
still be a large role for cameras and screens but
2:44:58
screens but
2:44:58
screens but the same way we're redefining what
2:45:00
the same way we're redefining what
2:45:00
the same way we're redefining what collaboration is we're going to redefine
2:45:03
collaboration is we're going to redefine
2:45:03
collaboration is we're going to redefine what a screen is does the screen mean a
2:45:05
what a screen is does the screen mean a
2:45:05
what a screen is does the screen mean a piece of glass
2:45:06
piece of glass
2:45:06
piece of glass it may it may just mean a piece of glass
2:45:08
it may it may just mean a piece of glass
2:45:08
it may it may just mean a piece of glass that's changed size
2:45:10
that's changed size
2:45:10
that's changed size and change location and has more
2:45:11
and change location and has more
2:45:11
and change location and has more flexibility or a screen may be
2:45:14
flexibility or a screen may be
2:45:14
flexibility or a screen may be i put on my ar goggles and there's my
2:45:16
i put on my ar goggles and there's my
2:45:16
i put on my ar goggles and there's my screen it may be projection it may be
2:45:18
screen it may be projection it may be
2:45:18
screen it may be projection it may be something projected onto my retina
2:45:20
something projected onto my retina
2:45:20
something projected onto my retina there are a lot of options of how you
2:45:22
there are a lot of options of how you
2:45:22
there are a lot of options of how you can describe a screen
2:45:24
can describe a screen
2:45:24
can describe a screen cameras are more interesting because the
2:45:27
cameras are more interesting because the
2:45:27
cameras are more interesting because the whole
2:45:28
whole
2:45:28
whole ability to capture we haven't figured
2:45:31
ability to capture we haven't figured
2:45:31
ability to capture we haven't figured out any way to do that without cameras
2:45:32
out any way to do that without cameras
2:45:32
out any way to do that without cameras and a lot of cameras
2:45:33
and a lot of cameras
2:45:33
and a lot of cameras and how do you make that not be
2:45:35
and how do you make that not be
2:45:35
and how do you make that not be obtrusive and that's actually
2:45:37
obtrusive and that's actually
2:45:37
obtrusive and that's actually the little teaser thing i was trying to
2:45:38
the little teaser thing i was trying to
2:45:38
the little teaser thing i was trying to show there that's something that my
2:45:39
show there that's something that my
2:45:40
show there that's something that my team's been working on is how do you
2:45:41
team's been working on is how do you
2:45:41
team's been working on is how do you enable these two-way interactions
2:45:43
enable these two-way interactions
2:45:43
enable these two-way interactions that are live and in real time and not
2:45:46
that are live and in real time and not
2:45:46
that are live and in real time and not make it feel like you have a wall of
2:45:47
make it feel like you have a wall of
2:45:47
make it feel like you have a wall of cameras in front of you and that's
2:45:49
cameras in front of you and that's
2:45:49
cameras in front of you and that's something we think we've solved
2:45:50
something we think we've solved
2:45:50
something we think we've solved but that's it's that's very challenging
2:45:53
but that's it's that's very challenging
2:45:53
but that's it's that's very challenging so i do
2:45:53
so i do
2:45:53
so i do i don't think they're going away but i
2:45:55
i don't think they're going away but i
2:45:55
i don't think they're going away but i think we're going to redefine
2:45:56
think we're going to redefine
2:45:56
think we're going to redefine what they mean yeah exactly
2:46:00
what they mean yeah exactly
2:46:00
what they mean yeah exactly and and in adding to that
2:46:03
and and in adding to that
2:46:03
and and in adding to that i saw another question from someone in
2:46:04
i saw another question from someone in
2:46:04
i saw another question from someone in the audience mentioning that
2:46:07
the audience mentioning that
2:46:07
the audience mentioning that now that we can use ai in computer
2:46:09
now that we can use ai in computer
2:46:09
now that we can use ai in computer vision and ai
2:46:11
vision and ai
2:46:11
vision and ai invoice generation actually
2:46:14
invoice generation actually
2:46:14
invoice generation actually we can actually enhance what we see we
2:46:17
we can actually enhance what we see we
2:46:17
we can actually enhance what we see we can
2:46:18
can
2:46:18
can we've seen all the negative samples with
2:46:20
we've seen all the negative samples with
2:46:20
we've seen all the negative samples with deep fakes with
2:46:22
deep fakes with
2:46:22
deep fakes with uh the president saying something else
2:46:25
uh the president saying something else
2:46:26
uh the president saying something else uh what do you think
2:46:29
uh what do you think
2:46:29
uh what do you think are we going to see more stuff in the
2:46:31
are we going to see more stuff in the
2:46:31
are we going to see more stuff in the future that actually uses
2:46:32
future that actually uses
2:46:32
future that actually uses a i uh to enhance the interaction
2:46:37
a i uh to enhance the interaction
2:46:37
a i uh to enhance the interaction that we have as humans most likely we're
2:46:40
that we have as humans most likely we're
2:46:40
that we have as humans most likely we're already seeing that in our meetings
2:46:41
already seeing that in our meetings
2:46:41
already seeing that in our meetings today we have things i'm going to
2:46:43
today we have things i'm going to
2:46:43
today we have things i'm going to do on product plug we have webex assist
2:46:46
do on product plug we have webex assist
2:46:46
do on product plug we have webex assist which is something that we've added to
2:46:47
which is something that we've added to
2:46:47
which is something that we've added to our webex meetings it's basically an ai
2:46:49
our webex meetings it's basically an ai
2:46:49
our webex meetings it's basically an ai running in the back of your meetings
2:46:50
running in the back of your meetings
2:46:50
running in the back of your meetings that's helping you take notes
2:46:54
follow-up meeting that kind of thing so
2:46:56
follow-up meeting that kind of thing so
2:46:56
follow-up meeting that kind of thing so i think you see a lot of that in the
2:46:57
i think you see a lot of that in the
2:46:57
i think you see a lot of that in the background
2:46:58
background
2:46:58
background you'll see and you're starting to see
2:47:01
you'll see and you're starting to see
2:47:01
you'll see and you're starting to see that i think in all of our tools you're
2:47:02
that i think in all of our tools you're
2:47:02
that i think in all of our tools you're also seeing
2:47:03
also seeing
2:47:03
also seeing you know snapchat filters something like
2:47:05
you know snapchat filters something like
2:47:05
you know snapchat filters something like that how do i enhance the way i look how
2:47:06
that how do i enhance the way i look how
2:47:06
that how do i enhance the way i look how the visual there of
2:47:07
the visual there of
2:47:08
the visual there of and that's because more and more people
2:47:09
and that's because more and more people
2:47:09
and that's because more and more people are on video that has become very
2:47:11
are on video that has become very
2:47:11
are on video that has become very important
2:47:12
important
2:47:12
important as well i i think that's going to be
2:47:15
as well i i think that's going to be
2:47:15
as well i i think that's going to be very
2:47:15
very
2:47:15
very different for the different use cases
2:47:18
different for the different use cases
2:47:18
different for the different use cases but
2:47:18
but
2:47:18
but we we use augmented reality as a buzz
2:47:20
we we use augmented reality as a buzz
2:47:20
we we use augmented reality as a buzz word all the time here but if you really
2:47:22
word all the time here but if you really
2:47:22
word all the time here but if you really think about what does augmented reality
2:47:23
think about what does augmented reality
2:47:23
think about what does augmented reality mean that it's reality that you have
2:47:25
mean that it's reality that you have
2:47:25
mean that it's reality that you have augmented you have improved it in some
2:47:27
augmented you have improved it in some
2:47:27
augmented you have improved it in some way by overlaying something on top of it
2:47:29
way by overlaying something on top of it
2:47:29
way by overlaying something on top of it and i think we are absolutely seeing
2:47:31
and i think we are absolutely seeing
2:47:31
and i think we are absolutely seeing that
2:47:31
that
2:47:32
that today in a lot of ways and i think the
2:47:34
today in a lot of ways and i think the
2:47:34
today in a lot of ways and i think the future will
2:47:35
future will
2:47:35
future will we'll definitely see more of that
2:47:38
we'll definitely see more of that
2:47:38
we'll definitely see more of that yeah sort of the advanced um snapchat
2:47:41
yeah sort of the advanced um snapchat
2:47:41
yeah sort of the advanced um snapchat filters where you can
2:47:42
filters where you can
2:47:42
filters where you can say well i'd like to make my face a
2:47:45
say well i'd like to make my face a
2:47:45
say well i'd like to make my face a little bit more sun-tanned
2:47:46
little bit more sun-tanned
2:47:46
little bit more sun-tanned um just to give the illusion that i had
2:47:49
um just to give the illusion that i had
2:47:49
um just to give the illusion that i had a great vacation
2:47:50
a great vacation
2:47:50
a great vacation uh those sort of things i think will
2:47:52
uh those sort of things i think will
2:47:52
uh those sort of things i think will well we'll be seeing soon in the future
2:47:54
well we'll be seeing soon in the future
2:47:54
well we'll be seeing soon in the future um given that we've seen so many
2:47:57
um given that we've seen so many
2:47:57
um given that we've seen so many snapchat filters with mustaches and
2:47:59
snapchat filters with mustaches and
2:47:59
snapchat filters with mustaches and dogs ears and those kind of things those
2:48:01
dogs ears and those kind of things those
2:48:01
dogs ears and those kind of things those are funny but
2:48:03
are funny but
2:48:03
are funny but i've i personally feel that we'll soon
2:48:05
i've i personally feel that we'll soon
2:48:06
i've i personally feel that we'll soon see a generation of
2:48:08
see a generation of
2:48:08
see a generation of video filters that are capable of
2:48:10
video filters that are capable of
2:48:10
video filters that are capable of capable of doing much more than that
2:48:12
capable of doing much more than that
2:48:12
capable of doing much more than that actually yeah i think beyond just things
2:48:14
actually yeah i think beyond just things
2:48:14
actually yeah i think beyond just things that are aesthetically
2:48:15
that are aesthetically
2:48:16
that are aesthetically pleasing the things that will be useful
2:48:17
pleasing the things that will be useful
2:48:18
pleasing the things that will be useful in an interaction like it would be great
2:48:19
in an interaction like it would be great
2:48:19
in an interaction like it would be great to have well and we're talking we can go
2:48:22
to have well and we're talking we can go
2:48:22
to have well and we're talking we can go back to our conversation we had on
2:48:23
back to our conversation we had on
2:48:23
back to our conversation we had on monday and i have this little field i
2:48:25
monday and i have this little field i
2:48:25
monday and i have this little field i can see in
2:48:25
can see in
2:48:25
can see in my view of like oh yeah we chatted about
2:48:27
my view of like oh yeah we chatted about
2:48:27
my view of like oh yeah we chatted about this you mentioned this about your kids
2:48:28
this you mentioned this about your kids
2:48:28
this you mentioned this about your kids and like
2:48:29
and like
2:48:29
and like a reminder of our last conversation it's
2:48:30
a reminder of our last conversation it's
2:48:30
a reminder of our last conversation it's enhancing our current interaction by
2:48:32
enhancing our current interaction by
2:48:32
enhancing our current interaction by being
2:48:33
being
2:48:33
being replacing my poor memory with the ai
2:48:36
replacing my poor memory with the ai
2:48:36
replacing my poor memory with the ai memory that remembers much better than i
2:48:38
memory that remembers much better than i
2:48:38
memory that remembers much better than i do
2:48:38
do
2:48:38
do that type of interaction you see how
2:48:40
that type of interaction you see how
2:48:40
that type of interaction you see how that can be phenomenally useful in
2:48:41
that can be phenomenally useful in
2:48:41
that can be phenomenally useful in meetings and enterprises in medical
2:48:43
meetings and enterprises in medical
2:48:43
meetings and enterprises in medical settings in
2:48:44
settings in
2:48:44
settings in a host of different industries and
2:48:47
a host of different industries and
2:48:47
a host of different industries and verticals
2:48:48
verticals
2:48:48
verticals yeah that sounds really interesting
2:48:50
yeah that sounds really interesting
2:48:50
yeah that sounds really interesting there's loads more questions coming in
2:48:52
there's loads more questions coming in
2:48:52
there's loads more questions coming in um so boaz has another question which is
2:48:55
um so boaz has another question which is
2:48:55
um so boaz has another question which is also interesting so we've talked about
2:48:57
also interesting so we've talked about
2:48:57
also interesting so we've talked about computer vision in this case
2:48:59
computer vision in this case
2:48:59
computer vision in this case how we use ai to enhance basically our
2:49:02
how we use ai to enhance basically our
2:49:02
how we use ai to enhance basically our meetings
2:49:03
meetings
2:49:03
meetings um how about haptic sensors
2:49:07
um how about haptic sensors
2:49:07
um how about haptic sensors maybe if we combine that with computer
2:49:09
maybe if we combine that with computer
2:49:09
maybe if we combine that with computer vision that that should
2:49:10
vision that that should
2:49:10
vision that that should yeah so we've played around some with
2:49:12
yeah so we've played around some with
2:49:12
yeah so we've played around some with haptics personally and
2:49:15
haptics personally and
2:49:15
haptics personally and the technology hasn't advanced quite as
2:49:17
the technology hasn't advanced quite as
2:49:17
the technology hasn't advanced quite as quickly in the haptic space
2:49:19
quickly in the haptic space
2:49:19
quickly in the haptic space as i would like to see and
2:49:22
as i would like to see and
2:49:22
as i would like to see and i think a lot less over the last year
2:49:24
i think a lot less over the last year
2:49:24
i think a lot less over the last year because
2:49:26
because
2:49:26
because copenhagen didn't do haptics any favor
2:49:28
copenhagen didn't do haptics any favor
2:49:28
copenhagen didn't do haptics any favor it's something that if the technology
2:49:30
it's something that if the technology
2:49:30
it's something that if the technology was already there this would be a great
2:49:31
was already there this would be a great
2:49:31
was already there this would be a great time to use it but
2:49:32
time to use it but
2:49:32
time to use it but when you're at a point in the technology
2:49:34
when you're at a point in the technology
2:49:34
when you're at a point in the technology where you need more people hands-on
2:49:36
where you need more people hands-on
2:49:36
where you need more people hands-on testing it
2:49:36
testing it
2:49:36
testing it it's really hard to do that in the
2:49:37
it's really hard to do that in the
2:49:38
it's really hard to do that in the current environment so i feel like
2:49:39
current environment so i feel like
2:49:39
current environment so i feel like haptics has almost stalled a bit in that
2:49:40
haptics has almost stalled a bit in that
2:49:40
haptics has almost stalled a bit in that regard
2:49:42
regard
2:49:42
regard but that's absolutely something i think
2:49:44
but that's absolutely something i think
2:49:44
but that's absolutely something i think we're going to see in the future
2:49:46
we're going to see in the future
2:49:46
we're going to see in the future is adding in the others the other senses
2:49:49
is adding in the others the other senses
2:49:49
is adding in the others the other senses right now so much is focused on audio
2:49:51
right now so much is focused on audio
2:49:51
right now so much is focused on audio and video that's
2:49:53
and video that's
2:49:53
and video that's how do we make those better how do we
2:49:54
how do we make those better how do we
2:49:54
how do we make those better how do we augment those how do we you know
2:49:56
augment those how do we you know
2:49:56
augment those how do we you know improve those but we have other senses
2:50:00
improve those but we have other senses
2:50:00
improve those but we have other senses we have multiple senses how do we layer
2:50:02
we have multiple senses how do we layer
2:50:02
we have multiple senses how do we layer in all those other things and there are
2:50:04
in all those other things and there are
2:50:04
in all those other things and there are some really great haptics companies out
2:50:05
some really great haptics companies out
2:50:06
some really great haptics companies out there that really focus on think of
2:50:08
there that really focus on think of
2:50:08
there that really focus on think of yourself in a physical world without the
2:50:09
yourself in a physical world without the
2:50:09
yourself in a physical world without the ability to touch
2:50:11
ability to touch
2:50:11
ability to touch how would you function like if it's so
2:50:13
how would you function like if it's so
2:50:13
how would you function like if it's so important in your physical why isn't
2:50:14
important in your physical why isn't
2:50:14
important in your physical why isn't that just as important in your virtual
2:50:16
that just as important in your virtual
2:50:16
that just as important in your virtual interactions
2:50:17
interactions
2:50:17
interactions and that's a pretty logical argument so
2:50:19
and that's a pretty logical argument so
2:50:19
and that's a pretty logical argument so i think adding haptics in the future
2:50:21
i think adding haptics in the future
2:50:21
i think adding haptics in the future will definitely be something with steam
2:50:23
will definitely be something with steam
2:50:23
will definitely be something with steam but i think it's going to be a lot
2:50:24
but i think it's going to be a lot
2:50:24
but i think it's going to be a lot longer on the time horizon just because
2:50:26
longer on the time horizon just because
2:50:26
longer on the time horizon just because of where we're at with the technology
2:50:29
of where we're at with the technology
2:50:29
of where we're at with the technology yeah so that yeah um
2:50:32
yeah so that yeah um
2:50:32
yeah so that yeah um so there's the answer boys it's a long
2:50:36
so there's the answer boys it's a long
2:50:36
so there's the answer boys it's a long way from home but i mean
2:50:38
way from home but i mean
2:50:38
way from home but i mean who knows what could happen next year
2:50:40
who knows what could happen next year
2:50:40
who knows what could happen next year once we get out of the uh pandemic and
2:50:42
once we get out of the uh pandemic and
2:50:42
once we get out of the uh pandemic and and people start thinking of new ways of
2:50:44
and people start thinking of new ways of
2:50:44
and people start thinking of new ways of interacting again
2:50:46
interacting again
2:50:46
interacting again or i mean we're in the pandemic i've
2:50:49
or i mean we're in the pandemic i've
2:50:49
or i mean we're in the pandemic i've seen some developments around spatial
2:50:51
seen some developments around spatial
2:50:51
seen some developments around spatial analysis for example which is also a
2:50:53
analysis for example which is also a
2:50:53
analysis for example which is also a computer vision application
2:50:54
computer vision application
2:50:54
computer vision application where uh the pandemic basically forced
2:50:57
where uh the pandemic basically forced
2:50:57
where uh the pandemic basically forced us to think about new ways of applying
2:51:00
us to think about new ways of applying
2:51:00
us to think about new ways of applying ai in this field so that's that's pretty
2:51:01
ai in this field so that's that's pretty
2:51:01
ai in this field so that's that's pretty interesting
2:51:04
interesting
2:51:04
interesting so talking about the social aspects of
2:51:06
so talking about the social aspects of
2:51:06
so talking about the social aspects of of
2:51:07
of
2:51:07
of those micro interactions um
2:51:10
those micro interactions um
2:51:10
those micro interactions um how does that work out currently in
2:51:11
how does that work out currently in
2:51:11
how does that work out currently in computer vision in your opinion
2:51:14
computer vision in your opinion
2:51:14
computer vision in your opinion so i'm gonna have an overly biased
2:51:16
so i'm gonna have an overly biased
2:51:16
so i'm gonna have an overly biased opinion here because of the technology
2:51:18
opinion here because of the technology
2:51:18
opinion here because of the technology that we're building
2:51:20
that we're building
2:51:20
that we're building we're building something that is in
2:51:24
we're building something that is in
2:51:24
we're building something that is in focus on photorealistic holograms so
2:51:27
focus on photorealistic holograms so
2:51:27
focus on photorealistic holograms so i believe that that is the future
2:51:31
i believe that that is the future
2:51:31
i believe that that is the future that is why i don't think avatar
2:51:33
that is why i don't think avatar
2:51:33
that is why i don't think avatar adoption will be
2:51:35
adoption will be
2:51:35
adoption will be huge overall in the long term because
2:51:37
huge overall in the long term because
2:51:37
huge overall in the long term because you lose those things
2:51:39
you lose those things
2:51:39
you lose those things having this sort of blanket yes i'm
2:51:41
having this sort of blanket yes i'm
2:51:41
having this sort of blanket yes i'm seeing my cartoonish version of me raise
2:51:43
seeing my cartoonish version of me raise
2:51:43
seeing my cartoonish version of me raise his eyebrows like
2:51:44
his eyebrows like
2:51:44
his eyebrows like that helps to get gross expression
2:51:47
that helps to get gross expression
2:51:48
that helps to get gross expression across but so much of human interaction
2:51:50
across but so much of human interaction
2:51:50
across but so much of human interaction is around these
2:51:51
is around these
2:51:51
is around these very fine-tuned micro expressions and
2:51:54
very fine-tuned micro expressions and
2:51:54
very fine-tuned micro expressions and you simply can't do that
2:51:55
you simply can't do that
2:51:55
you simply can't do that if you don't have photorealistic quality
2:51:58
if you don't have photorealistic quality
2:51:58
if you don't have photorealistic quality in real time that you're seeing how
2:51:59
in real time that you're seeing how
2:51:59
in real time that you're seeing how another person is reacting so
2:52:01
another person is reacting so
2:52:01
another person is reacting so that's my personal opinion i guess it's
2:52:04
that's my personal opinion i guess it's
2:52:04
that's my personal opinion i guess it's a
2:52:04
a
2:52:04
a it's a huge challenge to get that
2:52:06
it's a huge challenge to get that
2:52:06
it's a huge challenge to get that working with uh
2:52:07
working with uh
2:52:07
working with uh neural networks i've got some experience
2:52:09
neural networks i've got some experience
2:52:09
neural networks i've got some experience myself and i know that it takes an awful
2:52:12
myself and i know that it takes an awful
2:52:12
myself and i know that it takes an awful long time to train a neural network to
2:52:13
long time to train a neural network to
2:52:13
long time to train a neural network to even recognize
2:52:15
even recognize
2:52:15
even recognize the eyebrows on a person let alone those
2:52:17
the eyebrows on a person let alone those
2:52:17
the eyebrows on a person let alone those micro interactions of slightly raising
2:52:19
micro interactions of slightly raising
2:52:19
micro interactions of slightly raising your eyebrow when you're not agreeing
2:52:21
your eyebrow when you're not agreeing
2:52:21
your eyebrow when you're not agreeing with something
2:52:22
with something
2:52:22
with something yeah what does the challenge look like
2:52:25
yeah what does the challenge look like
2:52:26
yeah what does the challenge look like for you
2:52:27
for you
2:52:27
for you five years
2:52:30
it looks like a lot of work yeah i'm
2:52:33
it looks like a lot of work yeah i'm
2:52:33
it looks like a lot of work yeah i'm just thinking of the project we've been
2:52:34
just thinking of the project we've been
2:52:34
just thinking of the project we've been working on we've been working on it for
2:52:36
working on we've been working on it for
2:52:36
working on we've been working on it for five years to get to a point where it's
2:52:37
five years to get to a point where it's
2:52:37
five years to get to a point where it's like all right we're getting there
2:52:40
like all right we're getting there
2:52:40
like all right we're getting there and it's not even ready yet you're
2:52:42
and it's not even ready yet you're
2:52:42
and it's not even ready yet you're getting there so there's
2:52:43
getting there so there's
2:52:44
getting there so there's still a ways to go what do you think did
2:52:46
still a ways to go what do you think did
2:52:46
still a ways to go what do you think did we get this sort of stuff in five years
2:52:48
we get this sort of stuff in five years
2:52:48
we get this sort of stuff in five years ten years
2:52:49
ten years
2:52:49
ten years what sort of the protection i think
2:52:51
what sort of the protection i think
2:52:51
what sort of the protection i think you'll start seeing some pretty awesome
2:52:52
you'll start seeing some pretty awesome
2:52:52
you'll start seeing some pretty awesome stuff in the next two years quite
2:52:53
stuff in the next two years quite
2:52:53
stuff in the next two years quite frankly
2:52:54
frankly
2:52:54
frankly um but if you want to see everyone goes
2:52:57
um but if you want to see everyone goes
2:52:57
um but if you want to see everyone goes to this
2:52:58
to this
2:52:58
to this i want minority report i want these kind
2:53:00
i want minority report i want these kind
2:53:00
i want minority report i want these kind of crazy interactions like
2:53:02
of crazy interactions like
2:53:02
of crazy interactions like we're not quite at that point we're
2:53:03
we're not quite at that point we're
2:53:03
we're not quite at that point we're going to have these huge grandiose
2:53:05
going to have these huge grandiose
2:53:05
going to have these huge grandiose scales and you
2:53:06
scales and you
2:53:06
scales and you don't need anything other than a little
2:53:08
don't need anything other than a little
2:53:08
don't need anything other than a little chip where our contact lens you put on
2:53:10
chip where our contact lens you put on
2:53:10
chip where our contact lens you put on your eye like
2:53:11
your eye like
2:53:11
your eye like that's still going to be a ways away
2:53:12
that's still going to be a ways away
2:53:12
that's still going to be a ways away it'd be interesting to see
2:53:14
it'd be interesting to see
2:53:14
it'd be interesting to see what advancements we see in uh eyewear
2:53:18
what advancements we see in uh eyewear
2:53:18
what advancements we see in uh eyewear right now i mean if you want to get to
2:53:19
right now i mean if you want to get to
2:53:19
right now i mean if you want to get to these mixed reality type scenarios it's
2:53:21
these mixed reality type scenarios it's
2:53:21
these mixed reality type scenarios it's basically
2:53:22
basically
2:53:22
basically hololens or magically or basically the
2:53:24
hololens or magically or basically the
2:53:24
hololens or magically or basically the two out there that you can do these
2:53:25
two out there that you can do these
2:53:26
two out there that you can do these really complex interactions
2:53:28
really complex interactions
2:53:28
really complex interactions if you but they're both it's a heavy
2:53:30
if you but they're both it's a heavy
2:53:30
if you but they're both it's a heavy thing it's um
2:53:31
thing it's um
2:53:31
thing it's um you know you're wearing something so
2:53:33
you know you're wearing something so
2:53:33
you know you're wearing something so until we get to something closer to what
2:53:35
until we get to something closer to what
2:53:35
until we get to something closer to what you're wearing a set of eyeglasses that
2:53:37
you're wearing a set of eyeglasses that
2:53:37
you're wearing a set of eyeglasses that feels
2:53:37
feels
2:53:37
feels very lightweight and it's not intrusive
2:53:40
very lightweight and it's not intrusive
2:53:40
very lightweight and it's not intrusive i think uh
2:53:41
i think uh
2:53:41
i think uh there's a lot of limitations of what
2:53:44
there's a lot of limitations of what
2:53:44
there's a lot of limitations of what people are going to be willing to put up
2:53:45
people are going to be willing to put up
2:53:45
people are going to be willing to put up with
2:53:46
with
2:53:46
with so i think you'll start seeing you can
2:53:47
so i think you'll start seeing you can
2:53:47
so i think you'll start seeing you can see stuff in the very short term i think
2:53:49
see stuff in the very short term i think
2:53:49
see stuff in the very short term i think in
2:53:50
in
2:53:50
in enterprise settings in certain verticals
2:53:51
enterprise settings in certain verticals
2:53:51
enterprise settings in certain verticals where hey i can put on a headset for a
2:53:53
where hey i can put on a headset for a
2:53:53
where hey i can put on a headset for a short period of time
2:53:54
short period of time
2:53:54
short period of time but if you want to see mass adoption
2:53:57
but if you want to see mass adoption
2:53:57
but if you want to see mass adoption it's going to depend a lot
2:53:58
it's going to depend a lot
2:53:58
it's going to depend a lot on the form factor yeah i mean
2:54:01
on the form factor yeah i mean
2:54:01
on the form factor yeah i mean that one's pretty large still stefano
2:54:04
that one's pretty large still stefano
2:54:04
that one's pretty large still stefano talked about this earlier
2:54:06
talked about this earlier
2:54:06
talked about this earlier in the episode where he mentioned that
2:54:08
in the episode where he mentioned that
2:54:08
in the episode where he mentioned that the battery life of a hololens is pretty
2:54:10
the battery life of a hololens is pretty
2:54:10
the battery life of a hololens is pretty short
2:54:10
short
2:54:10
short and its weight is is is huge compared to
2:54:14
and its weight is is is huge compared to
2:54:14
and its weight is is is huge compared to normal glasses that
2:54:15
normal glasses that
2:54:15
normal glasses that alicia is wearing currently but i'm
2:54:17
alicia is wearing currently but i'm
2:54:17
alicia is wearing currently but i'm wearing my classes normally i don't put
2:54:19
wearing my classes normally i don't put
2:54:19
wearing my classes normally i don't put them on on stream because that's
2:54:20
them on on stream because that's
2:54:20
them on on stream because that's horrible to look at i've got those blue
2:54:22
horrible to look at i've got those blue
2:54:22
horrible to look at i've got those blue filters in there that's
2:54:23
filters in there that's
2:54:23
filters in there that's uh strange to look at um
2:54:26
uh strange to look at um
2:54:26
uh strange to look at um so alicia would you actually wear such
2:54:30
so alicia would you actually wear such
2:54:30
so alicia would you actually wear such such a pair of glasses if it would have
2:54:31
such a pair of glasses if it would have
2:54:31
such a pair of glasses if it would have a small camera inside and
2:54:33
a small camera inside and
2:54:33
a small camera inside and and allows you to communicate to people
2:54:36
and allows you to communicate to people
2:54:36
and allows you to communicate to people so
2:54:37
so
2:54:37
so you know i've i've been thinking about
2:54:39
you know i've i've been thinking about
2:54:39
you know i've i've been thinking about it
2:54:40
it
2:54:40
it and they have the face masks which
2:54:44
and they have the face masks which
2:54:44
and they have the face masks which which are a little bigger right so if
2:54:47
which are a little bigger right so if
2:54:47
which are a little bigger right so if you could get one of those just a little
2:54:48
you could get one of those just a little
2:54:48
you could get one of those just a little bit heavier
2:54:50
bit heavier
2:54:50
bit heavier um because people use those
2:54:54
um because people use those
2:54:54
um because people use those all over the place and you can see
2:54:55
all over the place and you can see
2:54:55
all over the place and you can see expression through the screen
2:54:57
expression through the screen
2:54:57
expression through the screen and it's still pretty protective right
2:55:00
and it's still pretty protective right
2:55:00
and it's still pretty protective right and um
2:55:02
and um
2:55:02
and um people are okay with that so i i think
2:55:04
people are okay with that so i i think
2:55:04
people are okay with that so i i think if they could get it down to that size
2:55:06
if they could get it down to that size
2:55:06
if they could get it down to that size and it looks like we're we're seeing the
2:55:08
and it looks like we're we're seeing the
2:55:08
and it looks like we're we're seeing the hardware improvements
2:55:09
hardware improvements
2:55:09
hardware improvements right and hardware improvements really
2:55:12
right and hardware improvements really
2:55:12
right and hardware improvements really drive
2:55:13
drive
2:55:13
drive a lot of these efficiencies sometimes
2:55:15
a lot of these efficiencies sometimes
2:55:16
a lot of these efficiencies sometimes which enable us
2:55:17
which enable us
2:55:17
which enable us to do these calculations at a quicker
2:55:20
to do these calculations at a quicker
2:55:20
to do these calculations at a quicker pace
2:55:21
pace
2:55:21
pace so um it sounds like maybe there's an
2:55:24
so um it sounds like maybe there's an
2:55:24
so um it sounds like maybe there's an opportunity for collaboration
2:55:28
right
2:55:34
yeah i guess the biggest challenge that
2:55:36
yeah i guess the biggest challenge that
2:55:36
yeah i guess the biggest challenge that we have currently with uh computer
2:55:38
we have currently with uh computer
2:55:38
we have currently with uh computer vision applications is that
2:55:40
vision applications is that
2:55:40
vision applications is that the smaller we make this form factor to
2:55:42
the smaller we make this form factor to
2:55:42
the smaller we make this form factor to put the
2:55:43
put the
2:55:43
put the models in and the other software that we
2:55:46
models in and the other software that we
2:55:46
models in and the other software that we need
2:55:47
need
2:55:47
need the more power constraint we become and
2:55:49
the more power constraint we become and
2:55:49
the more power constraint we become and the more compute power
2:55:50
the more compute power
2:55:50
the more compute power constraint we become so it's more of a
2:55:52
constraint we become so it's more of a
2:55:52
constraint we become so it's more of a matter of balancing
2:55:53
matter of balancing
2:55:54
matter of balancing the affordance of the technology we're
2:55:55
the affordance of the technology we're
2:55:56
the affordance of the technology we're using versus the power that it offers us
2:55:58
using versus the power that it offers us
2:55:58
using versus the power that it offers us so if it's a heavier device we get more
2:56:00
so if it's a heavier device we get more
2:56:00
so if it's a heavier device we get more compute power but it's also harder to
2:56:02
compute power but it's also harder to
2:56:02
compute power but it's also harder to use so that
2:56:03
use so that
2:56:03
use so that so finding the sweet spot i personally
2:56:06
so finding the sweet spot i personally
2:56:06
so finding the sweet spot i personally i don't think we will ever wear a a lens
2:56:09
i don't think we will ever wear a a lens
2:56:09
i don't think we will ever wear a a lens in our eye that basically projects
2:56:11
in our eye that basically projects
2:56:11
in our eye that basically projects images on our eye whenever it wants to
2:56:14
images on our eye whenever it wants to
2:56:14
images on our eye whenever it wants to because sometimes you just gotta get
2:56:17
because sometimes you just gotta get
2:56:17
because sometimes you just gotta get away from computers
2:56:18
away from computers
2:56:18
away from computers and that's even the case for me as a as
2:56:20
and that's even the case for me as a as
2:56:20
and that's even the case for me as a as a total
2:56:21
a total
2:56:21
a total deep learning geek uh that's important
2:56:24
deep learning geek uh that's important
2:56:24
deep learning geek uh that's important to still do
2:56:26
to still do
2:56:26
to still do uh it can be very overbearing all this
2:56:28
uh it can be very overbearing all this
2:56:28
uh it can be very overbearing all this technology
2:56:30
technology
2:56:30
technology you don't look forward to being plugged
2:56:32
you don't look forward to being plugged
2:56:32
you don't look forward to being plugged into the matrix
2:56:35
no no
2:56:39
who did not change any of their habits
2:56:41
who did not change any of their habits
2:56:41
who did not change any of their habits once
2:56:42
once
2:56:42
once coveted and quarantined started so
2:56:46
coveted and quarantined started so
2:56:46
coveted and quarantined started so any work from home friends
2:56:50
any work from home friends
2:56:50
any work from home friends yeah i guess we're lucky as i.t people
2:56:52
yeah i guess we're lucky as i.t people
2:56:52
yeah i guess we're lucky as i.t people that we get
2:56:53
that we get
2:56:54
that we get we do already a lot of stuff from home
2:56:55
we do already a lot of stuff from home
2:56:55
we do already a lot of stuff from home and remote so we're
2:56:57
and remote so we're
2:56:57
and remote so we're we're sort of used to this extra
2:56:59
we're sort of used to this extra
2:56:59
we're sort of used to this extra cognitive load
2:57:00
cognitive load
2:57:00
cognitive load um yeah yeah i'm just grateful that you
2:57:03
um yeah yeah i'm just grateful that you
2:57:04
um yeah yeah i'm just grateful that you know humanity still reaches out
2:57:05
know humanity still reaches out
2:57:05
know humanity still reaches out for for interpersonal interaction you
2:57:08
for for interpersonal interaction you
2:57:08
for for interpersonal interaction you know like i see people in my
2:57:10
know like i see people in my
2:57:10
know like i see people in my in my community taking walks and you
2:57:13
in my community taking walks and you
2:57:13
in my community taking walks and you know having conversations with each
2:57:14
know having conversations with each
2:57:14
know having conversations with each other six feet away
2:57:16
other six feet away
2:57:16
other six feet away and i i think you know regardless of the
2:57:20
and i i think you know regardless of the
2:57:20
and i i think you know regardless of the the technology um i don't think we're
2:57:22
the technology um i don't think we're
2:57:22
the technology um i don't think we're gonna get to a point where
2:57:23
gonna get to a point where
2:57:23
gonna get to a point where we're stuck in our homes um plugged in
2:57:26
we're stuck in our homes um plugged in
2:57:26
we're stuck in our homes um plugged in all the time because i think we
2:57:27
all the time because i think we
2:57:28
all the time because i think we we're still always going to need um
2:57:31
we're still always going to need um
2:57:31
we're still always going to need um that physical touch right sorry i'm a
2:57:34
that physical touch right sorry i'm a
2:57:34
that physical touch right sorry i'm a hugger
2:57:34
hugger
2:57:34
hugger i need my hugs i need to interact with
2:57:37
i need my hugs i need to interact with
2:57:37
i need my hugs i need to interact with people
2:57:37
people
2:57:37
people and hit my hug quota
2:57:42
and hit my hug quota
2:57:42
and hit my hug quota yeah i guess that that's one of the
2:57:44
yeah i guess that that's one of the
2:57:44
yeah i guess that that's one of the challenges that we have in technology i
2:57:46
challenges that we have in technology i
2:57:46
challenges that we have in technology i guess
2:57:47
guess
2:57:47
guess so one of the worries that i have is
2:57:51
so one of the worries that i have is
2:57:51
so one of the worries that i have is when you use a computer it's it's a
2:57:53
when you use a computer it's it's a
2:57:53
when you use a computer it's it's a decidedly different
2:57:55
decidedly different
2:57:55
decidedly different um form of interaction so it it places a
2:57:58
um form of interaction so it it places a
2:57:58
um form of interaction so it it places a cognitive load on you um is there any
2:58:03
cognitive load on you um is there any
2:58:03
cognitive load on you um is there any work that you know of elizabeth that
2:58:05
work that you know of elizabeth that
2:58:05
work that you know of elizabeth that that people are working on to reduce
2:58:07
that people are working on to reduce
2:58:07
that people are working on to reduce cognitive load of using tools like
2:58:09
cognitive load of using tools like
2:58:09
cognitive load of using tools like cisco webex and maybe other conferencing
2:58:12
cisco webex and maybe other conferencing
2:58:12
cisco webex and maybe other conferencing tools
2:58:13
tools
2:58:13
tools yeah there's a lot of uh all the
2:58:15
yeah there's a lot of uh all the
2:58:15
yeah there's a lot of uh all the technology companies
2:58:16
technology companies
2:58:16
technology companies i have right now have mlna ice groups
2:58:19
i have right now have mlna ice groups
2:58:19
i have right now have mlna ice groups that are looking at can we
2:58:21
that are looking at can we
2:58:21
that are looking at can we make this less
2:58:24
make this less
2:58:24
make this less painful maybe not the best word
2:58:27
painful maybe not the best word
2:58:27
painful maybe not the best word yes people are using things long longer
2:58:29
yes people are using things long longer
2:58:30
yes people are using things long longer and longer
2:58:30
and longer
2:58:30
and longer there's definitely a lot of teams that
2:58:32
there's definitely a lot of teams that
2:58:32
there's definitely a lot of teams that are looking at that i don't know of any
2:58:34
are looking at that i don't know of any
2:58:34
are looking at that i don't know of any amazing breakthroughs that
2:58:36
amazing breakthroughs that
2:58:36
amazing breakthroughs that have happened in that in that space i do
2:58:38
have happened in that in that space i do
2:58:38
have happened in that in that space i do know there's lots and lots of
2:58:40
know there's lots and lots of
2:58:40
know there's lots and lots of work and exploration but i would say
2:58:42
work and exploration but i would say
2:58:42
work and exploration but i would say that's not nearly as far along as some
2:58:44
that's not nearly as far along as some
2:58:44
that's not nearly as far along as some of the stuff we've been talking about in
2:58:45
of the stuff we've been talking about in
2:58:45
of the stuff we've been talking about in computer vision
2:58:47
computer vision
2:58:47
computer vision yeah i'm just looking at the questions
2:58:50
yeah i'm just looking at the questions
2:58:50
yeah i'm just looking at the questions from the audience
2:58:51
from the audience
2:58:51
from the audience there's a bunch more stuff going in so
2:58:56
one of the other things that i'm
2:58:57
one of the other things that i'm
2:58:57
one of the other things that i'm wondering about how do we make sure that
2:59:00
wondering about how do we make sure that
2:59:00
wondering about how do we make sure that this
2:59:00
this
2:59:00
this technology is inclusive for example for
2:59:02
technology is inclusive for example for
2:59:02
technology is inclusive for example for people who can't see
2:59:04
people who can't see
2:59:04
people who can't see we're talking about computer vision so
2:59:05
we're talking about computer vision so
2:59:05
we're talking about computer vision so it's visual and then
2:59:07
it's visual and then
2:59:07
it's visual and then i can't see what happens next
2:59:11
i can't see what happens next
2:59:11
i can't see what happens next yeah that's why i think adding the other
2:59:13
yeah that's why i think adding the other
2:59:13
yeah that's why i think adding the other senses is going to be really important
2:59:15
senses is going to be really important
2:59:15
senses is going to be really important and there's been a lot a lot more focus
2:59:18
and there's been a lot a lot more focus
2:59:18
and there's been a lot a lot more focus recently and some of the things i've
2:59:19
recently and some of the things i've
2:59:19
recently and some of the things i've seen of just the last few
2:59:21
seen of just the last few
2:59:21
seen of just the last few months actually where we're adding
2:59:24
months actually where we're adding
2:59:24
months actually where we're adding in alternate cues of i
2:59:27
in alternate cues of i
2:59:27
in alternate cues of i okay i'm doing something same i'm
2:59:28
okay i'm doing something same i'm
2:59:28
okay i'm doing something same i'm wearing a heads-up an ar headset or a vr
2:59:30
wearing a heads-up an ar headset or a vr
2:59:30
wearing a heads-up an ar headset or a vr headset
2:59:31
headset
2:59:31
headset and i i want to uh
2:59:34
and i i want to uh
2:59:34
and i i want to uh might have limited ability maybe i'm
2:59:36
might have limited ability maybe i'm
2:59:36
might have limited ability maybe i'm maybe the individual's colorblind or
2:59:38
maybe the individual's colorblind or
2:59:38
maybe the individual's colorblind or maybe there's some
2:59:39
maybe there's some
2:59:39
maybe there's some some sort of physical impairment there
2:59:41
some sort of physical impairment there
2:59:41
some sort of physical impairment there that they might not be able to see
2:59:43
that they might not be able to see
2:59:43
that they might not be able to see the full depth of certain things uh the
2:59:45
the full depth of certain things uh the
2:59:45
the full depth of certain things uh the other thing has been interesting to
2:59:45
other thing has been interesting to
2:59:45
other thing has been interesting to realize what percentage of the
2:59:47
realize what percentage of the
2:59:47
realize what percentage of the population
2:59:47
population
2:59:48
population actually cannot see 3d using a
2:59:50
actually cannot see 3d using a
2:59:50
actually cannot see 3d using a stereoscopic headset
2:59:52
stereoscopic headset
2:59:52
stereoscopic headset it's a double-digit percentage so
2:59:55
it's a double-digit percentage so
2:59:56
it's a double-digit percentage so that's a large factor as well uh but
2:59:58
that's a large factor as well uh but
2:59:58
that's a large factor as well uh but there's been a lot
2:59:59
there's been a lot
2:59:59
there's been a lot i've seen like i said just in the last
3:00:01
i've seen like i said just in the last
3:00:01
i've seen like i said just in the last few months of adding in a more maybe and
3:00:03
few months of adding in a more maybe and
3:00:03
few months of adding in a more maybe and an audio cue when
3:00:04
an audio cue when
3:00:04
an audio cue when when someone is doing this or maybe i
3:00:06
when someone is doing this or maybe i
3:00:06
when someone is doing this or maybe i add in some sort of
3:00:07
add in some sort of
3:00:07
add in some sort of haptic feedback depending on what kind
3:00:08
haptic feedback depending on what kind
3:00:08
haptic feedback depending on what kind of a rigor setup you have
3:00:10
of a rigor setup you have
3:00:10
of a rigor setup you have so that it might supplement the visual
3:00:13
so that it might supplement the visual
3:00:13
so that it might supplement the visual it's still
3:00:13
it's still
3:00:14
it's still very visual heavy uh i have seen a few
3:00:17
very visual heavy uh i have seen a few
3:00:17
very visual heavy uh i have seen a few cases where they've had some sort of
3:00:19
cases where they've had some sort of
3:00:19
cases where they've had some sort of virtual meeting
3:00:20
virtual meeting
3:00:20
virtual meeting and and the focus was primarily on audio
3:00:23
and and the focus was primarily on audio
3:00:24
and and the focus was primarily on audio and then visual visual was like a add-on
3:00:26
and then visual visual was like a add-on
3:00:26
and then visual visual was like a add-on so like if you have the ability to see
3:00:27
so like if you have the ability to see
3:00:28
so like if you have the ability to see as well but it was really focused on the
3:00:29
as well but it was really focused on the
3:00:29
as well but it was really focused on the audio
3:00:30
audio
3:00:30
audio and all the dimensionality of that of
3:00:31
and all the dimensionality of that of
3:00:31
and all the dimensionality of that of having the music and the cues and the
3:00:33
having the music and the cues and the
3:00:33
having the music and the cues and the prompts and i think you're seeing more
3:00:36
prompts and i think you're seeing more
3:00:36
prompts and i think you're seeing more of that but you see the same
3:00:37
of that but you see the same
3:00:37
of that but you see the same technology curve for everything first
3:00:39
technology curve for everything first
3:00:39
technology curve for everything first let's figure out how do we make
3:00:41
let's figure out how do we make
3:00:41
let's figure out how do we make is this technologically feasible can we
3:00:44
is this technologically feasible can we
3:00:44
is this technologically feasible can we can we make something that
3:00:45
can we make something that
3:00:45
can we make something that does this and then once we do that it's
3:00:47
does this and then once we do that it's
3:00:47
does this and then once we do that it's like how do we
3:00:48
like how do we
3:00:48
like how do we make it better faster cheaper smaller
3:00:51
make it better faster cheaper smaller
3:00:52
make it better faster cheaper smaller and and then in that process of better
3:00:54
and and then in that process of better
3:00:54
and and then in that process of better faster cheaper smaller it's also how to
3:00:56
faster cheaper smaller it's also how to
3:00:56
faster cheaper smaller it's also how to widen the usage for it and i think a lot
3:00:58
widen the usage for it and i think a lot
3:00:58
widen the usage for it and i think a lot of this technology it's still
3:00:59
of this technology it's still
3:00:59
of this technology it's still early in that curve so you're only
3:01:01
early in that curve so you're only
3:01:01
early in that curve so you're only seeing that
3:01:03
seeing that
3:01:03
seeing that inclusiveness in certain areas as
3:01:05
inclusiveness in certain areas as
3:01:05
inclusiveness in certain areas as they're sort of cusping that
3:01:06
they're sort of cusping that
3:01:06
they're sort of cusping that you know pressing that that curve there
3:01:08
you know pressing that that curve there
3:01:08
you know pressing that that curve there so it's i wouldn't say it's even across
3:01:10
so it's i wouldn't say it's even across
3:01:10
so it's i wouldn't say it's even across the board
3:01:11
the board
3:01:11
the board but there's definitely been a much
3:01:12
but there's definitely been a much
3:01:12
but there's definitely been a much bigger focus on it
3:01:14
bigger focus on it
3:01:14
bigger focus on it probably since march because more and
3:01:17
probably since march because more and
3:01:17
probably since march because more and more people are using it you know i
3:01:18
more people are using it you know i
3:01:18
more people are using it you know i guess you can notice the timeline
3:01:19
guess you can notice the timeline
3:01:20
guess you can notice the timeline relates very closely to what's going on
3:01:21
relates very closely to what's going on
3:01:21
relates very closely to what's going on in the world uh because those issues
3:01:23
in the world uh because those issues
3:01:23
in the world uh because those issues have been raised so the
3:01:24
have been raised so the
3:01:24
have been raised so the the acceleration of adding in that uh
3:01:27
the acceleration of adding in that uh
3:01:27
the acceleration of adding in that uh alternate usage
3:01:28
alternate usage
3:01:28
alternate usage and inclusivity i think is happening
3:01:30
and inclusivity i think is happening
3:01:30
and inclusivity i think is happening much more rapidly than it would normally
3:01:32
much more rapidly than it would normally
3:01:32
much more rapidly than it would normally which is maybe a silver lining to the
3:01:34
which is maybe a silver lining to the
3:01:34
which is maybe a silver lining to the current world situation
3:01:36
current world situation
3:01:36
current world situation yeah yeah we've seen that this
3:01:39
yeah yeah we've seen that this
3:01:39
yeah yeah we've seen that this across the board for for ai actually
3:01:42
across the board for for ai actually
3:01:42
across the board for for ai actually last week we had a discussion with
3:01:44
last week we had a discussion with
3:01:44
last week we had a discussion with uh richard campbell he's a famous guy
3:01:48
uh richard campbell he's a famous guy
3:01:48
uh richard campbell he's a famous guy he records a lot of podcasts uh on the
3:01:50
he records a lot of podcasts uh on the
3:01:50
he records a lot of podcasts uh on the internet um
3:01:51
internet um
3:01:51
internet um he's well known in the microsoft
3:01:52
he's well known in the microsoft
3:01:52
he's well known in the microsoft community as well um
3:01:54
community as well um
3:01:54
community as well um he is he actually showed us the history
3:01:57
he is he actually showed us the history
3:01:57
he is he actually showed us the history of ai last week
3:01:58
of ai last week
3:01:58
of ai last week and we had this discussion where we saw
3:02:00
and we had this discussion where we saw
3:02:00
and we had this discussion where we saw that in the 80s we had compute problems
3:02:03
that in the 80s we had compute problems
3:02:03
that in the 80s we had compute problems we couldn't possibly
3:02:05
we couldn't possibly
3:02:05
we couldn't possibly train a neural network that was too slow
3:02:07
train a neural network that was too slow
3:02:07
train a neural network that was too slow but nowadays we can easily train a
3:02:09
but nowadays we can easily train a
3:02:09
but nowadays we can easily train a neural network where
3:02:10
neural network where
3:02:10
neural network where heck we're training neural networks with
3:02:12
heck we're training neural networks with
3:02:12
heck we're training neural networks with one and a half billion parameters that's
3:02:14
one and a half billion parameters that's
3:02:14
one and a half billion parameters that's like
3:02:15
like
3:02:16
like i can't do that on my machine it doesn't
3:02:18
i can't do that on my machine it doesn't
3:02:18
i can't do that on my machine it doesn't fit
3:02:20
fit
3:02:20
fit what we've seen is that just now we're
3:02:22
what we've seen is that just now we're
3:02:22
what we've seen is that just now we're starting to learn that ai is actually
3:02:24
starting to learn that ai is actually
3:02:24
starting to learn that ai is actually something that's pretty dangerous and we
3:02:26
something that's pretty dangerous and we
3:02:26
something that's pretty dangerous and we should
3:02:27
should
3:02:27
should spend more time thinking about yes we
3:02:29
spend more time thinking about yes we
3:02:29
spend more time thinking about yes we can do that but why are we doing this
3:02:31
can do that but why are we doing this
3:02:31
can do that but why are we doing this is this the best option we can offer to
3:02:33
is this the best option we can offer to
3:02:33
is this the best option we can offer to a user is it inclusive is it fair
3:02:35
a user is it inclusive is it fair
3:02:35
a user is it inclusive is it fair to everybody um is this something that
3:02:39
to everybody um is this something that
3:02:39
to everybody um is this something that cisco is aware of
3:02:40
cisco is aware of
3:02:40
cisco is aware of and are you spending a lot of time on
3:02:41
and are you spending a lot of time on
3:02:41
and are you spending a lot of time on that
3:02:43
that
3:02:43
that i would say there is a group of people
3:02:45
i would say there is a group of people
3:02:45
i would say there is a group of people that is spending a lot of time on that
3:02:47
that is spending a lot of time on that
3:02:47
that is spending a lot of time on that uh it's a little bit at least for us
3:02:49
uh it's a little bit at least for us
3:02:49
uh it's a little bit at least for us it's a little bit of there is a group of
3:02:51
it's a little bit of there is a group of
3:02:51
it's a little bit of there is a group of people focus
3:02:51
people focus
3:02:52
people focus on those particular issues and problems
3:02:53
on those particular issues and problems
3:02:53
on those particular issues and problems rather than it necessarily being
3:02:55
rather than it necessarily being
3:02:55
rather than it necessarily being ingrained in everything
3:02:56
ingrained in everything
3:02:56
ingrained in everything i think at some point you will see that
3:02:59
i think at some point you will see that
3:02:59
i think at some point you will see that uh
3:03:00
uh
3:03:00
uh translate more broadly and people will
3:03:02
translate more broadly and people will
3:03:02
translate more broadly and people will be factoring that in from the get-go but
3:03:03
be factoring that in from the get-go but
3:03:04
be factoring that in from the get-go but again i think it's just where we are in
3:03:05
again i think it's just where we are in
3:03:05
again i think it's just where we are in the curve
3:03:06
the curve
3:03:06
the curve it's more now it's figuring out how do
3:03:08
it's more now it's figuring out how do
3:03:08
it's more now it's figuring out how do we do that and then once we figure out
3:03:09
we do that and then once we figure out
3:03:09
we do that and then once we figure out the how
3:03:10
the how
3:03:10
the how then we have the systems in place that
3:03:12
then we have the systems in place that
3:03:12
then we have the systems in place that we can roll that out more broadly but
3:03:14
we can roll that out more broadly but
3:03:14
we can roll that out more broadly but we're still working on the how well it
3:03:16
we're still working on the how well it
3:03:16
we're still working on the how well it sounds really interesting because i
3:03:18
sounds really interesting because i
3:03:18
sounds really interesting because i um i'm speaking to a lot of companies
3:03:21
um i'm speaking to a lot of companies
3:03:21
um i'm speaking to a lot of companies about ai and i don't know about you
3:03:23
about ai and i don't know about you
3:03:23
about ai and i don't know about you alicia but
3:03:25
alicia but
3:03:25
alicia but i haven't heard a lot of talk about
3:03:26
i haven't heard a lot of talk about
3:03:26
i haven't heard a lot of talk about model fairness or
3:03:28
model fairness or
3:03:28
model fairness or making computer efficient solutions
3:03:30
making computer efficient solutions
3:03:30
making computer efficient solutions compatible for people who can't see
3:03:33
compatible for people who can't see
3:03:33
compatible for people who can't see um how's that in your work
3:03:39
sorry who's that question director that
3:03:40
sorry who's that question director that
3:03:40
sorry who's that question director that uh sorry alicia
3:03:44
well i i think um you know i echo the
3:03:47
well i i think um you know i echo the
3:03:47
well i i think um you know i echo the sentiment that
3:03:48
sentiment that
3:03:48
sentiment that a lot of people are still talking about
3:03:50
a lot of people are still talking about
3:03:50
a lot of people are still talking about you know how and
3:03:52
you know how and
3:03:52
you know how and it's it's really the people who are
3:03:55
it's it's really the people who are
3:03:55
it's it's really the people who are um the architects
3:03:59
um the architects
3:03:59
um the architects that asked why you know and
3:04:03
that asked why you know and
3:04:03
that asked why you know and and you know when you start looking at
3:04:05
and you know when you start looking at
3:04:05
and you know when you start looking at um gdpr
3:04:06
um gdpr
3:04:06
um gdpr and where jada where data gets placed
3:04:09
and where jada where data gets placed
3:04:09
and where jada where data gets placed and data restrictions and
3:04:11
and data restrictions and
3:04:11
and data restrictions and the different privacy concerns and i
3:04:13
the different privacy concerns and i
3:04:13
the different privacy concerns and i think a lot of the
3:04:15
think a lot of the
3:04:15
think a lot of the um the rules and the governing
3:04:17
um the rules and the governing
3:04:17
um the rules and the governing restrictions
3:04:18
restrictions
3:04:18
restrictions of of the different countries drives a
3:04:21
of of the different countries drives a
3:04:21
of of the different countries drives a lot of these conversations
3:04:23
lot of these conversations
3:04:23
lot of these conversations but um they're they're just so relevant
3:04:26
but um they're they're just so relevant
3:04:26
but um they're they're just so relevant that when you're building these systems
3:04:29
that when you're building these systems
3:04:29
that when you're building these systems you you really need to start asking
3:04:30
you you really need to start asking
3:04:30
you you really need to start asking those questions up front nowadays
3:04:32
those questions up front nowadays
3:04:32
those questions up front nowadays you know and and um just with the court
3:04:35
you know and and um just with the court
3:04:35
you know and and um just with the court cases that came out this year
3:04:37
cases that came out this year
3:04:37
cases that came out this year um it's quite impactful to deploy a
3:04:41
um it's quite impactful to deploy a
3:04:41
um it's quite impactful to deploy a commercial
3:04:41
commercial
3:04:41
commercial solution that isn't
3:04:45
solution that isn't
3:04:45
solution that isn't kind of in the public's best interest
3:04:47
kind of in the public's best interest
3:04:47
kind of in the public's best interest and people find out about it
3:04:49
and people find out about it
3:04:49
and people find out about it and there there aren't as many rules out
3:04:52
and there there aren't as many rules out
3:04:52
and there there aren't as many rules out there that there should be
3:04:54
there that there should be
3:04:54
there that there should be but i i think that we're seeing that
3:04:57
but i i think that we're seeing that
3:04:57
but i i think that we're seeing that the response time is very very quick to
3:05:00
the response time is very very quick to
3:05:00
the response time is very very quick to rectify that so i
3:05:01
rectify that so i
3:05:01
rectify that so i i think that um you know
3:05:05
i think that um you know
3:05:05
i think that um you know the more we start using these systems
3:05:06
the more we start using these systems
3:05:06
the more we start using these systems the more these situations come up
3:05:08
the more these situations come up
3:05:08
the more these situations come up and the more opportunities we have to
3:05:11
and the more opportunities we have to
3:05:11
and the more opportunities we have to have those discussions
3:05:12
have those discussions
3:05:12
have those discussions and to make that change so definitely
3:05:15
and to make that change so definitely
3:05:15
and to make that change so definitely at the builder level and um
3:05:19
at the builder level and um
3:05:19
at the builder level and um but but i think everybody starts at the
3:05:21
but but i think everybody starts at the
3:05:21
but but i think everybody starts at the code and
3:05:22
code and
3:05:22
code and you know gets excited and think about
3:05:25
you know gets excited and think about
3:05:25
you know gets excited and think about all the things i can do with this
3:05:26
all the things i can do with this
3:05:26
all the things i can do with this and then after that it's um but should i
3:05:29
and then after that it's um but should i
3:05:29
and then after that it's um but should i be doing it right
3:05:34
cool cool so um
3:05:37
cool cool so um
3:05:38
cool cool so um the the the whole ethical aspect
3:05:41
the the the whole ethical aspect
3:05:41
the the the whole ethical aspect there's not a lot of legislation as you
3:05:43
there's not a lot of legislation as you
3:05:43
there's not a lot of legislation as you mentioned uh alicia
3:05:44
mentioned uh alicia
3:05:44
mentioned uh alicia and so elizabeth what do you think of
3:05:47
and so elizabeth what do you think of
3:05:47
and so elizabeth what do you think of initiatives around bringing in
3:05:49
initiatives around bringing in
3:05:49
initiatives around bringing in legislation
3:05:51
legislation
3:05:51
legislation around the use of ai
3:05:54
well i think we learned from terminator
3:05:56
well i think we learned from terminator
3:05:56
well i think we learned from terminator we don't want to wait until after we
3:05:58
we don't want to wait until after we
3:05:58
we don't want to wait until after we build skynet to then decide should we
3:06:00
build skynet to then decide should we
3:06:00
build skynet to then decide should we build skynet so
3:06:03
build skynet so
3:06:03
build skynet so i i think it's something that you do
3:06:05
i i think it's something that you do
3:06:05
i i think it's something that you do it's a unique area because i think it's
3:06:06
it's a unique area because i think it's
3:06:06
it's a unique area because i think it's something you do need to look at the
3:06:07
something you do need to look at the
3:06:07
something you do need to look at the legislation
3:06:08
legislation
3:06:08
legislation in the process of building it and a lot
3:06:11
in the process of building it and a lot
3:06:11
in the process of building it and a lot of times
3:06:12
of times
3:06:12
of times legislation is an after not necessarily
3:06:15
legislation is an after not necessarily
3:06:15
legislation is an after not necessarily an afterthought but something you do
3:06:16
an afterthought but something you do
3:06:16
an afterthought but something you do after the fact you you do
3:06:17
after the fact you you do
3:06:17
after the fact you you do build the technology first and then you
3:06:19
build the technology first and then you
3:06:19
build the technology first and then you look at how and where and why should we
3:06:21
look at how and where and why should we
3:06:21
look at how and where and why should we use it
3:06:22
use it
3:06:22
use it ai is unique in that and i feel like you
3:06:25
ai is unique in that and i feel like you
3:06:25
ai is unique in that and i feel like you are seeing that now
3:06:26
are seeing that now
3:06:26
are seeing that now and that there are more and more court
3:06:27
and that there are more and more court
3:06:27
and that there are more and more court cases around ai even though the
3:06:29
cases around ai even though the
3:06:29
cases around ai even though the technology is still quite nascent
3:06:31
technology is still quite nascent
3:06:31
technology is still quite nascent and i think it is necessary because the
3:06:33
and i think it is necessary because the
3:06:33
and i think it is necessary because the nature of what is capable with ai
3:06:35
nature of what is capable with ai
3:06:35
nature of what is capable with ai that you do need to be building your
3:06:37
that you do need to be building your
3:06:37
that you do need to be building your legislation along with building the
3:06:39
legislation along with building the
3:06:39
legislation along with building the technology
3:06:41
technology
3:06:41
technology yeah that's pretty unique for the two ai
3:06:44
yeah that's pretty unique for the two ai
3:06:44
yeah that's pretty unique for the two ai would you say that we are right on time
3:06:46
would you say that we are right on time
3:06:46
would you say that we are right on time with legislation around ai
3:06:50
all right do i want to answer that on
3:06:52
all right do i want to answer that on
3:06:52
all right do i want to answer that on the grounds i mean
3:06:54
the grounds i mean
3:06:54
the grounds i mean feel free not to answer that or um i'm
3:06:57
feel free not to answer that or um i'm
3:06:57
feel free not to answer that or um i'm wondering because i
3:06:58
wondering because i
3:06:58
wondering because i i personally feel that um so when i
3:07:01
i personally feel that um so when i
3:07:01
i personally feel that um so when i teach ai to
3:07:02
teach ai to
3:07:02
teach ai to to young students they already know that
3:07:05
to young students they already know that
3:07:05
to young students they already know that ai is in spotify
3:07:06
ai is in spotify
3:07:06
ai is in spotify in their snapchat in their instagram in
3:07:08
in their snapchat in their instagram in
3:07:08
in their snapchat in their instagram in their facebook if they use that at all
3:07:11
their facebook if they use that at all
3:07:11
their facebook if they use that at all anymore
3:07:11
anymore
3:07:11
anymore um it's everywhere and and since it's
3:07:15
um it's everywhere and and since it's
3:07:15
um it's everywhere and and since it's everywhere i would
3:07:16
everywhere i would
3:07:16
everywhere i would actually argue that it might be a little
3:07:18
actually argue that it might be a little
3:07:18
actually argue that it might be a little bit on the late side
3:07:20
bit on the late side
3:07:20
bit on the late side uh we might have to fix some of the
3:07:21
uh we might have to fix some of the
3:07:21
uh we might have to fix some of the damage that we've caused already with
3:07:23
damage that we've caused already with
3:07:23
damage that we've caused already with all the
3:07:23
all the
3:07:24
all the amazing cool code that we wrote
3:07:28
i won't argue with you
3:07:29
i won't argue with you
3:07:29
i won't argue with you [Laughter]
3:07:32
[Laughter]
3:07:32
[Laughter] cool let's keep it yeah that sounds like
3:07:34
cool let's keep it yeah that sounds like
3:07:34
cool let's keep it yeah that sounds like an awesome idea
3:07:35
an awesome idea
3:07:36
an awesome idea so there's one last question that i
3:07:37
so there's one last question that i
3:07:37
so there's one last question that i would like to ask um
3:07:39
would like to ask um
3:07:39
would like to ask um so um someone uh asked yesterday i was
3:07:43
so um someone uh asked yesterday i was
3:07:43
so um someone uh asked yesterday i was at a hacking conference in a
3:07:44
at a hacking conference in a
3:07:44
at a hacking conference in a ai hacking conference um
3:07:47
ai hacking conference um
3:07:47
ai hacking conference um uh how will you relate ai to hacking
3:07:51
uh how will you relate ai to hacking
3:07:52
uh how will you relate ai to hacking in this case what are you going to do
3:07:53
in this case what are you going to do
3:07:53
in this case what are you going to do against that for example
3:07:57
against that for example
3:07:57
against that for example why do anything against it yeah
3:08:00
why do anything against it yeah
3:08:00
why do anything against it yeah do you mean on the security side when it
3:08:02
do you mean on the security side when it
3:08:02
do you mean on the security side when it comes to on the creation side why would
3:08:03
comes to on the creation side why would
3:08:04
comes to on the creation side why would you do anything again today i mean the
3:08:05
you do anything again today i mean the
3:08:05
you do anything again today i mean the creation is creation it's great whether
3:08:07
creation is creation it's great whether
3:08:07
creation is creation it's great whether it comes from a human or an ai like
3:08:08
it comes from a human or an ai like
3:08:08
it comes from a human or an ai like let's let's get as much as we can
3:08:11
let's let's get as much as we can
3:08:11
let's let's get as much as we can on the security side and there's
3:08:14
on the security side and there's
3:08:14
on the security side and there's i know our security division is humming
3:08:15
i know our security division is humming
3:08:16
i know our security division is humming like nuts these days
3:08:17
like nuts these days
3:08:17
like nuts these days there's so much to be looking at so much
3:08:18
there's so much to be looking at so much
3:08:18
there's so much to be looking at so much to be exploring as you start getting
3:08:20
to be exploring as you start getting
3:08:20
to be exploring as you start getting into these really creative things like
3:08:22
into these really creative things like
3:08:22
into these really creative things like how do you detect and block deep fakes
3:08:24
how do you detect and block deep fakes
3:08:24
how do you detect and block deep fakes how do all there's so many different
3:08:25
how do all there's so many different
3:08:25
how do all there's so many different things that we need to do to ensure
3:08:27
things that we need to do to ensure
3:08:27
things that we need to do to ensure that our community where the
3:08:29
that our community where the
3:08:29
that our community where the communications collaboration team how do
3:08:30
communications collaboration team how do
3:08:30
communications collaboration team how do we ensure that our communications and
3:08:31
we ensure that our communications and
3:08:32
we ensure that our communications and collaborations are
3:08:34
collaborations are
3:08:34
collaborations are effectively reacting to every security
3:08:36
effectively reacting to every security
3:08:36
effectively reacting to every security threat uh ai is interesting in that it's
3:08:38
threat uh ai is interesting in that it's
3:08:38
threat uh ai is interesting in that it's throwing lots of security threats out
3:08:40
throwing lots of security threats out
3:08:40
throwing lots of security threats out there all at once so from that
3:08:41
there all at once so from that
3:08:41
there all at once so from that perspective i mean
3:08:42
perspective i mean
3:08:42
perspective i mean yeah absolutely it's something that we
3:08:43
yeah absolutely it's something that we
3:08:43
yeah absolutely it's something that we need to be focused on and i know
3:08:45
need to be focused on and i know
3:08:45
need to be focused on and i know our group is and i know other major
3:08:47
our group is and i know other major
3:08:47
our group is and i know other major technology companies are as well like
3:08:48
technology companies are as well like
3:08:48
technology companies are as well like if you just read the headlines for the
3:08:50
if you just read the headlines for the
3:08:50
if you just read the headlines for the last three months security security
3:08:53
last three months security security
3:08:53
last three months security security yeah it's been pretty important and
3:08:54
yeah it's been pretty important and
3:08:54
yeah it's been pretty important and we've seen actually
3:08:56
we've seen actually
3:08:56
we've seen actually uh a lot of samples where hackers
3:08:59
uh a lot of samples where hackers
3:08:59
uh a lot of samples where hackers attempted to fool a neural network uh
3:09:02
attempted to fool a neural network uh
3:09:02
attempted to fool a neural network uh that was using computer vision by just
3:09:04
that was using computer vision by just
3:09:04
that was using computer vision by just injecting some noise in there so that
3:09:06
injecting some noise in there so that
3:09:06
injecting some noise in there so that seems pretty interesting
3:09:07
seems pretty interesting
3:09:07
seems pretty interesting um to actually on one side try it out
3:09:11
um to actually on one side try it out
3:09:11
um to actually on one side try it out once
3:09:11
once
3:09:11
once i guess you need to do it once to
3:09:14
i guess you need to do it once to
3:09:14
i guess you need to do it once to discover what it's like and what you can
3:09:15
discover what it's like and what you can
3:09:16
discover what it's like and what you can do against it and on the other hand
3:09:17
do against it and on the other hand
3:09:17
do against it and on the other hand what are we going to do against that and
3:09:20
what are we going to do against that and
3:09:20
what are we going to do against that and do we need to worry
3:09:21
do we need to worry
3:09:21
do we need to worry at all about those kind of attacks when
3:09:23
at all about those kind of attacks when
3:09:23
at all about those kind of attacks when we we talk about collaboration or
3:09:26
we we talk about collaboration or
3:09:26
we we talk about collaboration or maybe other solutions in that field yeah
3:09:28
maybe other solutions in that field yeah
3:09:28
maybe other solutions in that field yeah on the security side absolutely
3:09:30
on the security side absolutely
3:09:30
on the security side absolutely and if i were in the product
3:09:32
and if i were in the product
3:09:32
and if i were in the product organization this would probably
3:09:33
organization this would probably
3:09:33
organization this would probably probably be the number one thing that
3:09:34
probably be the number one thing that
3:09:34
probably be the number one thing that kept me up at night but thankfully for
3:09:36
kept me up at night but thankfully for
3:09:36
kept me up at night but thankfully for my own sanity i am not the primary
3:09:38
my own sanity i am not the primary
3:09:38
my own sanity i am not the primary organization i'm a prototype
3:09:40
organization i'm a prototype
3:09:40
organization i'm a prototype organization so
3:09:41
organization so
3:09:41
organization so our view of this is much more on the not
3:09:44
our view of this is much more on the not
3:09:44
our view of this is much more on the not how do we make this safe for everyone
3:09:45
how do we make this safe for everyone
3:09:45
how do we make this safe for everyone how do we protect our tools against it
3:09:47
how do we protect our tools against it
3:09:47
how do we protect our tools against it it's much more on the
3:09:48
it's much more on the
3:09:48
it's much more on the cool what could we do here and and so
3:09:50
cool what could we do here and and so
3:09:50
cool what could we do here and and so it's more of the experimentation how do
3:09:52
it's more of the experimentation how do
3:09:52
it's more of the experimentation how do we break something like let's try 25
3:09:54
we break something like let's try 25
3:09:54
we break something like let's try 25 different ways to break and see what we
3:09:55
different ways to break and see what we
3:09:55
different ways to break and see what we could do and let's also like hey
3:09:57
could do and let's also like hey
3:09:57
could do and let's also like hey have you tried anytime we do something
3:09:58
have you tried anytime we do something
3:09:58
have you tried anytime we do something like have you tried like we have our
3:10:00
like have you tried like we have our
3:10:00
like have you tried like we have our resident ai experts like
3:10:02
resident ai experts like
3:10:02
resident ai experts like throw him in there see what he can do
3:10:03
throw him in there see what he can do
3:10:03
throw him in there see what he can do and see if he can come up with a
3:10:05
and see if he can come up with a
3:10:05
and see if he can come up with a different answer so
3:10:06
different answer so
3:10:06
different answer so on the creation experimentation side i
3:10:08
on the creation experimentation side i
3:10:08
on the creation experimentation side i think it's phenomenal because you can
3:10:09
think it's phenomenal because you can
3:10:10
think it's phenomenal because you can find some really interesting things that
3:10:12
find some really interesting things that
3:10:12
find some really interesting things that i think as as humans we would have a
3:10:13
i think as as humans we would have a
3:10:14
i think as as humans we would have a harder time making that mental beat but
3:10:15
harder time making that mental beat but
3:10:15
harder time making that mental beat but then you see
3:10:16
then you see
3:10:16
then you see let me interject some a oh this way oh
3:10:19
let me interject some a oh this way oh
3:10:19
let me interject some a oh this way oh yeah that's what a i would do in that
3:10:20
yeah that's what a i would do in that
3:10:20
yeah that's what a i would do in that scenario that's not what i would have
3:10:21
scenario that's not what i would have
3:10:21
scenario that's not what i would have thought of but that's pretty cool
3:10:22
thought of but that's pretty cool
3:10:22
thought of but that's pretty cool and it helps you learn much quicker yeah
3:10:25
and it helps you learn much quicker yeah
3:10:25
and it helps you learn much quicker yeah that sounds
3:10:25
that sounds
3:10:26
that sounds cool so actually now that i i come to
3:10:29
cool so actually now that i i come to
3:10:29
cool so actually now that i i come to think of it
3:10:30
think of it
3:10:30
think of it your job is actually at the beginning of
3:10:32
your job is actually at the beginning of
3:10:32
your job is actually at the beginning of the curve trying to
3:10:34
the curve trying to
3:10:34
the curve trying to explore what's possible and the other
3:10:36
explore what's possible and the other
3:10:36
explore what's possible and the other departments such as legal and security
3:10:38
departments such as legal and security
3:10:38
departments such as legal and security and product departments are actually
3:10:40
and product departments are actually
3:10:40
and product departments are actually working out
3:10:41
working out
3:10:41
working out is this actually an idea that's safe to
3:10:44
is this actually an idea that's safe to
3:10:44
is this actually an idea that's safe to bring on the market
3:10:45
bring on the market
3:10:45
bring on the market is that correct do the really hard jobs
3:10:47
is that correct do the really hard jobs
3:10:48
is that correct do the really hard jobs we just have to have all the fun
3:10:49
we just have to have all the fun
3:10:49
we just have to have all the fun i hope my management chain doesn't watch
3:10:51
i hope my management chain doesn't watch
3:10:51
i hope my management chain doesn't watch this
3:10:53
this
3:10:53
this sounds like an awesome job well thank
3:10:55
sounds like an awesome job well thank
3:10:55
sounds like an awesome job well thank you very much
3:10:57
you very much
3:10:57
you very much for talking to us so i'm actually quite
3:10:59
for talking to us so i'm actually quite
3:10:59
for talking to us so i'm actually quite happy with uh
3:11:00
happy with uh
3:11:00
happy with uh uh with the talk because i never
3:11:02
uh with the talk because i never
3:11:02
uh with the talk because i never actually spend a lot of time thinking
3:11:04
actually spend a lot of time thinking
3:11:04
actually spend a lot of time thinking about those micro interactions i'm i'm
3:11:06
about those micro interactions i'm i'm
3:11:06
about those micro interactions i'm i'm all about the code and worrying about
3:11:08
all about the code and worrying about
3:11:08
all about the code and worrying about how do i make this neural network work
3:11:10
how do i make this neural network work
3:11:10
how do i make this neural network work correctly
3:11:12
correctly
3:11:12
correctly so that sounds cool well uh alicia do
3:11:15
so that sounds cool well uh alicia do
3:11:15
so that sounds cool well uh alicia do you have anything to add to that i mean
3:11:18
you have anything to add to that i mean
3:11:18
you have anything to add to that i mean no you're one of my heroes now
3:11:21
no you're one of my heroes now
3:11:21
no you're one of my heroes now so i definitely enjoyed the talk
3:11:25
so i definitely enjoyed the talk
3:11:25
so i definitely enjoyed the talk and thank you for speaking with us today
3:11:28
and thank you for speaking with us today
3:11:28
and thank you for speaking with us today thank you very much it was a lot of fun
3:11:29
thank you very much it was a lot of fun
3:11:30
thank you very much it was a lot of fun yes i'll do it for my robot dog
3:11:34
yes i'll do it for my robot dog
3:11:34
yes i'll do it for my robot dog we'll see what we can do by christmas
3:11:37
we'll see what we can do by christmas
3:11:37
we'll see what we can do by christmas very very cool
3:11:38
very very cool
3:11:38
very very cool awesome thank you very much um
3:11:42
awesome thank you very much um
3:11:42
awesome thank you very much um the the uh point that uh we're at the
3:11:44
the the uh point that uh we're at the
3:11:44
the the uh point that uh we're at the end of this show
3:11:45
end of this show
3:11:45
end of this show and um i would like to thank uh all the
3:11:48
and um i would like to thank uh all the
3:11:48
and um i would like to thank uh all the speakers that have joined us today
3:11:50
speakers that have joined us today
3:11:50
speakers that have joined us today um they went out of their way to put
3:11:52
um they went out of their way to put
3:11:52
um they went out of their way to put together slides uh
3:11:54
together slides uh
3:11:54
together slides uh talk to us about their ideas and uh um
3:11:57
talk to us about their ideas and uh um
3:11:57
talk to us about their ideas and uh um uh i personally had a lot of meetings
3:12:00
uh i personally had a lot of meetings
3:12:00
uh i personally had a lot of meetings setting this up so i'm quite happy that
3:12:01
setting this up so i'm quite happy that
3:12:01
setting this up so i'm quite happy that this came out
3:12:02
this came out
3:12:02
this came out um thank you alicia for joining me as a
3:12:05
um thank you alicia for joining me as a
3:12:05
um thank you alicia for joining me as a co-host
3:12:06
co-host
3:12:06
co-host and we still need to talk about that dog
3:12:08
and we still need to talk about that dog
3:12:08
and we still need to talk about that dog i don't i'm not sure
3:12:09
i don't i'm not sure
3:12:09
i don't i'm not sure if i i want to get the mechanical one i
3:12:12
if i i want to get the mechanical one i
3:12:12
if i i want to get the mechanical one i guess
3:12:14
guess
3:12:14
guess no this one's cuter and it's it's not
3:12:16
no this one's cuter and it's it's not
3:12:16
no this one's cuter and it's it's not right that he snores
3:12:18
right that he snores
3:12:18
right that he snores it's it's not every night it's just you
3:12:20
it's it's not every night it's just you
3:12:20
it's it's not every night it's just you know the days i give him bacon i think
3:12:24
know the days i give him bacon i think
3:12:24
know the days i give him bacon i think oh it's probably self-imposed
3:12:27
oh it's probably self-imposed
3:12:27
oh it's probably self-imposed like i will start recording the days
3:12:31
like i will start recording the days
3:12:31
like i will start recording the days that i give him bacon
3:12:32
that i give him bacon
3:12:32
that i give him bacon and seeing if there's a direct
3:12:33
and seeing if there's a direct
3:12:33
and seeing if there's a direct correlation but i
3:12:35
correlation but i
3:12:35
correlation but i i'm pretty sure it's related to his
3:12:37
i'm pretty sure it's related to his
3:12:37
i'm pretty sure it's related to his level of happiness
3:12:39
level of happiness
3:12:39
level of happiness but i i do have a lot of trees in my
3:12:41
but i i do have a lot of trees in my
3:12:41
but i i do have a lot of trees in my backyard as well
3:12:43
backyard as well
3:12:43
backyard as well so i do have a beach as well
3:12:46
so i do have a beach as well
3:12:46
so i do have a beach as well no i don't have a beach oh might be time
3:12:49
no i don't have a beach oh might be time
3:12:49
no i don't have a beach oh might be time to build one
3:12:50
to build one
3:12:50
to build one i guess right but would that make him
3:12:53
i guess right but would that make him
3:12:53
i guess right but would that make him happier
3:12:53
happier
3:12:53
happier where he would snore more so does that
3:12:56
where he would snore more so does that
3:12:56
where he would snore more so does that help my situation with the snoring puppy
3:12:59
help my situation with the snoring puppy
3:13:00
help my situation with the snoring puppy i don't i don't think so so we'll save
3:13:02
i don't i don't think so so we'll save
3:13:02
i don't i don't think so so we'll save that for later
3:13:03
that for later
3:13:03
that for later we will have a chat about this robotic
3:13:05
we will have a chat about this robotic
3:13:05
we will have a chat about this robotic puppy uh later on
3:13:09
ai session is on october 29th
3:13:12
ai session is on october 29th
3:13:12
ai session is on october 29th so we will be answering deep questions
3:13:16
so we will be answering deep questions
3:13:16
so we will be answering deep questions then yes so that's that's a good point
3:13:19
then yes so that's that's a good point
3:13:19
then yes so that's that's a good point actually so we have
3:13:21
actually so we have
3:13:21
actually so we have another episode coming up next week
3:13:22
another episode coming up next week
3:13:22
another episode coming up next week about natural language processing
3:13:25
about natural language processing
3:13:25
about natural language processing uh we've got some awesome researchers
3:13:27
uh we've got some awesome researchers
3:13:27
uh we've got some awesome researchers who are working on tools like spacey
3:13:29
who are working on tools like spacey
3:13:29
who are working on tools like spacey a popular uh python framework for
3:13:33
a popular uh python framework for
3:13:33
a popular uh python framework for analyzing language we've got people
3:13:34
analyzing language we've got people
3:13:34
analyzing language we've got people working on
3:13:36
working on
3:13:36
working on large neural networks that identify text
3:13:39
large neural networks that identify text
3:13:39
large neural networks that identify text and generate text so that that's pretty
3:13:40
and generate text so that that's pretty
3:13:40
and generate text so that that's pretty cool
3:13:41
cool
3:13:41
cool we've got some awesome stuff coming up a
3:13:43
we've got some awesome stuff coming up a
3:13:43
we've got some awesome stuff coming up a deep stuff even
3:13:44
deep stuff even
3:13:44
deep stuff even um i guess some of the sessions today
3:13:47
um i guess some of the sessions today
3:13:47
um i guess some of the sessions today were a little bit more basic
3:13:49
were a little bit more basic
3:13:49
were a little bit more basic but expect more deep stuff uh next week
3:13:51
but expect more deep stuff uh next week
3:13:51
but expect more deep stuff uh next week and a week after that
3:13:53
and a week after that
3:13:53
and a week after that we've got all the a advanced ai
3:13:56
we've got all the a advanced ai
3:13:56
we've got all the a advanced ai stuff like multi-agent systems
3:13:59
stuff like multi-agent systems
3:13:59
stuff like multi-agent systems reinforcement learning
3:14:00
reinforcement learning
3:14:00
reinforcement learning uh those kind of crazy concepts that
3:14:03
uh those kind of crazy concepts that
3:14:03
uh those kind of crazy concepts that that might not be useful in production
3:14:05
that might not be useful in production
3:14:05
that might not be useful in production immediately
3:14:06
immediately
3:14:06
immediately but are really interesting to watch so
3:14:09
but are really interesting to watch so
3:14:09
but are really interesting to watch so thank you very much and for
3:14:10
thank you very much and for
3:14:10
thank you very much and for people who are still wanting to tweet to
3:14:12
people who are still wanting to tweet to
3:14:12
people who are still wanting to tweet to us with their pictures
3:14:14
us with their pictures
3:14:14
us with their pictures for the oculus rift too you have one
3:14:17
for the oculus rift too you have one
3:14:17
for the oculus rift too you have one more hour
3:14:18
more hour
3:14:18
more hour yes show the box
3:14:23
uh so you can win this actually if you
3:14:26
uh so you can win this actually if you
3:14:26
uh so you can win this actually if you send us your best photo of watching us
3:14:28
send us your best photo of watching us
3:14:28
send us your best photo of watching us uh talk live about ai um if you're not a
3:14:32
uh talk live about ai um if you're not a
3:14:32
uh talk live about ai um if you're not a photo person uh your camera shy or
3:14:34
photo person uh your camera shy or
3:14:34
photo person uh your camera shy or camera doesn't work
3:14:36
camera doesn't work
3:14:36
camera doesn't work uh that can happen too um uh feel free
3:14:39
uh that can happen too um uh feel free
3:14:39
uh that can happen too um uh feel free to tweet us your questions
3:14:40
to tweet us your questions
3:14:40
to tweet us your questions um uh you can ask your final question
3:14:43
um uh you can ask your final question
3:14:43
um uh you can ask your final question right now we will hold for a second or
3:14:45
right now we will hold for a second or
3:14:45
right now we will hold for a second or two
3:14:45
two
3:14:45
two um and after that we will choose a
3:14:47
um and after that we will choose a
3:14:47
um and after that we will choose a winner in the next uh a couple of hours
3:14:49
winner in the next uh a couple of hours
3:14:49
winner in the next uh a couple of hours and we will contact you personally if
3:14:51
and we will contact you personally if
3:14:51
and we will contact you personally if you want
3:14:53
you want
3:14:53
you want so thank you very much everyone and uh
3:14:56
so thank you very much everyone and uh
3:14:56
so thank you very much everyone and uh looking forward to next week
3:15:00
thank you see you later be well everyone
3:15:04
thank you see you later be well everyone
3:15:04
thank you see you later be well everyone yeah stay safe
#Computers & Electronics