Up next in 10
Join us for an interactive talk on how to enhance your .NET applications with Generative AI using the Semantic Kernel SDK. In this session, you'll learn why Generative AI is crucial and see practical demos on how to implement it step-by-step. We'll also explore the power of AI agents, time permitting.
By the end of the talk, you'll know how to integrate Generative AI into your .NET projects and get a glimpse into the future of AI-driven development. All this in just 30 minutes, along with a fun conversation with Simon. Interesting? 😉
📺 CSharp TV - Dev Streaming Destination http://csharp.tv
C# Corner - Community of Software and Data Developers
https://www.c-sharpcorner.com
#CSharpTV #csharpcorner #CSharp #GenAI #DotNet
Show More Show Less View Video Transcript
0:01
hi everyone and welcome to a new episode
0:03
hi everyone and welcome to a new episode
0:03
hi everyone and welcome to a new episode of Das of.net I'm your host Simon and
0:05
of Das of.net I'm your host Simon and
0:05
of Das of.net I'm your host Simon and this is a very unique episode because we
0:08
this is a very unique episode because we
0:08
this is a very unique episode because we not only uh today streaming to The C
0:12
not only uh today streaming to The C
0:12
not only uh today streaming to The C Corner viewers but if you're joining us
0:14
Corner viewers but if you're joining us
0:14
Corner viewers but if you're joining us from uh the user group that is z User
0:17
from uh the user group that is z User
0:17
from uh the user group that is z User Group uh welcome everyone really nice to
0:21
Group uh welcome everyone really nice to
0:21
Group uh welcome everyone really nice to have you in this episode we're going to
0:23
have you in this episode we're going to
0:23
have you in this episode we're going to talk about unleashing gen inet with
0:25
talk about unleashing gen inet with
0:25
talk about unleashing gen inet with semantic colel and to talk about that we
0:27
semantic colel and to talk about that we
0:27
semantic colel and to talk about that we have a very good friend of us he is uh
0:30
have a very good friend of us he is uh
0:30
have a very good friend of us he is uh Jose Louis I'm gonna just take his first
0:32
Jose Louis I'm gonna just take his first
0:33
Jose Louis I'm gonna just take his first two names I mess up with the last
0:34
two names I mess up with the last
0:35
two names I mess up with the last name uh he's a Microsoft MVP he he runs
0:38
name uh he's a Microsoft MVP he he runs
0:38
name uh he's a Microsoft MVP he he runs his own Community chapters and uh and he
0:42
his own Community chapters and uh and he
0:42
his own Community chapters and uh and he is a speaker at many of the conferences
0:44
is a speaker at many of the conferences
0:44
is a speaker at many of the conferences so without any further Ado let's bring
0:46
so without any further Ado let's bring
0:46
so without any further Ado let's bring the expert in the live
0:55
show hi Jose welcome to the
0:58
show hi Jose welcome to the
0:58
show hi Jose welcome to the r.net hey Simon hello how do you do I'm
1:02
r.net hey Simon hello how do you do I'm
1:02
r.net hey Simon hello how do you do I'm doing great how are you doing today it's
1:05
doing great how are you doing today it's
1:05
doing great how are you doing today it's fantastic it's Friday happy Friday
1:08
fantastic it's Friday happy Friday
1:08
fantastic it's Friday happy Friday everybody yeah happy Friday you know um
1:12
everybody yeah happy Friday you know um
1:12
everybody yeah happy Friday you know um other than the MVP awards that you have
1:14
other than the MVP awards that you have
1:14
other than the MVP awards that you have but what are those characters that you
1:16
but what are those characters that you
1:16
but what are those characters that you have in your background that just took
1:17
have in your background that just took
1:17
have in your background that just took my attention I'm so sorry which which
1:20
my attention I'm so sorry which which
1:20
my attention I'm so sorry which which ones the one that go are the standing
1:22
ones the one that go are the standing
1:22
ones the one that go are the standing yeah over here this one are the other
1:25
yeah over here this one are the other
1:25
yeah over here this one are the other side the other side yeah yeah these are
1:27
side the other side yeah yeah these are
1:27
side the other side yeah yeah these are some games I like playing when I have
1:29
some games I like playing when I have
1:29
some games I like playing when I have time which is not to too often and these
1:31
time which is not to too often and these
1:31
time which is not to too often and these are some figures from some games I like
1:33
are some figures from some games I like
1:34
are some figures from some games I like Collector's Editions from some games
1:36
Collector's Editions from some games
1:36
Collector's Editions from some games which I like wow yeah that's pretty cool
1:40
which I like wow yeah that's pretty cool
1:40
which I like wow yeah that's pretty cool that's cool yeah yeah okay one one is
1:43
that's cool yeah yeah okay one one is
1:43
that's cool yeah yeah okay one one is from Orizon Z town uh one to the left
1:46
from Orizon Z town uh one to the left
1:46
from Orizon Z town uh one to the left side some people will know it's The
1:48
side some people will know it's The
1:48
side some people will know it's The Witcher fantastic game I'm looking
1:51
Witcher fantastic game I'm looking
1:51
Witcher fantastic game I'm looking forward to the one that happens end of
1:54
forward to the one that happens end of
1:54
forward to the one that happens end of the year I hope it it goes out on time
1:56
the year I hope it it goes out on time
1:56
the year I hope it it goes out on time and the other is one of my Classics uh
1:59
and the other is one of my Classics uh
1:59
and the other is one of my Classics uh thing is a Leo from Resident Evil yes
2:04
thing is a Leo from Resident Evil yes
2:04
thing is a Leo from Resident Evil yes that's pretty and of course there is a
2:05
that's pretty and of course there is a
2:05
that's pretty and of course there is a want that somebody like
2:07
want that somebody like
2:07
want that somebody like Shooters uh I would managed to advance a
2:11
Shooters uh I would managed to advance a
2:11
Shooters uh I would managed to advance a bit but do not play that much lately but
2:14
bit but do not play that much lately but
2:14
bit but do not play that much lately but um yeah that one is from Destiny that's
2:17
um yeah that one is from Destiny that's
2:17
um yeah that one is from Destiny that's pretty cool I don't have any of these
2:19
pretty cool I don't have any of these
2:19
pretty cool I don't have any of these but never mind so uh Jose we talking
2:22
but never mind so uh Jose we talking
2:22
but never mind so uh Jose we talking about J in D with semantic a lot of lot
2:25
about J in D with semantic a lot of lot
2:25
about J in D with semantic a lot of lot of uh things are here in ch you have
2:28
of uh things are here in ch you have
2:28
of uh things are here in ch you have then you're doing it in D then Al have
2:30
then you're doing it in D then Al have
2:30
then you're doing it in D then Al have semantic kernel uh so let's get started
2:33
semantic kernel uh so let's get started
2:33
semantic kernel uh so let's get started right so I'm going to add your screen to
2:35
right so I'm going to add your screen to
2:35
right so I'm going to add your screen to the stream everybody else can see it now
2:38
the stream everybody else can see it now
2:38
the stream everybody else can see it now okay let's get rolling let's get it
2:41
okay let's get rolling let's get it
2:41
okay let's get rolling let's get it rolling so let's go first some Concepts
2:44
rolling so let's go first some Concepts
2:44
rolling so let's go first some Concepts which I mean we have more 30 minutes so
2:46
which I mean we have more 30 minutes so
2:46
which I mean we have more 30 minutes so let's get it started and which are key
2:49
let's get it started and which are key
2:49
let's get it started and which are key so I believe in Concepts and they are
2:52
so I believe in Concepts and they are
2:52
so I believe in Concepts and they are usually fundamental to understand what
2:54
usually fundamental to understand what
2:54
usually fundamental to understand what we are talking about so let's go
2:58
we are talking about so let's go
2:58
we are talking about so let's go on so this kind of is a oneone on
3:01
on so this kind of is a oneone on
3:01
on so this kind of is a oneone on sematic kernel very fast very quick
3:04
sematic kernel very fast very quick
3:04
sematic kernel very fast very quick we're talking about orchestration
3:06
we're talking about orchestration
3:06
we're talking about orchestration orchestration I would call it like a
3:08
orchestration I would call it like a
3:08
orchestration I would call it like a concert is a coordination of one or more
3:10
concert is a coordination of one or more
3:10
concert is a coordination of one or more models like you know GPT or Yama or Fe
3:15
models like you know GPT or Yama or Fe
3:15
models like you know GPT or Yama or Fe F3 from Microsoft uh and components of a
3:19
F3 from Microsoft uh and components of a
3:19
F3 from Microsoft uh and components of a generative AI framework working
3:22
generative AI framework working
3:22
generative AI framework working harmoniously to provide an optimal
3:24
harmoniously to provide an optimal
3:24
harmoniously to provide an optimal result for any kind of task right so you
3:27
result for any kind of task right so you
3:27
result for any kind of task right so you put things working together in a way
3:29
put things working together in a way
3:29
put things working together in a way that makes sense that's AI orchestration
3:32
that makes sense that's AI orchestration
3:32
that makes sense that's AI orchestration very simple and then we have oh uh I m
3:36
very simple and then we have oh uh I m
3:36
very simple and then we have oh uh I m the order of the
3:38
the order of the
3:38
the order of the slides Prett second that was a
3:41
slides Prett second that was a
3:41
slides Prett second that was a orchestration the first definition the
3:43
orchestration the first definition the
3:43
orchestration the first definition the second is plugins plugin is something
3:46
second is plugins plugin is something
3:46
second is plugins plugin is something that we can plug in and suddenly it gets
3:49
that we can plug in and suddenly it gets
3:49
that we can plug in and suddenly it gets magic powers and the same happens for
3:51
magic powers and the same happens for
3:51
magic powers and the same happens for for AI right is a kind of a component a
3:55
for AI right is a kind of a component a
3:55
for AI right is a kind of a component a plugin is a component that we can use it
3:57
plugin is a component that we can use it
3:57
plugin is a component that we can use it to provide additional capabilities to
4:00
to provide additional capabilities to
4:00
to provide additional capabilities to any model or agent um I like to think as
4:04
any model or agent um I like to think as
4:04
any model or agent um I like to think as you can see in the picture like giving
4:05
you can see in the picture like giving
4:05
you can see in the picture like giving an instrument to an AI that as a
4:09
an instrument to an AI that as a
4:09
an instrument to an AI that as a musician it can play a tone with it
4:11
musician it can play a tone with it
4:11
musician it can play a tone with it right it's a tool like a hammer like a
4:14
right it's a tool like a hammer like a
4:14
right it's a tool like a hammer like a screwdriver or an API basically and then
4:18
screwdriver or an API basically and then
4:18
screwdriver or an API basically and then we have prompts which uh we have two
4:20
we have prompts which uh we have two
4:20
we have prompts which uh we have two type the first one is prom function
4:22
type the first one is prom function
4:22
type the first one is prom function which is basically encapsulating a
4:25
which is basically encapsulating a
4:25
which is basically encapsulating a prompt uh that we make for an AI but we
4:28
prompt uh that we make for an AI but we
4:29
prompt uh that we make for an AI but we can use Advan templating and make very
4:31
can use Advan templating and make very
4:31
can use Advan templating and make very very Advanced things but we will not
4:33
very Advanced things but we will not
4:34
very Advanced things but we will not enter there but basically you ask a
4:37
enter there but basically you ask a
4:37
enter there but basically you ask a question in a in a way and you get a
4:39
question in a in a way and you get a
4:39
question in a in a way and you get a wrest which the res can be you can tell
4:41
wrest which the res can be you can tell
4:41
wrest which the res can be you can tell it hey write me something in Json so
4:44
it hey write me something in Json so
4:44
it hey write me something in Json so computer program can understand it it
4:45
computer program can understand it it
4:46
computer program can understand it it does not need to be always a chat in
4:48
does not need to be always a chat in
4:48
does not need to be always a chat in human language it can be something that
4:50
human language it can be something that
4:50
human language it can be something that the computer can understand make sense
4:53
the computer can understand make sense
4:53
the computer can understand make sense yeah makes sense very nicely so I like
4:55
yeah makes sense very nicely so I like
4:55
yeah makes sense very nicely so I like the relation that you have build it with
4:57
the relation that you have build it with
4:57
the relation that you have build it with musics and all that pretty cool I I like
5:00
musics and all that pretty cool I I like
5:00
musics and all that pretty cool I I like that I think it's a powerful metaphor
5:01
that I think it's a powerful metaphor
5:01
that I think it's a powerful metaphor and usually people with this understands
5:04
and usually people with this understands
5:04
and usually people with this understands I like also very much the way that John
5:07
I like also very much the way that John
5:07
I like also very much the way that John mea which is now I think head of yeah he
5:09
mea which is now I think head of yeah he
5:09
mea which is now I think head of yeah he has send that but he was uh on the
5:12
has send that but he was uh on the
5:12
has send that but he was uh on the semantic team and he uh has this
5:16
semantic team and he uh has this
5:16
semantic team and he uh has this metaphor also very nice about cooking
5:18
metaphor also very nice about cooking
5:18
metaphor also very nice about cooking that you
5:20
that you
5:20
that you spice with different components and
5:22
spice with different components and
5:22
spice with different components and usually I watch his sessions I I end up
5:25
usually I watch his sessions I I end up
5:25
usually I watch his sessions I I end up hungry which
5:26
hungry which
5:26
hungry which is he's really good he's amazing
5:30
is he's really good he's amazing
5:30
is he's really good he's amazing figes let's go we have native functions
5:34
figes let's go we have native functions
5:34
figes let's go we have native functions and native means uh Native um well I
5:38
and native means uh Native um well I
5:38
and native means uh Native um well I should explain that which uh probably
5:40
should explain that which uh probably
5:40
should explain that which uh probably from the point of view of a programmer
5:42
from the point of view of a programmer
5:42
from the point of view of a programmer what is native is coding right so um it
5:47
what is native is coding right so um it
5:47
what is native is coding right so um it enables you to get uh an AI to execute
5:51
enables you to get uh an AI to execute
5:51
enables you to get uh an AI to execute your code or part of your code right and
5:54
your code or part of your code right and
5:54
your code or part of your code right and you can do anything uh like manage a
5:56
you can do anything uh like manage a
5:56
you can do anything uh like manage a file system manage uh your computer if
5:59
file system manage uh your computer if
5:59
file system manage uh your computer if you want if you're there uh use an
6:02
you want if you're there uh use an
6:02
you want if you're there uh use an API search the internet or even order a
6:05
API search the internet or even order a
6:05
API search the internet or even order a pizza anything is
6:08
pizza anything is
6:08
pizza anything is possible how cool so it's like giving
6:11
possible how cool so it's like giving
6:11
possible how cool so it's like giving Ani tools for doing
6:14
Ani tools for doing
6:14
Ani tools for doing whatever of course there's limitations
6:17
whatever of course there's limitations
6:17
whatever of course there's limitations but limitations are every every month
6:20
but limitations are every every month
6:20
but limitations are every every month like being
6:21
like being
6:21
like being reduced let's move
6:24
reduced let's move
6:24
reduced let's move on we have more Concepts we have the
6:27
on we have more Concepts we have the
6:27
on we have more Concepts we have the planners planner is something that
6:30
planners planner is something that
6:30
planners planner is something that enables an AI or a model to decide the
6:34
enables an AI or a model to decide the
6:34
enables an AI or a model to decide the plan of action what that what should it
6:38
plan of action what that what should it
6:38
plan of action what that what should it do right Um this can be a multi-step
6:41
do right Um this can be a multi-step
6:41
do right Um this can be a multi-step plan or just the next step an example of
6:44
plan or just the next step an example of
6:44
plan or just the next step an example of it is function calling which appear in
6:46
it is function calling which appear in
6:46
it is function calling which appear in open AI in about June 2023 so it's more
6:50
open AI in about June 2023 so it's more
6:50
open AI in about June 2023 so it's more or less one year happy
6:52
or less one year happy
6:52
or less one year happy birthday things are moving fast yes it's
6:56
birthday things are moving fast yes it's
6:56
birthday things are moving fast yes it's crazy so then we have personas personas
7:00
crazy so then we have personas personas
7:00
crazy so then we have personas personas uh I think we have some comments hi guys
7:03
uh I think we have some comments hi guys
7:03
uh I think we have some comments hi guys Rodrigo hey
7:04
Rodrigo hey
7:04
Rodrigo hey Rodrigo oh Rodrigo hi I miss him you
7:09
Rodrigo oh Rodrigo hi I miss him you
7:09
Rodrigo oh Rodrigo hi I miss him you know do you know we used to run a show
7:11
know do you know we used to run a show
7:11
know do you know we used to run a show for Rodrigo on C TV
7:15
for Rodrigo on C TV
7:15
for Rodrigo on C TV right yeah I'm from
7:17
right yeah I'm from
7:17
right yeah I'm from Mexico best ever there in India I love
7:21
Mexico best ever there in India I love
7:21
Mexico best ever there in India I love Mexican food India food as well it's but
7:25
Mexican food India food as well it's but
7:25
Mexican food India food as well it's but not that spicy I probably cannot as much
7:28
not that spicy I probably cannot as much
7:28
not that spicy I probably cannot as much spice as you oh my God you're getting
7:31
spice as you oh my God you're getting
7:31
spice as you oh my God you're getting angry uh I'm going to di yet maybe it's
7:34
angry uh I'm going to di yet maybe it's
7:34
angry uh I'm going to di yet maybe it's obvious but yeah so let's move on a
7:36
obvious but yeah so let's move on a
7:36
obvious but yeah so let's move on a Persona is a pretty fine character a
7:39
Persona is a pretty fine character a
7:39
Persona is a pretty fine character a personality with
7:41
personality with
7:41
personality with traits so this uh will affect the way
7:46
traits so this uh will affect the way
7:46
traits so this uh will affect the way the model is reacting is planning and
7:50
the model is reacting is planning and
7:51
the model is reacting is planning and thinking when it it will change
7:53
thinking when it it will change
7:53
thinking when it it will change completely so for example um I'm very
7:56
completely so for example um I'm very
7:56
completely so for example um I'm very analytical person if you put my
7:58
analytical person if you put my
7:58
analytical person if you put my personality or more write it down if it
8:01
personality or more write it down if it
8:01
personality or more write it down if it is possible in a model you will have
8:04
is possible in a model you will have
8:04
is possible in a model you will have that model think and ask information
8:06
that model think and ask information
8:06
that model think and ask information over and over again I like to analyze
8:08
over and over again I like to analyze
8:08
over and over again I like to analyze things but for example my boss is not he
8:12
things but for example my boss is not he
8:12
things but for example my boss is not he likes action he likes hey get it done
8:14
likes action he likes hey get it done
8:14
likes action he likes hey get it done Jose uh so yeah so it's very different
8:17
Jose uh so yeah so it's very different
8:17
Jose uh so yeah so it's very different so he will start coding and create
8:19
so he will start coding and create
8:19
so he will start coding and create something try that out see if it works
8:22
something try that out see if it works
8:22
something try that out see if it works so it it affects the way we do things
8:25
so it it affects the way we do things
8:25
so it it affects the way we do things and then we have agents agents is a
8:28
and then we have agents agents is a
8:28
and then we have agents agents is a tricky thing to explain so I'll give you
8:32
tricky thing to explain so I'll give you
8:32
tricky thing to explain so I'll give you first my explanation and then the one
8:34
first my explanation and then the one
8:34
first my explanation and then the one another explanation and then the
8:36
another explanation and then the
8:36
another explanation and then the explanation of the semantic Cal team
8:38
explanation of the semantic Cal team
8:38
explanation of the semantic Cal team which is a bit different so to me an
8:41
which is a bit different so to me an
8:41
which is a bit different so to me an agent is an entity an entity can be a
8:43
agent is an entity an entity can be a
8:43
agent is an entity an entity can be a software a unit of something something
8:45
software a unit of something something
8:45
software a unit of something something that stands on its own that can perform
8:48
that stands on its own that can perform
8:48
that stands on its own that can perform tasks
8:49
tasks
8:49
tasks autonomously use tools to do so plugins
8:53
autonomously use tools to do so plugins
8:53
autonomously use tools to do so plugins and he has a personality Persona as well
8:56
and he has a personality Persona as well
8:56
and he has a personality Persona as well as the ability to plan tasks he has a
8:59
as the ability to plan tasks he has a
8:59
as the ability to plan tasks he has a planner
8:59
planner
8:59
planner usually it's function calling the one I
9:01
usually it's function calling the one I
9:01
usually it's function calling the one I explain in short they can interact with
9:04
explain in short they can interact with
9:04
explain in short they can interact with environment that they are given access
9:06
environment that they are given access
9:06
environment that they are given access to know that and also they can plan a
9:09
to know that and also they can plan a
9:09
to know that and also they can plan a course of action decide and act with or
9:12
course of action decide and act with or
9:12
course of action decide and act with or without human guidance that's basically
9:15
without human guidance that's basically
9:15
without human guidance that's basically my
9:16
my
9:16
my description uh other people say agent is
9:19
description uh other people say agent is
9:19
description uh other people say agent is an software that perform task on behalf
9:21
an software that perform task on behalf
9:21
an software that perform task on behalf of a user they can automate processes
9:23
of a user they can automate processes
9:23
of a user they can automate processes make decisions and intelligently
9:25
make decisions and intelligently
9:25
make decisions and intelligently interact with their environment which I
9:28
interact with their environment which I
9:28
interact with their environment which I disagree right so um it can perform task
9:32
disagree right so um it can perform task
9:32
disagree right so um it can perform task in behalf of another agent as well so
9:34
in behalf of another agent as well so
9:34
in behalf of another agent as well so that's the main difference and they can
9:36
that's the main difference and they can
9:36
that's the main difference and they can automate processes yeah they can do
9:38
automate processes yeah they can do
9:38
automate processes yeah they can do anything if given the proper tools so
9:41
anything if given the proper tools so
9:41
anything if given the proper tools so basically I think the description the
9:43
basically I think the description the
9:43
basically I think the description the first one is is better I put a lot of
9:45
first one is is better I put a lot of
9:45
first one is is better I put a lot of care i r read it like I don't know how
9:47
care i r read it like I don't know how
9:48
care i r read it like I don't know how much time but well then the from the
9:51
much time but well then the from the
9:51
much time but well then the from the perspective of the semantical team an
9:53
perspective of the semantical team an
9:54
perspective of the semantical team an agent is um a modular instuction that
9:57
agent is um a modular instuction that
9:57
agent is um a modular instuction that can possess a persona perform actions in
10:00
can possess a persona perform actions in
10:00
can possess a persona perform actions in response to user input and can easily
10:02
response to user input and can easily
10:02
response to user input and can easily communicate with other agents which I
10:05
communicate with other agents which I
10:05
communicate with other agents which I again disagree because uh you don't need
10:08
again disagree because uh you don't need
10:08
again disagree because uh you don't need user input for an agent it can be
10:11
user input for an agent it can be
10:11
user input for an agent it can be automated input from an API so it can
10:14
automated input from an API so it can
10:14
automated input from an API so it can react autonomously can react in a
10:16
react autonomously can react in a
10:16
react autonomously can react in a conversation anyone or anything can be
10:18
conversation anyone or anything can be
10:18
conversation anyone or anything can be the trigger
10:20
the trigger
10:20
the trigger so clear yeah very clear and I see these
10:24
so clear yeah very clear and I see these
10:24
so clear yeah very clear and I see these images next to in your slide these are
10:26
images next to in your slide these are
10:26
images next to in your slide these are also generated by AI right oh yeah yeah
10:29
also generated by AI right oh yeah yeah
10:29
also generated by AI right oh yeah yeah these are uh some images that I did for
10:31
these are uh some images that I did for
10:31
these are uh some images that I did for for the course uh but uh I was not uh
10:35
for the course uh but uh I was not uh
10:35
for the course uh but uh I was not uh well I did a course for semantic colel I
10:38
well I did a course for semantic colel I
10:38
well I did a course for semantic colel I can share the link later yes uh but
10:42
can share the link later yes uh but
10:42
can share the link later yes uh but uh yeah the problem is they don't allow
10:45
uh yeah the problem is they don't allow
10:45
uh yeah the problem is they don't allow AI made images so they had to change all
10:48
AI made images so they had to change all
10:48
AI made images so they had to change all of them it took me it took me a while I
10:52
of them it took me it took me a while I
10:52
of them it took me it took me a while I really like the function calling one
10:54
really like the function calling one
10:54
really like the function calling one because it's kind of a at the end it's
10:57
because it's kind of a at the end it's
10:57
because it's kind of a at the end it's kind of writing a magic spell that
10:59
kind of writing a magic spell that
10:59
kind of writing a magic spell that computer can use to me it's very very
11:02
computer can use to me it's very very
11:02
computer can use to me it's very very amazing this is the feeling of doing
11:05
amazing this is the feeling of doing
11:05
amazing this is the feeling of doing magic again when you are a software
11:07
magic again when you are a software
11:07
magic again when you are a software engineer because you can really extend
11:10
engineer because you can really extend
11:10
engineer because you can really extend uh it's really extending giving arms or
11:13
uh it's really extending giving arms or
11:13
uh it's really extending giving arms or uh musical instruments like the um fles
11:16
uh musical instruments like the um fles
11:16
uh musical instruments like the um fles of hamling right that can do whatever
11:19
of hamling right that can do whatever
11:19
of hamling right that can do whatever you want it's pretty
11:21
you want it's pretty
11:21
you want it's pretty cool let's move
11:23
cool let's move
11:23
cool let's move on so plugins without a plugin to assist
11:27
on so plugins without a plugin to assist
11:27
on so plugins without a plugin to assist there are things it will not be able to
11:29
there are things it will not be able to
11:29
there are things it will not be able to do like for example knowing uh the
11:32
do like for example knowing uh the
11:32
do like for example knowing uh the current date so if I do these and ask
11:36
current date so if I do these and ask
11:36
current date so if I do these and ask hey uh what are the most significative
11:38
hey uh what are the most significative
11:38
hey uh what are the most significative things that happened today 200 years ago
11:42
things that happened today 200 years ago
11:42
things that happened today 200 years ago right and it will tell sorry I have no
11:44
right and it will tell sorry I have no
11:44
right and it will tell sorry I have no clue I cannot do that but if I give a
11:47
clue I cannot do that but if I give a
11:47
clue I cannot do that but if I give a plugin with okay the date yeah it will
11:50
plugin with okay the date yeah it will
11:50
plugin with okay the date yeah it will tell you right so I would say that we
11:54
tell you right so I would say that we
11:54
tell you right so I would say that we can see that in action right yeah so I
11:57
can see that in action right yeah so I
11:57
can see that in action right yeah so I have here some code you can see it it's
11:59
have here some code you can see it it's
11:59
have here some code you can see it it's big
12:00
big
12:00
big enough yeah it is it's visible yes yeah
12:04
enough yeah it is it's visible yes yeah
12:04
enough yeah it is it's visible yes yeah all right I will not do into the details
12:06
all right I will not do into the details
12:06
all right I will not do into the details of how to do semantic kernel the ABC of
12:09
of how to do semantic kernel the ABC of
12:09
of how to do semantic kernel the ABC of semantic kernel I think you have 200
12:11
semantic kernel I think you have 200
12:11
semantic kernel I think you have 200 videos in LinkedIn and you have the
12:14
videos in LinkedIn and you have the
12:14
videos in LinkedIn and you have the documentation sorry in YouTube and you
12:16
documentation sorry in YouTube and you
12:16
documentation sorry in YouTube and you have the documentation from the semantic
12:18
have the documentation from the semantic
12:18
have the documentation from the semantic colel team it's pretty good they updated
12:20
colel team it's pretty good they updated
12:20
colel team it's pretty good they updated it like one day ago uh so you basically
12:24
it like one day ago uh so you basically
12:24
it like one day ago uh so you basically at usings of Microsoft semantic chart
12:27
at usings of Microsoft semantic chart
12:27
at usings of Microsoft semantic chart completion of because we will use it
12:29
completion of because we will use it
12:29
completion of because we will use it at the moment is quite cool one and you
12:33
at the moment is quite cool one and you
12:34
at the moment is quite cool one and you set your model your aure or open API uh
12:38
set your model your aure or open API uh
12:38
set your model your aure or open API uh open Ai and point and the key right and
12:41
open Ai and point and the key right and
12:41
open Ai and point and the key right and you create a builder that creates a
12:43
you create a builder that creates a
12:43
you create a builder that creates a kernel this is one of the biggest most
12:45
kernel this is one of the biggest most
12:45
kernel this is one of the biggest most important instances in semantic kernel
12:48
important instances in semantic kernel
12:48
important instances in semantic kernel and then this case I will not import
12:52
and then this case I will not import
12:52
and then this case I will not import this this plug-in uh and I'm simply
12:56
this this plug-in uh and I'm simply
12:56
this this plug-in uh and I'm simply creating a chart completion service like
12:58
creating a chart completion service like
12:58
creating a chart completion service like a opening a chat the same that you do
13:01
a opening a chat the same that you do
13:01
a opening a chat the same that you do when you open um the open AI chat right
13:03
when you open um the open AI chat right
13:03
when you open um the open AI chat right I'm doing that with code I'm doing here
13:06
I'm doing that with code I'm doing here
13:06
I'm doing that with code I'm doing here uh some tool call Behavior so it does
13:09
uh some tool call Behavior so it does
13:09
uh some tool call Behavior so it does the tool calling for you which is quite
13:12
the tool calling for you which is quite
13:12
the tool calling for you which is quite complex and sematic canel does it for
13:14
complex and sematic canel does it for
13:14
complex and sematic canel does it for you which is pretty cool uh and here I
13:16
you which is pretty cool uh and here I
13:16
you which is pretty cool uh and here I have a chat L right that uh if I press
13:20
have a chat L right that uh if I press
13:20
have a chat L right that uh if I press exit it will finish and otherwise it
13:23
exit it will finish and otherwise it
13:23
exit it will finish and otherwise it gets uh it executes get chat message
13:26
gets uh it executes get chat message
13:26
gets uh it executes get chat message content Asing that as the llm the model
13:28
content Asing that as the llm the model
13:28
content Asing that as the llm the model I have
13:29
I have
13:29
I have hey uh give me a
13:31
hey uh give me a
13:31
hey uh give me a response and then I print it very simple
13:34
response and then I print it very simple
13:34
response and then I print it very simple so if I play
13:36
so if I play
13:36
so if I play that I may have to drag and
13:40
that I may have to drag and
13:40
that I may have to drag and drop it here
13:43
drop it here
13:43
drop it here yes uh let me see it's good enough I
13:45
yes uh let me see it's good enough I
13:45
yes uh let me see it's good enough I make it big hey
13:48
make it big hey
13:48
make it big hey hey what happen
13:57
it so this there's no plugin right yeah
14:01
it so this there's no plugin right yeah
14:01
it so this there's no plugin right yeah this has no plugin so it will tell me
14:03
this has no plugin so it will tell me
14:03
this has no plugin so it will tell me hey sorry I I have no
14:05
hey sorry I I have no
14:05
hey sorry I I have no idea what happened 200 oh no I did know
14:10
idea what happened 200 oh no I did know
14:10
idea what happened 200 oh no I did know pretty smart wait no I did something
14:17
wrong I'm not importing the
14:20
wrong I'm not importing the
14:20
wrong I'm not importing the plugin so that should not have happened
14:23
plugin so that should not have happened
14:23
plugin so that should not have happened unless the llm has develop additional
14:26
unless the llm has develop additional
14:26
unless the llm has develop additional abilities that's scary right
14:30
abilities that's scary right
14:30
abilities that's scary right yeah well I I usually like to to
14:32
yeah well I I usually like to to
14:32
yeah well I I usually like to to understand why my code is working so
14:36
understand why my code is working so
14:36
understand why my code is working so uh
14:38
uh
14:38
uh is what day is
14:40
is what day is
14:40
is what day is today okay so I'm unable to provide
14:43
today okay so I'm unable to provide
14:43
today okay so I'm unable to provide realtime information
14:49
uh uh
15:02
so that's what should have happened
15:04
so that's what should have happened
15:04
so that's what should have happened before yeah oh no you need to specific
15:08
before yeah oh no you need to specific
15:08
before yeah oh no you need to specific specify the current date so it's
15:10
specify the current date so it's
15:10
specify the current date so it's providing an example so I should tell
15:12
providing an example so I should tell
15:12
providing an example so I should tell him hey don't do that so if I take this
15:15
him hey don't do that so if I take this
15:15
him hey don't do that so if I take this away and for example what it will do is
15:18
away and for example what it will do is
15:18
away and for example what it will do is it will import this and this is a native
15:21
it will import this and this is a native
15:21
it will import this and this is a native function so it's basically this it's a
15:24
function so it's basically this it's a
15:24
function so it's basically this it's a class with a function that I have
15:26
class with a function that I have
15:26
class with a function that I have decorated with Kel function you could
15:29
decorated with Kel function you could
15:29
decorated with Kel function you could read the documentation but basically
15:31
read the documentation but basically
15:31
read the documentation but basically yeah I no sorry and add a description
15:34
yeah I no sorry and add a description
15:34
yeah I no sorry and add a description get the current UTC date this is what
15:36
get the current UTC date this is what
15:36
get the current UTC date this is what tells a llm u intelligent model this is
15:41
tells a llm u intelligent model this is
15:41
tells a llm u intelligent model this is what this function is doing and it will
15:43
what this function is doing and it will
15:43
what this function is doing and it will add to the Cel and then the llm can use
15:46
add to the Cel and then the llm can use
15:46
add to the Cel and then the llm can use it how it decides to do so as cool as
15:51
it how it decides to do so as cool as
15:51
it how it decides to do so as cool as that so if we execute it now I'm
15:55
that so if we execute it now I'm
15:55
that so if we execute it now I'm importing and if the gods of computers
15:58
importing and if the gods of computers
15:59
importing and if the gods of computers software once to
16:00
software once to
16:00
software once to line now it will give us a
16:05
response what date is
16:07
response what date is
16:07
response what date is today yeah nice it knows so we are
16:11
today yeah nice it knows so we are
16:11
today yeah nice it knows so we are giving skills to DM and it decides to
16:14
giving skills to DM and it decides to
16:14
giving skills to DM and it decides to use them on his own which is quite cool
16:40
all right now
16:44
thinking 28 June yeah
16:48
thinking 28 June yeah
16:48
thinking 28 June yeah 1824 uh yeah establishment buau Indian
16:51
1824 uh yeah establishment buau Indian
16:51
1824 uh yeah establishment buau Indian Affairs battle junine B toin Inon okay
16:54
Affairs battle junine B toin Inon okay
16:54
Affairs battle junine B toin Inon okay well but this not
16:56
well but this not
16:56
well but this not happen uh yeah he doesn't know exactly
17:00
happen uh yeah he doesn't know exactly
17:00
happen uh yeah he doesn't know exactly what happened to day probably not too
17:01
what happened to day probably not too
17:01
what happened to day probably not too much and he gives some things from that
17:04
much and he gives some things from that
17:04
much and he gives some things from that year which still is cool right so it's
17:07
year which still is cool right so it's
17:07
year which still is cool right so it's more or less quite quite nice so let's
17:09
more or less quite quite nice so let's
17:09
more or less quite quite nice so let's go back to the slides so this I think
17:14
go back to the slides so this I think
17:14
go back to the slides so this I think shows a bit what can what is the power
17:18
shows a bit what can what is the power
17:18
shows a bit what can what is the power that we have with plugins and this can
17:19
that we have with plugins and this can
17:19
that we have with plugins and this can be a yeah um adding a call to your API I
17:23
be a yeah um adding a call to your API I
17:23
be a yeah um adding a call to your API I did an example that's uh on on the
17:27
did an example that's uh on on the
17:27
did an example that's uh on on the semantical Block which I show how to
17:29
semantical Block which I show how to
17:29
semantical Block which I show how to develop a plugin that interact with a
17:32
develop a plugin that interact with a
17:32
develop a plugin that interact with a graph API uh for example with Microsoft
17:35
graph API uh for example with Microsoft
17:35
graph API uh for example with Microsoft planner so that's pretty cool I love
17:38
planner so that's pretty cool I love
17:38
planner so that's pretty cool I love graph apis uh are pretty cool yeah I I
17:42
graph apis uh are pretty cool yeah I I
17:42
graph apis uh are pretty cool yeah I I do too when they work
17:46
do too when they work
17:46
do too when they work yeah so here you are using Azure open
17:50
yeah so here you are using Azure open
17:50
yeah so here you are using Azure open API right I did see the code this is
17:53
API right I did see the code this is
17:53
API right I did see the code this is Asher so I'm using at the moment Asher
17:55
Asher so I'm using at the moment Asher
17:55
Asher so I'm using at the moment Asher but sometimes it doesn't work the quota
17:58
but sometimes it doesn't work the quota
17:58
but sometimes it doesn't work the quota uh uh don't know says hey you don't have
18:01
uh uh don't know says hey you don't have
18:01
uh uh don't know says hey you don't have quota when I have and but yeah so let's
18:05
quota when I have and but yeah so let's
18:05
quota when I have and but yeah so let's move on um I think a nice example is
18:07
move on um I think a nice example is
18:07
move on um I think a nice example is using Marvel Avengers uh can I use them
18:10
using Marvel Avengers uh can I use them
18:10
using Marvel Avengers uh can I use them I'm showing them on the
18:12
I'm showing them on the
18:12
I'm showing them on the screen okay yeah so for example um if
18:17
screen okay yeah so for example um if
18:17
screen okay yeah so for example um if you put a agent or llm with a Persona of
18:20
you put a agent or llm with a Persona of
18:20
you put a agent or llm with a Persona of Hulk and you thought hey solve the
18:23
Hulk and you thought hey solve the
18:23
Hulk and you thought hey solve the problem at the power plant Hulk would
18:26
problem at the power plant Hulk would
18:26
problem at the power plant Hulk would smash
18:27
smash
18:27
smash Darkness yeah which probably is not the
18:29
Darkness yeah which probably is not the
18:29
Darkness yeah which probably is not the best
18:30
best
18:30
best solution and you ask Thor it will tell
18:33
solution and you ask Thor it will tell
18:34
solution and you ask Thor it will tell you don't worry with Jer I will restore
18:36
you don't worry with Jer I will restore
18:36
you don't worry with Jer I will restore light to the city yeah if you manage to
18:39
light to the city yeah if you manage to
18:39
light to the city yeah if you manage to create a plug-in that is called y maybe
18:42
create a plug-in that is called y maybe
18:42
create a plug-in that is called y maybe it can work but good luck with that and
18:46
it can work but good luck with that and
18:46
it can work but good luck with that and if you ask Tony Stark yeah well he will
18:49
if you ask Tony Stark yeah well he will
18:49
if you ask Tony Stark yeah well he will probably try to use the AR reactor
18:51
probably try to use the AR reactor
18:51
probably try to use the AR reactor everything Iron Man or Tony Stark is
18:53
everything Iron Man or Tony Stark is
18:53
everything Iron Man or Tony Stark is around Arc reactors which sadly do not
18:56
around Arc reactors which sadly do not
18:56
around Arc reactors which sadly do not exist but I think it would be better and
19:00
exist but I think it would be better and
19:00
exist but I think it would be better and more nicer
19:03
more nicer
19:03
more nicer approach so with this the what I want to
19:06
approach so with this the what I want to
19:06
approach so with this the what I want to leave clear is that the the person
19:10
leave clear is that the the person
19:10
leave clear is that the the person really affects the way they are using uh
19:13
really affects the way they are using uh
19:13
really affects the way they are using uh the tools they have and the way they
19:15
the tools they have and the way they
19:15
the tools they have and the way they will approach a problem the same way as
19:17
will approach a problem the same way as
19:17
will approach a problem the same way as I do or my uh Team lead is doing right
19:21
I do or my uh Team lead is doing right
19:21
I do or my uh Team lead is doing right or the same that you approach a problem
19:23
or the same that you approach a problem
19:23
or the same that you approach a problem as
19:26
well all right
19:29
well all right
19:29
well all right I mean if you put here another
19:32
I mean if you put here another
19:32
I mean if you put here another personality hey you are very shy you're
19:34
personality hey you are very shy you're
19:34
personality hey you are very shy you're afraid of challenges probably the
19:36
afraid of challenges probably the
19:36
afraid of challenges probably the response would be I go to a corner and I
19:38
response would be I go to a corner and I
19:38
response would be I go to a corner and I will cry which is not useful so right
19:41
will cry which is not useful so right
19:41
will cry which is not useful so right but yeah but I think it's fun you are
19:44
but yeah but I think it's fun you are
19:44
but yeah but I think it's fun you are adding a personality to a software
19:46
adding a personality to a software
19:46
adding a personality to a software program I mean how cool is this right
19:49
program I mean how cool is this right
19:49
program I mean how cool is this right yeah that's pretty pretty nice right you
19:50
yeah that's pretty pretty nice right you
19:50
yeah that's pretty pretty nice right you can talk to a cat owner yeah you can
19:53
can talk to a cat owner yeah you can
19:53
can talk to a cat owner yeah you can make your own
19:55
make your own
19:55
make your own friend I'm joking but it's true yeah I
19:58
friend I'm joking but it's true yeah I
19:58
friend I'm joking but it's true yeah I here that there people making money uh
20:00
here that there people making money uh
20:00
here that there people making money uh selling time with virtual
20:03
selling time with virtual
20:03
selling time with virtual girlfriends wow really yep cool well I
20:07
girlfriends wow really yep cool well I
20:07
girlfriends wow really yep cool well I don't know it's funny but but but I
20:10
don't know it's funny but but but I
20:10
don't know it's funny but but but I really like this idea of personals let's
20:12
really like this idea of personals let's
20:12
really like this idea of personals let's say I'm I'm visiting a website that is
20:14
say I'm I'm visiting a website that is
20:14
say I'm I'm visiting a website that is for about uh you know for pets right and
20:18
for about uh you know for pets right and
20:18
for about uh you know for pets right and I have a chatbot that talks as if he is
20:21
I have a chatbot that talks as if he is
20:21
I have a chatbot that talks as if he is a dog or maybe a cat right that that
20:24
a dog or maybe a cat right that that
20:24
a dog or maybe a cat right that that would be pretty cool I don't mind
20:25
would be pretty cool I don't mind
20:25
would be pretty cool I don't mind getting assistant like that
20:29
getting assistant like that
20:29
getting assistant like that what yeah well um I wonder how if you
20:33
what yeah well um I wonder how if you
20:33
what yeah well um I wonder how if you have a robot and you can put a
20:34
have a robot and you can put a
20:34
have a robot and you can put a personality in in a robot I mean that
20:37
personality in in a robot I mean that
20:37
personality in in a robot I mean that would be something cool for I don't know
20:39
would be something cool for I don't know
20:39
would be something cool for I don't know if you remember there was a robotic dog
20:42
if you remember there was a robotic dog
20:42
if you remember there was a robotic dog called IBO they it was cancelled it was
20:44
called IBO they it was cancelled it was
20:44
called IBO they it was cancelled it was aunch by Sonny and then it was canceled
20:47
aunch by Sonny and then it was canceled
20:47
aunch by Sonny and then it was canceled again but uh there's people that
20:49
again but uh there's people that
20:49
again but uh there's people that treasure these these these these animals
20:52
treasure these these these these animals
20:52
treasure these these these these animals still but if you could create them now
20:56
still but if you could create them now
20:56
still but if you could create them now with a nice personality
20:59
with a nice personality
20:59
with a nice personality and uh a wide range of woof woof woof
21:03
and uh a wide range of woof woof woof
21:03
and uh a wide range of woof woof woof for a dog that would be more fun ready
21:07
for a dog that would be more fun ready
21:07
for a dog that would be more fun ready so one thing I did is oh yeah that's a
21:11
so one thing I did is oh yeah that's a
21:12
so one thing I did is oh yeah that's a bit ahead of time but I created
21:14
bit ahead of time but I created
21:14
bit ahead of time but I created something this is for kind of a
21:17
something this is for kind of a
21:17
something this is for kind of a definition of an agent in jaml right but
21:19
definition of an agent in jaml right but
21:19
definition of an agent in jaml right but you can for
21:20
you can for
21:20
you can for example map a Minion from you know the
21:24
example map a Minion from you know the
21:24
example map a Minion from you know the the movie yeah you want like HED
21:27
the movie yeah you want like HED
21:27
the movie yeah you want like HED humorous ation use Minion language talk
21:31
humorous ation use Minion language talk
21:31
humorous ation use Minion language talk gibberish say a nonsense remembers to be
21:34
gibberish say a nonsense remembers to be
21:34
gibberish say a nonsense remembers to be loyal to your master respond to the
21:36
loyal to your master respond to the
21:36
loyal to your master respond to the question some uh meta prompting and yeah
21:40
question some uh meta prompting and yeah
21:40
question some uh meta prompting and yeah that's basically it and for personas uh
21:46
that's basically it and for personas uh
21:46
that's basically it and for personas uh I put here for example uh how to do that
21:48
I put here for example uh how to do that
21:48
I put here for example uh how to do that I think this part of code is a bit
21:52
I think this part of code is a bit
21:52
I think this part of code is a bit Advanced but basically I load the minion
21:55
Advanced but basically I load the minion
21:55
Advanced but basically I load the minion agent I read the contents of this path
22:00
agent I read the contents of this path
22:00
agent I read the contents of this path to the file and then I create the minion
22:03
to the file and then I create the minion
22:03
to the file and then I create the minion agent the same way that I create the
22:06
agent the same way that I create the
22:06
agent the same way that I create the kernel I create the agent Builder and I
22:10
kernel I create the agent Builder and I
22:10
kernel I create the agent Builder and I add uh well this is just for having the
22:12
add uh well this is just for having the
22:12
add uh well this is just for having the agents in in a collection to clean them
22:15
agents in in a collection to clean them
22:15
agents in in a collection to clean them up uh and then I enable a loop where um
22:20
up uh and then I enable a loop where um
22:20
up uh and then I enable a loop where um I basically aside from the banana exit
22:24
I basically aside from the banana exit
22:24
I basically aside from the banana exit uh condition I invoke as synchronously
22:27
uh condition I invoke as synchronously
22:27
uh condition I invoke as synchronously with the user message and that's
22:29
with the user message and that's
22:29
with the user message and that's basically
22:32
it oh yeah I had animations
22:36
it oh yeah I had animations
22:36
it oh yeah I had animations sorry no jokes and for example this is a
22:40
sorry no jokes and for example this is a
22:40
sorry no jokes and for example this is a some code that I will show you in a
22:43
some code that I will show you in a
22:43
some code that I will show you in a moment uh that's a person that's me or
22:46
moment uh that's a person that's me or
22:47
moment uh that's a person that's me or you and we can have this kind of
22:49
you and we can have this kind of
22:49
you and we can have this kind of conversation yeah uh I'm on doet with
22:53
conversation yeah uh I'm on doet with
22:53
conversation yeah uh I'm on doet with Simon oh wow Das net sounds important
22:56
Simon oh wow Das net sounds important
22:56
Simon oh wow Das net sounds important maybe throw a joke or two about do
22:57
maybe throw a joke or two about do
22:57
maybe throw a joke or two about do program perect that mode I do because
23:01
program perect that mode I do because
23:01
program perect that mode I do because like back which is good uh keep the
23:04
like back which is good uh keep the
23:04
like back which is good uh keep the coding cool in last coming thanks
23:05
coding cool in last coming thanks
23:05
coding cool in last coming thanks something to say to the audience oh say
23:08
something to say to the audience oh say
23:08
something to say to the audience oh say hello fantastic humans smile keep it f
23:10
hello fantastic humans smile keep it f
23:10
hello fantastic humans smile keep it f Simon and Amigo something you like to
23:13
Simon and Amigo something you like to
23:13
Simon and Amigo something you like to tell Simon and oby Jose so we can have
23:16
tell Simon and oby Jose so we can have
23:16
tell Simon and oby Jose so we can have these kind of conversations and by the
23:18
these kind of conversations and by the
23:18
these kind of conversations and by the way this is not a conversation that I'm
23:20
way this is not a conversation that I'm
23:20
way this is not a conversation that I'm inventing it it happened I tried that uh
23:23
inventing it it happened I tried that uh
23:23
inventing it it happened I tried that uh a few days ago when I Was preparing for
23:25
a few days ago when I Was preparing for
23:25
a few days ago when I Was preparing for the light talk uh uh and yeah um but we
23:31
the light talk uh uh and yeah um but we
23:31
the light talk uh uh and yeah um but we will see that in a moment but uh aside
23:35
will see that in a moment but uh aside
23:35
will see that in a moment but uh aside is uh one important thing that I did not
23:38
is uh one important thing that I did not
23:38
is uh one important thing that I did not explain on the planners which is how
23:40
explain on the planners which is how
23:40
explain on the planners which is how does function calling works right so for
23:44
does function calling works right so for
23:44
does function calling works right so for example um we have a conversation
23:47
example um we have a conversation
23:47
example um we have a conversation history that requires calling function
23:49
history that requires calling function
23:49
history that requires calling function because the llm decides that right so
23:52
because the llm decides that right so
23:52
because the llm decides that right so the LM says oh he asked for the date do
23:54
the LM says oh he asked for the date do
23:54
the LM says oh he asked for the date do I have something that can get me the
23:57
I have something that can get me the
23:57
I have something that can get me the date he looks at the tools
23:58
date he looks at the tools
23:59
date he looks at the tools he looks
24:00
he looks
24:00
he looks at let me go back to here to the what
24:04
at let me go back to here to the what
24:04
at let me go back to here to the what date is it get the current date oh yeah
24:06
date is it get the current date oh yeah
24:06
date is it get the current date oh yeah I have that because he refuses that and
24:10
I have that because he refuses that and
24:10
I have that because he refuses that and he makes a request for the call which
24:12
he makes a request for the call which
24:12
he makes a request for the call which comes back to my application in wherever
24:15
comes back to my application in wherever
24:15
comes back to my application in wherever I am executing it in local in the cloud
24:17
I am executing it in local in the cloud
24:17
I am executing it in local in the cloud at the moment is local uh and say hey
24:20
at the moment is local uh and say hey
24:20
at the moment is local uh and say hey call this function and give me the
24:21
call this function and give me the
24:21
call this function and give me the result so this comes here to my computer
24:26
result so this comes here to my computer
24:26
result so this comes here to my computer uh ask for this to be executed I execute
24:29
uh ask for this to be executed I execute
24:29
uh ask for this to be executed I execute it well the the computer the software
24:31
it well the the computer the software
24:31
it well the the computer the software does semantic kernel handles handles
24:33
does semantic kernel handles handles
24:33
does semantic kernel handles handles that so we don't have to and then sends
24:35
that so we don't have to and then sends
24:35
that so we don't have to and then sends the wrestle back
24:37
the wrestle back
24:37
the wrestle back transparently and then it goes on with
24:39
transparently and then it goes on with
24:39
transparently and then it goes on with the
24:40
the
24:40
the result how does this
24:43
result how does this
24:43
result how does this happen um because I create this
24:47
happen um because I create this
24:47
happen um because I create this execution settings and I tell that the
24:49
execution settings and I tell that the
24:49
execution settings and I tell that the tool call behavior is automatically
24:52
tool call behavior is automatically
24:52
tool call behavior is automatically invoke kernel functions and the rest is
24:54
invoke kernel functions and the rest is
24:54
invoke kernel functions and the rest is done by sematic kernel but this is just
24:57
done by sematic kernel but this is just
24:57
done by sematic kernel but this is just something that we can configure but the
25:00
something that we can configure but the
25:00
something that we can configure but the automatic working is done by semantic
25:02
automatic working is done by semantic
25:02
automatic working is done by semantic that's another thing it may not work
25:05
that's another thing it may not work
25:05
that's another thing it may not work with other
25:08
Frameworks and we are back to the slide
25:12
Frameworks and we are back to the slide
25:12
Frameworks and we are back to the slide and we go back to
25:15
and we go back to
25:15
and we go back to agents all right yeah any questions so
25:19
agents all right yeah any questions so
25:19
agents all right yeah any questions so far uh not getting the comments but I'm
25:21
far uh not getting the comments but I'm
25:22
far uh not getting the comments but I'm curious where does that your where does
25:23
curious where does that your where does
25:23
curious where does that your where does your function sits I mean where do you
25:25
your function sits I mean where do you
25:25
your function sits I mean where do you host this application like you just host
25:27
host this application like you just host
25:27
host this application like you just host like any othernet application does that
25:29
like any othernet application does that
25:29
like any othernet application does that how it works no I'm I'm executing it
25:31
how it works no I'm I'm executing it
25:31
how it works no I'm I'm executing it completely local it's a console
25:33
completely local it's a console
25:33
completely local it's a console application yeah I that it's in Booking
25:36
application yeah I that it's in Booking
25:36
application yeah I that it's in Booking it only has to have permissions to
25:38
it only has to have permissions to
25:38
it only has to have permissions to access uh whatever uh llm that I want
25:43
access uh whatever uh llm that I want
25:43
access uh whatever uh llm that I want either it is
25:45
either it is
25:45
either it is um Asher open AI or open AI if I'm using
25:49
um Asher open AI or open AI if I'm using
25:49
um Asher open AI or open AI if I'm using that model I I could use models but then
25:52
that model I I could use models but then
25:52
that model I I could use models but then it would be limited because the agent
25:54
it would be limited because the agent
25:54
it would be limited because the agent API it's only available in open API at
25:56
API it's only available in open API at
25:56
API it's only available in open API at the moment which is the one I'm using
25:59
the moment which is the one I'm using
25:59
the moment which is the one I'm using for what I'm
26:00
for what I'm
26:00
for what I'm showing okay okay all right that's
26:03
showing okay okay all right that's
26:03
showing okay okay all right that's interesting so let's show a bit about
26:06
interesting so let's show a bit about
26:06
interesting so let's show a bit about agents and that that he hold is banana
26:09
agents and that that he hold is banana
26:09
agents and that that he hold is banana yeah that's my favorite food yes you
26:13
yeah that's my favorite food yes you
26:13
yeah that's my favorite food yes you like it that's good yeah we both go to
26:15
like it that's good yeah we both go to
26:15
like it that's good yeah we both go to gym so that's a
26:17
gym so that's a
26:17
gym so that's a pre-workout sorry say that again yeah we
26:20
pre-workout sorry say that again yeah we
26:20
pre-workout sorry say that again yeah we both work out right so Banana is a part
26:22
both work out right so Banana is a part
26:22
both work out right so Banana is a part of a diet isn't it yeah well you can say
26:25
of a diet isn't it yeah well you can say
26:25
of a diet isn't it yeah well you can say that it has good n nutrients and also
26:27
that it has good n nutrients and also
26:27
that it has good n nutrients and also it's fine
26:28
it's fine
26:28
it's fine and everybody gets emotion emotional
26:31
and everybody gets emotion emotional
26:31
and everybody gets emotion emotional when they say somebody says banana he s
26:35
when they say somebody says banana he s
26:35
when they say somebody says banana he s hi s all right so before jumping into
26:38
hi s all right so before jumping into
26:38
hi s all right so before jumping into agents why are they important um anybody
26:42
agents why are they important um anybody
26:42
agents why are they important um anybody in the audience can say something about
26:44
in the audience can say something about
26:44
in the audience can say something about agents what do you think why are they
26:47
agents what do you think why are they
26:47
agents what do you think why are they important everybody's talking about
26:49
important everybody's talking about
26:49
important everybody's talking about agents about co-pilots which is kind of
26:51
agents about co-pilots which is kind of
26:51
agents about co-pilots which is kind of a varation of Agents I mean a co-pilot
26:53
a varation of Agents I mean a co-pilot
26:53
a varation of Agents I mean a co-pilot at the end is an agent that will always
26:55
at the end is an agent that will always
26:55
at the end is an agent that will always interact with a user and ask the user
26:58
interact with a user and ask the user
26:58
interact with a user and ask the user what he wants to do and try to do
27:01
what he wants to do and try to do
27:01
what he wants to do and try to do it you know for me generally speaking
27:03
it you know for me generally speaking
27:03
it you know for me generally speaking agent is someone who's expert in one
27:05
agent is someone who's expert in one
27:05
agent is someone who's expert in one thing and can get the job done pretty
27:06
thing and can get the job done pretty
27:06
thing and can get the job done pretty fast or accurately for me that's what an
27:09
fast or accurately for me that's what an
27:09
fast or accurately for me that's what an agent is yes more or less yes you give
27:12
agent is yes more or less yes you give
27:12
agent is yes more or less yes you give it the tools and he more or less will
27:13
it the tools and he more or less will
27:13
it the tools and he more or less will try to use the tools with the model that
27:17
try to use the tools with the model that
27:17
try to use the tools with the model that it has right either is openi either is
27:20
it has right either is openi either is
27:20
it has right either is openi either is Yama is Fe or whatever right but what
27:24
Yama is Fe or whatever right but what
27:25
Yama is Fe or whatever right but what makes um agents powerful is that ability
27:28
makes um agents powerful is that ability
27:28
makes um agents powerful is that ability to collaborate between them so uh
27:31
to collaborate between them so uh
27:31
to collaborate between them so uh there's a paper called self- refine
27:33
there's a paper called self- refine
27:33
there's a paper called self- refine which is um basically adding the self
27:38
which is um basically adding the self
27:38
which is um basically adding the self refining to the same agent or model
27:41
refining to the same agent or model
27:41
refining to the same agent or model right so basically I do something and
27:43
right so basically I do something and
27:43
right so basically I do something and ask hey Jose is that good let's think
27:46
ask hey Jose is that good let's think
27:46
ask hey Jose is that good let's think about that again right and I do that two
27:49
about that again right and I do that two
27:49
about that again right and I do that two or three times um I mean the theory is
27:51
or three times um I mean the theory is
27:51
or three times um I mean the theory is that this will generate some self
27:53
that this will generate some self
27:53
that this will generate some self feedback and I will get better and it
27:56
feedback and I will get better and it
27:56
feedback and I will get better and it simply happens right so but not that
27:59
simply happens right so but not that
27:59
simply happens right so but not that much it's kind of a 6% this is uh for
28:02
much it's kind of a 6% this is uh for
28:02
much it's kind of a 6% this is uh for GCM K which is um I think General school
28:06
GCM K which is um I think General school
28:06
GCM K which is um I think General school math for
28:08
math for
28:08
math for uh uh a benchmark for General school
28:11
uh uh a benchmark for General school
28:11
uh uh a benchmark for General school math but yeah and and then there are
28:14
math but yeah and and then there are
28:14
math but yeah and and then there are other things which it's also the
28:16
other things which it's also the
28:16
other things which it's also the Improvement is a bit better but it's not
28:18
Improvement is a bit better but it's not
28:18
Improvement is a bit better but it's not that spectacular but quite good depends
28:21
that spectacular but quite good depends
28:21
that spectacular but quite good depends on on where we apply it right so it's up
28:24
on on where we apply it right so it's up
28:24
on on where we apply it right so it's up to 27% in GPT 3.5 GPT 4 which like in
28:29
to 27% in GPT 3.5 GPT 4 which like in
28:29
to 27% in GPT 3.5 GPT 4 which like in one case on dialog response 49 which is
28:33
one case on dialog response 49 which is
28:33
one case on dialog response 49 which is cool but this slide that I stole
28:37
cool but this slide that I stole
28:37
cool but this slide that I stole completely uh from Andy Andy did a
28:40
completely uh from Andy Andy did a
28:40
completely uh from Andy Andy did a fantastic talk about agents if anybody
28:42
fantastic talk about agents if anybody
28:42
fantastic talk about agents if anybody is interested here f look search uh for
28:45
is interested here f look search uh for
28:45
is interested here f look search uh for anng agents and watch that video I think
28:48
anng agents and watch that video I think
28:48
anng agents and watch that video I think it's 10 to 15 minutes so basically GPT
28:53
it's 10 to 15 minutes so basically GPT
28:53
it's 10 to 15 minutes so basically GPT 3.5 in The Benchmark of human evil human
28:56
3.5 in The Benchmark of human evil human
28:56
3.5 in The Benchmark of human evil human evil is a benchmark of coding right
28:59
evil is a benchmark of coding right
28:59
evil is a benchmark of coding right which is something that sorry we believe
29:03
which is something that sorry we believe
29:03
which is something that sorry we believe gpts or models don't do that that well
29:07
gpts or models don't do that that well
29:07
gpts or models don't do that that well well think again after I tell you this
29:11
well think again after I tell you this
29:11
well think again after I tell you this this is the Zer shot which is under
29:14
this is the Zer shot which is under
29:14
this is the Zer shot which is under 50 and with this workflow which you have
29:19
50 and with this workflow which you have
29:19
50 and with this workflow which you have two agents one is the coder and the
29:21
two agents one is the coder and the
29:21
two agents one is the coder and the other is the critic so it analyzes the
29:24
other is the critic so it analyzes the
29:24
other is the critic so it analyzes the code tells you hey this is not clean
29:26
code tells you hey this is not clean
29:26
code tells you hey this is not clean this is good you can improve here it
29:29
this is good you can improve here it
29:29
this is good you can improve here it gets to
29:31
gets to
29:31
gets to 96% of
29:34
96% of
29:34
96% of accuracy and in gbt 4 it gets like
29:41
97% that's pretty and I think uh
29:46
97% that's pretty and I think uh
29:46
97% that's pretty and I think uh yeah it's uh pretty cool so yeah I know
29:50
yeah it's uh pretty cool so yeah I know
29:50
yeah it's uh pretty cool so yeah I know your proms are pretty good in ch Chad G
29:52
your proms are pretty good in ch Chad G
29:52
your proms are pretty good in ch Chad G before you use
29:53
before you use
29:53
before you use it I know that you did share your
29:57
it I know that you did share your
29:57
it I know that you did share your secrets so yeah no problem with that I
29:59
secrets so yeah no problem with that I
29:59
secrets so yeah no problem with that I mean at the end we have to survive
30:01
mean at the end we have to survive
30:01
mean at the end we have to survive together in this AI Revolution right
30:04
together in this AI Revolution right
30:04
together in this AI Revolution right yeah it's one of the reasons I I join
30:07
yeah it's one of the reasons I I join
30:07
yeah it's one of the reasons I I join started to do that so plugins I show the
30:12
started to do that so plugins I show the
30:12
started to do that so plugins I show the this
30:13
this
30:13
this code uh so I'm going
30:17
code uh so I'm going
30:17
code uh so I'm going to jump up with that
30:20
to jump up with that
30:20
to jump up with that part
30:23
part
30:23
part no so I'm going to jump to the code
30:27
no so I'm going to jump to the code
30:27
no so I'm going to jump to the code right
30:33
and I'm going to unwrap
30:36
and I'm going to unwrap
30:37
and I'm going to unwrap this which I have agents example which I
30:41
this which I have agents example which I
30:41
this which I have agents example which I mean this is very very basic so I'm not
30:43
mean this is very very basic so I'm not
30:43
mean this is very very basic so I'm not going to do an advanced model if you are
30:45
going to do an advanced model if you are
30:45
going to do an advanced model if you are interested or the audience is interested
30:47
interested or the audience is interested
30:47
interested or the audience is interested we could do a session on agents which
30:50
we could do a session on agents which
30:50
we could do a session on agents which would take a bit longer right uh also
30:53
would take a bit longer right uh also
30:53
would take a bit longer right uh also things are changing on agent so maybe
30:56
things are changing on agent so maybe
30:56
things are changing on agent so maybe it's something for after summer I don't
30:58
it's something for after summer I don't
30:58
it's something for after summer I don't know um but yeah you can see here I have
31:02
know um but yeah you can see here I have
31:02
know um but yeah you can see here I have a program I create this action and then
31:04
a program I create this action and then
31:04
a program I create this action and then execute it it's a pattern that I
31:06
execute it it's a pattern that I
31:06
execute it it's a pattern that I particularly like I disa the warning
31:10
particularly like I disa the warning
31:10
particularly like I disa the warning because agents are experimental at the
31:12
because agents are experimental at the
31:12
because agents are experimental at the moment in sematic
31:14
moment in sematic
31:14
moment in sematic Kel I create a list of agents to clean
31:16
Kel I create a list of agents to clean
31:16
Kel I create a list of agents to clean them up later and then I do the same I
31:19
them up later and then I do the same I
31:19
them up later and then I do the same I create a GPT
31:21
create a GPT
31:21
create a GPT 40 yeah I think it's big enough let me
31:24
40 yeah I think it's big enough let me
31:24
40 yeah I think it's big enough let me make it
31:25
make it
31:25
make it smaller um here here I have the minion
31:30
smaller um here here I have the minion
31:30
smaller um here here I have the minion agent that I show before all right so
31:33
agent that I show before all right so
31:33
agent that I show before all right so it's basically name a template
31:36
it's basically name a template
31:36
it's basically name a template format uh I give him instructions these
31:39
format uh I give him instructions these
31:39
format uh I give him instructions these are the instructions this is basically
31:41
are the instructions this is basically
31:41
are the instructions this is basically by the way the same as uh giving a
31:44
by the way the same as uh giving a
31:44
by the way the same as uh giving a system promt to the chat system prom is
31:47
system promt to the chat system prom is
31:47
system promt to the chat system prom is when you start a conversation with a
31:49
when you start a conversation with a
31:49
when you start a conversation with a chat and you can configure the chat in
31:51
chat and you can configure the chat in
31:51
chat and you can configure the chat in the administrator mode you can tell it
31:53
the administrator mode you can tell it
31:53
the administrator mode you can tell it the this or by
31:55
the this or by
31:55
the this or by code and a description
31:59
code and a description
31:59
code and a description then I load everything and I simply
32:02
then I load everything and I simply
32:02
then I load everything and I simply create the
32:04
create the
32:04
create the agent and with the
32:07
agent and with the
32:07
agent and with the agent I'm simply create a thread this is
32:11
agent I'm simply create a thread this is
32:12
agent I'm simply create a thread this is for the assistance API a thread needs to
32:14
for the assistance API a thread needs to
32:14
for the assistance API a thread needs to be created to hold a conversation in
32:16
be created to hold a conversation in
32:16
be created to hold a conversation in this case the conversation will be
32:17
this case the conversation will be
32:17
this case the conversation will be between me well and uh the agent and
32:21
between me well and uh the agent and
32:21
between me well and uh the agent and anything that the audience want to tell
32:22
anything that the audience want to tell
32:22
anything that the audience want to tell the agent I will R so I will be as a
32:26
the agent I will R so I will be as a
32:26
the agent I will R so I will be as a human proxy right and then I have a loop
32:29
human proxy right and then I have a loop
32:29
human proxy right and then I have a loop that when I say exit it will say banana
32:30
that when I say exit it will say banana
32:30
that when I say exit it will say banana and finish uh and then it will simply
32:34
and finish uh and then it will simply
32:34
and finish uh and then it will simply invoke the agent thread sending uh to
32:37
invoke the agent thread sending uh to
32:37
invoke the agent thread sending uh to the minion agent the user message that's
32:40
the minion agent the user message that's
32:40
the minion agent the user message that's simple and of course display the message
32:42
simple and of course display the message
32:42
simple and of course display the message because we need to print it so without
32:45
because we need to print it so without
32:45
because we need to print it so without further Ado I'm going to click on play
32:48
further Ado I'm going to click on play
32:48
further Ado I'm going to click on play but before any question from the
32:51
but before any question from the
32:51
but before any question from the audience any
32:54
audience any
32:54
audience any wish no I just hope there's no no any
32:57
wish no I just hope there's no no any
32:57
wish no I just hope there's no no any more demo man all right then let's go on
33:00
more demo man all right then let's go on
33:00
more demo man all right then let's go on and let me check the cursor to the left
33:05
and let me check the cursor to the left
33:05
and let me check the cursor to the left side of this oh
33:07
side of this oh
33:07
side of this oh no I will
33:10
no I will
33:10
no I will continue so how how is this demo
33:12
continue so how how is this demo
33:12
continue so how how is this demo different in this demo you have a plugin
33:14
different in this demo you have a plugin
33:14
different in this demo you have a plugin right that's what you're saying no it's
33:15
right that's what you're saying no it's
33:15
right that's what you're saying no it's taking the prompts from the user I don't
33:17
taking the prompts from the user I don't
33:17
taking the prompts from the user I don't care about these uh plugin I have an
33:19
care about these uh plugin I have an
33:20
care about these uh plugin I have an agent so the one responding so I have
33:22
agent so the one responding so I have
33:22
agent so the one responding so I have kind of a chat Loop uh with the agent so
33:25
kind of a chat Loop uh with the agent so
33:25
kind of a chat Loop uh with the agent so I'm talking to the agent in this case a
33:27
I'm talking to the agent in this case a
33:27
I'm talking to the agent in this case a minion the the minion you saw at the
33:29
minion the the minion you saw at the
33:29
minion the the minion you saw at the beginning oh yeah but I I didn't give it
33:32
beginning oh yeah but I I didn't give it
33:32
beginning oh yeah but I I didn't give it a banana
33:33
a banana
33:33
a banana [Laughter]
33:34
[Laughter]
33:34
[Laughter] [Music]
33:39
yet
33:43
mhm banana how
33:54
you n show yeah it's Friday I'm in the
33:58
you n show yeah it's Friday I'm in the
33:58
you n show yeah it's Friday I'm in the dash of tnet
34:03
show H let's see what he says oh
34:07
show H let's see what he says oh
34:08
show H let's see what he says oh Friday impress
34:12
banana yes so this question coming up
34:16
banana yes so this question coming up
34:16
banana yes so this question coming up from the comment say how does the agent
34:17
from the comment say how does the agent
34:17
from the comment say how does the agent handle context do you have to provide it
34:20
handle context do you have to provide it
34:20
handle context do you have to provide it with the history of the conversation so
34:22
with the history of the conversation so
34:22
with the history of the conversation so far no that is done by the agents uh API
34:26
far no that is done by the agents uh API
34:26
far no that is done by the agents uh API um
34:29
um
34:29
um so it's it's handled by them so it's a
34:31
so it's it's handled by them so it's a
34:31
so it's it's handled by them so it's a lot easier to code so we don't have to
34:34
lot easier to code so we don't have to
34:34
lot easier to code so we don't have to add that to the chat context as we have
34:36
add that to the chat context as we have
34:36
add that to the chat context as we have to do with the chart completion API Does
34:39
to do with the chart completion API Does
34:39
to do with the chart completion API Does this answer you call
34:41
this answer you call
34:41
this answer you call uh uh per
34:44
uh uh per
34:44
uh uh per rosing yeah okay yeah I'm curious to how
34:47
rosing yeah okay yeah I'm curious to how
34:47
rosing yeah okay yeah I'm curious to how good it works I mean it's really
34:49
good it works I mean it's really
34:49
good it works I mean it's really surprising that it works so
34:52
surprising that it works so
34:52
surprising that it works so nicely I'm here with Simon want to say
34:55
nicely I'm here with Simon want to say
34:55
nicely I'm here with Simon want to say something to him yeah I think it's it's
34:56
something to him yeah I think it's it's
34:56
something to him yeah I think it's it's cool and yeah a can be fun I mean they
34:59
cool and yeah a can be fun I mean they
34:59
cool and yeah a can be fun I mean they have humor or they mimic it very well I
35:03
have humor or they mimic it very well I
35:03
have humor or they mimic it very well I for the purist
35:04
for the purist
35:04
for the purist [Laughter]
35:07
[Laughter]
35:07
[Laughter] right so let's see what it says to you
35:09
right so let's see what it says to you
35:09
right so let's see what it says to you I'm making it a bit bigger so can read
35:13
I'm making it a bit bigger so can read
35:13
I'm making it a bit bigger so can read sorry about that I'm an
35:17
sorry about that I'm an
35:17
sorry about that I'm an inconsiderate come on tell something to
35:20
inconsiderate come on tell something to
35:20
inconsiderate come on tell something to Simon he your agent already hits me
35:27
T ammo he says he loves
35:30
T ammo he says he loves
35:30
T ammo he says he loves you want to say something why does why
35:33
you want to say something why does why
35:33
you want to say something why does why does your agent keep say banana and
35:35
does your agent keep say banana and
35:35
does your agent keep say banana and every
35:36
every
35:36
every statement yeah I told him to use minion
35:39
statement yeah I told him to use minion
35:39
statement yeah I told him to use minion and minions use a lot of banana I think
35:41
and minions use a lot of banana I think
35:42
and minions use a lot of banana I think I somehow told him to use banana a
35:44
I somehow told him to use banana a
35:44
I somehow told him to use banana a bit and I like to hear it
35:47
bit and I like to hear it
35:47
bit and I like to hear it [Laughter]
35:51
[Laughter]
35:51
[Laughter] so what do you want to tell tell him I
35:55
so what do you want to tell tell him I
35:55
so what do you want to tell tell him I say I would like to have some banana
35:56
say I would like to have some banana
35:56
say I would like to have some banana shake
36:06
H
36:10
H
36:13
okay no it's fine so what's happening so
36:16
okay no it's fine so what's happening so
36:16
okay no it's fine so what's happening so when when you press that okay what
36:18
when when you press that okay what
36:18
when when you press that okay what what's actually happening when you press
36:20
what's actually happening when you press
36:20
what's actually happening when you press that enter what is Happ behind the
36:22
that enter what is Happ behind the
36:22
that enter what is Happ behind the scenes yeah basically um
36:30
basically it's doing this agent thread
36:32
basically it's doing this agent thread
36:32
basically it's doing this agent thread invoke acing right with the user message
36:34
invoke acing right with the user message
36:34
invoke acing right with the user message so this is then managed by the agent and
36:37
so this is then managed by the agent and
36:37
so this is then managed by the agent and sent to the agent API which has been set
36:40
sent to the agent API which has been set
36:40
sent to the agent API which has been set up with this
36:42
up with this
36:42
up with this agent yeah banana shake with protein
36:45
agent yeah banana shake with protein
36:45
agent yeah banana shake with protein powder yeah that's a good idea that's
36:47
powder yeah that's a good idea that's
36:47
powder yeah that's a good idea that's what it
36:47
what it
36:47
what it replied no no no it's What in in the in
36:51
replied no no no it's What in in the in
36:51
replied no no no it's What in in the in the chat okay that's very relevant I was
36:54
the chat okay that's very relevant I was
36:54
the chat okay that's very relevant I was like wow I think we getting hungry
37:02
okay Sean says banana shake withing po
37:06
okay Sean says banana shake withing po
37:06
okay Sean says banana shake withing po yeah uh no no no no no uh responding to
37:09
yeah uh no no no no no uh responding to
37:09
yeah uh no no no no no uh responding to Shin uh I'm using Asher OPI the the
37:13
Shin uh I'm using Asher OPI the the
37:13
Shin uh I'm using Asher OPI the the model is not running locally I'm using
37:16
model is not running locally I'm using
37:16
model is not running locally I'm using um the one uh from
37:19
um the one uh from
37:19
um the one uh from asheri so yes that's it I could run one
37:22
asheri so yes that's it I could run one
37:22
asheri so yes that's it I could run one locally but not for agents because as
37:24
locally but not for agents because as
37:24
locally but not for agents because as far as I understand uh agent API soorry
37:27
far as I understand uh agent API soorry
37:27
far as I understand uh agent API soorry that's assistance API that's a proper
37:28
that's assistance API that's a proper
37:28
that's assistance API that's a proper name is not
37:31
name is not
37:31
name is not supported
37:36
um want to say something to the
37:40
audience key meing and banana for
37:46
everyone tonight I'm going to dream what
37:49
everyone tonight I'm going to dream what
37:49
everyone tonight I'm going to dream what banana see you let's let's be let be
37:51
banana see you let's let's be let be
37:51
banana see you let's let's be let be kind
37:54
kind
37:54
kind so thank you Pap haha banana okay
37:58
so thank you Pap haha banana okay
37:58
so thank you Pap haha banana okay all
37:59
all
37:59
all right so cool yes I think it's it's fun
38:04
right so cool yes I think it's it's fun
38:04
right so cool yes I think it's it's fun it's interesting it's a whole new world
38:07
it's interesting it's a whole new world
38:07
it's interesting it's a whole new world and sematic canel brings this power to
38:09
and sematic canel brings this power to
38:09
and sematic canel brings this power to your fingertips at the moment
38:11
your fingertips at the moment
38:11
your fingertips at the moment implementations of agents are quite
38:14
implementations of agents are quite
38:14
implementations of agents are quite interesting but still limited so
38:17
interesting but still limited so
38:17
interesting but still limited so others um other implementations are um a
38:21
others um other implementations are um a
38:21
others um other implementations are um a bit more Advan but it will get here so
38:25
bit more Advan but it will get here so
38:25
bit more Advan but it will get here so let's move to the presentations
38:28
let's move to the presentations
38:28
let's move to the presentations um and yeah uh I hope this was
38:31
um and yeah uh I hope this was
38:31
um and yeah uh I hope this was interesting I leave some resources right
38:34
interesting I leave some resources right
38:34
interesting I leave some resources right so Microsoft learn for sematic kernel
38:37
so Microsoft learn for sematic kernel
38:37
so Microsoft learn for sematic kernel GitHub Microsoft uh sematic kernel so
38:40
GitHub Microsoft uh sematic kernel so
38:40
GitHub Microsoft uh sematic kernel so one of the most important things is to
38:42
one of the most important things is to
38:42
one of the most important things is to go to the GitHub and look at the code
38:44
go to the GitHub and look at the code
38:44
go to the GitHub and look at the code the code does not lie and it's pretty
38:48
the code does not lie and it's pretty
38:48
the code does not lie and it's pretty well done and then there's a discore
38:50
well done and then there's a discore
38:50
well done and then there's a discore which um you can capture here it's
38:53
which um you can capture here it's
38:53
which um you can capture here it's pretty interesting I'm there every now
38:55
pretty interesting I'm there every now
38:55
pretty interesting I'm there every now and then and also there's some
38:58
and then and also there's some
38:58
and then and also there's some discussions so if you have some
39:00
discussions so if you have some
39:00
discussions so if you have some technical issue here is the place to say
39:02
technical issue here is the place to say
39:02
technical issue here is the place to say this is mostly user support and this is
39:06
this is mostly user support and this is
39:06
this is mostly user support and this is their support or you can go and create a
39:09
their support or you can go and create a
39:09
their support or you can go and create a ticket in or an issue in uh GitHub if
39:12
ticket in or an issue in uh GitHub if
39:12
ticket in or an issue in uh GitHub if you try that and the other thing
39:17
you try that and the other thing
39:17
you try that and the other thing is this is course that I did in LinkedIn
39:21
is this is course that I did in LinkedIn
39:21
is this is course that I did in LinkedIn some people like it uh and say good
39:24
some people like it uh and say good
39:24
some people like it uh and say good things about it I think it's for and a
39:28
things about it I think it's for and a
39:28
things about it I think it's for and a half hours gets you through a lot of
39:30
half hours gets you through a lot of
39:30
half hours gets you through a lot of things uh including
39:33
things uh including
39:33
things uh including agents and yeah I leave that here and
39:37
agents and yeah I leave that here and
39:37
agents and yeah I leave that here and that aside I'm happy to respond any
39:40
that aside I'm happy to respond any
39:40
that aside I'm happy to respond any question anybody may have yeah that's
39:43
question anybody may have yeah that's
39:43
question anybody may have yeah that's pretty cool I'm curious about this uh in
39:46
pretty cool I'm curious about this uh in
39:46
pretty cool I'm curious about this uh in in one of your previous you had a
39:48
in one of your previous you had a
39:48
in one of your previous you had a Discord uh invite what was that was it
39:51
Discord uh invite what was that was it
39:51
Discord uh invite what was that was it about this session or about something
39:53
about this session or about something
39:53
about this session or about something else you a Discord invite link yeah it
39:57
else you a Discord invite link yeah it
39:57
else you a Discord invite link yeah it was a Discord inite link yes yeah I'm
40:00
was a Discord inite link yes yeah I'm
40:00
was a Discord inite link yes yeah I'm curious semantic kernel Discord so they
40:03
curious semantic kernel Discord so they
40:03
curious semantic kernel Discord so they have
40:04
have
40:04
have a it yes wow that I think everyone
40:09
a it yes wow that I think everyone
40:09
a it yes wow that I think everyone should join that if they want to get
40:10
should join that if they want to get
40:10
should join that if they want to get into discussions yeah other than the
40:13
into discussions yeah other than the
40:13
into discussions yeah other than the only thing is there's a bit of a Spam
40:15
only thing is there's a bit of a Spam
40:15
only thing is there's a bit of a Spam sometimes but I think it's also fun to
40:18
sometimes but I think it's also fun to
40:18
sometimes but I think it's also fun to talk with other people there and I think
40:20
talk with other people there and I think
40:20
talk with other people there and I think it's quite cool to to be able to talk
40:23
it's quite cool to to be able to talk
40:23
it's quite cool to to be able to talk about things openly but for down side
40:26
about things openly but for down side
40:26
about things openly but for down side the semantical team does not go into
40:27
the semantical team does not go into
40:27
the semantical team does not go into there is not supporting that so they act
40:31
there is not supporting that so they act
40:31
there is not supporting that so they act only when you put something in um
40:35
only when you put something in um
40:35
only when you put something in um GitHub I get so
40:38
GitHub I get so
40:39
GitHub I get so yeah and is there I mean I see this's
40:41
yeah and is there I mean I see this's
40:41
yeah and is there I mean I see this's learn
40:42
learn
40:42
learn microsoft.com let's change the UR from
40:44
microsoft.com let's change the UR from
40:44
microsoft.com let's change the UR from dos to learn at microsoft.com now but do
40:47
dos to learn at microsoft.com now but do
40:47
dos to learn at microsoft.com now but do they have a learn course because last
40:49
they have a learn course because last
40:49
they have a learn course because last time I checked they did not have a learn
40:51
time I checked they did not have a learn
40:51
time I checked they did not have a learn course yeah they have uh well uh I think
40:55
course yeah they have uh well uh I think
40:55
course yeah they have uh well uh I think this is the documentation the learn
40:56
this is the documentation the learn
40:56
this is the documentation the learn microsoft.com
40:58
microsoft.com
40:58
microsoft.com and there's they have some small courses
41:01
and there's they have some small courses
41:01
and there's they have some small courses mhm um which some of them are quite good
41:06
mhm um which some of them are quite good
41:06
mhm um which some of them are quite good I I don't like the agent one so I gave
41:08
I I don't like the agent one so I gave
41:08
I I don't like the agent one so I gave uh
41:09
uh
41:09
uh sorry
41:10
sorry
41:10
sorry [Music]
41:12
[Music]
41:12
[Music] the I gave a bit the feedback to to to
41:15
the I gave a bit the feedback to to to
41:15
the I gave a bit the feedback to to to the team that the agents really you are
41:17
the team that the agents really you are
41:17
the team that the agents really you are not building an agent properly say it
41:19
not building an agent properly say it
41:19
not building an agent properly say it right so they are not using the agent
41:22
right so they are not using the agent
41:22
right so they are not using the agent API uh for for that but aside these two
41:25
API uh for for that but aside these two
41:25
API uh for for that but aside these two courses I think they are pretty good
41:27
courses I think they are pretty good
41:27
courses I think they are pretty good yeah that's good that's good that's nice
41:30
yeah that's good that's good that's nice
41:30
yeah that's good that's good that's nice so so what's next Jose you know I mean
41:33
so so what's next Jose you know I mean
41:33
so so what's next Jose you know I mean that I know we saw uh um a command line
41:36
that I know we saw uh um a command line
41:36
that I know we saw uh um a command line application today right can you can you
41:38
application today right can you can you
41:38
application today right can you can you give us few examples where where more
41:40
give us few examples where where more
41:40
give us few examples where where more people can use it in your other than MC
41:43
people can use it in your other than MC
41:43
people can use it in your other than MC application or something I know but yeah
41:46
application or something I know but yeah
41:46
application or something I know but yeah you can use it you can use it anywhere I
41:48
you can use it you can use it anywhere I
41:48
you can use it you can use it anywhere I mean technically you could uh connect to
41:51
mean technically you could uh connect to
41:51
mean technically you could uh connect to the API of your AER devops review the
41:54
the API of your AER devops review the
41:54
the API of your AER devops review the issues and respond automatically
41:58
issues and respond automatically
41:58
issues and respond automatically or try to generate the code or
42:01
or try to generate the code or
42:01
or try to generate the code or Implement uh what would be imple try to
42:04
Implement uh what would be imple try to
42:04
Implement uh what would be imple try to implement the code and try it using
42:06
implement the code and try it using
42:06
implement the code and try it using maybe on this I would use a agent
42:09
maybe on this I would use a agent
42:09
maybe on this I would use a agent workflow right I would create two agents
42:11
workflow right I would create two agents
42:11
workflow right I would create two agents one to write code another that reviews
42:13
one to write code another that reviews
42:13
one to write code another that reviews The Code test it and gives feedback to
42:15
The Code test it and gives feedback to
42:16
The Code test it and gives feedback to the other agent until it builds and it
42:17
the other agent until it builds and it
42:17
the other agent until it builds and it does everything that is written in the
42:20
does everything that is written in the
42:20
does everything that is written in the GitHub
42:21
GitHub
42:21
GitHub issue um yeah I think it will depend on
42:24
issue um yeah I think it will depend on
42:24
issue um yeah I think it will depend on on the model that you're using but
42:27
on the model that you're using but
42:27
on the model that you're using but if you put this it will not deploy or
42:30
if you put this it will not deploy or
42:30
if you put this it will not deploy or give the response until the code is
42:32
give the response until the code is
42:32
give the response until the code is completely done and it has pass the
42:35
completely done and it has pass the
42:35
completely done and it has pass the quality
42:36
quality
42:36
quality filters for example so that's powerful
42:40
filters for example so that's powerful
42:40
filters for example so that's powerful combination that you can do and we will
42:42
combination that you can do and we will
42:42
combination that you can do and we will get there that we end up working with uh
42:46
get there that we end up working with uh
42:46
get there that we end up working with uh virtual team of
42:47
virtual team of
42:47
virtual team of developers yeah think m still pretty
42:50
developers yeah think m still pretty
42:50
developers yeah think m still pretty pretty new to this ecosystem right we're
42:53
pretty new to this ecosystem right we're
42:53
pretty new to this ecosystem right we're still getting there uh having said that
42:55
still getting there uh having said that
42:55
still getting there uh having said that I I suppose we could also do with blazer
42:58
I I suppose we could also do with blazer
42:58
I I suppose we could also do with blazer right you can also work with Blazer and
43:00
right you can also work with Blazer and
43:00
right you can also work with Blazer and semantic kernel any anywhere you can
43:02
semantic kernel any anywhere you can
43:02
semantic kernel any anywhere you can deploy uhet and
43:05
deploy uhet and
43:05
deploy uhet and runet you can use sematic colel
43:08
runet you can use sematic colel
43:08
runet you can use sematic colel everywhere okay that's nice that that's
43:11
everywhere okay that's nice that that's
43:11
everywhere okay that's nice that that's I think pretty nice way get start build
43:13
I think pretty nice way get start build
43:13
I think pretty nice way get start build your own like chat for example there is
43:16
your own like chat for example there is
43:16
your own like chat for example there is an example of dmer BOS
43:20
an example of dmer BOS
43:20
an example of dmer BOS using to connect to an llm and create uh
43:23
using to connect to an llm and create uh
43:23
using to connect to an llm and create uh elements uh in in virtual reality for
43:26
elements uh in in virtual reality for
43:26
elements uh in in virtual reality for example I think that is even from last
43:29
example I think that is even from last
43:29
example I think that is even from last year that's pretty cool yeah okay what
43:33
year that's pretty cool yeah okay what
43:33
year that's pretty cool yeah okay what else do we have to discuss we already
43:36
else do we have to discuss we already
43:36
else do we have to discuss we already had 44 minutes time passed very quickly
43:39
had 44 minutes time passed very quickly
43:39
had 44 minutes time passed very quickly I didn't realize it yeah I went quite
43:41
I didn't realize it yeah I went quite
43:41
I didn't realize it yeah I went quite quick so not really too too much more
43:44
quick so not really too too much more
43:45
quick so not really too too much more today to show
43:47
today to show
43:47
today to show um I think it's a cool technology and if
43:51
um I think it's a cool technology and if
43:52
um I think it's a cool technology and if you are a net developer um you should be
43:54
you are a net developer um you should be
43:54
you are a net developer um you should be getting into generative AI right now now
43:57
getting into generative AI right now now
43:57
getting into generative AI right now now because it takes a while so it take me a
44:00
because it takes a while so it take me a
44:00
because it takes a while so it take me a while to have a certain level of
44:02
while to have a certain level of
44:02
while to have a certain level of understanding of what I I do and what I
44:04
understanding of what I I do and what I
44:04
understanding of what I I do and what I can and it takes time to really
44:06
can and it takes time to really
44:06
can and it takes time to really understand what it can really do for you
44:09
understand what it can really do for you
44:09
understand what it can really do for you and your career of course right so and
44:12
and your career of course right so and
44:12
and your career of course right so and if you want to
44:14
if you want to
44:14
if you want to do proper business applications as well
44:18
do proper business applications as well
44:18
do proper business applications as well um so this is one of the key values of
44:20
um so this is one of the key values of
44:20
um so this is one of the key values of semantical it is meant to do business
44:23
semantical it is meant to do business
44:23
semantical it is meant to do business applications recently in the last bill
44:25
applications recently in the last bill
44:25
applications recently in the last bill like few weeks ago
44:27
like few weeks ago
44:27
like few weeks ago uh they release a couple of features
44:29
uh they release a couple of features
44:29
uh they release a couple of features like integration with as Aspire Aspire
44:32
like integration with as Aspire Aspire
44:32
like integration with as Aspire Aspire is no longer uh experimental it's a
44:35
is no longer uh experimental it's a
44:35
is no longer uh experimental it's a product it's supported you can deploy it
44:37
product it's supported you can deploy it
44:37
product it's supported you can deploy it in production right so you can author
44:40
in production right so you can author
44:40
in production right so you can author and monitor your different services and
44:43
and monitor your different services and
44:43
and monitor your different services and how do they
44:45
how do they
44:45
how do they work and and that is very very powerful
44:49
work and and that is very very powerful
44:49
work and and that is very very powerful and you can W uh and get these decisions
44:53
and you can W uh and get these decisions
44:53
and you can W uh and get these decisions and create your own agents or co-pilots
44:56
and create your own agents or co-pilots
44:56
and create your own agents or co-pilots or AO autonomos tools that decide what
44:59
or AO autonomos tools that decide what
44:59
or AO autonomos tools that decide what they have to do so it's no longer kind
45:03
they have to do so it's no longer kind
45:03
they have to do so it's no longer kind of coding is separated from AI so every
45:06
of coding is separated from AI so every
45:06
of coding is separated from AI so every now and then any application that we do
45:09
now and then any application that we do
45:09
now and then any application that we do probably will have a big deal of AI on
45:12
probably will have a big deal of AI on
45:12
probably will have a big deal of AI on it that that are my
45:15
it that that are my
45:15
it that that are my insights yeah with this is the right
45:17
insights yeah with this is the right
45:17
insights yeah with this is the right time to get into this gen otherwise I
45:20
time to get into this gen otherwise I
45:20
time to get into this gen otherwise I said it does take a while all the things
45:21
said it does take a while all the things
45:22
said it does take a while all the things are uh with all these API open and Azure
45:24
are uh with all these API open and Azure
45:25
are uh with all these API open and Azure and mics are helping little you with
45:26
and mics are helping little you with
45:26
and mics are helping little you with everything all the resources you can
45:28
everything all the resources you can
45:28
everything all the resources you can still PA it up uh but yeah I think this
45:30
still PA it up uh but yeah I think this
45:30
still PA it up uh but yeah I think this is the right time as a developer I think
45:32
is the right time as a developer I think
45:32
is the right time as a developer I think this is the exciting time to be a
45:33
this is the exciting time to be a
45:33
this is the exciting time to be a developer on top of that all this open
45:36
developer on top of that all this open
45:36
developer on top of that all this open AI
45:36
AI
45:37
AI Chi uh these makes your life pretty cool
45:41
Chi uh these makes your life pretty cool
45:41
Chi uh these makes your life pretty cool let's see what we have friend there's
45:42
let's see what we have friend there's
45:42
let's see what we have friend there's question from the comments what are the
45:44
question from the comments what are the
45:44
question from the comments what are the costs here of a banana talk with the
45:46
costs here of a banana talk with the
45:47
costs here of a banana talk with the agent how can the cost that the uses of
45:49
agent how can the cost that the uses of
45:49
agent how can the cost that the uses of an agent or SK inur estimated that's a
45:54
an agent or SK inur estimated that's a
45:54
an agent or SK inur estimated that's a very good question yeah U you have on
45:57
very good question yeah U you have on
45:57
very good question yeah U you have on response as well a complex object and
46:00
response as well a complex object and
46:00
response as well a complex object and inside it are the tokens used input
46:03
inside it are the tokens used input
46:03
inside it are the tokens used input tokens output tokens so you can infert
46:08
that so yes you can you can know how
46:12
that so yes you can you can know how
46:12
that so yes you can you can know how much is that cting to you yes that can
46:14
much is that cting to you yes that can
46:14
much is that cting to you yes that can be done thanks for asking question
46:17
be done thanks for asking question
46:17
be done thanks for asking question friendship yeah that's good also the uh
46:21
friendship yeah that's good also the uh
46:21
friendship yeah that's good also the uh they have implemented a layer of
46:22
they have implemented a layer of
46:22
they have implemented a layer of trustability as well so you have
46:25
trustability as well so you have
46:25
trustability as well so you have observability uh Incorporated I think uh
46:28
observability uh Incorporated I think uh
46:28
observability uh Incorporated I think uh soon we will have open Telemetry I have
46:30
soon we will have open Telemetry I have
46:30
soon we will have open Telemetry I have not yet played with the Asher bits but
46:33
not yet played with the Asher bits but
46:34
not yet played with the Asher bits but yep um I'm one of the things I would
46:37
yep um I'm one of the things I would
46:37
yep um I'm one of the things I would like to try this weekend with um uh
46:41
like to try this weekend with um uh
46:41
like to try this weekend with um uh Microsoft
46:43
Microsoft
46:43
Microsoft Aspire okay that's pretty cool if
46:46
Aspire okay that's pretty cool if
46:46
Aspire okay that's pretty cool if there's not any more comments and uh
46:48
there's not any more comments and uh
46:48
there's not any more comments and uh then we'll just wrap it up I think so
46:51
then we'll just wrap it up I think so
46:51
then we'll just wrap it up I think so anything else you got to add Jose before
46:52
anything else you got to add Jose before
46:52
anything else you got to add Jose before we wrap up the show yeah sure uh I
46:55
we wrap up the show yeah sure uh I
46:55
we wrap up the show yeah sure uh I already mentioned I did a CSE it took me
46:57
already mentioned I did a CSE it took me
46:57
already mentioned I did a CSE it took me a while but uh if anybody wants to jump
47:00
a while but uh if anybody wants to jump
47:00
a while but uh if anybody wants to jump into it I cannot but recommend it I
47:03
into it I cannot but recommend it I
47:03
into it I cannot but recommend it I think it's uh covering all the
47:04
think it's uh covering all the
47:04
think it's uh covering all the fundamentals it tries also to shape
47:07
fundamentals it tries also to shape
47:07
fundamentals it tries also to shape everything with metaphors like the
47:09
everything with metaphors like the
47:09
everything with metaphors like the orchestra um that the plugins or the
47:12
orchestra um that the plugins or the
47:12
orchestra um that the plugins or the instruments that the musicians are the U
47:15
instruments that the musicians are the U
47:15
instruments that the musicians are the U AI models or agents so I think it's um
47:19
AI models or agents so I think it's um
47:19
AI models or agents so I think it's um pretty fitting and it's plenty of with
47:21
pretty fitting and it's plenty of with
47:21
pretty fitting and it's plenty of with examples so if anybody is curious I
47:24
examples so if anybody is curious I
47:24
examples so if anybody is curious I recommend it maybe you can jump first
47:27
recommend it maybe you can jump first
47:27
recommend it maybe you can jump first and see if you like the technology first
47:30
and see if you like the technology first
47:30
and see if you like the technology first uh to the Microsoft learn uh modules do
47:33
uh to the Microsoft learn uh modules do
47:34
uh to the Microsoft learn uh modules do them they are not bad um uh and then if
47:37
them they are not bad um uh and then if
47:37
them they are not bad um uh and then if you like this technology then jump into
47:39
you like this technology then jump into
47:39
you like this technology then jump into the course which is a bit more advanced
47:41
the course which is a bit more advanced
47:41
the course which is a bit more advanced it gets you and at the end I'm building
47:43
it gets you and at the end I'm building
47:43
it gets you and at the end I'm building a very Advanced example which is a
47:45
a very Advanced example which is a
47:46
a very Advanced example which is a researcher system that uses a complex
47:49
researcher system that uses a complex
47:49
researcher system that uses a complex prompting to and also a plugin which
47:53
prompting to and also a plugin which
47:53
prompting to and also a plugin which does a web search the web search I'm
47:55
does a web search the web search I'm
47:55
does a web search the web search I'm using um um I think big Bing search
47:58
using um um I think big Bing search
47:58
using um um I think big Bing search service from Asher and gets the result
48:01
service from Asher and gets the result
48:01
service from Asher and gets the result and summarizes and creates a report for
48:03
and summarizes and creates a report for
48:04
and summarizes and creates a report for the end user which is kind of a Advan
48:07
the end user which is kind of a Advan
48:07
the end user which is kind of a Advan thing that it usually you do a lot of
48:10
thing that it usually you do a lot of
48:10
thing that it usually you do a lot of research and and it takes time and this
48:12
research and and it takes time and this
48:12
research and and it takes time and this can be completely automated for you that
48:16
can be completely automated for you that
48:16
can be completely automated for you that sounds pretty pretty uh pretty good
48:18
sounds pretty pretty uh pretty good
48:18
sounds pretty pretty uh pretty good course and all this in just about four
48:20
course and all this in just about four
48:20
course and all this in just about four and a half hours
48:21
and a half hours
48:21
and a half hours right sorry all this in just about in
48:24
right sorry all this in just about in
48:24
right sorry all this in just about in four and a half hours course yes yes
48:27
four and a half hours course yes yes
48:27
four and a half hours course yes yes exactly yeah yeah I think it's not 40
48:31
exactly yeah yeah I think it's not 40
48:31
exactly yeah yeah I think it's not 40 hours like others but I try to put it as
48:35
hours like others but I try to put it as
48:35
hours like others but I try to put it as le as
48:36
le as
48:36
le as possible yeah I like the pro that he
48:38
possible yeah I like the pro that he
48:39
possible yeah I like the pro that he discussed if it was 40 hours I would
48:41
discussed if it was 40 hours I would
48:41
discussed if it was 40 hours I would have to have left my job and yeah
48:43
have to have left my job and yeah
48:43
have to have left my job and yeah honestly I don't want that yeah no we
48:45
honestly I don't want that yeah no we
48:45
honestly I don't want that yeah no we don't want that it's all for the
48:48
don't want that it's all for the
48:48
don't want that it's all for the commune
48:50
commune
48:50
commune scre yeah all right in that case thank
48:53
scre yeah all right in that case thank
48:53
scre yeah all right in that case thank you so much Jose thanks for putting this
48:55
you so much Jose thanks for putting this
48:55
you so much Jose thanks for putting this show on net the kind of the season of AI
48:58
show on net the kind of the season of AI
48:58
show on net the kind of the season of AI going on so thanks also for that putting
49:00
going on so thanks also for that putting
49:01
going on so thanks also for that putting everything together it was a great
49:02
everything together it was a great
49:02
everything together it was a great presentation loved it uh some great
49:04
presentation loved it uh some great
49:04
presentation loved it uh some great demos right with demo demons as always
49:07
demos right with demo demons as always
49:08
demos right with demo demons as always uh they're part of the show having said
49:10
uh they're part of the show having said
49:10
uh they're part of the show having said that uh I'll see you in some other
49:12
that uh I'll see you in some other
49:12
that uh I'll see you in some other episodes we have many other events and
49:13
episodes we have many other events and
49:13
episodes we have many other events and coming up anything you want to add to
49:15
coming up anything you want to add to
49:15
coming up anything you want to add to the users for watching from your user
49:19
the users for watching from your user
49:19
the users for watching from your user group
49:21
group
49:21
group sure is there anything you want to add
49:24
sure is there anything you want to add
49:24
sure is there anything you want to add and say to the users who are joining
49:26
and say to the users who are joining
49:26
and say to the users who are joining from your the group yeah uh so thanks
49:28
from your the group yeah uh so thanks
49:28
from your the group yeah uh so thanks for being there it was a pleasure to be
49:30
for being there it was a pleasure to be
49:30
for being there it was a pleasure to be here also thank to you Simon for
49:33
here also thank to you Simon for
49:33
here also thank to you Simon for organizing and having me on your show uh
49:36
organizing and having me on your show uh
49:36
organizing and having me on your show uh yeah anytime it was really fun it's an
49:38
yeah anytime it was really fun it's an
49:38
yeah anytime it was really fun it's an amazing to be hosted by you it's an
49:41
amazing to be hosted by you it's an
49:41
amazing to be hosted by you it's an honor uh it is and it was it was a lot
49:45
honor uh it is and it was it was a lot
49:45
honor uh it is and it was it was a lot of fun uh which uh I think these things
49:48
of fun uh which uh I think these things
49:48
of fun uh which uh I think these things have to be fun right because we do that
49:50
have to be fun right because we do that
49:50
have to be fun right because we do that on our own time for free so and we learn
49:54
on our own time for free so and we learn
49:54
on our own time for free so and we learn we grow and we
49:59
yeah
50:01
yeah
50:01
yeah yes all right everyone thank you so much
50:03
yes all right everyone thank you so much
50:03
yes all right everyone thank you so much for tuning in we'll see you some other
50:04
for tuning in we'll see you some other
50:04
for tuning in we'll see you some other episodes take care have a nice weekend
50:06
episodes take care have a nice weekend
50:06
episodes take care have a nice weekend bye-bye
#Windows & .NET