Up next in 10
Azure AI Foundry has quickly become a hub for building, deploying, and scaling AI models — but what if we explored it through the lens of parody?
In this light-hearted session, we’ll take a playful dive into:
• How Azure AI Foundry builds and deploys AI agents
• The quirks of working with third-party models (yes, even the mysterious “zero-humped Llama”)
• The humorous side of text generation, image creation, and AI experiments
• What parody can teach us about the very real power and risks of AI
Whether you’re an Azure developer, AI enthusiast, or just here for the fun, you’ll walk away with insights into Azure AI Foundry — and a few laughs.
-------------------------------------------------------------------
⚡Download Sharp Rewards Wallet to earn Rewards: (200 bonus on using this link) http://invite.sharpplatform.com/VirtualConf
📺 CSharp TV - Dev Streaming Destination http://csharp.tv
🌎 C# Corner - Community of Software and Data Developers
Show More Show Less View Video Transcript
0:00
So hi everyone, welcome to a new episode
0:02
So hi everyone, welcome to a new episode
0:02
So hi everyone, welcome to a new episode of AI Dev Tools. I am Simon and today we
0:05
of AI Dev Tools. I am Simon and today we
0:05
of AI Dev Tools. I am Simon and today we have Andy who is a human being. Right.
0:08
have Andy who is a human being. Right.
0:08
have Andy who is a human being. Right. So Andy, welcome to the AI Dev Tools
0:11
So Andy, welcome to the AI Dev Tools
0:11
So Andy, welcome to the AI Dev Tools today.
0:12
today.
0:12
today. [Music]
0:17
Okay, thank you very much Simon. So
0:20
Okay, thank you very much Simon. So
0:20
Okay, thank you very much Simon. So yeah, indeed uh since we're going to
0:22
yeah, indeed uh since we're going to
0:22
yeah, indeed uh since we're going to talk about AI, it's definitely a good
0:24
talk about AI, it's definitely a good
0:24
talk about AI, it's definitely a good idea to mention that I am a human being.
0:26
idea to mention that I am a human being.
0:26
idea to mention that I am a human being. I'm real uh etc. you know, all the usual
0:29
I'm real uh etc. you know, all the usual
0:29
I'm real uh etc. you know, all the usual things. And uh that we can uh that we
0:33
things. And uh that we can uh that we
0:33
things. And uh that we can uh that we can use AI for interesting things
0:35
can use AI for interesting things
0:35
can use AI for interesting things without actually being absorbed by
0:38
without actually being absorbed by
0:38
without actually being absorbed by Skynets or I don't know anything like
0:40
Skynets or I don't know anything like
0:40
Skynets or I don't know anything like that.
0:42
that.
0:42
that. >> Yeah, that's true. That's true. Yeah.
0:44
>> Yeah, that's true. That's true. Yeah.
0:44
>> Yeah, that's true. That's true. Yeah. And I I'll take a moment, you know, to
0:45
And I I'll take a moment, you know, to
0:45
And I I'll take a moment, you know, to ask you Andy as you know, what's up with
0:48
ask you Andy as you know, what's up with
0:48
ask you Andy as you know, what's up with your virtual background? Why why do you
0:50
your virtual background? Why why do you
0:50
your virtual background? Why why do you have a world map and what are those
0:52
have a world map and what are those
0:52
have a world map and what are those green signals and everything?
0:54
green signals and everything?
0:54
green signals and everything? Okay. So, I think that this one comes
0:56
Okay. So, I think that this one comes
0:56
Okay. So, I think that this one comes from Skynet. And I do apologize that
0:59
from Skynet. And I do apologize that
0:59
from Skynet. And I do apologize that there is some lag happening uh with the
1:01
there is some lag happening uh with the
1:01
there is some lag happening uh with the background. I'm not sure. It has to do
1:03
background. I'm not sure. It has to do
1:03
background. I'm not sure. It has to do something with my hardware. Uh but some
1:06
something with my hardware. Uh but some
1:06
something with my hardware. Uh but some things aren't perfect with uh using
1:08
things aren't perfect with uh using
1:08
things aren't perfect with uh using Streamyard here. Uh this is on me. This
1:11
Streamyard here. Uh this is on me. This
1:11
Streamyard here. Uh this is on me. This is not on you. Uh so yeah. Uh but the
1:14
is not on you. Uh so yeah. Uh but the
1:14
is not on you. Uh so yeah. Uh but the background is supposed to be AI
1:16
background is supposed to be AI
1:16
background is supposed to be AI generated. You can remove the background
1:19
generated. You can remove the background
1:19
generated. You can remove the background around my face. There's actually very
1:21
around my face. There's actually very
1:21
around my face. There's actually very easy Python scripts to do that frame by
1:23
easy Python scripts to do that frame by
1:23
easy Python scripts to do that frame by frame. But then of course you have some
1:25
frame. But then of course you have some
1:25
frame. But then of course you have some better tools using these kind of
1:28
better tools using these kind of
1:28
better tools using these kind of products to to actually do this better.
1:30
products to to actually do this better.
1:30
products to to actually do this better. It's not perfect in this case.
1:33
It's not perfect in this case.
1:33
It's not perfect in this case. >> Well, it's okay. We're humans at the end
1:35
>> Well, it's okay. We're humans at the end
1:35
>> Well, it's okay. We're humans at the end of the day. Things doesn't have to be
1:37
of the day. Things doesn't have to be
1:37
of the day. Things doesn't have to be perfect, right? We're not AI, right? So
1:39
perfect, right? We're not AI, right? So
1:40
perfect, right? We're not AI, right? So So that's how it is, right? So Andy,
1:41
So that's how it is, right? So Andy,
1:41
So that's how it is, right? So Andy, what are we talking? Are we focusing
1:43
what are we talking? Are we focusing
1:43
what are we talking? Are we focusing more on the, you know, uh the studio,
1:47
more on the, you know, uh the studio,
1:47
more on the, you know, uh the studio, right?
1:47
right?
1:48
right? AI studio what let's go somewhere there.
1:52
AI studio what let's go somewhere there.
1:52
AI studio what let's go somewhere there. >> Yeah. So let's let's have a look at some
1:54
>> Yeah. So let's let's have a look at some
1:54
>> Yeah. So let's let's have a look at some slides and then we can talk about um
1:57
slides and then we can talk about um
1:57
slides and then we can talk about um nowadays it's no longer called AI
1:59
nowadays it's no longer called AI
1:59
nowadays it's no longer called AI studio. It's called AI foundry which is
2:02
studio. It's called AI foundry which is
2:02
studio. It's called AI foundry which is a crazy name but uh let's let's try to
2:05
a crazy name but uh let's let's try to
2:05
a crazy name but uh let's let's try to explain all of this to to humans so that
2:08
explain all of this to to humans so that
2:08
explain all of this to to humans so that we can understand each other. So first
2:10
we can understand each other. So first
2:10
we can understand each other. So first of all let's let's move on. This is of
2:12
of all let's let's move on. This is of
2:12
of all let's let's move on. This is of course my running joke. If you know uh
2:15
course my running joke. If you know uh
2:15
course my running joke. If you know uh basic then you know what this means but
2:17
basic then you know what this means but
2:17
basic then you know what this means but let's move on.
2:19
let's move on.
2:19
let's move on. This is my cat or used to be my cat. So
2:22
This is my cat or used to be my cat. So
2:22
This is my cat or used to be my cat. So she sadly passed away last month but she
2:25
she sadly passed away last month but she
2:25
she sadly passed away last month but she was a very nice companion that for like
2:27
was a very nice companion that for like
2:27
was a very nice companion that for like 10 years doing these remote sessions she
2:30
10 years doing these remote sessions she
2:30
10 years doing these remote sessions she would always come here and uh complain
2:33
would always come here and uh complain
2:33
would always come here and uh complain every time I said something wrong which
2:35
every time I said something wrong which
2:35
every time I said something wrong which happened a lot. Uh and she was
2:37
happened a lot. Uh and she was
2:37
happened a lot. Uh and she was particularly fond of Azour. though she
2:39
particularly fond of Azour. though she
2:39
particularly fond of Azour. though she would actually hop on as soon as I
2:42
would actually hop on as soon as I
2:42
would actually hop on as soon as I showed something related to Azour.
2:44
showed something related to Azour.
2:44
showed something related to Azour. Sadly, uh she passed away last month.
2:46
Sadly, uh she passed away last month.
2:46
Sadly, uh she passed away last month. This uh this happens, but she was 17
2:49
This uh this happens, but she was 17
2:49
This uh this happens, but she was 17 years old at the time. So, I think she
2:51
years old at the time. So, I think she
2:51
years old at the time. So, I think she has a really nice technological life. Um
2:56
has a really nice technological life. Um
2:56
has a really nice technological life. Um sharing her meows about technology. Uh
2:59
sharing her meows about technology. Uh
2:59
sharing her meows about technology. Uh and this is me. This is um I'm Andrea. I
3:03
and this is me. This is um I'm Andrea. I
3:03
and this is me. This is um I'm Andrea. I usually do crazy sessions a bit all over
3:05
usually do crazy sessions a bit all over
3:05
usually do crazy sessions a bit all over the world whenever I can. And this one
3:08
the world whenever I can. And this one
3:08
the world whenever I can. And this one is not that much of a difference.
3:10
is not that much of a difference.
3:10
is not that much of a difference. Although I will only focus on demos. So
3:12
Although I will only focus on demos. So
3:12
Although I will only focus on demos. So no more slides. Uh but this is um
3:16
no more slides. Uh but this is um
3:16
no more slides. Uh but this is um melting AI foundry. So of course they
3:19
melting AI foundry. So of course they
3:19
melting AI foundry. So of course they came up with such a weird name. It used
3:21
came up with such a weird name. It used
3:21
came up with such a weird name. It used to be called Azour a open AI studio and
3:24
to be called Azour a open AI studio and
3:24
to be called Azour a open AI studio and then they renamed it to Azour AI studio
3:27
then they renamed it to Azour AI studio
3:27
then they renamed it to Azour AI studio and then they called it AI foundry like
3:29
and then they called it AI foundry like
3:29
and then they called it AI foundry like uh you know almost a year ago and I'm
3:31
uh you know almost a year ago and I'm
3:31
uh you know almost a year ago and I'm like why would you do that? The other
3:33
like why would you do that? The other
3:33
like why would you do that? The other name was so nice. it was so Microsoft
3:35
name was so nice. it was so Microsoft
3:35
name was so nice. it was so Microsoft style and then they came up with a name
3:37
style and then they came up with a name
3:37
style and then they came up with a name that no one actually understands how to
3:39
that no one actually understands how to
3:39
that no one actually understands how to say it which is really really bad. But
3:42
say it which is really really bad. But
3:42
say it which is really really bad. But anyway, let's ignore that part and let's
3:44
anyway, let's ignore that part and let's
3:44
anyway, let's ignore that part and let's let's show some demos. And pretty much
3:48
let's show some demos. And pretty much
3:48
let's show some demos. And pretty much uh the only demos that I have are
3:50
uh the only demos that I have are
3:50
uh the only demos that I have are similar to this one. This seems to be a
3:52
similar to this one. This seems to be a
3:52
similar to this one. This seems to be a real sandwich, but don't trust AI.
3:55
real sandwich, but don't trust AI.
3:55
real sandwich, but don't trust AI. Actually, this is real. If you ask AI,
3:57
Actually, this is real. If you ask AI,
3:57
Actually, this is real. If you ask AI, they will say that this is a perfectly
3:59
they will say that this is a perfectly
3:59
they will say that this is a perfectly normal sandwich. Okay? because it has
4:01
normal sandwich. Okay? because it has
4:01
normal sandwich. Okay? because it has breads and then ham and then cheese or
4:04
breads and then ham and then cheese or
4:04
breads and then ham and then cheese or vice versa and then um you know bread
4:06
vice versa and then um you know bread
4:06
vice versa and then um you know bread again. So technically this is a sandwich
4:09
again. So technically this is a sandwich
4:09
again. So technically this is a sandwich although it's not very practical for a
4:11
although it's not very practical for a
4:11
although it's not very practical for a human being to eat it. So we have to
4:13
human being to eat it. So we have to
4:13
human being to eat it. So we have to understand the difference between AI
4:16
understand the difference between AI
4:16
understand the difference between AI using AI things that feel normal for AI
4:19
using AI things that feel normal for AI
4:19
using AI things that feel normal for AI and things that feel normal for us. But
4:21
and things that feel normal for us. But
4:21
and things that feel normal for us. But let's let's first understand how these
4:24
let's let's first understand how these
4:24
let's let's first understand how these things work and let's let's do this
4:27
things work and let's let's do this
4:27
things work and let's let's do this interactively. So let's uh let's talk
4:30
interactively. So let's uh let's talk
4:30
interactively. So let's uh let's talk about all of this. So uh I think you've
4:32
about all of this. So uh I think you've
4:32
about all of this. So uh I think you've had sessions before about uh about Azour
4:35
had sessions before about uh about Azour
4:35
had sessions before about uh about Azour probably the Azour portal. Many people
4:37
probably the Azour portal. Many people
4:37
probably the Azour portal. Many people already have this and I'm sure you have
4:39
already have this and I'm sure you have
4:40
already have this and I'm sure you have a lot of students that can actually use
4:42
a lot of students that can actually use
4:42
a lot of students that can actually use this. So, if you go for Azour uh free uh
4:47
this. So, if you go for Azour uh free uh
4:47
this. So, if you go for Azour uh free uh credits to something, if you search for
4:49
credits to something, if you search for
4:49
credits to something, if you search for something like this, you'll find a way
4:51
something like this, you'll find a way
4:52
something like this, you'll find a way to get some free credits um that are
4:54
to get some free credits um that are
4:54
to get some free credits um that are around $100 per year that you can use
4:58
around $100 per year that you can use
4:58
around $100 per year that you can use specifically the things that I'm going
4:59
specifically the things that I'm going
4:59
specifically the things that I'm going to show you. Now, once you get to the
5:01
to show you. Now, once you get to the
5:01
to show you. Now, once you get to the Azor portal at portal.zor.com,
5:04
Azor portal at portal.zor.com,
5:04
Azor portal at portal.zor.com, uh you we can start to use AI to do
5:07
uh you we can start to use AI to do
5:07
uh you we can start to use AI to do that. We can go and search for AI
5:09
that. We can go and search for AI
5:09
that. We can go and search for AI services or AI foundry. Let's search for
5:12
services or AI foundry. Let's search for
5:12
services or AI foundry. Let's search for AI foundry now because that's the new
5:15
AI foundry now because that's the new
5:15
AI foundry now because that's the new name of everything AI services related.
5:18
name of everything AI services related.
5:18
name of everything AI services related. AI services what I
5:20
AI services what I
5:20
AI services what I >> I still remember Andy know back in the
5:22
>> I still remember Andy know back in the
5:22
>> I still remember Andy know back in the days 2014 it was called machine learning
5:25
days 2014 it was called machine learning
5:25
days 2014 it was called machine learning studio but not right so long years back
5:29
studio but not right so long years back
5:29
studio but not right so long years back >> uh I think you can still use that. So
5:32
>> uh I think you can still use that. So
5:32
>> uh I think you can still use that. So machine learning studio is still here. I
5:34
machine learning studio is still here. I
5:34
machine learning studio is still here. I still have machine learning studios but
5:36
still have machine learning studios but
5:36
still have machine learning studios but now it's getting more and more
5:38
now it's getting more and more
5:38
now it's getting more and more integrated with AI foundry as well. I
5:40
integrated with AI foundry as well. I
5:40
integrated with AI foundry as well. I don't think it's perfect yet because
5:43
don't think it's perfect yet because
5:43
don't think it's perfect yet because this still requires a bit of work but
5:45
this still requires a bit of work but
5:45
this still requires a bit of work but machine learning studio is going to be
5:48
machine learning studio is going to be
5:48
machine learning studio is going to be integrated with AI foundry and at some
5:51
integrated with AI foundry and at some
5:51
integrated with AI foundry and at some point in time it will all be you know
5:53
point in time it will all be you know
5:53
point in time it will all be you know integrated into it the same way that you
5:55
integrated into it the same way that you
5:56
integrated into it the same way that you have language services and uh a few
5:58
have language services and uh a few
5:58
have language services and uh a few others face computer vision etc. Some of
6:01
others face computer vision etc. Some of
6:01
others face computer vision etc. Some of these are already there and you can use
6:04
these are already there and you can use
6:04
these are already there and you can use them directly into the AI foundry and I
6:07
them directly into the AI foundry and I
6:07
them directly into the AI foundry and I keep trying to say AI studio but okay.
6:11
keep trying to say AI studio but okay.
6:11
keep trying to say AI studio but okay. So what do we need to start this? Well,
6:13
So what do we need to start this? Well,
6:13
So what do we need to start this? Well, we need to go here and create an object
6:17
we need to go here and create an object
6:17
we need to go here and create an object similar to some of these. Um some some
6:21
similar to some of these. Um some some
6:21
similar to some of these. Um some some things apply. So the subscription is
6:24
things apply. So the subscription is
6:24
things apply. So the subscription is whatever uh you have there that pays for
6:26
whatever uh you have there that pays for
6:26
whatever uh you have there that pays for things. The resource group is whatever
6:28
things. The resource group is whatever
6:28
things. The resource group is whatever you want to create. Just it's like a box
6:31
you want to create. Just it's like a box
6:31
you want to create. Just it's like a box to hold things. So, it's easy enough.
6:33
to hold things. So, it's easy enough.
6:33
to hold things. So, it's easy enough. Give it a name. Give it a region. There
6:35
Give it a name. Give it a region. There
6:35
Give it a name. Give it a region. There are some regions that make more sense
6:37
are some regions that make more sense
6:37
are some regions that make more sense than others. I would say that Sweden
6:40
than others. I would say that Sweden
6:40
than others. I would say that Sweden central is currently the best region
6:42
central is currently the best region
6:42
central is currently the best region because it has everything that you want
6:44
because it has everything that you want
6:44
because it has everything that you want in terms of AI. Well, not everything.
6:47
in terms of AI. Well, not everything.
6:47
in terms of AI. Well, not everything. Sora is not available there. Something
6:49
Sora is not available there. Something
6:49
Sora is not available there. Something it's I don't know, but it's one of the
6:52
it's I don't know, but it's one of the
6:52
it's I don't know, but it's one of the best. And then you have to figure out
6:54
best. And then you have to figure out
6:54
best. And then you have to figure out where everything else is. Um there's uh
6:57
where everything else is. Um there's uh
6:57
where everything else is. Um there's uh one or two more regions that have
6:59
one or two more regions that have
6:59
one or two more regions that have whatever's left, but Sweden central is
7:01
whatever's left, but Sweden central is
7:02
whatever's left, but Sweden central is the best for you because mostly it's on
7:04
the best for you because mostly it's on
7:04
the best for you because mostly it's on top of Europe. No one actually cares.
7:05
top of Europe. No one actually cares.
7:05
top of Europe. No one actually cares. It's very far away. Really cold by the
7:07
It's very far away. Really cold by the
7:07
It's very far away. Really cold by the way also. And uh pretty much empty, so
7:10
way also. And uh pretty much empty, so
7:10
way also. And uh pretty much empty, so everything AI is actually there. So it
7:13
everything AI is actually there. So it
7:13
everything AI is actually there. So it doesn't occupy any other resources that
7:15
doesn't occupy any other resources that
7:15
doesn't occupy any other resources that we need for real world things. Anyway,
7:18
we need for real world things. Anyway,
7:18
we need for real world things. Anyway, you choose the region pretty much. You
7:19
you choose the region pretty much. You
7:19
you choose the region pretty much. You click on next a whole bunch of times
7:21
click on next a whole bunch of times
7:21
click on next a whole bunch of times until you accept it. I'm not going to go
7:23
until you accept it. I'm not going to go
7:23
until you accept it. I'm not going to go into security stuff like uh networking
7:27
into security stuff like uh networking
7:27
into security stuff like uh networking etc. Let's just assume that this would
7:28
etc. Let's just assume that this would
7:28
etc. Let's just assume that this would be for a a different session altogether
7:31
be for a a different session altogether
7:31
be for a a different session altogether but please be careful with it of course.
7:33
but please be careful with it of course.
7:33
but please be careful with it of course. And then once you are done you have uh
7:36
And then once you are done you have uh
7:36
And then once you are done you have uh the uh the objects here. So these
7:39
the uh the objects here. So these
7:39
the uh the objects here. So these services these are still called AI
7:41
services these are still called AI
7:41
services these are still called AI services and this is not going to change
7:43
services and this is not going to change
7:44
services and this is not going to change because that kind thing is an internal
7:47
because that kind thing is an internal
7:47
because that kind thing is an internal definition which they cannot change. So
7:48
definition which they cannot change. So
7:48
definition which they cannot change. So they will not going to rename it to AI
7:50
they will not going to rename it to AI
7:50
they will not going to rename it to AI foundaries. So that's fine.
7:52
foundaries. So that's fine.
7:52
foundaries. So that's fine. I like they're keeping something from
7:54
I like they're keeping something from
7:54
I like they're keeping something from the legacy.
7:56
the legacy.
7:56
the legacy. >> So yeah, you can also go directly to
7:58
>> So yeah, you can also go directly to
7:58
>> So yeah, you can also go directly to Azor OpenAI and create some of these
8:00
Azor OpenAI and create some of these
8:00
Azor OpenAI and create some of these OpenAI ones which I have created a few.
8:03
OpenAI ones which I have created a few.
8:03
OpenAI ones which I have created a few. The limitations with this one is Azour
8:05
The limitations with this one is Azour
8:05
The limitations with this one is Azour OpenAI services only allow you to see
8:08
OpenAI services only allow you to see
8:08
OpenAI services only allow you to see models that are OpenAI. Okay, so that is
8:11
models that are OpenAI. Okay, so that is
8:11
models that are OpenAI. Okay, so that is a limitation. Uh now um with AI foundry
8:14
a limitation. Uh now um with AI foundry
8:14
a limitation. Uh now um with AI foundry this is easy. You can choose one and
8:16
this is easy. You can choose one and
8:16
this is easy. You can choose one and then eventually you either click
8:19
then eventually you either click
8:19
then eventually you either click somewhere here if this pops up you click
8:22
somewhere here if this pops up you click
8:22
somewhere here if this pops up you click on one of these go to AI foundry portal
8:25
on one of these go to AI foundry portal
8:25
on one of these go to AI foundry portal or you just go directly to ai.zour.com.
8:28
or you just go directly to ai.zour.com.
8:28
or you just go directly to ai.zour.com. Now, uh, some considerations. First of
8:30
Now, uh, some considerations. First of
8:30
Now, uh, some considerations. First of all, if you have more than one of these
8:32
all, if you have more than one of these
8:32
all, if you have more than one of these services, you can choose different
8:34
services, you can choose different
8:34
services, you can choose different directories, which typically you only
8:36
directories, which typically you only
8:36
directories, which typically you only have one. Um, so this is your tenant.
8:39
have one. Um, so this is your tenant.
8:39
have one. Um, so this is your tenant. You typically only have one from one
8:41
You typically only have one from one
8:41
You typically only have one from one company subscription and then working
8:43
company subscription and then working
8:43
company subscription and then working in. And I have quite a few. I'm not
8:45
in. And I have quite a few. I'm not
8:45
in. And I have quite a few. I'm not going to change the current one. I know
8:47
going to change the current one. I know
8:47
going to change the current one. I know I'm in the right one, so I'm not going
8:49
I'm in the right one, so I'm not going
8:49
I'm in the right one, so I'm not going to touch it, otherwise I'm going to
8:51
to touch it, otherwise I'm going to
8:51
to touch it, otherwise I'm going to screw up something. Now,
8:54
screw up something. Now,
8:54
screw up something. Now, important things. Um when you get here
8:58
important things. Um when you get here
8:58
important things. Um when you get here uh you know that you are in the Azour
9:00
uh you know that you are in the Azour
9:00
uh you know that you are in the Azour OpenAI um you know uh service if you
9:04
OpenAI um you know uh service if you
9:04
OpenAI um you know uh service if you only see something like 30 or 40 models
9:07
only see something like 30 or 40 models
9:07
only see something like 30 or 40 models here. Okay. If you see something like
9:09
here. Okay. If you see something like
9:09
here. Okay. If you see something like 11,000 that means you are in the generic
9:12
11,000 that means you are in the generic
9:12
11,000 that means you are in the generic let's go back in the generic where is
9:14
let's go back in the generic where is
9:14
let's go back in the generic where is this AI foundry or AI services model
9:18
this AI foundry or AI services model
9:18
this AI foundry or AI services model which pretty much accepts anything. So
9:20
which pretty much accepts anything. So
9:20
which pretty much accepts anything. So that's the one you always want to go
9:22
that's the one you always want to go
9:22
that's the one you always want to go for. So here you have a lot of different
9:25
for. So here you have a lot of different
9:25
for. So here you have a lot of different types of models. These are models like
9:27
types of models. These are models like
9:27
types of models. These are models like Omni3 uh 03. These are DeepSeek. You
9:31
Omni3 uh 03. These are DeepSeek. You
9:31
Omni3 uh 03. These are DeepSeek. You probably heard of this. Uh Sora for
9:33
probably heard of this. Uh Sora for
9:33
probably heard of this. Uh Sora for video generation. Um Grock. So Grock is
9:37
video generation. Um Grock. So Grock is
9:37
video generation. Um Grock. So Grock is like the ugliest name you can find for a
9:40
like the ugliest name you can find for a
9:40
like the ugliest name you can find for a uh for large language model and probably
9:43
uh for large language model and probably
9:43
uh for large language model and probably invented by by Ellen Musk which uh
9:47
invented by by Ellen Musk which uh
9:47
invented by by Ellen Musk which uh nowadays it's a groggy character as we
9:49
nowadays it's a groggy character as we
9:49
nowadays it's a groggy character as we know. Um, some people say that this is a
9:53
know. Um, some people say that this is a
9:53
know. Um, some people say that this is a very good model. I tend to disagree.
9:56
very good model. I tend to disagree.
9:56
very good model. I tend to disagree. Apparently, according to the trends,
9:57
Apparently, according to the trends,
9:57
Apparently, according to the trends, this is a fast model, but it's not a
10:00
this is a fast model, but it's not a
10:00
this is a fast model, but it's not a good model in terms of contents. Okay. I
10:03
good model in terms of contents. Okay. I
10:04
good model in terms of contents. Okay. I would say the same for Llama from
10:05
would say the same for Llama from
10:05
would say the same for Llama from Facebook. Facebook or Meta if you
10:07
Facebook. Facebook or Meta if you
10:07
Facebook. Facebook or Meta if you prefer. So, it's here. Llama is here as
10:10
prefer. So, it's here. Llama is here as
10:10
prefer. So, it's here. Llama is here as well. Why is this uh a bigger problem?
10:13
well. Why is this uh a bigger problem?
10:13
well. Why is this uh a bigger problem? Because I don't think that they trained
10:14
Because I don't think that they trained
10:14
Because I don't think that they trained them correctly. Okay. They train them on
10:17
them correctly. Okay. They train them on
10:17
them correctly. Okay. They train them on the information from Twitter, from
10:19
the information from Twitter, from
10:19
the information from Twitter, from Facebook, whichever a bit more of
10:20
Facebook, whichever a bit more of
10:20
Facebook, whichever a bit more of course, but um not all of that
10:23
course, but um not all of that
10:23
course, but um not all of that information is ideal in terms of text,
10:25
information is ideal in terms of text,
10:25
information is ideal in terms of text, in terms of content. So it is quite
10:27
in terms of content. So it is quite
10:28
in terms of content. So it is quite biased in terms of you know, in terms of
10:31
biased in terms of you know, in terms of
10:31
biased in terms of you know, in terms of conspiracy theories, etc. Um so it's not
10:35
conspiracy theories, etc. Um so it's not
10:35
conspiracy theories, etc. Um so it's not perfect. So I would say that both of
10:37
perfect. So I would say that both of
10:37
perfect. So I would say that both of these types of models, and there's a
10:39
these types of models, and there's a
10:39
these types of models, and there's a lot, there's a lot of GRs, there's a lot
10:41
lot, there's a lot of GRs, there's a lot
10:41
lot, there's a lot of GRs, there's a lot of llamas, um these are not ideal. I
10:44
of llamas, um these are not ideal. I
10:44
of llamas, um these are not ideal. I would still prefer open AAI. So GP
10:46
would still prefer open AAI. So GP
10:46
would still prefer open AAI. So GP models like um like uh um GPT 4.1 which
10:51
models like um like uh um GPT 4.1 which
10:51
models like um like uh um GPT 4.1 which is trending at the moment. 03 there was
10:53
is trending at the moment. 03 there was
10:53
is trending at the moment. 03 there was 4.5 which is crazy expensive. So still
10:57
4.5 which is crazy expensive. So still
10:57
4.5 which is crazy expensive. So still not uh currently ideal. Let's put it
11:00
not uh currently ideal. Let's put it
11:00
not uh currently ideal. Let's put it like this. Uh but maybe in the future.
11:04
like this. Uh but maybe in the future.
11:04
like this. Uh but maybe in the future. So anyway,
11:05
So anyway,
11:05
So anyway, >> no 11,000 is not it's not a small
11:07
>> no 11,000 is not it's not a small
11:07
>> no 11,000 is not it's not a small number. 11,000 is too big, right? I mean
11:10
number. 11,000 is too big, right? I mean
11:10
number. 11,000 is too big, right? I mean it's I'm glad you know you you you put
11:12
it's I'm glad you know you you you put
11:12
it's I'm glad you know you you you put out the difference between the open AI
11:14
out the difference between the open AI
11:14
out the difference between the open AI services that you have in the Azure and
11:16
services that you have in the Azure and
11:16
services that you have in the Azure and the Azure foundry because someone who's
11:18
the Azure foundry because someone who's
11:18
the Azure foundry because someone who's getting started right they may get
11:20
getting started right they may get
11:20
getting started right they may get confused right so good to know that uh
11:23
confused right so good to know that uh
11:23
confused right so good to know that uh yeah and yeah group three I think it was
11:25
yeah and yeah group three I think it was
11:25
yeah and yeah group three I think it was uh recently you know added into the
11:27
uh recently you know added into the
11:27
uh recently you know added into the Azure foundry I saw it was a conference
11:30
Azure foundry I saw it was a conference
11:30
Azure foundry I saw it was a conference I think ignite or build uh Musk and SATA
11:33
I think ignite or build uh Musk and SATA
11:33
I think ignite or build uh Musk and SATA they were both on the same screen they
11:35
they were both on the same screen they
11:35
they were both on the same screen they were talking about you know it is coming
11:37
were talking about you know it is coming
11:37
were talking about you know it is coming to Azure foundry So yeah, you're right.
11:39
to Azure foundry So yeah, you're right.
11:39
to Azure foundry So yeah, you're right. Can we by any chance also find Gemini? I
11:41
Can we by any chance also find Gemini? I
11:41
Can we by any chance also find Gemini? I don't see it in the in the list here.
11:44
don't see it in the in the list here.
11:44
don't see it in the in the list here. >> Find uh Gemini.
11:46
>> Find uh Gemini.
11:46
>> Find uh Gemini. >> I don't think so. Gemini,
11:48
>> I don't think so. Gemini,
11:48
>> I don't think so. Gemini, >> that would be interesting if Gemini is
11:50
>> that would be interesting if Gemini is
11:50
>> that would be interesting if Gemini is not there.
11:51
not there.
11:51
not there. >> I don't think so. The the thing is uh
11:54
>> I don't think so. The the thing is uh
11:54
>> I don't think so. The the thing is uh and we need to understand how this
11:55
and we need to understand how this
11:55
and we need to understand how this works. So these are large language
11:57
works. So these are large language
11:57
works. So these are large language models. Uh there are also small language
12:00
models. Uh there are also small language
12:00
models. Uh there are also small language models like fi uh fi models like fi3
12:05
models like fi uh fi models like fi3
12:05
models like fi uh fi models like fi3 etc. Maybe 54 which is a bit bigger. uh
12:08
etc. Maybe 54 which is a bit bigger. uh
12:08
etc. Maybe 54 which is a bit bigger. uh but these are language models for
12:09
but these are language models for
12:09
but these are language models for instance from Microsoft 3.5 etc. These
12:12
instance from Microsoft 3.5 etc. These
12:12
instance from Microsoft 3.5 etc. These are small. So these would fit in some
12:15
are small. So these would fit in some
12:15
are small. So these would fit in some some IoT device maybe like one of these
12:18
some IoT device maybe like one of these
12:18
some IoT device maybe like one of these Raspberry Pies something like that. This
12:21
Raspberry Pies something like that. This
12:21
Raspberry Pies something like that. This would actually fit on some of these.
12:25
would actually fit on some of these.
12:25
would actually fit on some of these. >> Okay, that's good to know. That's good
12:26
>> Okay, that's good to know. That's good
12:26
>> Okay, that's good to know. That's good to know. I mean if you have all these
12:28
to know. I mean if you have all these
12:28
to know. I mean if you have all these models running on the edge devices,
12:30
models running on the edge devices,
12:30
models running on the edge devices, right, the possibilities are endless,
12:32
right, the possibilities are endless,
12:32
right, the possibilities are endless, right? And I don't think you even these
12:34
right? And I don't think you even these
12:34
right? And I don't think you even these days you need a continuous network
12:36
days you need a continuous network
12:36
days you need a continuous network connection to run this. Once you have it
12:38
connection to run this. Once you have it
12:38
connection to run this. Once you have it on your edge devices, you know, you can
12:40
on your edge devices, you know, you can
12:40
on your edge devices, you know, you can do wonders.
12:41
do wonders.
12:41
do wonders. >> Exactly. So if they are small enough
12:43
>> Exactly. So if they are small enough
12:43
>> Exactly. So if they are small enough like 53 uh 54 is a bit more complicated
12:46
like 53 uh 54 is a bit more complicated
12:46
like 53 uh 54 is a bit more complicated because it's much bigger but anything
12:49
because it's much bigger but anything
12:49
because it's much bigger but anything small you can actually download you put
12:51
small you can actually download you put
12:51
small you can actually download you put it in an app on your phone something
12:53
it in an app on your phone something
12:53
it in an app on your phone something like that that would be fine. Uh but if
12:55
like that that would be fine. Uh but if
12:55
like that that would be fine. Uh but if it is big like GP4 etc these would need
13:00
it is big like GP4 etc these would need
13:00
it is big like GP4 etc these would need not even a computer at home but you
13:02
not even a computer at home but you
13:02
not even a computer at home but you would actually need something in the
13:04
would actually need something in the
13:04
would actually need something in the cloud because you are in the area of
13:06
cloud because you are in the area of
13:06
cloud because you are in the area of hundreds of servers uh you know
13:09
hundreds of servers uh you know
13:09
hundreds of servers uh you know answering questions for you at the same
13:10
answering questions for you at the same
13:10
answering questions for you at the same time. So you need big computational
13:12
time. So you need big computational
13:12
time. So you need big computational power. Um the way this works is you have
13:16
power. Um the way this works is you have
13:16
power. Um the way this works is you have um so obviously I'm I need to unfilter
13:19
um so obviously I'm I need to unfilter
13:19
um so obviously I'm I need to unfilter this. Uh you have 100,000 models and
13:22
this. Uh you have 100,000 models and
13:22
this. Uh you have 100,000 models and these are not just large language
13:24
these are not just large language
13:24
these are not just large language models. Uh you also have Dolly. You also
13:27
models. Uh you also have Dolly. You also
13:27
models. Uh you also have Dolly. You also have a few others. Oh, Dolly, etc. By
13:31
have a few others. Oh, Dolly, etc. By
13:31
have a few others. Oh, Dolly, etc. By the way, this is beeping because this is
13:33
the way, this is beeping because this is
13:34
the way, this is beeping because this is my caps lock and I memorized
13:36
my caps lock and I memorized
13:36
my caps lock and I memorized >> the audio. Uh but no because I I want to
13:40
>> the audio. Uh but no because I I want to
13:40
>> the audio. Uh but no because I I want to use it in a second for a demo. Uh so if
13:43
use it in a second for a demo. Uh so if
13:43
use it in a second for a demo. Uh so if you hear it, sorry about that, but uh
13:45
you hear it, sorry about that, but uh
13:45
you hear it, sorry about that, but uh it's only meant to be done in a second.
13:48
it's only meant to be done in a second.
13:48
it's only meant to be done in a second. Okay. But anyway, Dali, you also have
13:50
Okay. But anyway, Dali, you also have
13:50
Okay. But anyway, Dali, you also have Dali, you also have Sora, you have a few
13:53
Dali, you also have Sora, you have a few
13:53
Dali, you also have Sora, you have a few others. Sor is for generating video. So
13:56
others. Sor is for generating video. So
13:56
others. Sor is for generating video. So let's actually have a look at some of
13:58
let's actually have a look at some of
13:58
let's actually have a look at some of these. And we can uh first start by
14:00
these. And we can uh first start by
14:00
these. And we can uh first start by looking at the playground. And then you
14:02
looking at the playground. And then you
14:02
looking at the playground. And then you can click on each of these models. Let's
14:03
can click on each of these models. Let's
14:04
can click on each of these models. Let's try Dolli first for instance. Um, I've
14:07
try Dolli first for instance. Um, I've
14:07
try Dolli first for instance. Um, I've already generated images and a few
14:10
already generated images and a few
14:10
already generated images and a few things because this would take a bit of
14:11
things because this would take a bit of
14:11
things because this would take a bit of time and we don't want to waste that
14:13
time and we don't want to waste that
14:13
time and we don't want to waste that much time. Uh, so a few things to
14:16
much time. Uh, so a few things to
14:16
much time. Uh, so a few things to consider. There are going to be models
14:17
consider. There are going to be models
14:18
consider. There are going to be models that are going to retire this. I'm still
14:20
that are going to retire this. I'm still
14:20
that are going to retire this. I'm still not sure which one would be the best
14:22
not sure which one would be the best
14:22
not sure which one would be the best recommended one because there are GPT
14:25
recommended one because there are GPT
14:25
recommended one because there are GPT models like GPT4.1 and a few others that
14:27
models like GPT4.1 and a few others that
14:27
models like GPT4.1 and a few others that are the D the image ones that are being
14:30
are the D the image ones that are being
14:30
are the D the image ones that are being considered as a replacement. I'm not
14:32
considered as a replacement. I'm not
14:32
considered as a replacement. I'm not sure but you can actually learn about uh
14:35
sure but you can actually learn about uh
14:35
sure but you can actually learn about uh the deployments etc. you looking at the
14:38
the deployments etc. you looking at the
14:38
the deployments etc. you looking at the documentation let's not do that. Yeah,
14:41
documentation let's not do that. Yeah,
14:41
documentation let's not do that. Yeah, that's good to know. It's good to know
14:43
that's good to know. It's good to know
14:43
that's good to know. It's good to know they have a retirement date. And you
14:44
they have a retirement date. And you
14:44
they have a retirement date. And you know what I also loved, Andy, from your
14:46
know what I also loved, Andy, from your
14:46
know what I also loved, Andy, from your previous screen, what was really
14:48
previous screen, what was really
14:48
previous screen, what was really tempting to me was at at somewhere at
14:50
tempting to me was at at somewhere at
14:50
tempting to me was at at somewhere at the corner, right? You had compare
14:52
the corner, right? You had compare
14:52
the corner, right? You had compare models, right?
14:53
models, right?
14:54
models, right? >> Oh, yeah. Let's let's go for that if you
14:55
>> Oh, yeah. Let's let's go for that if you
14:55
>> Oh, yeah. Let's let's go for that if you want to. Okay,
14:56
want to. Okay,
14:56
want to. Okay, >> that was very tempting. You know, I
14:57
>> that was very tempting. You know, I
14:57
>> that was very tempting. You know, I could compare models with just a click.
14:58
could compare models with just a click.
14:58
could compare models with just a click. That would be so nice.
15:00
That would be so nice.
15:00
That would be so nice. >> Okay, let's let's go down a bit. So,
15:03
>> Okay, let's let's go down a bit. So,
15:03
>> Okay, let's let's go down a bit. So, actually, yeah. Uh, so model
15:05
actually, yeah. Uh, so model
15:05
actually, yeah. Uh, so model leaderboard. So you have quite a few of
15:07
leaderboard. So you have quite a few of
15:07
leaderboard. So you have quite a few of these and then you can choose a few
15:09
these and then you can choose a few
15:09
these and then you can choose a few models and say compare. Now compare is
15:11
models and say compare. Now compare is
15:11
models and say compare. Now compare is not perfect because you can start with a
15:13
not perfect because you can start with a
15:14
not perfect because you can start with a few models here and they will give you
15:16
few models here and they will give you
15:16
few models here and they will give you some information about so this one is
15:18
some information about so this one is
15:18
some information about so this one is quality for instance this one is costs
15:21
quality for instance this one is costs
15:21
quality for instance this one is costs and they will tell you which one is nice
15:23
and they will tell you which one is nice
15:23
and they will tell you which one is nice for certain things but let's say
15:25
for certain things but let's say
15:25
for certain things but let's say question answering so some of these are
15:28
question answering so some of these are
15:28
question answering so some of these are here in terms of cost so mostly cheap
15:30
here in terms of cost so mostly cheap
15:30
here in terms of cost so mostly cheap but then in terms of quality the ones on
15:33
but then in terms of quality the ones on
15:33
but then in terms of quality the ones on top would be 03 from from from openai 04
15:38
top would be 03 from from from openai 04
15:38
top would be 03 from from from openai 04 mini here it's 03 pro So you can see
15:41
mini here it's 03 pro So you can see
15:41
mini here it's 03 pro So you can see that the best ones in terms of quality
15:42
that the best ones in terms of quality
15:42
that the best ones in terms of quality are still open AI. Deepseek is here. I
15:45
are still open AI. Deepseek is here. I
15:45
are still open AI. Deepseek is here. I have no idea where Grock is, but Llama
15:48
have no idea where Grock is, but Llama
15:48
have no idea where Grock is, but Llama is down here. GT4.5 is incredibly
15:51
is down here. GT4.5 is incredibly
15:51
is down here. GT4.5 is incredibly expensive. That's why it's in that
15:52
expensive. That's why it's in that
15:52
expensive. That's why it's in that corner. So this is expected. And maybe a
15:56
corner. So this is expected. And maybe a
15:56
corner. So this is expected. And maybe a few others. Okay. So let's No idea where
15:59
few others. Okay. So let's No idea where
15:59
few others. Okay. So let's No idea where Grock is. If it is even here on the list
16:01
Grock is. If it is even here on the list
16:01
Grock is. If it is even here on the list at all. I have no idea. So but anyway,
16:04
at all. I have no idea. So but anyway,
16:04
at all. I have no idea. So but anyway, uh let's ignore it. So but you can
16:06
uh let's ignore it. So but you can
16:06
uh let's ignore it. So but you can actually go and compare. You can compare
16:08
actually go and compare. You can compare
16:08
actually go and compare. You can compare other things here. You can compare other
16:10
other things here. You can compare other
16:10
other things here. You can compare other things here. So you can analyze which
16:13
things here. So you can analyze which
16:13
things here. So you can analyze which one is best. This is previously done for
16:15
one is best. This is previously done for
16:15
one is best. This is previously done for you. So this is not something that you
16:17
you. So this is not something that you
16:17
you. So this is not something that you are trying out for the first time. They
16:19
are trying out for the first time. They
16:19
are trying out for the first time. They have this information already. But they
16:21
have this information already. But they
16:21
have this information already. But they will give you this information. Now very
16:23
will give you this information. Now very
16:23
will give you this information. Now very very quickly once you select a model you
16:26
very quickly once you select a model you
16:26
very quickly once you select a model you need to go down here and sometimes it
16:28
need to go down here and sometimes it
16:28
need to go down here and sometimes it has a different name. Sometimes it says
16:30
has a different name. Sometimes it says
16:30
has a different name. Sometimes it says models endpoint. Sometimes it says
16:32
models endpoint. Sometimes it says
16:32
models endpoint. Sometimes it says something slightly different because
16:34
something slightly different because
16:34
something slightly different because this environment on the left changes if
16:37
this environment on the left changes if
16:37
this environment on the left changes if it is your pure AI foundry kind of thing
16:40
it is your pure AI foundry kind of thing
16:40
it is your pure AI foundry kind of thing or Azour open AAI kind of thing. So they
16:43
or Azour open AAI kind of thing. So they
16:43
or Azour open AAI kind of thing. So they have slightly different names. They are
16:45
have slightly different names. They are
16:45
have slightly different names. They are in the process of uniforming this but
16:47
in the process of uniforming this but
16:47
in the process of uniforming this but it's not perfect. So I have a few models
16:50
it's not perfect. So I have a few models
16:50
it's not perfect. So I have a few models deployed. I have Sora here. Um I have uh
16:54
deployed. I have Sora here. Um I have uh
16:54
deployed. I have Sora here. Um I have uh GP4 Omni. I have a few others. Some of
16:57
GP4 Omni. I have a few others. Some of
16:57
GP4 Omni. I have a few others. Some of these say disabled because they've
16:59
these say disabled because they've
16:59
these say disabled because they've expired. they've been deprecated since.
17:02
expired. they've been deprecated since.
17:02
expired. they've been deprecated since. Um I also have 3.5 which should have
17:05
Um I also have 3.5 which should have
17:05
Um I also have 3.5 which should have been deprecated one week ago something
17:08
been deprecated one week ago something
17:08
been deprecated one week ago something like that but they extended this until
17:10
like that but they extended this until
17:10
like that but they extended this until the beginning of September. So you know
17:13
the beginning of September. So you know
17:13
the beginning of September. So you know uh we can try it if you want to but this
17:15
uh we can try it if you want to but this
17:15
uh we can try it if you want to but this is supposed to be deprecated
17:17
is supposed to be deprecated
17:17
is supposed to be deprecated in theory. Anyway you can say deploy
17:20
in theory. Anyway you can say deploy
17:20
in theory. Anyway you can say deploy model you choose the model and then you
17:22
model you choose the model and then you
17:22
model you choose the model and then you select limitations for the model like
17:24
select limitations for the model like
17:24
select limitations for the model like maximum number of tokens which is a
17:27
maximum number of tokens which is a
17:27
maximum number of tokens which is a token is not a word but it's part of a
17:28
token is not a word but it's part of a
17:28
token is not a word but it's part of a word. So you limit it and basically a
17:31
word. So you limit it and basically a
17:31
word. So you limit it and basically a token is how you pay for things. But now
17:33
token is how you pay for things. But now
17:33
token is how you pay for things. But now once you deploy these let's let's
17:35
once you deploy these let's let's
17:35
once you deploy these let's let's actually try them out. So we have a few
17:37
actually try them out. So we have a few
17:37
actually try them out. So we have a few playgrounds and again playgrounds might
17:39
playgrounds and again playgrounds might
17:39
playgrounds and again playgrounds might show up in a slightly different way
17:41
show up in a slightly different way
17:41
show up in a slightly different way depending on the branding for the AI
17:44
depending on the branding for the AI
17:44
depending on the branding for the AI foundry. Hopefully in the future it will
17:46
foundry. Hopefully in the future it will
17:46
foundry. Hopefully in the future it will be more standard but it's still a bit
17:48
be more standard but it's still a bit
17:48
be more standard but it's still a bit different because of the way that it
17:50
different because of the way that it
17:50
different because of the way that it comes from you know historical reasons.
17:53
comes from you know historical reasons.
17:53
comes from you know historical reasons. So we'll try the chat playground in the
17:55
So we'll try the chat playground in the
17:55
So we'll try the chat playground in the end and mention agents and chat because
17:57
end and mention agents and chat because
17:57
end and mention agents and chat because these are very similar. Um but let's
18:00
these are very similar. Um but let's
18:00
these are very similar. Um but let's let's first try images. So the images
18:02
let's first try images. So the images
18:02
let's first try images. So the images playground here. So I've done this
18:05
playground here. So I've done this
18:05
playground here. So I've done this before because most of these will take
18:07
before because most of these will take
18:07
before because most of these will take about 20 seconds as you see here. So I
18:11
about 20 seconds as you see here. So I
18:11
about 20 seconds as you see here. So I actually decided to create one that says
18:13
actually decided to create one that says
18:13
actually decided to create one that says elephant in the room to talk about
18:15
elephant in the room to talk about
18:15
elephant in the room to talk about pricing because pricing is scary. So let
18:18
pricing because pricing is scary. So let
18:18
pricing because pricing is scary. So let me mention a few things about pricing.
18:21
me mention a few things about pricing.
18:21
me mention a few things about pricing. So apparently it's like €2 or something
18:23
So apparently it's like €2 or something
18:23
So apparently it's like €2 or something to to uh to get something like 5 seconds
18:27
to to uh to get something like 5 seconds
18:27
to to uh to get something like 5 seconds of video with Sora. Uh not the best
18:30
of video with Sora. Uh not the best
18:30
of video with Sora. Uh not the best quality. It is a bit expensive and
18:32
quality. It is a bit expensive and
18:32
quality. It is a bit expensive and especially because you have to try it
18:34
especially because you have to try it
18:34
especially because you have to try it multiple times until you get whatever
18:36
multiple times until you get whatever
18:36
multiple times until you get whatever you want because the prompt needs to be
18:38
you want because the prompt needs to be
18:38
you want because the prompt needs to be very good. Uh if you want prompts like
18:41
very good. Uh if you want prompts like
18:41
very good. Uh if you want prompts like this, uh it is a bit cheaper, but it's
18:43
this, uh it is a bit cheaper, but it's
18:44
this, uh it is a bit cheaper, but it's still a bit of money. go to
18:45
still a bit of money. go to
18:45
still a bit of money. go to wzour.com/pricing
18:48
wzour.com/pricing
18:48
wzour.com/pricing and then search for you know OpenAI etc
18:51
and then search for you know OpenAI etc
18:51
and then search for you know OpenAI etc and try to find out how much it costs
18:53
and try to find out how much it costs
18:53
and try to find out how much it costs but this is the elephant in the room we
18:55
but this is the elephant in the room we
18:55
but this is the elephant in the room we need to talk about and to generate this
18:57
need to talk about and to generate this
18:57
need to talk about and to generate this you can see family guy here well kind of
19:00
you can see family guy here well kind of
19:00
you can see family guy here well kind of anyway
19:01
anyway
19:01
anyway you can even see a kangaroo there but
19:04
you can even see a kangaroo there but
19:04
you can even see a kangaroo there but what I used was create a family guy
19:06
what I used was create a family guy
19:06
what I used was create a family guy style cartoon depicting elephant in the
19:09
style cartoon depicting elephant in the
19:09
style cartoon depicting elephant in the room with elephant hanging in reverse
19:11
room with elephant hanging in reverse
19:11
room with elephant hanging in reverse precarious position from the chandelier
19:13
precarious position from the chandelier
19:14
precarious position from the chandelier it didn't do a very good job at it, but
19:16
it didn't do a very good job at it, but
19:16
it didn't do a very good job at it, but let's ignore it. Uh, anyway, I tried
19:18
let's ignore it. Uh, anyway, I tried
19:18
let's ignore it. Uh, anyway, I tried this one as well. Didn't work that much.
19:20
this one as well. Didn't work that much.
19:20
this one as well. Didn't work that much. And I tried this one because I decided
19:22
And I tried this one because I decided
19:22
And I tried this one because I decided to do something very hateful, which is I
19:25
to do something very hateful, which is I
19:25
to do something very hateful, which is I decided to create something which um
19:27
decided to create something which um
19:28
decided to create something which um actually has copyrights. Okay, so this
19:30
actually has copyrights. Okay, so this
19:30
actually has copyrights. Okay, so this one has copyrights or or at least it
19:32
one has copyrights or or at least it
19:32
one has copyrights or or at least it did. This was u create a human
19:34
did. This was u create a human
19:34
did. This was u create a human equivalent of Steamboat Willie, which
19:37
equivalent of Steamboat Willie, which
19:37
equivalent of Steamboat Willie, which used to be the first name that they gave
19:39
used to be the first name that they gave
19:39
used to be the first name that they gave to a mouse commonly referred to as
19:42
to a mouse commonly referred to as
19:42
to a mouse commonly referred to as Mickey. Okay, that's how I'm saying it.
19:44
Mickey. Okay, that's how I'm saying it.
19:44
Mickey. Okay, that's how I'm saying it. So we don't get censored on YouTube or
19:47
So we don't get censored on YouTube or
19:47
So we don't get censored on YouTube or something like that. But uh nowadays
19:49
something like that. But uh nowadays
19:49
something like that. But uh nowadays this is out of copyright.
19:52
this is out of copyright.
19:52
this is out of copyright. >> Interesting.
19:54
>> Interesting.
19:54
>> Interesting. >> But this is the the possible result. I
19:56
>> But this is the the possible result. I
19:56
>> But this is the the possible result. I tried a few others that all got blocked
19:58
tried a few others that all got blocked
19:58
tried a few others that all got blocked because of copyright. And I've tried
20:00
because of copyright. And I've tried
20:00
because of copyright. And I've tried this one also because this one is a love
20:03
this one also because this one is a love
20:03
this one also because this one is a love lace in a L looney tunes kind of style
20:06
lace in a L looney tunes kind of style
20:06
lace in a L looney tunes kind of style with the first computer engines and
20:09
with the first computer engines and
20:09
with the first computer engines and things like that. Didn't go very well.
20:11
things like that. Didn't go very well.
20:11
things like that. Didn't go very well. She feels like a witch. But okay, let's
20:14
She feels like a witch. But okay, let's
20:14
She feels like a witch. But okay, let's >> I like the hairstyle. It's very vintage.
20:17
>> I like the hairstyle. It's very vintage.
20:17
>> I like the hairstyle. It's very vintage. >> It's very Princess Leia or something
20:19
>> It's very Princess Leia or something
20:19
>> It's very Princess Leia or something like that.
20:21
like that.
20:21
like that. >> But anyway, so images is fine. Uh we
20:24
>> But anyway, so images is fine. Uh we
20:24
>> But anyway, so images is fine. Uh we have a few others. So for speech, we
20:26
have a few others. So for speech, we
20:26
have a few others. So for speech, we also have a few others. And um you can
20:29
also have a few others. And um you can
20:29
also have a few others. And um you can go for speech playgrounds. I've decided
20:31
go for speech playgrounds. I've decided
20:31
go for speech playgrounds. I've decided to try the text speech avatar which is
20:34
to try the text speech avatar which is
20:34
to try the text speech avatar which is here. You can create your own avatar and
20:36
here. You can create your own avatar and
20:36
here. You can create your own avatar and change the image for something that you
20:38
change the image for something that you
20:38
change the image for something that you can create which is fine. But this one
20:41
can create which is fine. But this one
20:41
can create which is fine. But this one uh is already done for you and you can
20:42
uh is already done for you and you can
20:42
uh is already done for you and you can try it. So first of all you choose the
20:45
try it. So first of all you choose the
20:45
try it. So first of all you choose the one you want and because this takes a
20:47
one you want and because this takes a
20:47
one you want and because this takes a few seconds to generate. I've decided to
20:49
few seconds to generate. I've decided to
20:49
few seconds to generate. I've decided to generate it first. English Ireland Emily
20:52
generate it first. English Ireland Emily
20:52
generate it first. English Ireland Emily and then how are you blah blah blah. But
20:54
and then how are you blah blah blah. But
20:54
and then how are you blah blah blah. But adding gestures. So you have the button
20:56
adding gestures. So you have the button
20:56
adding gestures. So you have the button here to insert gesture. It's not
20:59
here to insert gesture. It's not
20:59
here to insert gesture. It's not available anymore because for some
21:01
available anymore because for some
21:01
available anymore because for some reason I've generated this. I need to
21:03
reason I've generated this. I need to
21:03
reason I've generated this. I need to refresh. I'm not going to do that
21:05
refresh. I'm not going to do that
21:05
refresh. I'm not going to do that otherwise this stops working. But let's
21:07
otherwise this stops working. But let's
21:07
otherwise this stops working. But let's see the generated video.
21:09
see the generated video.
21:09
see the generated video. So, let's see if this works.
21:11
So, let's see if this works.
21:11
So, let's see if this works. >> Hello, how are you?
21:13
>> Hello, how are you?
21:13
>> Hello, how are you? >> This is an example of an AI generated
21:16
>> This is an example of an AI generated
21:16
>> This is an example of an AI generated video with me showing some gestures.
21:20
>> How do I prompt? Should I just write
21:23
>> How do I prompt? Should I just write
21:23
>> How do I prompt? Should I just write here, you know, speak the sentence? And
21:24
here, you know, speak the sentence? And
21:24
here, you know, speak the sentence? And this is how how does this work?
21:27
this is how how does this work?
21:27
this is how how does this work? >> So, this one is just you say whatever
21:29
>> So, this one is just you say whatever
21:29
>> So, this one is just you say whatever you want the the avatar to say literally
21:31
you want the the avatar to say literally
21:31
you want the the avatar to say literally like this and then you add some gestures
21:34
like this and then you add some gestures
21:34
like this and then you add some gestures which maybe I can refresh by now. Um and
21:37
which maybe I can refresh by now. Um and
21:37
which maybe I can refresh by now. Um and then you add some gestures which is
21:39
then you add some gestures which is
21:39
then you add some gestures which is these you know kind of macro style
21:41
these you know kind of macro style
21:41
these you know kind of macro style things. Uh but basically it's just
21:43
things. Uh but basically it's just
21:43
things. Uh but basically it's just adding to it and then it moves around
21:46
adding to it and then it moves around
21:46
adding to it and then it moves around visually in terms in terms of image.
21:49
visually in terms in terms of image.
21:49
visually in terms in terms of image. It's not perfect. I would say that the
21:51
It's not perfect. I would say that the
21:51
It's not perfect. I would say that the most interesting one is um for instance
21:54
most interesting one is um for instance
21:54
most interesting one is um for instance let's say I say hello.
21:55
let's say I say hello.
21:55
let's say I say hello. >> Oh I see.
21:56
>> Oh I see.
21:56
>> Oh I see. >> Insert gesture. Uh break is also fine
21:58
>> Insert gesture. Uh break is also fine
21:58
>> Insert gesture. Uh break is also fine but insert gesture and then you can see
22:00
but insert gesture and then you can see
22:00
but insert gesture and then you can see some of these gestures. Uh and then you
22:03
some of these gestures. Uh and then you
22:03
some of these gestures. Uh and then you can say two. You can say one. There's
22:06
can say two. You can say one. There's
22:06
can say two. You can say one. There's some other gestures here and then you
22:08
some other gestures here and then you
22:08
some other gestures here and then you can add these. Okay, so this is all very
22:10
can add these. Okay, so this is all very
22:10
can add these. Okay, so this is all very fine. Okay, but uh you can also download
22:13
fine. Okay, but uh you can also download
22:13
fine. Okay, but uh you can also download the video by the way if you really want
22:15
the video by the way if you really want
22:15
the video by the way if you really want to. You can download the video but as
22:17
to. You can download the video but as
22:17
to. You can download the video but as soon as I refresh the video was lost.
22:19
soon as I refresh the video was lost.
22:19
soon as I refresh the video was lost. Now I have a copy of the video if I
22:21
Now I have a copy of the video if I
22:21
Now I have a copy of the video if I needed it but I didn't need it. Uh so
22:23
needed it but I didn't need it. Uh so
22:23
needed it but I didn't need it. Uh so anyway there is much more like realtime
22:25
anyway there is much more like realtime
22:25
anyway there is much more like realtime transcription. You receive audio, you
22:28
transcription. You receive audio, you
22:28
transcription. You receive audio, you translate the audio as audio as well or
22:31
translate the audio as audio as well or
22:31
translate the audio as audio as well or as text. It's up to you. So there's a
22:33
as text. It's up to you. So there's a
22:33
as text. It's up to you. So there's a whole ton of things that you can do
22:35
whole ton of things that you can do
22:35
whole ton of things that you can do here. Now, um there's documentation also
22:37
here. Now, um there's documentation also
22:37
here. Now, um there's documentation also in learn.microsoft.com and in a few
22:40
in learn.microsoft.com and in a few
22:40
in learn.microsoft.com and in a few other places. Uh because the the most
22:42
other places. Uh because the the most
22:42
other places. Uh because the the most interesting part about all of this is to
22:44
interesting part about all of this is to
22:44
interesting part about all of this is to be able to uh you know try this out. So
22:47
be able to uh you know try this out. So
22:47
be able to uh you know try this out. So view code here should be available for
22:49
view code here should be available for
22:49
view code here should be available for some reason in my case in this one is
22:52
some reason in my case in this one is
22:52
some reason in my case in this one is not available for but for all of the
22:54
not available for but for all of the
22:54
not available for but for all of the others it is. Maybe it's a bug when I
22:55
others it is. Maybe it's a bug when I
22:55
others it is. Maybe it's a bug when I refreshed. I don't know something like
22:57
refreshed. I don't know something like
22:57
refreshed. I don't know something like that. Now I want to show you a few more.
22:59
that. Now I want to show you a few more.
22:59
that. Now I want to show you a few more. So for things like um wait a minute
23:03
So for things like um wait a minute
23:03
So for things like um wait a minute playgrounds
23:04
playgrounds
23:04
playgrounds and actually it's not this one it's AI
23:07
and actually it's not this one it's AI
23:07
and actually it's not this one it's AI services. Give me a second. So for
23:09
services. Give me a second. So for
23:10
services. Give me a second. So for things like vision and this is not new
23:12
things like vision and this is not new
23:12
things like vision and this is not new this is old school Azour cognitive
23:15
this is old school Azour cognitive
23:15
this is old school Azour cognitive services. You can do things like process
23:17
services. You can do things like process
23:17
services. You can do things like process documents like invoices. This is the old
23:20
documents like invoices. This is the old
23:20
documents like invoices. This is the old uh forms recognizer which is now called
23:23
uh forms recognizer which is now called
23:23
uh forms recognizer which is now called document intelligence. You can do a lot
23:25
document intelligence. You can do a lot
23:25
document intelligence. You can do a lot of these with all of these kind of
23:27
of these with all of these kind of
23:27
of these with all of these kind of documents. You can process faces. Faces
23:30
documents. You can process faces. Faces
23:30
documents. You can process faces. Faces is nice. So, let me just go and try the
23:33
is nice. So, let me just go and try the
23:33
is nice. So, let me just go and try the demo. See if I it works. And then you
23:36
demo. See if I it works. And then you
23:36
demo. See if I it works. And then you can try out some of these faces. And
23:39
can try out some of these faces. And
23:39
can try out some of these faces. And just to make sure that this is real. So,
23:41
just to make sure that this is real. So,
23:41
just to make sure that this is real. So, we have Jason detects the face on a
23:44
we have Jason detects the face on a
23:44
we have Jason detects the face on a certain position in terms of pixels. If
23:47
certain position in terms of pixels. If
23:47
certain position in terms of pixels. If you don't believe me, I know that
23:49
you don't believe me, I know that
23:49
you don't believe me, I know that yesterday we had Aussie Osborne sadly
23:52
yesterday we had Aussie Osborne sadly
23:52
yesterday we had Aussie Osborne sadly pass away. So, we're just going to go
23:54
pass away. So, we're just going to go
23:54
pass away. So, we're just going to go and say copy image link. Hopefully this
23:57
and say copy image link. Hopefully this
23:57
and say copy image link. Hopefully this is fine and we are actually gonna try
24:00
is fine and we are actually gonna try
24:00
is fine and we are actually gonna try out and I know he usually has a very
24:03
out and I know he usually has a very
24:03
out and I know he usually has a very let's say weird face. Sometimes it's
24:06
let's say weird face. Sometimes it's
24:06
let's say weird face. Sometimes it's scary but let's let's try this out. And
24:11
scary but let's let's try this out. And
24:11
scary but let's let's try this out. And uh okay image is not valid. So give me a
24:14
uh okay image is not valid. So give me a
24:14
uh okay image is not valid. So give me a second. Let me try one of these uh
24:17
second. Let me try one of these uh
24:17
second. Let me try one of these uh images. So I need because these were
24:19
images. So I need because these were
24:19
images. So I need because these were internal to the page. So let's Okay,
24:21
internal to the page. So let's Okay,
24:22
internal to the page. So let's Okay, this one is fine. This one is amazing.
24:24
this one is fine. This one is amazing.
24:24
this one is fine. This one is amazing. Okay. So let's say copy image link.
24:27
Okay. So let's say copy image link.
24:27
Okay. So let's say copy image link. Okay. Detects
24:29
Okay. Detects
24:29
Okay. Detects uh browse for video files. Say yes.
24:33
uh browse for video files. Say yes.
24:33
uh browse for video files. Say yes. Enter. And then it takes a few seconds.
24:36
Enter. And then it takes a few seconds.
24:36
Enter. And then it takes a few seconds. And then we have a very scary evil image
24:40
And then we have a very scary evil image
24:40
And then we have a very scary evil image which has some weird pixels. I don't
24:42
which has some weird pixels. I don't
24:42
which has some weird pixels. I don't know why but it it seems to be true. So
24:45
know why but it it seems to be true. So
24:45
know why but it it seems to be true. So apparently it also works for vampires as
24:47
apparently it also works for vampires as
24:47
apparently it also works for vampires as well.
24:49
well.
24:49
well. Anyway, but you can detect images uh of
24:52
Anyway, but you can detect images uh of
24:52
Anyway, but you can detect images uh of people. you can detect um a few more
24:56
people. you can detect um a few more
24:56
people. you can detect um a few more things like objects. So you have a few
24:58
things like objects. So you have a few
24:58
things like objects. So you have a few more examples here of I don't know
25:01
more examples here of I don't know
25:01
more examples here of I don't know bicycles and a few other things. So you
25:03
bicycles and a few other things. So you
25:03
bicycles and a few other things. So you have a bicycle, you have wheels, you
25:05
have a bicycle, you have wheels, you
25:06
have a bicycle, you have wheels, you have bicycle wheel as well, etc. So you
25:08
have bicycle wheel as well, etc. So you
25:08
have bicycle wheel as well, etc. So you have all of these things that you can
25:10
have all of these things that you can
25:10
have all of these things that you can just upload and get more information.
25:13
just upload and get more information.
25:13
just upload and get more information. Now I want to to follow up with
25:15
Now I want to to follow up with
25:15
Now I want to to follow up with something which is the language
25:16
something which is the language
25:16
something which is the language playgrounds. So you can still do
25:19
playgrounds. So you can still do
25:19
playgrounds. So you can still do playground language playground and then
25:20
playground language playground and then
25:20
playground language playground and then you select one of these which is
25:23
you select one of these which is
25:23
you select one of these which is somewhere extract personal identifiable
25:25
somewhere extract personal identifiable
25:25
somewhere extract personal identifiable information from texts. Okay. And again
25:28
information from texts. Okay. And again
25:28
information from texts. Okay. And again um we haven't talked much about codes
25:30
um we haven't talked much about codes
25:30
um we haven't talked much about codes but the important part about code is you
25:32
but the important part about code is you
25:32
but the important part about code is you can get this from codes. Okay. Uh so you
25:35
can get this from codes. Okay. Uh so you
25:35
can get this from codes. Okay. Uh so you define the configuration here you say
25:37
define the configuration here you say
25:37
define the configuration here you say run and then the result will look
25:39
run and then the result will look
25:39
run and then the result will look something like this. So all of the
25:41
something like this. So all of the
25:41
something like this. So all of the information that's personally
25:43
information that's personally
25:43
information that's personally identifiable you can ask to hide it if
25:45
identifiable you can ask to hide it if
25:45
identifiable you can ask to hide it if you want to. Okay. You can see the JSON
25:48
you want to. Okay. You can see the JSON
25:48
you want to. Okay. You can see the JSON that will be returned as the API and
25:51
that will be returned as the API and
25:51
that will be returned as the API and then you can call this API. You can call
25:53
then you can call this API. You can call
25:53
then you can call this API. You can call it from C. You can call it from Python.
25:56
it from C. You can call it from Python.
25:56
it from C. You can call it from Python. Um these are the only two languages that
25:58
Um these are the only two languages that
25:58
Um these are the only two languages that I can pretty much guarantee that they
26:00
I can pretty much guarantee that they
26:00
I can pretty much guarantee that they will work. C. So basically anything
26:03
will work. C. So basically anything
26:03
will work. C. So basically anything doesn't have to be C#. It can be Visual
26:05
doesn't have to be C#. It can be Visual
26:05
doesn't have to be C#. It can be Visual Basic and usually Python. Okay? Because
26:09
Basic and usually Python. Okay? Because
26:09
Basic and usually Python. Okay? Because everyone seems to want Python. I really
26:11
everyone seems to want Python. I really
26:11
everyone seems to want Python. I really prefer C and Visual Basic, but people
26:13
prefer C and Visual Basic, but people
26:13
prefer C and Visual Basic, but people like Python for some reason. this
26:16
like Python for some reason. this
26:16
like Python for some reason. this terrible eh ugly based language. Ah,
26:20
terrible eh ugly based language. Ah,
26:20
terrible eh ugly based language. Ah, okay. But anyway,
26:21
okay. But anyway,
26:21
okay. But anyway, >> Python, yeah,
26:23
>> Python, yeah,
26:23
>> Python, yeah, >> but seems to be a thing. Uh, so you can
26:25
>> but seems to be a thing. Uh, so you can
26:25
>> but seems to be a thing. Uh, so you can try these two languages, but sometimes
26:26
try these two languages, but sometimes
26:26
try these two languages, but sometimes there's a few other languages like Go,
26:29
there's a few other languages like Go,
26:29
there's a few other languages like Go, JavaScript or Node.js and a few others.
26:31
JavaScript or Node.js and a few others.
26:32
JavaScript or Node.js and a few others. So sometimes you can use them. Now, how
26:34
So sometimes you can use them. Now, how
26:34
So sometimes you can use them. Now, how do you find examples for this? So in
26:35
do you find examples for this? So in
26:36
do you find examples for this? So in this case, this will be the extracting
26:38
this case, this will be the extracting
26:38
this case, this will be the extracting personal identifiable information from
26:39
personal identifiable information from
26:39
personal identifiable information from this. You can say view codes. Okay,
26:42
this. You can say view codes. Okay,
26:42
this. You can say view codes. Okay, sometimes this view code button is here
26:44
sometimes this view code button is here
26:44
sometimes this view code button is here and it is available. So you immediately
26:46
and it is available. So you immediately
26:46
and it is available. So you immediately have an example on how to do this. And
26:48
have an example on how to do this. And
26:48
have an example on how to do this. And in this case you just call sometimes you
26:51
in this case you just call sometimes you
26:51
in this case you just call sometimes you call uh a function sometimes you call
26:53
call uh a function sometimes you call
26:53
call uh a function sometimes you call just an API rest API but in this case uh
26:57
just an API rest API but in this case uh
26:57
just an API rest API but in this case uh there is an actual library for this
26:58
there is an actual library for this
26:58
there is an actual library for this azour.ai
27:00
azour.ai
27:00
azour.ai something in this case text analytics
27:03
something in this case text analytics
27:03
something in this case text analytics that allows you to go there and say look
27:05
that allows you to go there and say look
27:05
that allows you to go there and say look I have a client I specify the endpoints
27:08
I have a client I specify the endpoints
27:08
I have a client I specify the endpoints etc and then recognize whatever uh
27:11
etc and then recognize whatever uh
27:11
etc and then recognize whatever uh whatever is here. So string document
27:13
whatever is here. So string document
27:13
whatever is here. So string document something obviously you can pass this on
27:15
something obviously you can pass this on
27:15
something obviously you can pass this on as a uh you know as um as a parameter
27:19
as a uh you know as um as a parameter
27:19
as a uh you know as um as a parameter into the function to make it properly
27:21
into the function to make it properly
27:21
into the function to make it properly usable but uh the simple examples here
27:24
usable but uh the simple examples here
27:24
usable but uh the simple examples here you can just then use it in a very easy
27:26
you can just then use it in a very easy
27:26
you can just then use it in a very easy way. Remember all of this should be a
27:28
way. Remember all of this should be a
27:28
way. Remember all of this should be a synchronous because you still have to
27:30
synchronous because you still have to
27:30
synchronous because you still have to wait like one second for the results. So
27:32
wait like one second for the results. So
27:32
wait like one second for the results. So always keep it a synchronous. Okay they
27:35
always keep it a synchronous. Okay they
27:35
always keep it a synchronous. Okay they even tell you which service to connect
27:37
even tell you which service to connect
27:37
even tell you which service to connect to. So that's even easy. Okay, you can
27:39
to. So that's even easy. Okay, you can
27:39
to. So that's even easy. Okay, you can change this but they they will help you
27:41
change this but they they will help you
27:41
change this but they they will help you out with this. Okay, now uh there is
27:44
out with this. Okay, now uh there is
27:44
out with this. Okay, now uh there is much others but we do have limited time.
27:46
much others but we do have limited time.
27:46
much others but we do have limited time. So let me show you just the final thing
27:49
So let me show you just the final thing
27:49
So let me show you just the final thing that I want to show you. So of course
27:51
that I want to show you. So of course
27:51
that I want to show you. So of course chat playground and view code. You can
27:53
chat playground and view code. You can
27:53
chat playground and view code. You can definitely do this and I'm going to do
27:56
definitely do this and I'm going to do
27:56
definitely do this and I'm going to do something very very simple which is I'm
27:59
something very very simple which is I'm
27:59
something very very simple which is I'm going to say a system message that says
28:01
going to say a system message that says
28:01
going to say a system message that says something like this. You are the famous
28:03
something like this. You are the famous
28:03
something like this. You are the famous Aussie Osborne. You talk about Aussie
28:05
Aussie Osborne. You talk about Aussie
28:05
Aussie Osborne. You talk about Aussie Osborne stuff. Uh, please don't be
28:08
Osborne stuff. Uh, please don't be
28:08
Osborne stuff. Uh, please don't be violent. I didn't say that. But hello.
28:12
violent. I didn't say that. But hello.
28:12
violent. I didn't say that. But hello. Okay. And maybe we can answer. Hello,
28:14
Okay. And maybe we can answer. Hello,
28:14
Okay. And maybe we can answer. Hello, mate. How are you? Ready to talk about
28:16
mate. How are you? Ready to talk about
28:16
mate. How are you? Ready to talk about rock and roll? Your your favorite
28:21
rock and roll? Your your favorite
28:21
rock and roll? Your your favorite favorite song. Okay. And maybe we'll
28:25
favorite song. Okay. And maybe we'll
28:25
favorite song. Okay. And maybe we'll tough one. So, Paranoid. Yeah, most
28:27
tough one. So, Paranoid. Yeah, most
28:27
tough one. So, Paranoid. Yeah, most people will go for that one, but I'm not
28:29
people will go for that one, but I'm not
28:29
people will go for that one, but I'm not sure if the real Aussie Osborne would go
28:31
sure if the real Aussie Osborne would go
28:31
sure if the real Aussie Osborne would go for that one or not. I'm not sure. Um,
28:35
for that one or not. I'm not sure. Um,
28:35
for that one or not. I'm not sure. Um, but anyway, we could spend a lot of time
28:37
but anyway, we could spend a lot of time
28:37
but anyway, we could spend a lot of time here. One of the things that we can add
28:39
here. One of the things that we can add
28:39
here. One of the things that we can add is that we can add speech to text and
28:41
is that we can add speech to text and
28:41
is that we can add speech to text and text to speech and I'm going to go for I
28:45
text to speech and I'm going to go for I
28:45
text to speech and I'm going to go for I like English and this would be United
28:48
like English and this would be United
28:48
like English and this would be United Kingdom. So, this would pretty much be a
28:50
Kingdom. So, this would pretty much be a
28:50
Kingdom. So, this would pretty much be a dude, right? So, this is Aussie Osborne.
28:53
dude, right? So, this is Aussie Osborne.
28:53
dude, right? So, this is Aussie Osborne. Uh, and I haven't tried this one, but
28:55
Uh, and I haven't tried this one, but
28:55
Uh, and I haven't tried this one, but let's try Ethan. I don't know. Maybe
28:58
let's try Ethan. I don't know. Maybe
28:58
let's try Ethan. I don't know. Maybe this is a nice one. Play.
29:01
this is a nice one. Play.
29:01
this is a nice one. Play. Oh, by the way, this is a bug here. You
29:03
Oh, by the way, this is a bug here. You
29:03
Oh, by the way, this is a bug here. You have to play stop and then play again.
29:10
>> Uh there is a bug with this. So
29:12
>> Uh there is a bug with this. So
29:12
>> Uh there is a bug with this. So sometimes
29:12
sometimes
29:12
sometimes >> sometimes you know Andy I wonder you
29:14
>> sometimes you know Andy I wonder you
29:14
>> sometimes you know Andy I wonder you know.
29:14
know.
29:14
know. >> Hi Ethan. Hi I'm Ethan. How can I help
29:17
>> Hi Ethan. Hi I'm Ethan. How can I help
29:17
>> Hi Ethan. Hi I'm Ethan. How can I help you today?
29:17
you today?
29:17
you today? >> How can I help you today?
29:19
>> How can I help you today?
29:20
>> How can I help you today? >> Okay. So sorry about the the the extra
29:22
>> Okay. So sorry about the the the extra
29:22
>> Okay. So sorry about the the the extra feedback because this repeated a bit. My
29:24
feedback because this repeated a bit. My
29:24
feedback because this repeated a bit. My fault. Uh but yes, there is a bug with
29:27
fault. Uh but yes, there is a bug with
29:27
fault. Uh but yes, there is a bug with this start and stop button because it
29:29
this start and stop button because it
29:29
this start and stop button because it delays until the previous one. Okay. So
29:31
delays until the previous one. Okay. So
29:31
delays until the previous one. Okay. So you have to click on it twice. It's a
29:33
you have to click on it twice. It's a
29:33
you have to click on it twice. It's a It's a weird thing, but it doesn't
29:35
It's a weird thing, but it doesn't
29:35
It's a weird thing, but it doesn't affect anything else. You click on save.
29:37
affect anything else. You click on save.
29:37
affect anything else. You click on save. You just say
29:40
You just say
29:40
You just say hello.
29:46
Hello.
29:47
Hello.
29:47
Hello. Hello. Testing. Hello.
29:51
Hello. Testing. Hello.
29:51
Hello. Testing. Hello. Okay, something happened. I'm just going
29:52
Okay, something happened. I'm just going
29:52
Okay, something happened. I'm just going to say hello. Uh,
29:56
to say hello. Uh,
29:56
to say hello. Uh, right. Demo demon going to happen.
29:59
right. Demo demon going to happen.
29:59
right. Demo demon going to happen. >> Hello there, my friend.
30:01
>> Hello there, my friend.
30:01
>> Hello there, my friend. What's on your mind today? Rock and
30:04
What's on your mind today? Rock and
30:04
What's on your mind today? Rock and roll. I hope.
30:05
roll. I hope.
30:05
roll. I hope. >> Oh, that's lovely.
30:06
>> Oh, that's lovely.
30:06
>> Oh, that's lovely. >> It's not that different from Aussie
30:08
>> It's not that different from Aussie
30:08
>> It's not that different from Aussie Osborne. I think
30:12
Osborne. I think
30:12
Osborne. I think >> maybe a bit more desperation would be
30:14
>> maybe a bit more desperation would be
30:14
>> maybe a bit more desperation would be perfect for Aussie Osborne, but apart
30:15
perfect for Aussie Osborne, but apart
30:16
perfect for Aussie Osborne, but apart from that, it's still still a good idea.
30:18
from that, it's still still a good idea.
30:18
from that, it's still still a good idea. Anyway, what we do with this? Okay, so
30:20
Anyway, what we do with this? Okay, so
30:20
Anyway, what we do with this? Okay, so by the way, what we didn't see here with
30:22
by the way, what we didn't see here with
30:22
by the way, what we didn't see here with other capabilities was to add things
30:25
other capabilities was to add things
30:25
other capabilities was to add things like um audio, not audio, um image. So I
30:28
like um audio, not audio, um image. So I
30:28
like um audio, not audio, um image. So I could add image but this model GP3.5
30:31
could add image but this model GP3.5
30:31
could add image but this model GP3.5 which is kind of deprecated for all
30:33
which is kind of deprecated for all
30:33
which is kind of deprecated for all purposes does not have it. Some of these
30:35
purposes does not have it. Some of these
30:35
purposes does not have it. Some of these other ones might have it but let's not
30:37
other ones might have it but let's not
30:37
other ones might have it but let's not try it out. So view code first of all we
30:40
try it out. So view code first of all we
30:40
try it out. So view code first of all we have access to codes Python C and a few
30:44
have access to codes Python C and a few
30:44
have access to codes Python C and a few others. You can even try with curl or
30:46
others. You can even try with curl or
30:46
others. You can even try with curl or wget on a command line if you want to as
30:48
wget on a command line if you want to as
30:48
wget on a command line if you want to as well. Uh so you can actually try this
30:51
well. Uh so you can actually try this
30:51
well. Uh so you can actually try this out. So they tell you exactly what to
30:53
out. So they tell you exactly what to
30:53
out. So they tell you exactly what to do. Notice the extra thing which
30:55
do. Notice the extra thing which
30:55
do. Notice the extra thing which corresponds to speech. speech is not
30:57
corresponds to speech. speech is not
30:57
corresponds to speech. speech is not easy to program. So this already gives
31:00
easy to program. So this already gives
31:00
easy to program. So this already gives you all of the things that you need. So
31:02
you all of the things that you need. So
31:02
you all of the things that you need. So that that helps a bit. Okay. And then
31:05
that that helps a bit. Okay. And then
31:05
that that helps a bit. Okay. And then you write whatever you want to write and
31:07
you write whatever you want to write and
31:07
you write whatever you want to write and then you get an answer. So this is easy
31:08
then you get an answer. So this is easy
31:08
then you get an answer. So this is easy enough. Basically in the end it's just
31:10
enough. Basically in the end it's just
31:10
enough. Basically in the end it's just let's execute this and then you give it
31:12
let's execute this and then you give it
31:12
let's execute this and then you give it a try. Send the stuff there. If it
31:14
a try. Send the stuff there. If it
31:14
a try. Send the stuff there. If it works, here's the output. If it doesn't
31:16
works, here's the output. If it doesn't
31:16
works, here's the output. If it doesn't work, uh you say there's an exception.
31:19
work, uh you say there's an exception.
31:19
work, uh you say there's an exception. Here's a message. Goodbye. Okay. So this
31:21
Here's a message. Goodbye. Okay. So this
31:21
Here's a message. Goodbye. Okay. So this is how you can do it. So easy enough.
31:23
is how you can do it. So easy enough.
31:23
is how you can do it. So easy enough. But okay. So one last thing. What if I'm
31:26
But okay. So one last thing. What if I'm
31:26
But okay. So one last thing. What if I'm really really lazy because I am and I
31:29
really really lazy because I am and I
31:29
really really lazy because I am and I just want to deploy something that looks
31:31
just want to deploy something that looks
31:31
just want to deploy something that looks like this as fast as possible. I can go
31:33
like this as fast as possible. I can go
31:33
like this as fast as possible. I can go to deploy. Okay, click on deploy as a
31:36
to deploy. Okay, click on deploy as a
31:36
to deploy. Okay, click on deploy as a web app and very very quickly enter the
31:39
web app and very very quickly enter the
31:39
web app and very very quickly enter the name of a subscription, resource group,
31:41
name of a subscription, resource group,
31:41
name of a subscription, resource group, name of a web app, etc. Pricing plan
31:44
name of a web app, etc. Pricing plan
31:44
name of a web app, etc. Pricing plan which can start at free by the way. So
31:46
which can start at free by the way. So
31:46
which can start at free by the way. So it can be free. The one I'm trying is
31:48
it can be free. The one I'm trying is
31:48
it can be free. The one I'm trying is free. You can even add chat history
31:50
free. You can even add chat history
31:50
free. You can even add chat history using uh using um uh Cosmos DB as well.
31:54
using uh using um uh Cosmos DB as well.
31:54
using uh using um uh Cosmos DB as well. Okay. So, easy enough. And then you can
31:56
Okay. So, easy enough. And then you can
31:56
Okay. So, easy enough. And then you can go here and see the result. This one is
31:58
go here and see the result. This one is
31:58
go here and see the result. This one is on Richard Fineman. So, if say if you
32:01
on Richard Fineman. So, if say if you
32:01
on Richard Fineman. So, if say if you say hello, this would be Richard Fineman
32:03
say hello, this would be Richard Fineman
32:03
say hello, this would be Richard Fineman the famous physicist. All I said was
32:05
the famous physicist. All I said was
32:05
the famous physicist. All I said was something similar. You are Richard
32:07
something similar. You are Richard
32:07
something similar. You are Richard Fineman the famous physicist. That's it.
32:09
Fineman the famous physicist. That's it.
32:09
Fineman the famous physicist. That's it. This is GPT4.0 like the first one. And
32:12
This is GPT4.0 like the first one. And
32:12
This is GPT4.0 like the first one. And it just works. And then you can ask for
32:14
it just works. And then you can ask for
32:14
it just works. And then you can ask for things and get a result. Notice that
32:16
things and get a result. Notice that
32:16
things and get a result. Notice that this is almost public. But if you try it
32:19
this is almost public. But if you try it
32:19
this is almost public. But if you try it publicly, it doesn't work because it
32:21
publicly, it doesn't work because it
32:21
publicly, it doesn't work because it still requires you to log in with a user
32:24
still requires you to log in with a user
32:24
still requires you to log in with a user from my uh from my um tenants which most
32:28
from my uh from my um tenants which most
32:28
from my uh from my um tenants which most people don't have. So they will ask you
32:30
people don't have. So they will ask you
32:30
people don't have. So they will ask you to sign in. This is how you can secure
32:32
to sign in. This is how you can secure
32:32
to sign in. This is how you can secure it so that people don't abuse it. But
32:34
it so that people don't abuse it. But
32:34
it so that people don't abuse it. But you can turn off this limitation on your
32:37
you can turn off this limitation on your
32:37
you can turn off this limitation on your Azour web app so that anyone can access
32:39
Azour web app so that anyone can access
32:39
Azour web app so that anyone can access it. You can do that if you want to. I
32:41
it. You can do that if you want to. I
32:41
it. You can do that if you want to. I just didn't for obvious reasons. Anyway,
32:43
just didn't for obvious reasons. Anyway,
32:43
just didn't for obvious reasons. Anyway, there's uh there's quite a lot here. So
32:45
there's uh there's quite a lot here. So
32:45
there's uh there's quite a lot here. So let me just mention that there's quite a
32:47
let me just mention that there's quite a
32:47
let me just mention that there's quite a lot here that you can try in terms of
32:50
lot here that you can try in terms of
32:50
lot here that you can try in terms of playgrounds also these AI services
32:52
playgrounds also these AI services
32:52
playgrounds also these AI services there's also a lot more one of the
32:54
there's also a lot more one of the
32:54
there's also a lot more one of the things that's uh that you can try we
32:57
things that's uh that you can try we
32:57
things that's uh that you can try we tried vision a bit more things about
32:59
tried vision a bit more things about
32:59
tried vision a bit more things about language like the personal identifiable
33:01
language like the personal identifiable
33:01
language like the personal identifiable information that we tried even more
33:03
information that we tried even more
33:03
information that we tried even more things about speech uh there was getting
33:06
things about speech uh there was getting
33:06
things about speech uh there was getting started there's a few other things that
33:08
started there's a few other things that
33:08
started there's a few other things that you can start learning from Microsoft
33:10
you can start learning from Microsoft
33:10
you can start learning from Microsoft learn which are actually quite nice but
33:12
learn which are actually quite nice but
33:12
learn which are actually quite nice but let me mention one last thing I'm not
33:14
let me mention one last thing I'm not
33:14
let me mention one last thing I'm not going to try it I'm just going to
33:15
going to try it I'm just going to
33:15
going to try it I'm just going to mention it which is you can try the
33:17
mention it which is you can try the
33:17
mention it which is you can try the video playground. This one is expensive.
33:19
video playground. This one is expensive.
33:20
video playground. This one is expensive. And from here you can just write some
33:22
And from here you can just write some
33:22
And from here you can just write some text here and say please uh create some
33:25
text here and say please uh create some
33:25
text here and say please uh create some 5-second video and then you can specify
33:28
5-second video and then you can specify
33:28
5-second video and then you can specify sizes 5 20 seconds. Uh you can specify
33:31
sizes 5 20 seconds. Uh you can specify
33:32
sizes 5 20 seconds. Uh you can specify oh by the way these sizes are also
33:33
oh by the way these sizes are also
33:33
oh by the way these sizes are also different. Okay. So if you specify full
33:36
different. Okay. So if you specify full
33:36
different. Okay. So if you specify full HD here it's going to be more expensive
33:38
HD here it's going to be more expensive
33:38
HD here it's going to be more expensive as well. Okay. So, uh, two or3 for a
33:43
as well. Okay. So, uh, two or3 for a
33:43
as well. Okay. So, uh, two or3 for a full HD video that takes 5 seconds.
33:46
full HD video that takes 5 seconds.
33:46
full HD video that takes 5 seconds. >> Yeah. Two or three. Yeah. A little
33:48
>> Yeah. Two or three. Yeah. A little
33:48
>> Yeah. Two or three. Yeah. A little expensive, you know.
33:50
expensive, you know.
33:50
expensive, you know. >> Yeah, it is expensive. Quality is nice,
33:52
>> Yeah, it is expensive. Quality is nice,
33:52
>> Yeah, it is expensive. Quality is nice, but you have to be very explicit what
33:55
but you have to be very explicit what
33:55
but you have to be very explicit what you want to show. Otherwise, it will
33:58
you want to show. Otherwise, it will
33:58
you want to show. Otherwise, it will just, you know, start to hallucinate a
34:00
just, you know, start to hallucinate a
34:00
just, you know, start to hallucinate a bit. It will still be nice, but it's
34:02
bit. It will still be nice, but it's
34:02
bit. It will still be nice, but it's probably not going to be what you want.
34:04
probably not going to be what you want.
34:04
probably not going to be what you want. So, be very clear about the kind of
34:06
So, be very clear about the kind of
34:06
So, be very clear about the kind of things that you want to try. Okay. So um
34:10
things that you want to try. Okay. So um
34:10
things that you want to try. Okay. So um I guess this is pretty much what I
34:11
I guess this is pretty much what I
34:11
I guess this is pretty much what I wanted to show for now. Let me let me
34:14
wanted to show for now. Let me let me
34:14
wanted to show for now. Let me let me know if you have anything that you would
34:15
know if you have anything that you would
34:16
know if you have anything that you would like me to show anything else.
34:17
like me to show anything else.
34:17
like me to show anything else. >> Yeah, this is this is great. You covered
34:19
>> Yeah, this is this is great. You covered
34:19
>> Yeah, this is this is great. You covered yeah you covered a lot Andy you know in
34:21
yeah you covered a lot Andy you know in
34:21
yeah you covered a lot Andy you know in in in last like 30ish minutes you know
34:24
in in last like 30ish minutes you know
34:24
in in last like 30ish minutes you know and you know support someone who's using
34:26
and you know support someone who's using
34:26
and you know support someone who's using this aure foundry right uh let alone
34:29
this aure foundry right uh let alone
34:29
this aure foundry right uh let alone this you know it can be really
34:30
this you know it can be really
34:30
this you know it can be really overwhelming right because there's so
34:32
overwhelming right because there's so
34:32
overwhelming right because there's so many things you know and I already see
34:33
many things you know and I already see
34:33
many things you know and I already see on your sidebar you have a AI agents you
34:36
on your sidebar you have a AI agents you
34:36
on your sidebar you have a AI agents you have fine tune you have prompt there's
34:38
have fine tune you have prompt there's
34:38
have fine tune you have prompt there's so much more you can do here right and
34:41
so much more you can do here right and
34:41
so much more you can do here right and uh and it turns out right if I'm not
34:43
uh and it turns out right if I'm not
34:43
uh and it turns out right if I'm not wrong to use Azure foundry you have this
34:46
wrong to use Azure foundry you have this
34:46
wrong to use Azure foundry you have this web portal but you also have the SDKs
34:49
web portal but you also have the SDKs
34:49
web portal but you also have the SDKs and libraries and endpoints. So now more
34:52
and libraries and endpoints. So now more
34:52
and libraries and endpoints. So now more or less right you you need to have a
34:54
or less right you you need to have a
34:54
or less right you you need to have a decent knowledge of all these models and
34:56
decent knowledge of all these models and
34:56
decent knowledge of all these models and fine-tuning and everything and all you
34:58
fine-tuning and everything and all you
34:58
fine-tuning and everything and all you need to do is learn how to you know work
34:59
need to do is learn how to you know work
34:59
need to do is learn how to you know work with these API and SDKs because
35:01
with these API and SDKs because
35:01
with these API and SDKs because everything all the tough part has been
35:03
everything all the tough part has been
35:03
everything all the tough part has been taken care of by by the Azure.
35:07
taken care of by by the Azure.
35:07
taken care of by by the Azure. >> Exactly. So it's it's incredibly easy
35:09
>> Exactly. So it's it's incredibly easy
35:09
>> Exactly. So it's it's incredibly easy nowadays to use this kind of thing. Uh
35:11
nowadays to use this kind of thing. Uh
35:11
nowadays to use this kind of thing. Uh what I would say is that um even the the
35:14
what I would say is that um even the the
35:14
what I would say is that um even the the wizards here already create things for
35:16
wizards here already create things for
35:16
wizards here already create things for you. So they entered whatever you wanted
35:18
you. So they entered whatever you wanted
35:18
you. So they entered whatever you wanted already. So it's incredibly easy for you
35:20
already. So it's incredibly easy for you
35:20
already. So it's incredibly easy for you to just copy paste, change a few things
35:23
to just copy paste, change a few things
35:23
to just copy paste, change a few things and start getting this done. Uh a few
35:25
and start getting this done. Uh a few
35:26
and start getting this done. Uh a few disclaimers here. So yes, the elephant
35:27
disclaimers here. So yes, the elephant
35:28
disclaimers here. So yes, the elephant in the room that I mentioned before,
35:29
in the room that I mentioned before,
35:29
in the room that I mentioned before, there is a price to this. But in the
35:32
there is a price to this. But in the
35:32
there is a price to this. But in the case of AI services like this, it's
35:34
case of AI services like this, it's
35:34
case of AI services like this, it's incredibly cheap. So let's say we we
35:36
incredibly cheap. So let's say we we
35:36
incredibly cheap. So let's say we we were here for 15 20 minutes. Uh in the
35:39
were here for 15 20 minutes. Uh in the
35:39
were here for 15 20 minutes. Uh in the end, what I used and what you saw me use
35:42
end, what I used and what you saw me use
35:42
end, what I used and what you saw me use will cost me about 15 cents. That's the
35:45
will cost me about 15 cents. That's the
35:45
will cost me about 15 cents. That's the most
35:46
most
35:46
most It's not bad. Probably even less. Okay.
35:48
It's not bad. Probably even less. Okay.
35:48
It's not bad. Probably even less. Okay. If some things might be cheap or free,
35:51
If some things might be cheap or free,
35:51
If some things might be cheap or free, the free tier, but if it is AI stuff
35:53
the free tier, but if it is AI stuff
35:53
the free tier, but if it is AI stuff like LLMs, etc., they will always have a
35:56
like LLMs, etc., they will always have a
35:56
like LLMs, etc., they will always have a cost. But these are incredibly cheap,
35:58
cost. But these are incredibly cheap,
35:58
cost. But these are incredibly cheap, especially uh you know, GP 3.5.
36:02
especially uh you know, GP 3.5.
36:02
especially uh you know, GP 3.5. That was incredibly cheap. But okay. Uh
36:04
That was incredibly cheap. But okay. Uh
36:04
That was incredibly cheap. But okay. Uh I would recommend 4 something or you
36:07
I would recommend 4 something or you
36:07
I would recommend 4 something or you know 03 something like that because
36:09
know 03 something like that because
36:09
know 03 something like that because these are clearly better. But it was the
36:12
these are clearly better. But it was the
36:12
these are clearly better. But it was the default one that was selected and I
36:14
default one that was selected and I
36:14
default one that was selected and I didn't even pay attention. but it is
36:16
didn't even pay attention. but it is
36:16
didn't even pay attention. but it is going to be discontinued in two months.
36:17
going to be discontinued in two months.
36:18
going to be discontinued in two months. The good thing is you can change from
36:19
The good thing is you can change from
36:19
The good thing is you can change from one deployment to another from uh 3.5 to
36:22
one deployment to another from uh 3.5 to
36:22
one deployment to another from uh 3.5 to something else. It's like changing a
36:24
something else. It's like changing a
36:24
something else. It's like changing a connection string. Okay, it's just as
36:26
connection string. Okay, it's just as
36:26
connection string. Okay, it's just as easy as this from a programming point of
36:28
easy as this from a programming point of
36:28
easy as this from a programming point of view. You have some kind of variable
36:30
view. You have some kind of variable
36:30
view. You have some kind of variable that you put as a you know external uh
36:33
that you put as a you know external uh
36:33
that you put as a you know external uh configuration and then it's like a you
36:36
configuration and then it's like a you
36:36
configuration and then it's like a you know a configuration string um that that
36:39
know a configuration string um that that
36:39
know a configuration string um that that you choose a new deployment every time
36:41
you choose a new deployment every time
36:42
you choose a new deployment every time you know something becomes obsolete. So
36:43
you know something becomes obsolete. So
36:43
you know something becomes obsolete. So it's even easier. It's like connecting
36:45
it's even easier. It's like connecting
36:45
it's even easier. It's like connecting to a database. It's like a connection
36:47
to a database. It's like a connection
36:47
to a database. It's like a connection string. Uh in terms of in terms of
36:49
string. Uh in terms of in terms of
36:49
string. Uh in terms of in terms of programming, if you go to
36:50
programming, if you go to
36:50
programming, if you go to learn.microsoft.com,
36:52
learn.microsoft.com,
36:52
learn.microsoft.com, everything is there. Uh so you have
36:54
everything is there. Uh so you have
36:54
everything is there. Uh so you have examples, getting started demos, which
36:57
examples, getting started demos, which
36:57
examples, getting started demos, which are incredible. So it's ready to use. So
36:59
are incredible. So it's ready to use. So
36:59
are incredible. So it's ready to use. So they tell you exactly what to do. You
37:01
they tell you exactly what to do. You
37:01
they tell you exactly what to do. You copy paste, it's ready to go. You still
37:03
copy paste, it's ready to go. You still
37:03
copy paste, it's ready to go. You still need the the thing. So if you don't have
37:06
need the the thing. So if you don't have
37:06
need the the thing. So if you don't have Azour, you can still get it for free.
37:08
Azour, you can still get it for free.
37:08
Azour, you can still get it for free. Even if you're not a student, you can go
37:09
Even if you're not a student, you can go
37:10
Even if you're not a student, you can go to azour.comfree.
37:12
to azour.comfree.
37:12
to azour.comfree. You can get it for free for one month.
37:13
You can get it for free for one month.
37:13
You can get it for free for one month. And after that, the value is incredibly
37:16
And after that, the value is incredibly
37:16
And after that, the value is incredibly cheap. So, you know, at the end of the
37:17
cheap. So, you know, at the end of the
37:18
cheap. So, you know, at the end of the month, uh, if I just do demos like this,
37:20
month, uh, if I just do demos like this,
37:20
month, uh, if I just do demos like this, I would probably pay like one or two per
37:22
I would probably pay like one or two per
37:22
I would probably pay like one or two per month. I'm perfectly okay with that. So,
37:25
month. I'm perfectly okay with that. So,
37:25
month. I'm perfectly okay with that. So, this is very cheap. If you start to add
37:27
this is very cheap. If you start to add
37:27
this is very cheap. If you start to add a lot more images, if you start to add a
37:29
a lot more images, if you start to add a
37:30
a lot more images, if you start to add a lot more videos, then that becomes a
37:32
lot more videos, then that becomes a
37:32
lot more videos, then that becomes a fortune. Okay, there is a solution for
37:34
fortune. Okay, there is a solution for
37:34
fortune. Okay, there is a solution for that one. You use copilots
37:37
that one. You use copilots
37:38
that one. You use copilots or the free chatgpt. So you use, you
37:40
or the free chatgpt. So you use, you
37:40
or the free chatgpt. So you use, you know, copilot.microsoft.com,
37:42
know, copilot.microsoft.com,
37:42
know, copilot.microsoft.com, which is the the former one that used to
37:45
which is the the former one that used to
37:45
which is the the former one that used to be called bing copilot. Just, you know,
37:47
be called bing copilot. Just, you know,
37:47
be called bing copilot. Just, you know, just use that one. It's free. Generate
37:49
just use that one. It's free. Generate
37:49
just use that one. It's free. Generate the images there. It's the same thing.
37:51
the images there. It's the same thing.
37:51
the images there. It's the same thing. And in the end, it doesn't cost you any
37:53
And in the end, it doesn't cost you any
37:53
And in the end, it doesn't cost you any money. But if you want to program
37:55
money. But if you want to program
37:55
money. But if you want to program something that generates the images for
37:57
something that generates the images for
37:57
something that generates the images for you, then that's a different story. Then
37:59
you, then that's a different story. Then
37:59
you, then that's a different story. Then you need to use one of these services.
38:01
you need to use one of these services.
38:01
you need to use one of these services. By the way, you can use the services.
38:03
By the way, you can use the services.
38:03
By the way, you can use the services. Uh, obviously you need to have
38:05
Uh, obviously you need to have
38:05
Uh, obviously you need to have connection to the to the internet to be
38:07
connection to the to the internet to be
38:07
connection to the to the internet to be able to access this. uh but you can use
38:10
able to access this. uh but you can use
38:10
able to access this. uh but you can use this in whatever application you want on
38:12
this in whatever application you want on
38:12
this in whatever application you want on your phone, anywhere else there. This is
38:14
your phone, anywhere else there. This is
38:14
your phone, anywhere else there. This is just uh a uh you know a REST API. It's
38:17
just uh a uh you know a REST API. It's
38:17
just uh a uh you know a REST API. It's like a web service that you just connect
38:20
like a web service that you just connect
38:20
like a web service that you just connect to and get something back. In most
38:22
to and get something back. In most
38:22
to and get something back. In most cases, it's going to be JSON format
38:24
cases, it's going to be JSON format
38:24
cases, it's going to be JSON format being returned, sent and returned. Um if
38:27
being returned, sent and returned. Um if
38:27
being returned, sent and returned. Um if it is an image, they will just return a
38:30
it is an image, they will just return a
38:30
it is an image, they will just return a JSON that says here's the address of the
38:31
JSON that says here's the address of the
38:31
JSON that says here's the address of the image, https, some crazy uh letters and
38:35
image, https, some crazy uh letters and
38:35
image, https, some crazy uh letters and numbers, and then you can just go and
38:37
numbers, and then you can just go and
38:37
numbers, and then you can just go and download that one. So it's incredibly
38:39
download that one. So it's incredibly
38:39
download that one. So it's incredibly easy to use. Okay. So the idea here is
38:41
easy to use. Okay. So the idea here is
38:42
easy to use. Okay. So the idea here is if you feel like oh it might be
38:43
if you feel like oh it might be
38:43
if you feel like oh it might be difficult for me to create some AI
38:45
difficult for me to create some AI
38:45
difficult for me to create some AI service AI um you know program that I
38:49
service AI um you know program that I
38:49
service AI um you know program that I want to use for my company for you know
38:51
want to use for my company for you know
38:51
want to use for my company for you know just for fun something like that. Now
38:53
just for fun something like that. Now
38:53
just for fun something like that. Now you have a chance to to use these
38:55
you have a chance to to use these
38:55
you have a chance to to use these services. Yes there is a cost but you
38:57
services. Yes there is a cost but you
38:58
services. Yes there is a cost but you know it still works really really well.
39:00
know it still works really really well.
39:00
know it still works really really well. Okay.
39:01
Okay.
39:01
Okay. >> Yeah that is great. You know I so I
39:03
>> Yeah that is great. You know I so I
39:03
>> Yeah that is great. You know I so I think the call to action is you know
39:04
think the call to action is you know
39:04
think the call to action is you know Azure is free to get started. Go ahead
39:07
Azure is free to get started. Go ahead
39:07
Azure is free to get started. Go ahead and create a free account, right? As you
39:09
and create a free account, right? As you
39:09
and create a free account, right? As you can literally do anything with Azure A
39:11
can literally do anything with Azure A
39:12
can literally do anything with Azure A Foundry and there's so much you can do.
39:13
Foundry and there's so much you can do.
39:13
Foundry and there's so much you can do. So go ahead and explore and yeah, all
39:16
So go ahead and explore and yeah, all
39:16
So go ahead and explore and yeah, all you need to know is how to work with
39:18
you need to know is how to work with
39:18
you need to know is how to work with some SDKs and endpoints, right? That
39:20
some SDKs and endpoints, right? That
39:20
some SDKs and endpoints, right? That that's all you need and you made things
39:22
that's all you need and you made things
39:22
that's all you need and you made things look very easy any so thank you so much.
39:25
look very easy any so thank you so much.
39:25
look very easy any so thank you so much. >> Thank you very much.
39:27
>> Thank you very much.
39:27
>> Thank you very much. >> Yeah.
39:27
>> Yeah.
39:27
>> Yeah. >> Any final thing you want to plug in
39:29
>> Any final thing you want to plug in
39:29
>> Any final thing you want to plug in before we wrap up the show?
39:30
before we wrap up the show?
39:30
before we wrap up the show? >> Okay. So um you know for me I would just
39:33
>> Okay. So um you know for me I would just
39:33
>> Okay. So um you know for me I would just say have fun with AI. uh AI is there uh
39:37
say have fun with AI. uh AI is there uh
39:37
say have fun with AI. uh AI is there uh if you are a developer uh it is a good
39:39
if you are a developer uh it is a good
39:39
if you are a developer uh it is a good idea to know how to use AI you don't
39:41
idea to know how to use AI you don't
39:41
idea to know how to use AI you don't have to be an expert and understand
39:43
have to be an expert and understand
39:43
have to be an expert and understand what's behind the AI you know a lot of
39:46
what's behind the AI you know a lot of
39:46
what's behind the AI you know a lot of very intelligent people are doing that
39:49
very intelligent people are doing that
39:49
very intelligent people are doing that we just need to use it in terms of you
39:51
we just need to use it in terms of you
39:51
we just need to use it in terms of you know uh web services as a developer uh
39:54
know uh web services as a developer uh
39:54
know uh web services as a developer uh so you need to know at least this part
39:57
so you need to know at least this part
39:57
so you need to know at least this part whatever is behind it if you want to
39:58
whatever is behind it if you want to
39:58
whatever is behind it if you want to find out more of course have a look at
40:00
find out more of course have a look at
40:00
find out more of course have a look at documentation and information but from
40:03
documentation and information but from
40:03
documentation and information but from from a practical point of you at least
40:06
from a practical point of you at least
40:06
from a practical point of you at least uh you know learn how to call it, learn
40:08
uh you know learn how to call it, learn
40:08
uh you know learn how to call it, learn how to use it in some some kind of you
40:11
how to use it in some some kind of you
40:11
how to use it in some some kind of you know work environment. Okay. And that is
40:14
know work environment. Okay. And that is
40:14
know work environment. Okay. And that is expected for you in the future.
40:17
expected for you in the future.
40:17
expected for you in the future. >> Perfect. Andy, thank you so much for
40:18
>> Perfect. Andy, thank you so much for
40:18
>> Perfect. Andy, thank you so much for sharing all your pearls of wisdom and
40:20
sharing all your pearls of wisdom and
40:20
sharing all your pearls of wisdom and thank you so much everyone for joining
40:22
thank you so much everyone for joining
40:22
thank you so much everyone for joining us today. That was it in this episode of
40:24
us today. That was it in this episode of
40:24
us today. That was it in this episode of AI Dev Tools. We'll see you next time.
40:25
AI Dev Tools. We'll see you next time.
40:25
AI Dev Tools. We'll see you next time. See you. Bye- bye.
40:26
See you. Bye- bye.
40:26
See you. Bye- bye. >> Thank you very much. See you next time.
40:27
>> Thank you very much. See you next time.
40:27
>> Thank you very much. See you next time. Thank you for having me.