Up next in 10
What’s New in .NET: AI, Aspire & Azure for Cloud-Native Devs
0 views
Aug 7, 2025
Explore what's new in the .NET ecosystem with a focus on AI integration, .NET Aspire, and Azure-native development. In this session, we cover: • The latest updates and enhancements in .NET • How AI capabilities are integrated into the .NET platform • What .NET Aspire is and how it simplifies cloud-native development • How developers can deploy seamlessly to Azure using the new stack Whether you're a .NET developer or exploring how to bring AI and cloud-native practices into your applications, this session delivers actionable insights for building modern, scalable systems. ⚡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 https://www.c-sharpcorner.com
View Video Transcript
0:00
explore the AI capabilities integrated
0:02
explore the AI capabilities integrated
0:02
explore the AI capabilities integrated within the net ecosystem in this episode
0:04
within the net ecosystem in this episode
0:04
within the net ecosystem in this episode of AI dev tools today and we have today
0:07
of AI dev tools today and we have today
0:07
of AI dev tools today and we have today with us Bruno who is a cloud advocate at
0:10
with us Bruno who is a cloud advocate at
0:10
with us Bruno who is a cloud advocate at Microsoft. Hi Pruno and welcome to the
0:11
Microsoft. Hi Pruno and welcome to the
0:12
Microsoft. Hi Pruno and welcome to the show.
0:14
show.
0:14
show. [Music]
0:17
[Music]
0:17
[Music] Hey Simon, hello everyone and hey super
0:19
Hey Simon, hello everyone and hey super
0:19
Hey Simon, hello everyone and hey super happy to be here. It's always a pleasure
0:21
happy to be here. It's always a pleasure
0:21
happy to be here. It's always a pleasure to talk about AI and more. So hey super
0:23
to talk about AI and more. So hey super
0:24
to talk about AI and more. So hey super happy to be here. Long story short, uh
0:26
happy to be here. Long story short, uh
0:26
happy to be here. Long story short, uh we've been thinking a lot how we can
0:28
we've been thinking a lot how we can
0:28
we've been thinking a lot how we can help uh our .NET community, users and
0:31
help uh our .NET community, users and
0:31
help uh our .NET community, users and more to learn about genai and we created
0:34
more to learn about genai and we created
0:34
more to learn about genai and we created we have this main repo which is called
0:36
we have this main repo which is called
0:36
we have this main repo which is called generative AI for beginners.net net and
0:39
generative AI for beginners.net net and
0:39
generative AI for beginners.net net and it has basically has a lot of se
0:41
it has basically has a lot of se
0:41
it has basically has a lot of se sessions lessons here where you have an
0:43
sessions lessons here where you have an
0:43
sessions lessons here where you have an intro to zenai you have how you can
0:46
intro to zenai you have how you can
0:46
intro to zenai you have how you can setting up the environment basically c
0:49
setting up the environment basically c
0:49
setting up the environment basically c techniques that you can do with an AI
0:50
techniques that you can do with an AI
0:50
techniques that you can do with an AI calling functions doing rag a lot of
0:52
calling functions doing rag a lot of
0:52
calling functions doing rag a lot of stuff that you can learn here the whole
0:54
stuff that you can learn here the whole
0:54
stuff that you can learn here the whole idea is that you can start here without
0:58
idea is that you can start here without
0:58
idea is that you can start here without installing anything we are going to use
1:00
installing anything we are going to use
1:00
installing anything we are going to use GitHub models which are models AI models
1:03
GitHub models which are models AI models
1:03
GitHub models which are models AI models that GitHub provides you for free so you
1:05
that GitHub provides you for free so you
1:05
that GitHub provides you for free so you only need a GitHub account you can run
1:08
only need a GitHub account you can run
1:08
only need a GitHub account you can run this in code spaces. So you don't need
1:10
this in code spaces. So you don't need
1:10
this in code spaces. So you don't need to install anything. Everything is going
1:11
to install anything. Everything is going
1:11
to install anything. Everything is going to be run in the cloud. So the whole
1:13
to be run in the cloud. So the whole
1:13
to be run in the cloud. So the whole idea is that we set up everything to
1:16
idea is that we set up everything to
1:16
idea is that we set up everything to give everybody the chance to you can
1:18
give everybody the chance to you can
1:18
give everybody the chance to you can learn generative AI for vision, chat,
1:21
learn generative AI for vision, chat,
1:21
learn generative AI for vision, chat, audio and more. And then you can also do
1:24
audio and more. And then you can also do
1:24
audio and more. And then you can also do a lot of different other stuffs like
1:26
a lot of different other stuffs like
1:26
a lot of different other stuffs like example when we talk about techniques we
1:29
example when we talk about techniques we
1:29
example when we talk about techniques we have here a solution that include I
1:32
have here a solution that include I
1:32
have here a solution that include I don't know there is probably 30 sample
1:35
don't know there is probably 30 sample
1:35
don't know there is probably 30 sample projects here. how how you can use local
1:37
projects here. how how you can use local
1:37
projects here. how how you can use local modies with AI foundry, locals modeling
1:39
modies with AI foundry, locals modeling
1:39
modies with AI foundry, locals modeling with AI toolkit, how you can do some
1:41
with AI toolkit, how you can do some
1:41
with AI toolkit, how you can do some agents activities, chat activities, rag
1:44
agents activities, chat activities, rag
1:44
agents activities, chat activities, rag activities, docker models. There are
1:46
activities, docker models. There are
1:46
activities, docker models. There are plenty plenty of scenes that you can do.
1:48
plenty plenty of scenes that you can do.
1:48
plenty plenty of scenes that you can do. But today I want to show something
1:50
But today I want to show something
1:50
But today I want to show something different that we also add here because
1:52
different that we also add here because
1:52
different that we also add here because this is still generative AI but today we
1:56
this is still generative AI but today we
1:56
this is still generative AI but today we are also start to see that people are
1:58
are also start to see that people are
1:58
are also start to see that people are creating applications using tools like
2:01
creating applications using tools like
2:01
creating applications using tools like GitHub copilot. big fan here of Copilot
2:04
GitHub copilot. big fan here of Copilot
2:04
GitHub copilot. big fan here of Copilot and hey people are starting to do this
2:06
and hey people are starting to do this
2:06
and hey people are starting to do this bite coding starting to bite the code a
2:09
bite coding starting to bite the code a
2:09
bite coding starting to bite the code a lot by the way we do that a lot so when
2:12
lot by the way we do that a lot so when
2:12
lot by the way we do that a lot so when I was thinking what we can do what can
2:14
I was thinking what we can do what can
2:14
I was thinking what we can do what can create we decided to create kind of a
2:17
create we decided to create kind of a
2:17
create we decided to create kind of a standard retro game in the console that
2:21
standard retro game in the console that
2:21
standard retro game in the console that it's this little battle ship whatever
2:23
it's this little battle ship whatever
2:23
it's this little battle ship whatever can't remember the name you know the you
2:25
can't remember the name you know the you
2:25
can't remember the name you know the you know probably the game and the whole
2:27
know probably the game and the whole
2:27
know probably the game and the whole idea is that we created this using a
2:29
idea is that we created this using a
2:29
idea is that we created this using a very very big prompt that is here and
2:31
very very big prompt that is here and
2:32
very very big prompt that is here and Remember this is all in the repo. So you
2:33
Remember this is all in the repo. So you
2:33
Remember this is all in the repo. So you can take a look at the prompt, you can
2:36
can take a look at the prompt, you can
2:36
can take a look at the prompt, you can take a look at the
2:39
take a look at the
2:39
take a look at the everything that we have there, the final
2:40
everything that we have there, the final
2:40
everything that we have there, the final code and more. So we have here a very
2:42
code and more. So we have here a very
2:42
code and more. So we have here a very big prompt that is going to say, "Hey,
2:44
big prompt that is going to say, "Hey,
2:44
big prompt that is going to say, "Hey, you know what? This is a C# console
2:46
you know what? This is a C# console
2:46
you know what? This is a C# console project. We want to build a modular
2:48
project. We want to build a modular
2:48
project. We want to build a modular flicker game. We are going to have these
2:50
flicker game. We are going to have these
2:50
flicker game. We are going to have these files. Everything is styled here. We
2:53
files. Everything is styled here. We
2:53
files. Everything is styled here. We created the prompt using GitHub copilot.
2:55
created the prompt using GitHub copilot.
2:55
created the prompt using GitHub copilot. And at the end, what we have is, as we
2:58
And at the end, what we have is, as we
2:58
And at the end, what we have is, as we mentioned, we have this game. And the
3:00
mentioned, we have this game. And the
3:00
mentioned, we have this game. And the game is kind of nice because yes, you
3:01
game is kind of nice because yes, you
3:01
game is kind of nice because yes, you can play the game. So, let me this is
3:03
can play the game. So, let me this is
3:03
can play the game. So, let me this is the final game. Let me show you a quick
3:06
the final game. Let me show you a quick
3:06
the final game. Let me show you a quick overview of how we can how we we
3:09
overview of how we can how we we
3:09
overview of how we can how we we organize the game here. So, we have the
3:13
organize the game here. So, we have the
3:13
organize the game here. So, we have the main console project and then we have
3:16
main console project and then we have
3:16
main console project and then we have two others which one is to process each
3:19
two others which one is to process each
3:19
two others which one is to process each one of the frames and another one is to
3:22
one of the frames and another one is to
3:22
one of the frames and another one is to have a screenshot of the frames. And we
3:24
have a screenshot of the frames. And we
3:24
have a screenshot of the frames. And we have some Aspire for telemetry not going
3:26
have some Aspire for telemetry not going
3:26
have some Aspire for telemetry not going to go there right now. And the whole
3:28
to go there right now. And the whole
3:28
to go there right now. And the whole idea is that we have a main game manager
3:31
idea is that we have a main game manager
3:31
idea is that we have a main game manager which is the one that controls
3:32
which is the one that controls
3:32
which is the one that controls everything and then we have entities for
3:34
everything and then we have entities for
3:34
everything and then we have entities for the players like the player which is
3:37
the players like the player which is
3:37
the players like the player which is going to have a location in the screen
3:39
going to have a location in the screen
3:39
going to have a location in the screen saying from the enemy or a bullet and
3:42
saying from the enemy or a bullet and
3:42
saying from the enemy or a bullet and more. Hey again this is all this is all
3:46
more. Hey again this is all this is all
3:46
more. Hey again this is all this is all this was created using a big prompt. If
3:48
this was created using a big prompt. If
3:48
this was created using a big prompt. If I run the
3:50
I run the
3:50
I run the if I run the program we are going to see
3:52
if I run the program we are going to see
3:52
if I run the program we are going to see here that this is the program up and
3:54
here that this is the program up and
3:54
here that this is the program up and running. We have the controls. We have
3:56
running. We have the controls. We have
3:56
running. We have the controls. We have three game speeds. Let's start in
3:57
three game speeds. Let's start in
3:57
three game speeds. Let's start in default one and I am moving left, moving
4:00
default one and I am moving left, moving
4:00
default one and I am moving left, moving right and if I press a space I'm going
4:02
right and if I press a space I'm going
4:02
right and if I press a space I'm going to start to to shoot. So that's kind of
4:05
to start to to shoot. So that's kind of
4:05
to start to to shoot. So that's kind of nice. This is what we have and what we
4:07
nice. This is what we have and what we
4:07
nice. This is what we have and what we do. But we also have one of the feature
4:09
do. But we also have one of the feature
4:09
do. But we also have one of the feature which is if I press the letter A and I
4:11
which is if I press the letter A and I
4:12
which is if I press the letter A and I am going to show you how it works. This
4:13
am going to show you how it works. This
4:14
am going to show you how it works. This right now is going to start to go to the
4:16
right now is going to start to go to the
4:16
right now is going to start to go to the cloud and it's going to start to play
4:19
cloud and it's going to start to play
4:19
cloud and it's going to start to play with using AI. And the way that is
4:22
with using AI. And the way that is
4:22
with using AI. And the way that is working is we are using Microsoft
4:24
working is we are using Microsoft
4:24
working is we are using Microsoft extension for AI and we are passing each
4:27
extension for AI and we are passing each
4:27
extension for AI and we are passing each one of the frames to the game. So the
4:30
one of the frames to the game. So the
4:30
one of the frames to the game. So the game knows how to is going to know what
4:32
game knows how to is going to know what
4:32
game knows how to is going to know what is happening in the frame and it's going
4:34
is happening in the frame and it's going
4:34
is happening in the frame and it's going to took a decision to say okay I need to
4:35
to took a decision to say okay I need to
4:36
to took a decision to say okay I need to do this or I need to do that or I need
4:37
do this or I need to do that or I need
4:37
do this or I need to do that or I need to move I need to fire I need to stop.
4:39
to move I need to fire I need to stop.
4:39
to move I need to fire I need to stop. It's going to happen all that. Right now
4:41
It's going to happen all that. Right now
4:41
It's going to happen all that. Right now it's not working because I haven't
4:43
it's not working because I haven't
4:43
it's not working because I haven't connected this to Ashure OpenAI. I don't
4:45
connected this to Ashure OpenAI. I don't
4:45
connected this to Ashure OpenAI. I don't have my cloud models up and running. But
4:48
have my cloud models up and running. But
4:48
have my cloud models up and running. But we have in the provider section because
4:50
we have in the provider section because
4:50
we have in the provider section because we depend on this. We also have the
4:53
we depend on this. We also have the
4:53
we depend on this. We also have the chance of run this yuka using local
4:55
chance of run this yuka using local
4:55
chance of run this yuka using local models and this becomes interesting. I'm
4:58
models and this becomes interesting. I'm
4:58
models and this becomes interesting. I'm testing I'm a big fan of you. You
5:01
testing I'm a big fan of you. You
5:01
testing I'm a big fan of you. You probably if you follow me online you
5:02
probably if you follow me online you
5:02
probably if you follow me online you probably see that I do a lot of I like
5:05
probably see that I do a lot of I like
5:05
probably see that I do a lot of I like as a developer environment because you
5:07
as a developer environment because you
5:07
as a developer environment because you have kind of the power of models locally
5:11
have kind of the power of models locally
5:11
have kind of the power of models locally and you can do a lot. If I have to
5:13
and you can do a lot. If I have to
5:13
and you can do a lot. If I have to choose the whether I have lama and this
5:15
choose the whether I have lama and this
5:15
choose the whether I have lama and this is lama remember that you can install
5:17
is lama remember that you can install
5:17
is lama remember that you can install lama in windows, linux, mac in all of
5:20
lama in windows, linux, mac in all of
5:20
lama in windows, linux, mac in all of OS. Uh I like to run in a container
5:23
OS. Uh I like to run in a container
5:23
OS. Uh I like to run in a container because I have kind of more control of
5:25
because I have kind of more control of
5:25
because I have kind of more control of the container. I can store stop start
5:27
the container. I can store stop start
5:27
the container. I can store stop start stop and more. And this is my container
5:30
stop and more. And this is my container
5:30
stop and more. And this is my container and if I open my terminal I am going to
5:33
and if I open my terminal I am going to
5:33
and if I open my terminal I am going to show you here that I have a couple of
5:36
show you here that I have a couple of
5:36
show you here that I have a couple of models installed. So yes, I know I have
5:38
models installed. So yes, I know I have
5:38
models installed. So yes, I know I have 20 plus models here and there are small
5:41
20 plus models here and there are small
5:41
20 plus models here and there are small models, big models and more. But the
5:43
models, big models and more. But the
5:43
models, big models and more. But the whole idea is that I like to have these
5:45
whole idea is that I like to have these
5:45
whole idea is that I like to have these models locally because once my container
5:48
models locally because once my container
5:48
models locally because once my container is up and running, I can test my games
5:52
is up and running, I can test my games
5:52
is up and running, I can test my games in example like this one because I don't
5:54
in example like this one because I don't
5:54
in example like this one because I don't have access to ashure right now, but I
5:56
have access to ashure right now, but I
5:56
have access to ashure right now, but I can test the game running locally. So if
5:59
can test the game running locally. So if
5:59
can test the game running locally. So if I change here, if I go back here, the
6:02
I change here, if I go back here, the
6:02
I change here, if I go back here, the game that the way that the game works is
6:04
game that the way that the game works is
6:04
game that the way that the game works is we have a provider that basically is an
6:07
we have a provider that basically is an
6:07
we have a provider that basically is an interface and interface is going to
6:08
interface and interface is going to
6:08
interface and interface is going to create a client a client chat. The
6:11
create a client a client chat. The
6:11
create a client a client chat. The client chat will be created using Olama
6:13
client chat will be created using Olama
6:13
client chat will be created using Olama locally or maybe ashure OpenAI or if you
6:16
locally or maybe ashure OpenAI or if you
6:16
locally or maybe ashure OpenAI or if you want to use model from I don't know
6:18
want to use model from I don't know
6:18
want to use model from I don't know another model from another provider
6:20
another model from another provider
6:20
another model from another provider OpenAI APIs or entropic or whatever you
6:22
OpenAI APIs or entropic or whatever you
6:22
OpenAI APIs or entropic or whatever you can use them the extensibility is there
6:24
can use them the extensibility is there
6:24
can use them the extensibility is there and then when we have the game what we
6:26
and then when we have the game what we
6:26
and then when we have the game what we are going to do is we are going to have
6:28
are going to do is we are going to have
6:28
are going to do is we are going to have a very very big prompt here switch I
6:31
a very very big prompt here switch I
6:31
a very very big prompt here switch I don't want to switch my style that is
6:33
don't want to switch my style that is
6:33
don't want to switch my style that is going to say hey you know what this is
6:36
going to say hey you know what this is
6:36
going to say hey you know what this is you are you need to act as a video game
6:37
you are you need to act as a video game
6:38
you are you need to act as a video game player with a lot of expertise do this
6:40
player with a lot of expertise do this
6:40
player with a lot of expertise do this and at analyze the frame, analyze the
6:44
and at analyze the frame, analyze the
6:44
and at analyze the frame, analyze the what is going to happen in the game and
6:46
what is going to happen in the game and
6:46
what is going to happen in the game and then and then what we can do is define
6:49
then and then what we can do is define
6:49
then and then what we can do is define the next action and explain why are you
6:52
the next action and explain why are you
6:52
the next action and explain why are you doing this and this is in the game what
6:53
doing this and this is in the game what
6:53
doing this and this is in the game what we see at the bottom we are going to see
6:54
we see at the bottom we are going to see
6:54
we see at the bottom we are going to see at the bottom. So examples we do a few a
6:58
at the bottom. So examples we do a few a
6:58
at the bottom. So examples we do a few a few short learning here where we provide
7:00
few short learning here where we provide
7:00
few short learning here where we provide a couple of samples like hey next action
7:02
a couple of samples like hey next action
7:02
a couple of samples like hey next action move right and moving right will help
7:04
move right and moving right will help
7:04
move right and moving right will help avoid we have everything here and then
7:06
avoid we have everything here and then
7:06
avoid we have everything here and then we have the way and then we have right
7:09
we have the way and then we have right
7:09
we have the way and then we have right now uh basically how we are going to
7:12
now uh basically how we are going to
7:12
now uh basically how we are going to implement this we implemented this in
7:15
implement this we implemented this in
7:15
implement this we implemented this in two phases initially what I was doing I
7:17
two phases initially what I was doing I
7:18
two phases initially what I was doing I was taking a screenshots of the game of
7:19
was taking a screenshots of the game of
7:19
was taking a screenshots of the game of the console application and I was
7:21
the console application and I was
7:21
the console application and I was passing the image to the model to be
7:23
passing the image to the model to be
7:23
passing the image to the model to be analyzed and this was great so we have
7:25
analyzed and this was great so we have
7:25
analyzed and this was great so we have here a frame which is a bite which is a
7:29
here a frame which is a bite which is a
7:29
here a frame which is a bite which is a a bite representation of the of the
7:31
a bite representation of the of the
7:31
a bite representation of the of the image and then I pass this I create an
7:33
image and then I pass this I create an
7:33
image and then I pass this I create an image PNG and I send this to the model
7:35
image PNG and I send this to the model
7:35
image PNG and I send this to the model but talking with a couple of friends and
7:37
but talking with a couple of friends and
7:38
but talking with a couple of friends and they said hey you know what a bite
7:40
they said hey you know what a bite
7:40
they said hey you know what a bite representation and arrive by an image is
7:42
representation and arrive by an image is
7:42
representation and arrive by an image is kind of heavy maybe we can pass directly
7:45
kind of heavy maybe we can pass directly
7:45
kind of heavy maybe we can pass directly the frame as a string remember this is a
7:47
the frame as a string remember this is a
7:47
the frame as a string remember this is a console application and when we pass the
7:49
console application and when we pass the
7:50
console application and when we pass the frame as a string we can have the option
7:53
frame as a string we can have the option
7:53
frame as a string we can have the option to see if the model is going to use this
7:55
to see if the model is going to use this
7:55
to see if the model is going to use this is going to basically analyze this. So
7:57
is going to basically analyze this. So
7:58
is going to basically analyze this. So if I start again and I have here the
8:00
if I start again and I have here the
8:00
if I start again and I have here the frame as a string
8:02
frame as a string
8:02
frame as a string what we can do is the same. So we are
8:05
what we can do is the same. So we are
8:05
what we can do is the same. So we are here we have the starting
8:09
here we have the starting
8:09
here we have the starting start there which like it and more. So
8:11
start there which like it and more. So
8:11
start there which like it and more. So let's start the game. I can play
8:12
let's start the game. I can play
8:12
let's start the game. I can play pressing the letters or if I press the O
8:16
pressing the letters or if I press the O
8:16
pressing the letters or if I press the O this is the game up and running. And
8:17
this is the game up and running. And
8:17
this is the game up and running. And here in the frame let's take a look at
8:19
here in the frame let's take a look at
8:19
here in the frame let's take a look at the frame. What you can see this is the
8:22
the frame. What you can see this is the
8:22
the frame. What you can see this is the literally the representation of the
8:24
literally the representation of the
8:24
literally the representation of the game. You can see the main ship in the
8:25
game. You can see the main ship in the
8:25
game. You can see the main ship in the bottom. You can see the enemies in the
8:27
bottom. You can see the enemies in the
8:27
bottom. You can see the enemies in the top. This is everything that we have in
8:29
top. This is everything that we have in
8:29
top. This is everything that we have in the console. So right now let me stop
8:32
the console. So right now let me stop
8:32
the console. So right now let me stop the let me I'm sorry remove the the
8:34
the let me I'm sorry remove the the
8:34
the let me I'm sorry remove the the break point. I am going to keep using
8:37
break point. I am going to keep using
8:37
break point. I am going to keep using the game. Keep playing. We are going to
8:39
the game. Keep playing. We are going to
8:39
the game. Keep playing. We are going to see that it's going to take some time
8:41
see that it's going to take some time
8:41
see that it's going to take some time because basically warming up in the back
8:44
because basically warming up in the back
8:44
because basically warming up in the back here. Oh, not this one. This one. So we
8:48
here. Oh, not this one. This one. So we
8:48
here. Oh, not this one. This one. So we are going to start to see what we see in
8:50
are going to start to see what we see in
8:50
are going to start to see what we see in green in the bottom are the decision
8:52
green in the bottom are the decision
8:52
green in the bottom are the decision that Olama is taking here to play the
8:54
that Olama is taking here to play the
8:54
that Olama is taking here to play the game. It's firing it. Move left a little
8:57
game. It's firing it. Move left a little
8:57
game. It's firing it. Move left a little more and it's working. If I open my task
9:01
more and it's working. If I open my task
9:01
more and it's working. If I open my task manager, if I go here to see the
9:03
manager, if I go here to see the
9:03
manager, if I go here to see the processes, we are going to see in the
9:05
processes, we are going to see in the
9:05
processes, we are going to see in the bottom that the GPU is working. Don't
9:09
bottom that the GPU is working. Don't
9:09
bottom that the GPU is working. Don't care about internet and more. This is
9:11
care about internet and more. This is
9:11
care about internet and more. This is literally using my local resources to do
9:14
literally using my local resources to do
9:14
literally using my local resources to do with the CPU. Right now it's not playing
9:16
with the CPU. Right now it's not playing
9:16
with the CPU. Right now it's not playing very well. izing when I was playing with
9:18
very well. izing when I was playing with
9:18
very well. izing when I was playing with the prompt, I tweak the prompt a little
9:20
the prompt, I tweak the prompt a little
9:20
the prompt, I tweak the prompt a little to be kind of a more defensive. I should
9:22
to be kind of a more defensive. I should
9:22
to be kind of a more defensive. I should probably be say, "Hey, be more
9:24
probably be say, "Hey, be more
9:24
probably be say, "Hey, be more aggressive. Try to follow the enemies
9:25
aggressive. Try to follow the enemies
9:26
aggressive. Try to follow the enemies and kill the enemies. The enemies is are
9:28
and kill the enemies. The enemies is are
9:28
and kill the enemies. The enemies is are this." But this is kind of nice. This is
9:30
this." But this is kind of nice. This is
9:30
this." But this is kind of nice. This is kind of great. This is literally created
9:33
kind of great. This is literally created
9:33
kind of great. This is literally created the game with a big prompt and then with
9:35
the game with a big prompt and then with
9:35
the game with a big prompt and then with a couple of other extra prompts. We play
9:37
a couple of other extra prompts. We play
9:37
a couple of other extra prompts. We play this, we switch this to make our game
9:40
this, we switch this to make our game
9:40
this, we switch this to make our game play locally. And yes, it get stuck in
9:42
play locally. And yes, it get stuck in
9:42
play locally. And yes, it get stuck in the bottom there. So let me disable
9:46
the bottom there. So let me disable
9:46
the bottom there. So let me disable this. Let's move this a little by hand
9:49
this. Let's move this a little by hand
9:49
this. Let's move this a little by hand and press the amma model again. So it's
9:53
and press the amma model again. So it's
9:53
and press the amma model again. So it's running again with lama at any time. And
9:55
running again with lama at any time. And
9:55
running again with lama at any time. And hey, let me stop it and we are going to
9:57
hey, let me stop it and we are going to
9:57
hey, let me stop it and we are going to see how the GPU is going to get back to
10:00
see how the GPU is going to get back to
10:00
see how the GPU is going to get back to normal. But this is cool. This is nice
10:02
normal. But this is cool. This is nice
10:02
normal. But this is cool. This is nice because in example remember I have a lot
10:04
because in example remember I have a lot
10:04
because in example remember I have a lot of models here. I have JMA models, llama
10:07
of models here. I have JMA models, llama
10:07
of models here. I have JMA models, llama models, five models. I think I'm running
10:09
models, five models. I think I'm running
10:09
models, five models. I think I'm running with 54 right now. We have a new model
10:12
with 54 right now. We have a new model
10:12
with 54 right now. We have a new model which is the reasoning model.
10:14
which is the reasoning model.
10:14
which is the reasoning model. which is very very good. So if I want to
10:17
which is very very good. So if I want to
10:17
which is very very good. So if I want to change this, I only go here and of
10:19
change this, I only go here and of
10:19
change this, I only go here and of course you need to have this in your
10:22
course you need to have this in your
10:22
course you need to have this in your config somewhere there. But I am going
10:24
config somewhere there. But I am going
10:24
config somewhere there. But I am going to change and play the game using 54
10:26
to change and play the game using 54
10:26
to change and play the game using 54 mini reasoning and where is my game?
10:29
mini reasoning and where is my game?
10:29
mini reasoning and where is my game? There is there it is my game. So if I go
10:31
There is there it is my game. So if I go
10:31
There is there it is my game. So if I go back here, let me open clear the
10:35
back here, let me open clear the
10:35
back here, let me open clear the terminal. put the game back online. And
10:38
terminal. put the game back online. And
10:38
terminal. put the game back online. And if I start to play and I ask for the
10:42
if I start to play and I ask for the
10:42
if I start to play and I ask for the reasoning, you can see Lama working in
10:44
reasoning, you can see Lama working in
10:44
reasoning, you can see Lama working in the back. And the reasoning model is
10:46
the back. And the reasoning model is
10:46
the back. And the reasoning model is probably not the best one to do like
10:47
probably not the best one to do like
10:47
probably not the best one to do like this. I will probably go for a big I'm
10:49
this. I will probably go for a big I'm
10:49
this. I will probably go for a big I'm sorry, for a short small game. And yes,
10:52
sorry, for a short small game. And yes,
10:52
sorry, for a short small game. And yes, the reasoning model doesn't make sense.
10:54
the reasoning model doesn't make sense.
10:54
the reasoning model doesn't make sense. It's not going to know what to do
10:55
It's not going to know what to do
10:55
It's not going to know what to do because the instructions are not for a
10:57
because the instructions are not for a
10:57
because the instructions are not for a reasoning task. Are literally for a
11:00
reasoning task. Are literally for a
11:00
reasoning task. Are literally for a gaming task. And oh, almost dead there.
11:03
gaming task. And oh, almost dead there.
11:03
gaming task. And oh, almost dead there. But hey, this is kind of the idea. Oh,
11:05
But hey, this is kind of the idea. Oh,
11:05
But hey, this is kind of the idea. Oh, sorry. Yeah, it's not giving not even
11:08
sorry. Yeah, it's not giving not even
11:08
sorry. Yeah, it's not giving not even giving me a response. So, let's try
11:10
giving me a response. So, let's try
11:10
giving me a response. So, let's try another model. The last model. Let's try
11:12
another model. The last model. Let's try
11:12
another model. The last model. Let's try an example lama 3.2 vision. This is the
11:17
an example lama 3.2 vision. This is the
11:17
an example lama 3.2 vision. This is the one that I used to do locally for when I
11:20
one that I used to do locally for when I
11:20
one that I used to do locally for when I was doing images. But this is also good
11:23
was doing images. But this is also good
11:23
was doing images. But this is also good enough to process this in string mode.
11:25
enough to process this in string mode.
11:25
enough to process this in string mode. So, go back here.
11:27
So, go back here.
11:27
So, go back here. Let me go back to let's clean the
11:32
Let me go back to let's clean the
11:32
Let me go back to let's clean the terminal. Play it again.
11:35
terminal. Play it again.
11:35
terminal. Play it again. It's playing there. Enable the lama part
11:38
It's playing there. Enable the lama part
11:38
It's playing there. Enable the lama part and you can see in the back that is
11:40
and you can see in the back that is
11:40
and you can see in the back that is starting
11:41
starting
11:41
starting and it's going to take some time until
11:43
and it's going to take some time until
11:43
and it's going to take some time until basically load the model in memory and
11:45
basically load the model in memory and
11:45
basically load the model in memory and start to work with the model. But at any
11:47
start to work with the model. But at any
11:47
start to work with the model. But at any time soon we should start to get
11:49
time soon we should start to get
11:49
time soon we should start to get resources responses there. So it seem
11:51
resources responses there. So it seem
11:51
resources responses there. So it seem that the model is already loaded and a
11:53
that the model is already loaded and a
11:53
that the model is already loaded and a couple of seconds later we should start
11:55
couple of seconds later we should start
11:55
couple of seconds later we should start to okay. So hey, this model seems to
11:57
to okay. So hey, this model seems to
11:57
to okay. So hey, this model seems to also work nice. And again, I think there
12:00
also work nice. And again, I think there
12:00
also work nice. And again, I think there is getting stuck in the bottom right
12:02
is getting stuck in the bottom right
12:02
is getting stuck in the bottom right because I was playing around with the
12:03
because I was playing around with the
12:03
because I was playing around with the prompt and I need to improve my prompt.
12:06
prompt and I need to improve my prompt.
12:06
prompt and I need to improve my prompt. But you get the idea. If I want, by the
12:07
But you get the idea. If I want, by the
12:07
But you get the idea. If I want, by the way, if I want to change this, I usually
12:09
way, if I want to change this, I usually
12:09
way, if I want to change this, I usually go here, open my project, go to
12:12
go here, open my project, go to
12:12
go here, open my project, go to compilot, go to my game provider base,
12:15
compilot, go to my game provider base,
12:15
compilot, go to my game provider base, and I will do something using the
12:18
and I will do something using the
12:18
and I will do something using the current file, which is game provider
12:21
current file, which is game provider
12:21
current file, which is game provider base. update the prompt to be more
12:26
base. update the prompt to be more
12:26
base. update the prompt to be more aggressive.
12:30
I focus on early and focus on uh
12:35
I focus on early and focus on uh
12:35
I focus on early and focus on uh [Music]
12:37
[Music]
12:37
[Music] firing the enemies. So I am going to ask
12:40
firing the enemies. So I am going to ask
12:40
firing the enemies. So I am going to ask a pilot. I'm using GPT for one. Use
12:43
a pilot. I'm using GPT for one. Use
12:43
a pilot. I'm using GPT for one. Use whatever model you are feeling more
12:44
whatever model you are feeling more
12:44
whatever model you are feeling more comfortable here. If you want to use uh
12:47
comfortable here. If you want to use uh
12:47
comfortable here. If you want to use uh clouds on it or Germany or more
12:50
clouds on it or Germany or more
12:50
clouds on it or Germany or more depending on what you have there. I'm
12:52
depending on what you have there. I'm
12:52
depending on what you have there. I'm going to ask this and it's probably
12:53
going to ask this and it's probably
12:53
going to ask this and it's probably going to suggest me. Oh, there it is.
12:55
going to suggest me. Oh, there it is.
12:55
going to suggest me. Oh, there it is. So, it's going to suggest me to a couple
12:57
So, it's going to suggest me to a couple
12:57
So, it's going to suggest me to a couple of changes here in the in the prompt.
13:00
of changes here in the in the prompt.
13:00
of changes here in the in the prompt. And hey, oh, a big change. Not a couple
13:02
And hey, oh, a big change. Not a couple
13:02
And hey, oh, a big change. Not a couple of changes, kind of a big one to
13:05
of changes, kind of a big one to
13:05
of changes, kind of a big one to basically say that, hey, you know what?
13:07
basically say that, hey, you know what?
13:07
basically say that, hey, you know what? This is the way that we can be more
13:08
This is the way that we can be more
13:08
This is the way that we can be more aggressive hitting everything. Play
13:11
aggressive hitting everything. Play
13:11
aggressive hitting everything. Play around with this. I like it and it works
13:13
around with this. I like it and it works
13:13
around with this. I like it and it works great. And once again, if you want to
13:15
great. And once again, if you want to
13:15
great. And once again, if you want to take a look at the code of everything,
13:17
take a look at the code of everything,
13:17
take a look at the code of everything, you can take a look here at in the
13:20
you can take a look here at in the
13:20
you can take a look here at in the generative AI for beginners.NET
13:23
generative AI for beginners.NET
13:23
generative AI for beginners.NET repository, you can take a look at the
13:24
repository, you can take a look at the
13:24
repository, you can take a look at the mega prompt that we use at the beginning
13:26
mega prompt that we use at the beginning
13:26
mega prompt that we use at the beginning to create this. You can take a look, we
13:28
to create this. You can take a look, we
13:28
to create this. You can take a look, we have a onehour video in the .NET channel
13:30
have a onehour video in the .NET channel
13:30
have a onehour video in the .NET channel where we explain all of the details. And
13:33
where we explain all of the details. And
13:33
where we explain all of the details. And then you have also you also have here
13:35
then you have also you also have here
13:35
then you have also you also have here kind of the requirements if you want to
13:37
kind of the requirements if you want to
13:37
kind of the requirements if you want to use locally or in the cloud. If someone
13:40
use locally or in the cloud. If someone
13:40
use locally or in the cloud. If someone wants to implement this using another
13:41
wants to implement this using another
13:42
wants to implement this using another provider happy to please add the new
13:45
provider happy to please add the new
13:45
provider happy to please add the new code or add the the necessary changes
13:49
code or add the the necessary changes
13:49
code or add the the necessary changes submit the PR I will be happy to add it
13:51
submit the PR I will be happy to add it
13:51
submit the PR I will be happy to add it here the PR and what we have here is
13:54
here the PR and what we have here is
13:54
here the PR and what we have here is basically the steps that you need the
13:55
basically the steps that you need the
13:55
basically the steps that you need the step that you need to do how it works a
13:57
step that you need to do how it works a
13:57
step that you need to do how it works a couple of notes and more it's kind of a
13:59
couple of notes and more it's kind of a
13:59
couple of notes and more it's kind of a nice one but going back to the beginning
14:02
nice one but going back to the beginning
14:02
nice one but going back to the beginning if you want to know more how to set up
14:04
if you want to know more how to set up
14:04
if you want to know more how to set up your environment to start to do this
14:06
your environment to start to do this
14:06
your environment to start to do this type of activities games and more the
14:08
type of activities games and more the
14:08
type of activities games and more the generative AI
14:09
generative AI
14:09
generative AI uh repository is a nice place to start
14:11
uh repository is a nice place to start
14:11
uh repository is a nice place to start and you have here the short link which
14:13
and you have here the short link which
14:13
and you have here the short link which is going to also probably going to be
14:15
is going to also probably going to be
14:15
is going to also probably going to be available so everybody can start so
14:17
available so everybody can start so
14:17
available so everybody can start so happy to get some feedback I know a game
14:19
happy to get some feedback I know a game
14:20
happy to get some feedback I know a game is just a sample it's just a simple
14:21
is just a sample it's just a simple
14:22
is just a sample it's just a simple sample but we have also end to end
14:24
sample but we have also end to end
14:24
sample but we have also end to end activities here using a spark SQL server
14:27
activities here using a spark SQL server
14:27
activities here using a spark SQL server SQL server 2025 you can find a lot of
14:29
SQL server 2025 you can find a lot of
14:29
SQL server 2025 you can find a lot of also enterprise scenarios as part of the
14:32
also enterprise scenarios as part of the
14:32
also enterprise scenarios as part of the repository so I hope you like it feel
14:34
repository so I hope you like it feel
14:34
repository so I hope you like it feel free to contact me if you have any
14:36
free to contact me if you have any
14:36
free to contact me if you have any question and hey Simon what do you think
14:38
question and hey Simon what do you think
14:38
question and hey Simon what do you think about this.
14:39
about this.
14:39
about this. That was absolutely great. I mean, you
14:41
That was absolutely great. I mean, you
14:41
That was absolutely great. I mean, you know, you made things look so simple and
14:43
know, you made things look so simple and
14:43
know, you made things look so simple and it is actually, you know, building AI
14:46
it is actually, you know, building AI
14:46
it is actually, you know, building AI applications with C# has never been so
14:48
applications with C# has never been so
14:48
applications with C# has never been so easy until now. And what I really loved,
14:51
easy until now. And what I really loved,
14:51
easy until now. And what I really loved, you know, is as you mentioned in one of
14:53
you know, is as you mentioned in one of
14:53
you know, is as you mentioned in one of your slides or I think in the GitHub is
14:55
your slides or I think in the GitHub is
14:55
your slides or I think in the GitHub is how beautiful the Azure Azure AI founder
14:59
how beautiful the Azure Azure AI founder
14:59
how beautiful the Azure Azure AI founder is, right? You can go there, you can
15:01
is, right? You can go there, you can
15:01
is, right? You can go there, you can find all the different models and it
15:02
find all the different models and it
15:02
find all the different models and it it's a very intuitive way. It tells you
15:04
it's a very intuitive way. It tells you
15:04
it's a very intuitive way. It tells you step by step on how you can build
15:06
step by step on how you can build
15:06
step by step on how you can build applications and then it may be C# or
15:08
applications and then it may be C# or
15:08
applications and then it may be C# or whichever programming language you like.
15:10
whichever programming language you like.
15:10
whichever programming language you like. You can go ahead and build it. For
15:11
You can go ahead and build it. For
15:11
You can go ahead and build it. For everyone's watching, you'll find the
15:13
everyone's watching, you'll find the
15:13
everyone's watching, you'll find the GitHub link down in the video
15:15
GitHub link down in the video
15:15
GitHub link down in the video description. I'll also put Bruno's uh
15:17
description. I'll also put Bruno's uh
15:17
description. I'll also put Bruno's uh LinkedIn and Twitter handles in the
15:20
LinkedIn and Twitter handles in the
15:20
LinkedIn and Twitter handles in the description. Please feel free to go
15:21
description. Please feel free to go
15:22
description. Please feel free to go ahead and reach out to him on
15:23
ahead and reach out to him on
15:23
ahead and reach out to him on everywhere. I think he's active on all
15:25
everywhere. I think he's active on all
15:25
everywhere. I think he's active on all the platforms. So, thank you so much,
15:27
the platforms. So, thank you so much,
15:27
the platforms. So, thank you so much, Bruno. Any final thing you want to plug
15:28
Bruno. Any final thing you want to plug
15:28
Bruno. Any final thing you want to plug in before we wrap up the show?
15:30
in before we wrap up the show?
15:30
in before we wrap up the show? No, no. Uh we are I am looking for
15:34
No, no. Uh we are I am looking for
15:34
No, no. Uh we are I am looking for feedback. I am looking all the time to
15:36
feedback. I am looking all the time to
15:36
feedback. I am looking all the time to see how we can do better samples, how we
15:39
see how we can do better samples, how we
15:39
see how we can do better samples, how we can help communities. If you have
15:40
can help communities. If you have
15:40
can help communities. If you have question and more, feel free to contact
15:42
question and more, feel free to contact
15:42
question and more, feel free to contact me. I will be happy to try to answer or
15:44
me. I will be happy to try to answer or
15:44
me. I will be happy to try to answer or at least pass the question to the right
15:46
at least pass the question to the right
15:46
at least pass the question to the right person. And I'm taking some time to
15:49
person. And I'm taking some time to
15:49
person. And I'm taking some time to answer. There are a lot of questions and
15:51
answer. There are a lot of questions and
15:51
answer. There are a lot of questions and more, but I'm really really looking
15:52
more, but I'm really really looking
15:52
more, but I'm really really looking forward to see what people are building,
15:54
forward to see what people are building,
15:54
forward to see what people are building, how we can help, and how we can do
15:55
how we can help, and how we can do
15:56
how we can help, and how we can do AI.NET the whole ecosystem a better one.
15:59
AI.NET the whole ecosystem a better one.
15:59
AI.NET the whole ecosystem a better one. So no, that's it for me. And hey, once
16:01
So no, that's it for me. And hey, once
16:01
So no, that's it for me. And hey, once again, thanks everyone and thanks Simon
16:03
again, thanks everyone and thanks Simon
16:03
again, thanks everyone and thanks Simon Simon for inviting me.
16:04
Simon for inviting me.
16:04
Simon for inviting me. Yeah, thanks everyone for watching and
16:06
Yeah, thanks everyone for watching and
16:06
Yeah, thanks everyone for watching and thanks to Bruno and we'll see you in the
16:07
thanks to Bruno and we'll see you in the
16:07
thanks to Bruno and we'll see you in the next episode of AI Deep Tooth today.
16:09
next episode of AI Deep Tooth today.
16:09
next episode of AI Deep Tooth today. Bye-bye.