[00:00:00] Speaker 1: In this video, we're going to be talking about Vertex AI from Google and how it empowers and enables businesses and developers to thrive with AI. I've been using Vertex AI for just over two years and we're just going to scratch the surface of some of the cool features and cool little things that you can do within Vertex AI. It's been announced at the Google Cloud Next and Vertex AI is now the only platform with generative media models across all the modalities, so video, image, speech, and even music. In this video, we're just going to expand and we're just going to go into a little bit of detail on some of the cool features and how to navigate Vertex AI. If you enjoy this video, I've got a great little course which shows you how to set up Google Cloud, how to build your first agents, and how to build robust enterprise applications using Vertex AI. So if you want to support me, get that course and let's get going. As we can see over here, Vertex AI is now the only platform with generative models and a great little post on LinkedIn over here. So we've got VO2, which is a video generation model. We've got Imagine 3, which is allowing you to create images and some in-painting. So in-painting is really when you've got an image and you want to add something to it, like you can see in the picture over there. So Chirp 3 is about creating some audio from your own voice or from other voices, just pop in some text. And then Lyria is text to music, which is pretty fantastic actually. So we're going to go through all of these at the moment, and we're also going to go through some of the existing features, the models, the agent guard, and building agents, and just really quickly just giving you an overview of it. So if you don't already have Google Cloud and Vertex AI, I've got another video which will show you exactly how to set it up. Or if you'd like a full course, go down to the description below and hit that enroll button and it'll show you exactly how to set up everything you need to do. All right, so let's just jump into Vertex AI, and you're going to be presented with a screen very similar to this once you've landed in. All right, it looks quite complicated and quite busy, but really, really simple, really easy. On the left-hand side, you're going to have your dashboard, your model garden, and all of the options that you may or may want to use for your AI development or exploration. All right, so you've got your dashboard, you've got your model garden, and the model garden is a really fantastic little tool to show you exactly what models are available within the Vertex AI environment. So you've got Gemini, you've got Imagine, you've got Weather, you've got Llama, you've got Anthropic, Meta, Hugging Face, Mistral. So you've basically got all the models that you could possibly want at your fingertips, and really simple to spin up and to use. You've also got different type of modalities, different types of tasks that you might want to use. So again, really fantastic for your model garden. I'm just going to jump down into your Vertex AI Studio, and Vertex AI Studio is really about creating your text outputs or to create any of your generative AI chatbots, and basically a beautiful, fantastic environment to test, prototype, and ultimately refine how you work with your generative AI and large language models. Again, really fantastic, and I've got a huge amount of videos in my collection. Go and have a look at those or have a look at my course. All right, we've spoken previously about streaming and how to use the streaming function. Go and have a look at my previous videos to figure that out. And again, really, really powerful. We've got some awesome things around building AI agents. All right, so we've got AI Agent Builder that is now part of the ADK software development kit, and you can now start using programmatic agents being built and some samples and templates to get best practice going over here. You've also got all these different connectors that you can work with. Again, extremely, extremely powerful. All right, you've got AI Agent Engine, which is coming soon. All right, we'll basically talk about that when it comes out. You've got the Vertex AI Search. So if you want to create your own little RAG application, which is a Retrieval Augmented Generative AI, you can start creating it over here, or enterprise-grade, or very robust, or very secure. And there's, again, some videos on how to actually do that in my previous videos, as well as in my course. So again, have a look at it, and it's really, really powerful. But let's just jump into a couple of things that we wanted to talk about today, and that's really around the Media Studio. So these are some of the cool new features that have been created within the Vertex AI environment. And you've got your documentation, you've got your API reference information, if you wanted to. And here's a couple of samples over here. But what we're going to do is we're going to build our own. So let's do a quick little test over here. So let's do a first little image, and we click on it. And a really simple, easy canvas over here. So you can upload an image if you want to do your in-painting. So that's where you've got an image of yourself, or whatever you want, and you wanted to add something to it. That's what the upload image would be for. You can now choose your models. Imagine 3 is the best model. And you can choose the others just to try and see which ones they are that you like. And we'll just do 16.9, so it's nice and wide. How many results would you like to create? We can do 4. We can do references. So if you wanted to add a reference, you can add an image as a base image, and you can put in a description. And this would now use it as a reference point, so you can try and get character consistency. So if you've got a person, you can get consistency in that particular person, which is an absolutely fantastic feature. And I hope you use it, because I think it's great. And we'll go into a bit more detail of how to use it in subsequent videos. What you can do over here, safety. Obviously, don't allow all ages. All right. Do we want to block anything? Let's just move that away. We don't want to block something. Most things will block few. We don't want to block much, right? Advanced options, you can choose the region where you want to be created. So let's pop in a prompt over here, and I've got a prompt ready. And we're going to just go there, and let's hit the Enter button. So we want to create a time lapse of a futuristic cityscape where people live and work in giant inflatable structures. Okay. So it's made very, very short work of that. And now you can click on it. You can do some AI actions of it. You can inpaint it. You can outpaint it. You can even generate a video using this as the initial key frame. Go into too much detail in this particular video. This is just a quick snippet of how easy and how quick and fast things are here. All right. Let's go back to the media studio, and let's go create some audio. All right. So we're going to do a... Oh, let's just do something really simple, like today is a magnificent day in the age of AI. Are you excited? On the right-hand side, you'll notice that there's different models that you can choose. At the moment, it's only the CHIRP 3. You can choose your different languages, English, United States, and then the voice that'll be determined by the type you want, so male, female, et cetera. And let's just go enter over here. All right. So if you've never used it before, it's going to ask you to just go in and enable it. So let's go in and enable it. And let's pop in the developer link over here. And you'll probably have to just enable an API. All right. So we want to enable the cloud text to speech API. And let's just go enable. All right. You're only going to be charged for what you use. So don't feel that you can't try things out. So let's go back to Vertex AI. Let's go close and we can try that again. All right. So now we're going to generate the audio, and it's going to make really, really quick work of that. Today is a magnificent day in the age of AI. Are you excited? All right. So really quick, really easy to create text to speech. And you can obviously change and you can modify this to your own voice as well. So that's going to be quite good. And it's an awesome little feature coming out. All right. Let's do some music. So we always want a bit of music. So let's do a bit of EDM music. So let's go an EDM energetic with slight tribal undertone, dreamlike beat. All right. We want to use the Lyria model. All right. We can choose some advanced options over here. I won't choose anything at this stage. This is just to quickly check. I don't think I've been enabled for this just yet, but let's just give it a go. And yep. So it's basically failed at the moment. So the MusicGen API is basically being developed and worked on. It should be available in the next couple of days. That's what Google told me. So we'll just keep that on the side for the moment. All right. How about we go in and create a video? So let's go create a video real quick and let's see what it does. So we'll take that same image that we used previously for the bubbles, and we can now see which model we want to use, the VO2, choose your aspect ratio, number of results, how long you want the video to be, eight, seven, six, five. If you think about the way videos are created, there's usually breaks in between. So every three to four seconds is a change in scenery. So having something longer than eight seconds is not actually great because you need to be changing and telling the story as you're moving along. So eight seconds for the time being is actually okay. You can choose where you want to save the file. So you can put it into a storage bucket if you've got storage buckets already created, prompt enhancement. So this is actually quite a great feature. So if you put in a prompt, it'll actually make your prompt even better to make sure that the video generates correctly. And you can again do some advanced options and you can choose the safety. All right. So let's just go ahead, enter, and let's see what it does. All right. So as you can see, it's created four different videos over here and it's just played this one over here and let's see what it looks like. So eight seconds, as you can see, it's moving through these little bubble pods and that looks pretty good. Let's try this one over here and yeah, it looks pretty good. I mean, it's eight seconds of some pretty fantastic. So, all right. And then a couple of other things that I wanted to show you really quickly is that we've got some awesome other features inside Google just showing you AI applications, which is basically like an agent builder. So Vertex AI offers you the environment to create your own agents. And it's a really powerful, really awesome space to go in and create your little applications, which are production ready that you can deploy. So you can create custom searches, media searches, healthcare, conversational agents, chat agents, and a whole bunch of different types of yes. If you want to learn a little bit more about this, go check out my other videos, check out my course, and it'll ultimately give you all the information you need to start creating real life, real world applications for using generative AI. That was a quick whirlwind tour. Hopefully you enjoyed it. If you'd like to see anything specific or a little bit more detail, let me know in the comments below and thanks for watching.