Google AI Creative Essentials - Gemini, Flow Veo 3.1, Nano Banana Pro, Notebook LM, Music FX

1. AI Flow state Course Intro : Hi, my name is Greg Kong, a Canadian creator, and welcome to the flow state Mastering Google's AI tools. So in this course, we're going to look at several groundbreaking tools that Google has released either through their lab or mainstream products. We're going to be covering Google Flow, which has their groundbreaking AI video model called Vo three that we're going to use to learn how to create really cinematic and believable video but with sound effects and audio to complement it. And then we're going to look at Notebook lm, one of the products of the year created by Google to feed its sources on a specialized topic to use for a study guide for groundbreaking research, study topics, and also to create audio podcasts and not only an audio podcast, but an interactive podcast that you can ask questions of the AI and the actual podcast host. You can also use Google Gemini Ultra plan to further enhance Notebook alm using discovered sources right from the Internet, so we're going to be covering that in the course. We're going to tie in how you can use Notebook ALM to tie it in with Google's flow. But some other tools that we're going to cover in this course include WISC, which uses a groundbreaking image model called Imagen to create not only amazing images, but how to animate them and to use images to reverse engineer into a text prompt, which you can use to generate and refine further images. We're also going to look at music tools. So music effects to create music using text prompts to generate beats for your YouTube videos, for your online courses. So this course is going to be really interesting for content creators, digital marketers, entrepreneurs, creative professionals who want to learn the cutting edge tools that are available on the market to do things like research, create audio podcasts, to create videos and images. To drive your business to new levels using the help of AI to accelerate what you would normally take a longer time to do just on your own or with a video production team. Next step is to enroll and we'll see you inside the course. 2. Accessing the apps: Okay, so let's get started. The main address, URL for accessing these applications is labs do Google fordslashFX. Labs dot Google fordslaFX. Login with your Google account, and you'll get access to these four applications. No Book M is going to be accessed from a different URL. So Notebook LM is access from notebookmggle.com. But all the other applications are at labsgooglelash FX. Alright, so let's just go through one by one before we deep dive into each of these applications. So WISC is the application you use to create images and animate your images. All right. And I don't have a subscription on this account, and you can tell that I can actually use it without a paid plan. So you have zero AI credits on this one. Okay, now Flo, this one does require a subscription to use it because it's saying you must be a Google AI subscriber to generate. And if you want to subscribe, they do have a promotion one month for free, and then becomes $27 Canadian. So that's good for getting your feet wet. But if you want to deep dive, I recommend going to the Google AI Ultra, but first use the Google EI Pro, and you'll quickly use up your credits if you're really into it, like myself. And then you may have to upgrade. So I think the Google AI Ultra right now, 170 for three months is a good deal. You get YouTube premium 30 terabytes of Gmail storage, Google Drive storage for photos, which is great. Also on flow, the filmmaking tool, you get access to VO three. And what they don't tell you here is they remove the water marking on the videos. And you also get a lot of credits for use in flowing with 12,500. Okay, going back to our dashboard, Image FX. We're not actually going to go into image FX. We're going to use WIS to generate images with their imaging or imagine AI model. Okay, Music effac, this is an exciting one. So we're going to use music effac to generate tracks and music Ffcs DJ to generate some beats on the fly. Alright, so that's how we access the applications that are gonna be covered in this course. 3. Prompting Best Practices Flow: What are some prompting best practices for Google's flow with VO three? Well, it's basically being creative and as detailed as you can be and experimenting in a nutshell. But what I would say is there's a lot of different types of videos that can be created. So at the beginning of the course, you went through flow TV. That's obviously a very good starting point to find the style video that you like and reverse engineer it by getting the prompt and then using that yourself or remixing the prompt to see what you can create. I'll share some of my video projects here, and then I'll show you how you can actually use another application by Google, Geminiz AI to create the prompt for flow because the actual prompt within flow doesn't have any well, a lot of AI intelligence, so you can actually enhance your existing prompt like you can in some applications. So let's take a look at this project here. This is the Yeti. So I'm just going to share some of the prompts I've used for my projects. So we have the Yadi here. So this is the prompt self camera, Angleshaw from an extended arm, a white Yeti holding up a golf ball in the course, dialogue. Welcome back to Uncle Greg's Golf Adventures. Today, we're going to learn how to play boogie golf with Kirkland balls. Let's play through one of these videos, and I'll turn the sound on. Welcome back to Uncle Greg's golf Adventures. Today, we're learning how to play bogie golf with Kirkland balls. Welcome back to Uncle Greg. So you can see here the important points are describing what the character is, what color they are. What is the angle? So self camera angle shot from an extended arm. And then for dialogue, actually put dialogue semicolon. And then in the quotes, the actual dialogue that I wanted the character to say. So being descriptive with the prompt there, and I use the Video three model. So the actual model that you use actually matters. So let's just go to something else here. Let's look at the pickleball video project. So if you're very broad and not very descriptive, you're going to leave it up to the AI model to create the videos. So you can see four different outputs for a simple prompt, pick a wall coach explaining what the kitchen zone is. So we have three different characters here. Let's just give you an example. So this is the kitchen. It is a seven foot area from the net back to this line. So this is the kitchen. It is a seven foot area from the net back. Okay, now let's listen to another output for the same prompt. So the kitchen, this front area up here at the neck. You can't volley when your feet are inside the kitchen. So the kitchen, this front area. So, completely different character, slightly different message. So that's what happens when you're very broad and general with your prompt. The more specific, the better. Here. And let's go back out to another project. Actually find for cinematic sci fi videos. It does a pretty good job here. Here's an example where I put several Jedi Mandalorians with dark sabers battling Neo from the matrix before they fight the leader. And the leader says, This is the way, and Neo says, I know Kung fu. So let's just take a look at one of these outputs. This is the way. I know Kung fu. This is the way. I know Kung Fu. This is the way. So that's an example of a prompt that's a bit more specific. I describe the type of character, some of the dialogue, and makes some assumptions there, of course. As you can see, each output is slightly slightly different. This is the way. I know Kung fu. This is the way. Alright, let me just show you one more. So this project here actually put quite a lot of work into the prompt. So I asked it to recreate the scene of operation Midnight Hammer, a US military operation that struck the Iranian nuclear facilities with seven B two spirit bombers, show the BTs dropping a total of 14 massive buster bunker bombs. So let's watch this. So I was pretty happy with that. That was pretty good. But let's just say if you actually want to create a cinematic scene, you can just see how long this prompt is here. It's basically setting the scene. It's creating the characters, and it's actually specifying the dialogue for these characters. And I'll just show you an example of this eight second clip of how detailed you can get. Spector one, this is Oracle. Target locked. Munitions armed. Ready for release on your command. Inspector one, this is Oracle. Target locked. Munitions arm. Alright, so I didn't use the I mean, I didn't create this by myself. I used the help of AI, Google's Gemini, using one of the latest models, 2.5 to actually create this script. So my original prompt was create a prompt to re enact the scene of the US Army dropping the bombs over Iran's nuclear facility and created this prompt here. So I had to modify it slightly. But, you know, this really helped me create a super detailed prompt. And you can just see the difference in the output. It's very specific. The dialogue was there. It actually created the characters from the script. So yeah, there's different ways you can create these videos, and your prompts can be super detailed like this with, like, a mini movie script to having your own detail prompt describing the characters, what they say. Or you can be really broad and general and let the model make the assumptions, or you can reverse engineer using Google Flow TV. So those are some good general best practices for prompts in Google's flow with VO three. 4. Veo 3 walk through: Hi, my name is Greg, and I'm going to give you a walk through of Google Flow. Which has their groundbreaking AI model called Vo three, used to generate really cinematic videos or any type of video you can imagine, really. And the really cool feature is not just the video, but the audio and sound effects that can complement the videos. So we're on Google's lab site right now, labs do Google furlashFX. And this is where you can access their other creative tools like Wisk, flow, image FX, music effecs. But we're going to take a look at flow here. And this is not a free tool, but if you get the Gemini Pro or the Gemini Ultra plans, you can access flow. So I'm going to give you a tour of flow, and then we'll get through generating our very first video. So here's the main page where you can see different projects that you have and organize by date and the way I like to do my projects is I just like to group them by certain categories. For example, I've got more of my Star Wars types of videos, so I can just rename those there, or we've got more of the comic books superhero genre videos over here. And I've got a whole lot of other videos. I've got war series videos, sports. So once you click in, so I'm actually just going to click into my Star Wars project area. And once you click into a project, you'll see the different variations that you've had with different prompts. For example, over here, if I click in, I can see the prompt here. Several Jedi Mandalorians with dark sabers battling Neo from the matrix. Before they fight the leader, Mandalorian says, This is the way, and Neo says, I know Kung fu. So this prompt that's highlighted, this has created four different videos over here. And each of these videos, we have the ability to download them. We can full screen it. We can add them to a scene which will cover in another video in depth. And if you click the three dots, you can delete it. You can flag the output. This is the way. And, of course, you can play the video. I know Kung fu. To actually see what it looks like. So on the project page, it defaults to a four square grid view. But there's another option here which will give you a bigger overview of all the projects that you have within a project or all the videos you've created within a project. So let's just back out. Let's go back to the main page. On the top right, we have flow TV. So flow TV is basically an idea where you can watch different videos that are already created to generate inspiration. So these are different channels, and I'll go into it in another video, but you can check out these other channels, and you can actually reverse engineer the prompt. You can show the prompt, and then you can bring these prompts into your own video projects. So that's flow TV, and again, we'll cover that in another video. So let's just go back to flow here. And on the top right, another important area is the plan that you're on. So you can see here I'm on the ultra plan on the top right, and you can see the number of EI credits. So if you click onto the EI credits, you can see your current balance as well as how your credits have been used. For example, I just created some videos today that were worth 20 credits, and if I just change the page, I can see I created some yesterday that were worth ten credits. So just know that with certain plans, for example, I'm on the Gemini Ultra plan. You get something like 12 to 13,000 credits, and then each video generation uses up a certain amount of credits. For example, if you use the VO three fast model, that'll use up 20 credits. If you use the VO three regular model, that'll use up 100 credits. And I'll show you where you can see the cost of the video generation. So I just wanted to point that out to you. And you can manage your membership over here. And you can see your plan. So I've got the Google AI Ultra and all the benefits that comes along with it. So in addition to flow with VO three, I've also got WIS, 12,500 monthly AI credits, notebook um, with the highest limits, et cetera, for currently $170 Canadian for three months. So these plans are going to be constantly changing. So let's go back to flow, and let's look at creating our very first video. I'll actually show you how you can reverse engineer a quick prompte from flow TV. So I'm going to flip through the different channels. This one's called Fit Check. And I'm just going to go into the grid view. So I'm basically looking for one that I find pretty interesting. Okay, so I'll click in here. I can copy the prompt. And these are based on VO two. So if you put it in Vo three, you may get a different result. So I'm back on flow right now. I'm going to click New Project. And this is our main prompt area. So by default, it goes to text to video, but you can also go from frames to video ingredients to video, which we'll cover in a separate session. The default model is VOT fast that you can see on the right side. But if you click settings, you have the ability to change the outputs per prompt. So I like to do four because there's no additional cost and you get more choices. See if I click two, it's still ten credits to generate this. So I'm going to leave it on four. It's currently at Vo two FAST, and it'll tell you each generation uses ten credits based on your current settings. If we change it to Vo three FAST, that changes to 20 credits. View two quality. That's 100 credits, and Vo three quality is 100 credits as well. So we're just going to go to Vo three, fast text video, which costs 20 credits. I'm going to paste I'm going to paste that prompt in high fashion editorial address made of living jellyfish underwater shot. Okay, I'm just going to click the prompt, generate our video. And because this is the fast model, it's going to be a bit faster, and you'll see the percentage over here while it's being generated. So you can actually go somewhere else and come back. So I'll just continue to scroll through the different channels here. Time to reflect. And you can see the different channels and an overview. So you get an idea of the really vast range of videos that you can create with flow. Okay, let's go back here. Let's see how we're doing. We're at 58%. Alright, so our video has generated, and I'm actually just going to go go to go back to the original original video that inspired this. I think it was this lady here with the jellyfish. So take a look at this. This is generated on Video two. And I'm going to play this So it's pretty cool, slightly different. So I have an eight second video that has audio, and I've got four choices here. Okay. Notice that some of the videos are longer than others. This one's 2 seconds. This is eight. This is eight, or the first one's five. All right, so if I want to go ahead and download this, I can click on the Download button, and then I have three options. I can download the original size, 720 P. I can upscale it to ten ADP. Or do an animated Gift 270 P. So animated Gift maybe you want to share that through WhatsApp. It's not going to take up a lot of space. 720 P is just going to be the video as you see it. It's not going to be too big, upscale. If I click this, it's going to bring it up to ten ADP, higher quality, larger file size, slightly larger. But one thing I want to note is that it doesn't cost any additional credits to upscale your video. And then lastly, you can add this to a scene, and we'll cover the Scene Builder in another video. So we covered quite a bit. We've covered Google Flow's product. We've covered flow TV. We did a quick reverse prompt engineering, and we generated some videos. We upscaled it to ten ADP. And I've introduced you to the Scene Builder. So the next set of lessons, we're going to drill into the Scene Builder. I'll share a variety of projects more advanced prompts to give you different ideas of how you can create these videos. So those will be in the next lessons. And lastly, once you finish upscaling the project on the top right, it'll give you a download link that you can download your video, and then you can use it on social media, YouTube, wherever you want. And if you're on the Gemini Ultra plan, it'll download the video without a watermark. If you're on the pro plan, you're going to get that via watermark on the top right. This is the video that I've downloaded in ten ADP. So we'll end the video here. 5. Intro to Flow Tv & the scene builder: Okay, this is the flow TV product from Google. So basically, this is like television, except you're watching videos created with flow. And the reason why you would want to come here is to get some inspiration. You could also reverse engineer videos that you like. So for example, let's just go to the main screen. So over here, it's like your remote control, so you can see all the channels. So if we see category we like, we can click into Okay, this is like no one's watching. So from this view, we could go onto the Next clip. We can full screen the videos. We can unmute you can see we got a variety of different videos. So let's just say we like that one. We can click this radio button to show the prompt. I'm just going to mute it right now just for ease of recording. So you can hear me properly. We have a grid view here if this is easier just to look at a glance at different types of videos. So I'm just taking a look at some of these. They look cool. That's skeleton, so you can change channels over here. So this is called Time to reflect. And you can also just search for a category, right? So let's just say we like drone videos. So we can get a nice filter here. Let's click on this hockey one. This looks pretty cool. So this is a drone tracking video of people playing hockey somewhere. So we can actually just go back, click Show Prompt. And the magic sauce is this prompt, so I can copy that. And let me just pause this. So that's how we can use flow TV. So now I'm going to show you how we can use that inspiration to create something similar. So we're going to flow now. And while flow is loading, I'll just show you another browser tab where I use the prompt to create a video and I'll just show you the prompt right over here. So it's a really long one, as you can see, I'm not going to read it out for you, but I'll just leave it on the screen. And what I may do is I'll just include this in the prompt pack. Let's just play this. So that gives you an idea of how we can use or reverse engineer from flow TV to creating our own creation. Okay, here we are. So this is my flow, and I've got my different products. So when I create Krojax I like to try to keep them together. Let's just say I've got ones with drones or ones with pickleball or ones that are golf related. I want to try to keep them together. So this one I'm going to This is like my reverse engineer area. So now we are in flow. Now I got to select my model here. So V two FAST. See, this will use ten credits. If I go to Vo three, quality, this will use up the most credits, 100 credits. So I'm going to go Vo three Fast, which uses 20 credits, five times less. So that's a good way to start out. Maybe you want to experiment with the lower models first. And then if you really like something, you can change to Vo three quality, which is the highest model. Now, things are continually changing. You can see here that if you use Vo three, the audio may not work. So I'm basically going to just copy and paste that prompt from flow TV into here. And because we're using Vo three fast, it's going to be faster than Vo three quality. So we get four different squares or variations that we get to choose from. And if you click the settings here, this is where you can change not only the model, but outputs per prompt. We also have a different view here. So this allows us to see multiple projects on the same page, and we can see here on the top right which model or which Gemini plan we have. I'm on the ultra plan right now, and I can see the number of credits. Now, if you want to see how your credits are doing, how they've been spent, you can see here my credits have been spent. So this is just from today. And I actually had some generations that didn't work because I use popular characters. I use the US presidents and that got refunded because they couldn't generate it. So if I go back from yesterday, you can see a lot use a lot of VO three FAST credits because those are about 20. Okay, let's go back. We are at 90%. So we have one generation here. Let's check it out. Okay. Interesting. So not exactly like the flow TV one. Let's just go back here. And you can see this was done with the VO two model. So maybe we got to change the model to get something the same. Okay. So these are a bit of a wider angle. They're not as close up as the flow TV video, even though it's the same prompt. So let's just say I like this third one here. I have the option to download it. I can upscale it to ten ADP, create a 720 P, which is what it is now. This is not DVD. It's higher quality than DVD, but it's not quite full HD, and the file size will be smaller. So let me just upscale it. And then when it's ready, it'll tell you on the top right and you can download it. It doesn't use any extra credits. So you might as well upscale it if there's a clip that you really like. Okay, so let's just try the prompt again except I'll do the same model as flow TV. So that's VO two. So VO two fast is ten credits, VO two, quality is 100 credits. So let's just do VO two fast. Let's just do two. Actually, let's do four since it's the same cost. And we'll just take a look at the difference for the same prompt, but comparing Vo three fast versus VO two fast. And you can see my upscaling requests in the top right. That's still happening. Okay. So we have our generation. Let's take a look. So this angle, even though it's the same prompt, it's pretty similar pretty similar to what we saw on Flow TV. Alright, let's try to build a scene with this VO two generation. So we're going to go in the top left of the video we like, select Add to scene. And what are we going to do is say frame as assets. So we can build a thumbnail. And then if we click Plus, beside it, we can select jump to or extend. So what I want is this to jump to a wider aerial angle orbiting the game. And we'll use the VO two FAST model. Alright, finish generating, so let's play it. So we have a regular scene. Yeah, that didn't work too well. Let's try it with VO three. Might be better in Vo three. Let's go add to scene. Okay, so you can see it's added this different scene over to the Video three and the Vo two clips to the same project. So I can arrange, so I can delete the second one and then move that in there. Now let's play it. Okay. And then we can actually can highlight this one. And close up of the goalie saving a puck. Okay. Okay, it's done. Let's play this back. Okay, so I asked it for a close up of the goalie, and it's sort of, it's done it, but it's a completely different scene. All right, so that's the Scene Builder. Definitely is easier to edit in Premiere Pro. But the Scene Builder, you can do it right within flow. It's a little bit finicky, in my opinion. There we go. So we demonstrated creating a video with flow, reverse engineering flow TV into flow. And I also demonstrated the Scene Builder. It doesn't work that well from my experience, but that may change. And we also upscale the video. We downloaded it. We compared doing VO three FAST versus VO two. I demonstrated how many credits are used between different models where you can check your AI credits. So we cover quite a bit. A 6. Frames to video veo 3: All right, so text to video is the most popular type of prompt. But there's also two different ways that you can also use the prompt to create video. So the second is called frames to video, which is basically screenshots that you use to create the video in addition to a prompt. And ingredients to video, you can use different files, or you can actually generate an image with text to create a video. So we're going to start off with frames to video. And I've actually already created a project using two screenshots of pickleball, and I've used the VO two model. So let me just show you how it did the job. There's no sound here. It's a little bit choppy. But you know what? It actually did a somewhat decent job. There's actually one where it actually changes the camera angle, and I wanted to show you this one. Okay, let's just expand this. I thought this was very interesting. So it changes the camera angle there. So it's basically taking the two screenshots and making an assumption and building its own video clip. Alright, here are some examples, other examples of how you can use frame to video. I'm using an AI generated image with another platform, Adobe Firefly. And what I've done here is I've fed a frames to video with a still, and I've asked her to do a flyover drone video at night over the Arce Triumph to the Eiffel Tower. So you can see here in the Scene Builder, which is covered in another lesson, you can save any of the frames as an asset for a future still and now I'm extending the scene just a little bit longer to include the Eiffel Tower. So this will give you a slightly longer clip, but it uses the VO two model. So this is another example of how you can use the frames to video. And this will also allow you to use a consistent type of visual or even a character. The frames to video is very powerful in your arsenal to creating videos using flow. So that's an idea of how you can use frames to video to use image stills to create a video using Google Flow and the video model. 7. Advanced Google Veo 3 : Get acquired weapons armed. Confirming release sequence. Standby. Bombs away. This is a bit more of an advanced tutorial on how I created the B two stealth bomber video. So the video was actually created in Premiere Pro. I created several scenes. That I'm going to share with you. But first, I'm gonna play the video, and then I'm going to show you how it was made from beginning to end. Spector I, this is Oracle. Target locked. Munitions armed. Ready for release on your command. Target acquired weapons armed. Confirming release sequence. Standby. Bombs away. Okay, so I actually used Gemini, Google Gemini to create the prompt. So I started with in Gemini 2.5 flash, create a prompt to re enact the scene of the US Army dropping the bombs over Iran's nuclear facility. So it came up with this quite detailed prompt actually with a resembles a movie script with different characters. And so it's got Major Thorn, which is the pilot, Captain Lina, Sergeant first class, David, and actually has a very detailed script. So I use the script in flow to create the scene, which I'll show you in this project. So it's this project here. You can just see how long it is, right? I'll just cover it off a little bit. It's a Mullis night over the Iranian desert. You board a B two Spirit stealth bomber, call sign Spector one, part of a small elite strike package. You targets, the heavily fortified Natans Enrichment facility. So it actually has the dialogue of Pilot one. So I just want to show you the amount of detail and a prompt you need to build something like this. Spector one, this is Oracle. Target locked. Munitions armed. Ready for release on your command. Okay, I'll just play another video. Target acquired weapons armed. Confirming release sequence. Standby. Bombs away. Target Acquired weapons armed. Confirming release sequence. Standby. Okay, so what I did is I upscaled these videos. I downloaded them, and then I put them into Adobe Premiere, which you can use other video editors like final Cap Pro. But basically, you just want to create a video project, and you may need to trim it and just edit it so it runs as a smooth video. Okay, so here we go. We are in Premiere Pro. And let me just play this video. It's quite a short one. Spector one, this is Oracle. Target locked. Munitions armed. Ready for release on your command. Target acquired weapons armed. Confirming release sequence. Standby. Bombs away. So I probably could have tightened up the edit a little bit more. But the idea was to put all the generated clips into the timeline, and basically that the result is a 22nd video. And if we export it, it's not going to be too large because this is only full HD, so we end up with a file that's 42 megabytes. So I just wanted to show you what is behind creating a short cinematic video using Video. So to summarize, we use Gemini to help us with the prompt to basically create a mini script. That script was used as the first prompt in Vo flow, and we used the Vio three model to create the video clips. I upscaled the clips in flow, downloaded them, and then I imported those clips into Adobe Premiere Pro, where I put it into the timeline to create a 22nd video. 8. Gemini Various use cases: Alright, Gemini 2.5 P. That's great for all purpose things from creating apps to asking questions. But one really useful use case is customizing a resume based on a job poster. So I've uploaded a current resume. Download it straight from LinkedIn, which is this file here, Profile 29 PDF. And I downloaded a job poster straight from the website. I turned it into a PDF. And my prompt was adjust my resume for this job poster. And what it did is it came up with came up with a resume, but it gave me the option to bring this to Canva. And then this is basically the Canvas on the right. And it just gives you, like, a better editing experience. And from here, I can basically copy and paste. I can turn this into a webpage, an infographic, a quiz. But I just pasted it to Google Docs and created a PDF resume from it. Just quickly, some prompts I've created in the pass on Google Gemini Pro 2.5. I've looked for barbecue parts. Basically, taking a photo, asking what barbecue model is this, and where can I find the replacement screw for the lid and the leg? It did a great job of identifying the barbecue, the parts, and potential website I can I can contact to get it. Alright, I've asked it to do research on the Pickleball algorithm, and it gave me a really great answer. Here's an example of an actual app to create beats. And if you highlight the canvas, it'll actually create the app for you. In this case, I think it didn't quite work. But let me just quickly show you create a Pacman game in all pink. So I've highlighted Canvas. And, you know, this is basically a vibe coding feature. It's not as advanced as vibe coding platforms, but it's getting good, and it's only going to get better. And you can actually use the code that it generates. I'll show you the code in the preview. And you can take that code. You can plug it in to a vibe coding platform to create a more advanced version. But the code that Gemini creates is a good starting point if you're looking to create an app and need a bit of assistance. So it's still thinking through. So this is the code view. But once it's done, it's going to actually have a playable version of the game. Well, that's the plan. There we go. Pink pack man, there was a bit of an error, so I can just I can just ask it to fix the game. Let's just create a new one. Create a tetris game. So Gemini Pro, you can think of it as a hat GBT. It's basically your all purpose AI assistant that can connect to the Internet. It's a powerful model. You can create games, you can ask questions, you can feed it photos, you can do deep research. You can even create videos. If you look down at the bottom, there's Canvas, which we're on now. There's deep research. There's the video. But if you want to create a more advanced video, you can go to Flow and you can add attachments if you hit the plus icon here. So here's the playable version of Tetris. And yeah, I can see it's buggy. It says GameOver and the play game button doesn't work, but you can reiterate. You can get the code. You can put it into vibe coding platform. And, of course, you can change the models here. Yeah, and that's Gemini 2.5 Pro. A really quick example. You can see all these prompts here, I've been using it a lot, and it's become my all purpose assistant. Cha GBT is not the only game in town. 9. Gemini pro more use cases pt 2: I'm going to show you a few more different applications for Gemini. So remember, this is the all purpose AI prompt AI system, so it's really limited by your creativity. So over here, I'm looking to actually create a similar beat to a mainstream artist, Sabrina Carpenter, Bedkem. So I asked, How would you describe this beat? So I got my first response, but the response wasn't what the actual AI application prompt wanted. So in this case, I want to use Google's music eft to create the prompt. So I did a second prompt. And you don't need to put the link again because it understands the context, which means it remembers you sent the link earlier. So I just said, describe the beat to a music prompter. So then it gave me exactly what I wanted the tempo, beats per minutes, time signature, genre and feel, rhythmic breakdown, instrumentation and groove, artists and era references. And so I plugged that response in. I took out I took out some references to mainstream artist names because I think that was a trigger to discontinue generating the beat. But this one worked. And here's an example. Actually, that wasn't from the exact track. I think it's this one here. Pretty good, right? So I think that's a creative way to create music with Gemini using favorite artists. Let me give you some other examples of how you can use Gemini. So you can feed it YouTube videos. And because this is a Google product, one of the key advantages Gemini has over all the other AI applications is it plays nice with Google's applications like YouTube. So summarize the key points of the video. So I got the key points of this long video. And then I did a fall prompt, write a freestyle about the mid 40 years midlife crisis. And so knows I'm in Vancouver. So it actually is smart enough to localize this freestyle using the context of that mid 40s video and I try to take it a step further, create a short highlights reel of that video. I can't do that. But what I did here I used the canvas products down here to build an application that can find the highlights of the video for any YouTube video. So this is where Gemini becomes powerful. So I can feed it a video, and it can analyze it. It can find the highlights. So, I'll just show you here. Let's just use another video. Copy the link. Go to Gemini. Okay, so it's not working. So I'm going to try fix it. This app is not working for different YouTube videos. Fix it, please. So yeah, this is all within Gemini. I just want to show you an example, so you can use Gemini for summarization, for analysis, for music, for building basic applications. And remember, Plays Nice with Google's other apps. So Google Maps, Google Music EX. Even Vo three, you can create videos from here, as well as YouTube videos. So yeah. So yeah, hopefully this gives you some ideas. You can't cover the full breadth of Gemini in one video. It's just too broad of an application, but hopefully I've demonstrated some unique features that Gemini can do over ChatBT, other AI apps as well as the Canvas, the vibe coder. Let's try this again. Okay, I think this is better. So gave a short summary. And what I can do is I got the code here, and I can actually share this. All right. So that's it. Thank you. 10. Notebook LM Pro: All right. This is Greg here, and we're going to look at Notebook M. This was one of the products of the year by Google. This product allows you to feed its sources, whether it's YouTube videos, PDFs. It could be audio files. And basically, it takes all these sources. It uses AI. And with this notebook, you can ask it questions. You can create study guides. You can share your notebook with other friends. And the killer feature, which went viral was the ability to create a podcast, but not only a podcast, a podcast that you can customize to focus on a particular topic in your notebook, and you can actually interact with it now. You can ask questions of the podcast host, and you can download this podcast if you want to listen to your particular topic while on the go or in your car. So in this lesson, we're going to cover the pro version of Notebook LM. And the major difference between the free version and the pro version is being able to not only add more sources and get analytics, but you can also use the Discover source function to search the Internet for some of the top articles and add it to your notebook. The very cool thing about notebook lm is that if you ask it an answer, it's not going to hallucinate. It's only going to use the sources that you have approved for your notebook to come up with the answers. The answers are going to be linked to a citation which linked to your source. So that in a nutshell is notebook allem. Now, I'm just going to show you a notebook that I've created for pickleball paddles, actually. You can see one that I have for VO three, but I've also created one just for pickabllPaddles. So I've asked it to find some sources. So if you click Discover, I asked it to find the top pickabll paddles of 2025, and what are some best practices for customization and maintenance? And after I submitted that, it came up with some of the best sources using Google search technology to come up with blog articles, PDFs, videos, and I think I have close up to a whole page of sources, 20 sources, in fact. And this is enough to create a study guide. So I can click study guide, so it's generating one based on the 20 sources. So these sources in some cases are handbooks, so they're really really in depth, long in depth topics on a particular topic. But Notebook LM is able to really synthesize this information and become a knowledge base. I can ask questions right here from the prompt, and you can see that it actually comes up with some prompts that I can ask it. What different paddle materials and construction elements influence paddle design, so I can click that. It's going to come up with the answer using the sources. If you've actually used Microsoft products, they're calling they're calling this technology, their AI agents. Basically, you're creating an AI specialist, and you're feeding it the sources, so it becomes really good at this one topic that you can use it to ask questions at a prompt. So it came up with the answer here. The overall design of a pickleball paddle is significantly influenced by its materials. I'm not going to read it all, but and came up with a very detailed answer with a lot of citations. And if I click one of those citations, it goes to the source guide, right? And that actually shows you the source of that reply. On the right hand side in the studio, this is where you can create your podcast. And interactive mode is where you can actually join the podcast, ask it a question. So we can refresh the chat here. Let's go over to the studio. We can load the conversation. You can create an FAQ. But the podcast is probably my favorite feature of Notebook alum. Some other things I wanted to note is you can share this out. So if you click Share in the top right, you can share this to other people with a Gmail address. You can share this publicly. And then once you share it to a certain amount of people, you can actually get analytics of your Notebook alum. So my podcast has generated. Welcome to the Deep Dive. If you've dipped even a toe into the world of pickleball and I'll just show you how I can join, it's really become a phenomenon. Oh, absolutely. It's everywhere. I'm gonna click Interactime and join the pocas. Literally. Hello. Welcome to the Deep Dive. If you've dipped even a toe. No, hey, our listener wants to join in. What's up? Hi, can you tell me what the top pickleball powder of 2025 is and the cost and why? That is a fantastic question, and we love that you're jumping right in with it. It's definitely on a lot of people's minds. So when we were talking about the top paddle for 2025, especially considering cost, it really depends on what you're looking for. Right, because top can mean different things. If you're eyeing something under $100, the warping point paddle, which B Pickle Baller actually labeled the hashtag one control paddle for 2025 is now so that's the interactive mode. So that's a really powerful feature. I can see podcasters using this tool. I can see people that want to study, so students if they want to study for an exam or certification, this is a great resource. And also, as a creator, you can basically focus on a particular topic, let's say, pickleball paddles, and you can actually create a video out of this. So one idea is you can download the audio. You can put some video footage over it to create a video, but you can also use your NOPA Calm to create prompts and scripts that you can then feed into Google's flow using VO three, to create some educational videos using AI. So I'm just scratching the surface of what's possible, but I wanted to make that link between Google's different products. So Google Gemini is really good for general purpose questions, but you can also use it to create prompts for Google's flow. But Notebook LM, you can make it very specialized and smart on a particular topic, like pickleball paddles is example. And then also use this as a script for your Google flow. So let's do an example, create a short prompt script for flow vio three on the top, pick a ball paddle and Y in 8 seconds. All right, so I got a short prompt script. I can then bring it to Google Flow. Go to my Pickleball project. I'll just change this model to Vo three so we can get some audio. This we'll use 20 credits. Okay, I'm basically just going to paste this in. So that is a full workflow for Notebook LM, how we can use it to discover sources on the Internet, how we can feed it our own sources using PDFs and YouTube videos, and then use that information to create study guides, FAQs, to ask questions using the prompt, to generate an audio podcast, how to interact with the podcasters, and also how we can use NobookM to create a script for videos and Google's flow using VO three. Discover the 60 double Black Diamond control, the ultimate pickleball paddle. It delivers unrivaled power, control, and insane spin to elevate your game. 11. Intro to Music FX: All right. Welcome to the Google AI Test Kitchen. Let's go to Music effXs DJ. So here we're going to prompt. A tempo in the style of Scott of just trying to think of a rapper. Oh, Kendrick Lamar. Turn down the Celtic folk. Add some kicks and drums. Okay, we can turn out the bass. Speed it up. I like the orchestral score. Now we can share this. So I can download it and now I have the ability to remix it. That's cool. They've changed they've changed a little bit. Now I got a wayfle I can use. Let's try and add some guitar. Less hip hop. Will you add random prompts? Less orchestral. Base seats. There we go. Speed it up. Let's just change the key. G Major. Very interesting. Add some orchestral score. I like this. So we can share Download. Okay, let's just see what the downloaded wayfle looks like. Let's see what we get. 28 seconds. Hopefully you can hear that. Let's try the other one. Not bad. Not bad. Okay, so this is the old one. You can actually use a prompt. Create a beat in the style of Scott Storch. Generate of Drake. Call me on my cell phone. And for each generation, you get two different tracks. Okay, so we get another option. You can change the track length. You can loop it. You can add more options here. Yeah, so there's Music effac DJ and music effcs. So Music effac is more you can create longer tracks. You can download them, you can loop them. You can see I've used this prompt in the style of Drake, but it's it's not very similar to Drake's song. But it still is a hip hop truck. Okay, I like this one. It's got good bass. Not bad. I like this one better. This needs to be a bit faster. So let's put it to medium tempo. There we go. I like this one. It's got a good bass. So you can see this is pretty fun, right? And you can remix. So this is called Music effets on Google Labs. 12. Music FX udpate: Alright, here's an update on the music effects. It used to be able to accept prompts like this where you create a beat similar to you would name the artist and the song. But now it'll actually prevent you from doing that. You can try one more here. Remember that time Yeah, so I got a message here. We couldn't create what you asked for. Please review our FAQ. So yeah, there's a lot of good information here about the creations, the quotas, content policies. So music AX and music FX DJ features precautions, also known as guardrails to protect artists voices and styles. So certain queries that mention specific artists or include vocals will not be generated. So that's not all encompassing. I got away with using that one prompt mentioning Drake used to call me on my cell phone, but generally, you won't be able to get away with that. But on Google Flow, you'll be able to create mainstream characters like Superman and Star Wars. So it's not super consistent right now. Can I use these outputs for of these tools for commercial purposes. Well, you have to go into the terms of service, but some of the services allow you to generate original content that Google won't claim ownership of. So yeah, have a good read over the FAQ, but I just wanted to mention an update to the music effects when you're trying to mention mainstream artists. So here's an example of how I've mentioned a mainstream artist, but they're not as popular as Drake and Michael Jackson. So the group is Cannons fire for you. And I've literally put in the prompt, create a beat in the style of Cannons fire for you. And it didn't complain. Now, the beat's pretty good, but it doesn't resemble the beat, but it's kind of in the same genre, I would say, so very loosely similar. So I can download that. That's an example. 13. How to use whisk to create and animate images: Here's a quick tour of WISC. So on the front page, you can enter the tool here. If you want to see your existing creations, you can go to the My Library on the top right. And from My library, you can actually switch between different applications from music to flow, and the images. And this is a project view. This is the Images view. And these are videos that are created with WISC. And you can see here that I've got the project that we went through. We're going to go through it, actually, how we created it. So let's go back. Now, if we enter the tool, you can see you've got your prompt here at the bottom. So this is your main prompt for entering what you want to see using natural language. We've got a sample prompt right over here. We can change our aspect ratio from landscape, square, and portrait, and we can change the model. So right now it's using Image and four, and we can adjust the seating to give more variety. On the left, this is where we can describe the subject, the scene, and the style. So, for example, create a beautiful golf beautiful female golf player, and we'll just generate that. Okay, for the scene, I'm actually gonna upload an image from a golf course that I played at. Let's just try this. And the style. Keep it realistic. All right, so we can create multiple subjects. All right, so we set up our subject scene and our style. I'm actually just going to delete this Okay. Have the woman golf player hit a driver on the fairway. And we'll just do this in a vertical style. So by default crates, two outputs. There we go. You can see that sky looks similar to the scene image that I provided. So if we click in, so you just click in the image, you get the prompt. You can see how detailed it is a wide angle eye level outdoor shot that shows a beautiful female golf player. Let's change it to a close up angle. Close, Let's regenerate this. Okay, this looks good. I like the one on the right. I'm gonna save it as a favorite. I could download this or I could animate. So I'm going to animate this, and it says we can only animate them on a landscape. So what we're going to do is readjust it This looks good. Let's animate this. Okay, so now I have to describe the animation. You can see a 422 generations, I am on the Gemini ultraplne. Have her hit the ball and have it land deep on the fairway. Okay, it is finished. You can see it's watermark with VO, most likely using the VO two model. There you go. We got a nice shot, and it's actually tracking the ball. So from here, we can flag it, we can delete it. We can download it and also put this on full screen. So we have the option for an animated gift, which is lower quality or we can download it as an MP four. So there we go. We've created an image with whisk and we've animated it. Okay, let's just go back. So we just created that video. So on the top there, we can see our videos. We can see our images. And because I gave it a heart, it's also the filter helps me find my favorite images a lot easier. You can also create a variation by specifying the style after the prompt. So in the style of Gibli we'll create an animated Japanese look with the same prompt. Just another idea to add a twist to your creativity. H 14. Create a pickleball animated video using whisk: Alright. This is Imaging, Google's cutting edge image model. And we're going to try this model out using the application called WISC, which is still experimental. So with this application, we can create images. We can create short videos that animate those images from a text prompt. So we're just starting off here with the basic image. Putting a prompt, create beautiful, diverse pickabll couples on the court, and you can see that we can adjust the settings for quality. I'm going to get an output here for two images. I'm also going to use the same prompt in Google Gemini and then compare it. Okay, back to MIS, you can see we've got two images, and we have the option to refine it. So with refine, we can actually adjust and reiterate the image. And I'm looking to create some images for a pickable dating app. So I want to adjust the images to reflect a younger audience. So I'm just going to ask to remove the ladies and add more younger people. So this is the image that Gemini created. You can see it's very different. And we didn't specify the style, but by default, you can see WISC is creating a very photo realistic type of image. So I'm just downloading the images just in case, and I'm further reiterating it. Replace the guy in the blue T shirt with an Asian Chinese guy. You have to be very specific. Don't worry about being politically correct here. People are just going to be seeing the image. They're not going to be seeing the prompt behind it if you're concerned about that. But only you have access to view the prompt, and I'll show you later on how detailed it can get after you get out of the area that you can refine it. But right now I'm just trying to refine my image, and then we'll get into more detail. So that image didn't work out. So the download button is on the top right, and you can see that we have a refine area on the top left which we'll get to. The Google Gemini image, I'm also going to download Okay, so now I'm just going to click the Edit button on the top left. You can see how detailed the prompt is. And you can also do this if you upload an existing image and you want whisk to convert that image to a detail prompt, which you can then edit the details if you want to be super specific. Let's just say you want a specific type of person in there, particular type of clothing, color clothing, particular type of hair. The more detailed you are, the more specific whisk will be. So I'm trying to get another Asian girl in there with a tennis outfit. This is pickleball. And I want it to reflect that vibe. So being very specific here, you can see I'm adjusting the wording. And I would like an Asian guy in there to reflect the potential audience, which could be in Asia, could be a North American market. Okay, I'm just going to skip ahead, and let's generate the image. It does take a little bit of time. And the plan after generating this image is to eventually animate it. And then I'll show you some other examples of how we can use whisk. Okay, so for this image, we have two Asian guys there, but we've lost the diversity. So now I got to reiterate the image again. Okay, so we're editing the image again, and we're adding additional details here. So I have a Caucasian and an African American guy behind the court. And I want an audience watching them from, like, a restaurant bar type of background at the beach. Let's see what we get. Okay, a little bit better. Not quite the beach, but we're getting closer. I do like the characters that are in the image. So now I'm just going to focus on emphasizing the beach background for the final image. I like that, so I'm going to download it from the top right. You can also share it. You can save it as a favorite if you want to recall the image in the future with Wisk. Alright, I'm going to try to animate some of the images I created. So this one here on the beach, I'm going to click Animate on the top left. Have the Asian guy in front hit a ball and people laughing. Alright, here's the finished video. As you can see, it looks very realistic, but the animation isn't that believable, but that's just the category for pickleball. Here are some other examples of how you can create images and animate them. So here I'm refining an image of an attractive Asian golfer with a pickleball cord on a power cart. And from here, I can download the image on the bottom right. I can also animate it. And hopefully you enjoyed this lesson, and it gives you some inspiration to create your first image using whisk. 15. Nano banana AI Image Editing tutorial: Okay, we're going to be giving you a tutorial of nano Banana. This is Google Gemini's latest image editing AI powered program, and it's September 2025, so things will change depending on when you see this. And you can see here I'm at gemini doggle.com. Now you see there at the top, they've got the tr image editing with their best image model, nano Banana. You can see here with this icon. So you can either prompt an image. Create a let's just see a future istic Kuala Lumpur, Malaysia, 3,000. So let's just try this. This is a prompt. And then what I'm going to do is use existing images and show you some ideas that showcase the power of this image model. As you can see here, generated this image. It's got some spaceships. And we can download the full size of this image. So it's got a watermark at the bottom right. So people clearly know that this is generated by Gemini. Okay, so let's just say we want to use an existing image. I'll click the plus and actually, let's just go back. Okay, this is more of an image editing now. So we're going to we're going to upload an image. So we're going to click Plus, Upload Files. Now I've got a nano Banana area here. So this is a blurry image screenshot from Superman. Can you sharpen this image up drastically. And use natural colors. So this is just a basic sample of you can do for any image. There we go. So here's the first one. And there's the EI edited image, and it did exactly what I wanted. So the cool thing about this, once you have the image, you can actually create a video from. I'll just show you. You can click Plus or click Tools, create videos with VO. And then what you could do because you need to start a new chat. Add that photo, and then you can actually animate it. But I've used up all my videos for today, so that's out of scope. But I just want to point it out to you, once you've created this image, using nano Banana, you can actually create a video from it. Okay, let's use another example. Okay. So we've got this car here. Okay. So add a pretty long gal driving car in the car. Change the angle. So it's facing the car, but you can still see those buildings. Okay. So I'm going to add a girl, and we're going to change the angle. There we go. Now, she's driving. Actually, it didn't help me. It made me less handsome. Make the guy a lot more handsome. Just giving you ideas, giving you ideas. See what we got. Alright. So let's look at the original. I think that actually did a really good job because this building over here, you can't really see it, but it manages to just make up the difference. Okay, let's do one more example here. I want to illustrate the character consistency. So let's put these two people on a pickle bowl bowl court in the appropriate clothing near the beach. So this is what we call character consistency. So it's taking the same two people but put them in a different scene. So yeah, more or less. Zoom in a bit. Because it's a bit of a wide angle shot. Okay. And then we could, um, Zoom close to the lady. So you can you're basically prompt editing here using prompts to edit the image. There we go. Good response. Okay. So, yeah, that's an example of nano banana using Google Gemini image editing. 16. Nano banana pro & flow: Alright, everyone, we're at the end of 2025. We now have Nano Banana Pro and Google Veo 3.1, which we can use together to create some interesting image to video possibilities. So I'll just show you an example of a more or less finished video that we've started from an image. So that was going from an indoor court to an outdoor court. Here's another example. So I think it looks pretty seamless. So I'm going to show you how to do this today, and I have the pro plan of Google Flow. So we're going to start off with a new project. And we're going to start off in the images, and this is going to use the Nano Banana. So on the drop down here, we want to go to create image. So this is going to use the Nano Banana Pro. So what I'm going to do is I'm going to upload a I'm going to upload a new image. What I'm going to do is I'm going to ask Nano Banana to change the pickleball setting to at the beach and wearing different outfits to match the weather Play. All right. So first, we're going to use our source image, and we're going to create another scene using that image. So another image. And then what we're going to do is use Google Flow, then create a before and after or a transition scene. Okay, so this one This one is good. Yeah, I like that one. We could also use a real pickleball court at the beach change to swimwear. Okay, so you can actually make additional changes or alterations. But if the alteration isn't what we like, we'll just go with this image. It's not too bad, actually. Let's go to Let's exit. We have our before and after images, and now we're going to click the video section on the top left, and we're going to switch to frames video. And I'm going to leave it on Veo 3.1 FAST. We're going to do four outputs using Vertical, and now I'm going to use my image before. And then I'm going to add the new image. Now I'm going to work on my video prompts. Okay. To finish their pickleball game and take a short flight to Phuket, they end up playing pickleball at the beach. On the Bangkok rooftop. Okay, so let's see how this goes. So we're trying to create one scene to another. And we're using, again, we're using frames of video, Veo 3.1 FAS and four vertical outputs, and we're using two images as for the frames, the video. So before and after. And this all started with one real image. But you can actually if you want to be creative, you can actually use AI generated images if that works out better for you. Alright, we have finish the videos. Let's check them out. Good game. Cheers. Next stop the beach courts. Good game. Cheers. Next stop the beach courts. That's pretty good. I like that. So I'm going to heart that and we can further edit it if we want. We can download this. And once you do download the video, you can upscale it. Let's just play this one here. Okay. Okay, interesting. Double fister drinks and showed some walking and voice. Okay, that was a weird hello. Okay, and that actually had an airplane seen, so I like that. So yeah, you can download this. You can edit it. You can change the camera position. And then regenerate. But there's the idea before we get a little carried away. So just to summarize, we've created a before and after video scene, starting with an original image, and we use the powerful Nano Banana Pro. To create the end image. We then use those two images using the video section of Google Flow, using the frames to video option to create a before and after scene. And yeah, there's a lot of possibilities here, but hopefully you can use this technique to create something really cool. 17. Google ai summary summary: Congratulations on finishing the course. You've learned some cutting edge skills using Google's various tools from Google Gemini for all purpose creation. You've used Google flows with the cutting edge Vo three model to do text to video, image stills to video. And this is only going to continue to improve, whether it's longer video clips or better dialogue or even higher resolution videos. We've also covered different types of media forms like creating images with whisk, animating those images. We've also covered how to create beats with music effects, and to create your music from scratch with various instruments or from inspiration with creative prompts. It's up to you how you want to put this all together. One way I can think of applying all this is towards YouTube videos or short form media for Tik Tok or for Instagram, to find music beats or even just creating your own short videos on your Instagram reels. I've already started experimenting with this on TikTok and have had some small wins that I'll share with you. It's really up to you how you want to use these cutting edge skills with these cutting edge tools. And the technology is out there, so I can only see other players coming into the space and more improvements. I'm most excited about Google Flow because I think video has the most power to catch attention from audiences. And this is a big disruptor for other industries like film, stock footage, digital animation, and a whole lot more. Thanks for taking the course, and I hope you will create your own project that you'll share with the class.