AI Videos: Create recordings with artificial intelligence

1. Introduction and what you will learn: Ai videos are making bigger waves than ever. For example, Opmei announces Sura leaving entire industries stunted. According to Opmayi, Sa is not just great for videos, but can also understand physics. In the future, Sa is expected to serve as a world simulator for robotics. And even the gaming industry is moving in this direction. Whole sectors will be transformed by such technologies in the future, be it the front end and shape the future. Do you want to understand which AI tools are right now a viable, how you can use them and how to create such videos? Then this course is for you. Imagine how it feels like to generate AI videos that have the potential to reach a cinema screen or simply generate some attention on social media. I'm sure your friends would be envious. In this course, we will look at everything, the basic invo on AI videos with the technology behind them. Creating videos with Moon Valley, stable diffusion, runway, hyper, and many more. Everything about I avatars and their applications in marketing, sales, customer acquisition, and for explainer videos or online content using tools like Cubicle and technologies that can even turn you into Spider Man. Many case studies and of course, practical applications with video editing. We will even look into the future of I, data protection and ethical concerns of the technologies. And by the way, if you ask yourself who I am, my name is Arnie and I teach I classes before GBT even was a thing. I am relatively long in the game. I also have a small German Youtube channel, and that's what I do. The sooner you sign up, the greater the chance to be ahead of competition. And of course, I can answer questions quicker. So don't test to date. Click the button now, because your future starts today and not tomorrow. 2. Silly Question: What Is Actually a Video: In this video, I want to talk about how all of this works. A lot of people already know that AI can make pictures. We have a lot of different tools out there. We have mid journey. We have stable diffusion. We have Ali, we have Adobe Firefly. And I am relatively sure a lot of other tools will come around the corner. All of these tools, yes, they can make pictures. We make these pictures with a technique that's called diffusion. We use a diffusion model to generate pictures. And that's the one part, because if we can make pictures, we can also make videos. In order to get that, we need to understand what the video is. Yes, that's a stupid question. A video is simply a lot of pictures in a row. One picture and another picture and another picture. Maybe you remember, like from the old, old films. If you take, for example, a paper, okay, I don't have a paper right now, but you will get it. Sometimes you have some paper and you do it like this. If you see all these papers one after another, the picture starts to move. So maybe you saw something like this. Maybe you see 30 FPS or 60 FPS, or maybe also 25 FPS. What all of this means, this simply means frames per second. If you see, for example, a video with 30 FPS, it simply means that there is a sequence where you see 30 different pictures in one single second. That's a nice little video. The video that you see right now, this is in 25 frames per second. That stuff that you already see right now in this video. These are simply pictures. Yes, you see simply 25 pictures per second and that's why I am moving. These are just pictures that are moving. So if we understand that AI can make picture out of the noise with a diffusion model, don't worry, we take a look at the diffusion model. Later in the next lecture, we will take a deep dive. We just need to understand that the AI can make pictures. And if the AI can make picture, the AI can also make videos. And yes, we have a lot of different things with we can make pictures and also videos. And you need to understand that the hard part is of course to make a video that is consistent. Maybe you have tried it a few times. Maybe you tried to make a picture and then you try to make a picture again, but a little bit different. It's tricky sometimes, it takes you a lot of effort to make a picture that is somewhat similar. Now think about it. How hard is it to make 25 pictures that are nearly completely the same so that we can make all of this frame one after another that this video moves. Yes, this is hard. But we have tools. And these tools, they work better and better over time. I also have to admit, we don't only have diffusion that works here, we have also things that are called guns. These guns work relatively similar than diffusion, like I said. In the next video, we will talk about the applications in more detail. But just remember we can make pictures out of the noise and we can also use guns. And these things can do a lot of really cool stuff. We can really dialed down videos if we know what we do. We need to understand what Lauras are, what sets are, how all of this works, and then we can make really standing videos. I have to tell you in some tools, we don't even need to know the seat and we don't need Alora. It depends a little bit on the tool that we are using. If you ask yourself what we can make with all of these tools, that possibilities are completely endless. We can make AI avatars. We can make stuff that is called lip sync. So we can animate our lips, for example, to a different language. We can animate pictures to different voices. We can take other videos, and we can animate all over them with lip sync. We can create videos out of thin air. So we can make videos out of text, but we can also make videos out of videos. And we can make videos out of pictures. It's called text to video, Image to video, and video to video. You will learn how all of this works. You just need to understand that a video is simply a lot of different pictures, one after another. And that's why all of this works. If we use, for example, video to video, we use all the frames from our video, all the different pictures from our video. We simply throw a prompt on these frames and we can make them like a little bit nicer. If we go with text to video. We make these pictures one after another out of thin air and then we have our video. If we start with a picture, we use this picture as a reference picture and we make similar pictures. That's the concept that what you need to understand in this video, you have simply learned that the video is just a lot of pictures. This is important to understand because we know that AI can make pictures. And if I can make pictures, it can also make videos. Yes, sometimes it can be a little bit tricky, but I promise you this will also get easier over time. In the next video, we will take a closer look behind the doors of difusion. Laura's checkpoints, seats, and much more, because we will need it over the course. 3. Diffusion Models, Loras, Seeds, and Checkpoints: You have learned what a video is. Yes question. In this video, we will dive a little bit deeper. Because we will learn what the diffusion model is, what sets are and why we need the sets, what Lauras and Checkpoints are, and how we can make consistent videos. All of this will make sense after this video. Maybe we can also talk a little bit about gas. We start, of course, with the diffusion model. In order to explain the diffusion model, I like this article. We make it really damn simple because we go onto this picture and we just use this single picture to explain it. Yes, we have also the possibility to dive into code and do whatsoever, but I like to make it simple. What's a diffusion model? A diffusion model is simply a computer. This computer is trained on pictures and text. You can see it. Here you have a picture and you have text. In this text stands exactly what you can see in this picture. For example, you see a beach with sand, blue ocean. There's a mountain. There's some green on the mountain. Maybe the sky is blue. This is how I would describe this picture with my bad English. After that comes the magic, because we add some noise. We add some noise, but we still tell the computer what's in this picture. A beach with sand, blue ocean, and so on. And we do it again. A beach with sand, blue ocean, and so on. And we do it again and again and again until we have just noise. We simply start with a picture and a description. We give the description and the picture over and over and over again, and we add noise over this picture. We do this until we only have noise and the description. In this process, this computer learned how this picture look like. Then you can simply take this prompt, take this text and you tell the computer, like a beach with sand, blue ocean, a mountain with some green, and the sky is blue. And you can simply throw this into the computer, into the noise, and the computer will give you pictures back because the computer has learned how this picture has to look like. And now comes the fun part, because we don't do this with one picture, but we do this with like a gazillion pictures. We do this over and over and over again with like nearly every single picture that we can find on the web. In this process, the diffusion model learns how these pictures look like. I will show you one quick example in mid journey. This right here is the example of a diffusion model. And yes, I have made a lot of pictures with these diffusion models. We have here like cat woman and some older pictures. So we can make the pictures however we want. You also see that these pictures are really consistent because of the tricks that you will learn in this video. But we start with something like this. A small white dog sits on a wood floor. If we simply tell this the computer, the computer will start to calculate how this needs to look like. And he will calculate this out of the noise in no time. You will see how this will look. You can see it right here. We start with noise and the computer starts to denoise the picture. We get our pixels where they should be. So you can see we have a small white dog that sits on a wood floor. We have every single thing that we told the computer into this picture. This only works because the computer is trained on such a large amount of pictures. And he has also the descriptions of these pictures and he understands how these things need to look. I always like to tell all the people something like this. Just imagine you look up into the sky. The guy next to you simply tells you, can you see the apple in the sky? But first you can see the apple. And you say, no, I don't see an apple in the sky, Dude, like what's wrong with you? Then the sky shows you see this cloud, cloud, this looks like an apple. And then you start to see the apple because your brain is also like trained on apples. Your brain knows how apples need to look like. Now you start to see the apple. Maybe you see a red apple because our brain is trained on red apples. Apples are normally red, but there are also yellow apples. If the guy next to you doesn't tell you that you need to see a yellow apple, you will most likely see a red apple. You need to understand this because you also need to be specific with your prompts. If you don't tell the diffusion model, you want to have a yellow apple. The apple will be most likely red. But don't worry. We will talk about prompt engineering for AI videos, of course, in the next video, because you understand what a diffusion model is. So that's checked. Now you need to understand what a set is. I can explain the set, really simple. The set is simply the first starting noise of the diffusion model. The set is always a number. It can be a number between one and like a real big number, something like this. The first starting point of this diffusion model is always the set If we tell, for example, our diffusion model like this white dog and so on. And then we add a seat, maybe the seat 55. The difusion model will start with a specific focus point of this picture generation. You know what happens if we do that, If we add a seat over and over and over again and simply play a little bit with the text that we give the diffusion model, right? We get more character consistency. That's how easy all of this works because we have always the same starting point. You see, we start with a picture that is already a little bit the noit. That's our first starting point. Because we do that, our characters get more consistent. Don't worry, you will see all of this, of course, and action over the course. The seat is simply the first starting point of the picture and you get more character consistency if you use a seat. And then we should go to look what Lauras and checkpoints are. In order to do that, we go into paint. Because I like to paint like a real pro. I need to tell you this is like not 100% right how I show it. But we want to make this simple. We have a lot of different difusion models. One diffusion model is stable difusion. Stable diffusion is gigantic. And we will use stable diffusion to generate pictures over this course. Not only pictures but also videos. Of course, stable diffusion has different models. We have stable difusion 1.5 for example, but we have also stable diviusion three, this is the newest version. We have also stable difusion X L. You see we have a lot of different stable difusion models. If we create a picture, for example, stable difusion three, the pictures will get a little bit random, just like if we don't use a seat. But we can add a seat and get more character consistency. But we have also more options we can use like a small piece of stable difusion three. This right here, I, I want to call it checkpoint. Not only I like to call this checkpoint because these are checkpoints. Now what is a checkpoint? A checkpoint is simply a smaller part of stable difusion three, for example. This checkpoint is specifically trained, for example, on people, on men, on women, on cars, on what you ever can think of. So this checkpoint, this is a little bit fine tuned. We will also, by default, make more consistent characters if you use a checkpoint in stable difusion, the gigantic stable difusion three model, this is trained on like a lot of different pictures. If you use a checkpoint, the checkpoints are trained on more specific things. There is a checkpoint that is called Jagger out. You don't need to know this, I'd just like to throw it in here. This chagger out checkpoint can make really, really realistic pictures. And the coolest thing is that we can even go down smaller. This right here is called Laura and disluras. You can also train them. You can train luras. Really specific. One last time. Stable difusion three is the gigantic program with stable diffusion. You can make pictures, of course, on stable diffusion you can use a checkpoint. And one checkpoint is, for example, this chugga out checkpoint. And a chugga out checkpoint can make people really, really good. And then we have also Laura's. Laura's. You can really dial down Laura's in detail. You can train Allora for example, on one single person, you can train Allura on my face. Allura that is trained on my face will make pictures that look something like this. You can see it's really easy to create character consistency if we really dial down our things. If we use Allura, we will make character consistency really easy because it's trained on a lot of pictures that are nailed down to the character that we want to create. So what is a diffusion model? Check, What are seats checked? Lauras and checkpoints. Checkpoints, of course, are also checked. How can we make consistent videos? Of course you understand it. Now we need to use a diffusion model in order to be more consistent. We can't take a seat because it's the first starting point out of the noise If we want to nail this down even more. And if we work with stable diffusion, we can use checkpoints and in optimal cases As, and we can make consistent characters with all of these. Before I talk about gains, I will also talk about consistent characters. If we have other options, of course, we can use reference pictures, and stable diffusion is perfect for this. We can use a reference picture and we can forget all of these. Yeah, maybe the diffusion model not, but all of the other stuff we can forget if we simply use a reference picture. It's called image to image in stable diffusion and you will learn about it. We have also some tools that do all of this automatical. That's also something that I want to show you in this course because we start with the easy stuff. And the easy stuff does all of this automatically. Before this lecture is over, I want to talk just a brief moment also about Gans. Because Gans, Gans are simply like, similar to diffusion models. The diffusion model creates the picture out of the noise. You have already learned that. And again, does it a little bit different with again you also vet computer pictures, but you tap always to pictures. The computer tells you, hey, this picture is not similar than this picture. And then you feed other and other and other pictures until you have two pictures that match. As soon as the two pictures match, the computer understands how it works and it can create similar stuff with these guns. We can make really, really enormous good deep fakes for Lip. Think this is perfect. Something from Alibaba that uses guns look like this. When I was a kid, I feel like you heard the thing. You heard the term don't cry. You don't need to cry. Crying is the most beautiful thing you can do. I encourage people to cry. I cry all the time, and I think it's the most healthy expression of how you're feeling. And I sometimes wish I just could have been told you can cry. There's no shame in that. There's no in how you're feeling. And also you don't need to always be justifying it. Because I think I was constantly trying to come up with reasons why rather than just being accepted for what it was. You see the works really, really amazing. One last time. You have learned in this video what a diffusion model is. It creates pictures out of the noise. You can use sets because the seat is the first starting point of the difusion model and you will get better character consistency if you use a fixed set, remember always use the same set. You can also use checkpoints and lower as these are fine tuned models of stable refusion. You need this to make consistent videos, but you have also the possibility to use stuff like image to image or if you have the option to use automatical solutions this is perfect. Then you have also Gans and gangs are like similar to diffusion models but a little bit different. You have learned a lot in this video. In the next video, we will talk about the prompt engineering for AI videos because this is really important. And then we have all the theory done and we start to create our own videos. Come on, this was the hardest part. It only gets better from here on. 4. Prompt Engineering for AI Videos: In this video, we talk about prompt engineering for AI videos. And of course, you already know it. If you want to make good videos, you need to have a good prompt, good input, that you will get a good output. You also understand that pictures and video, they are nearly the same. We need to use the same prompts for pictures and for videos, but you have to be specific. Let's just take a look at these. You need to consider a few things. You need to consider the subject, the medium, the setting, the lightning, the color, the mood, and the compositions. For example, the subject could be a person, an animal, a character, a place, a object, or something else. The medium could be a photo, a painting, an illustration, a sculpture, a doodle, or whatever you want. The setting could be indoors, outdoors, on the moon, underwater, or wherever you want to see the stuff happening. The lightning could be soft, ambient, cloudy, neon studio lite or something else. The colors could be vibrant, bright, colorful, black and white, pastel, and so on. The mood could be also, of course, calm, merry, energetic, also angry if you want the compositions. It could be a portray a close up, a bird's eye view and much, much more. Of course, also a point of view. And then you can also include a movement, for example, a person is walking and so on. An example would look something like this. An illustration of a dog on the moon. Neon lights colorful. The dog is energetic, full body view. Then you can also include a movement. What does this dog do? Illustration of a dog, let's just say walking on the moon. I hope you get it. Of course. We can also make pictures out of this prompt just to show it. I copy this prompt. And now I will go into Microsoft Copilot. And you can do that too, because in Copilot we can make pictures completely for free. First, we start with a picture of later, we will also make videos. This is just that I can show you this in more detail. This right here is Microsoft Co Pilot. Right now I am in Co Pilot Pro. Yes, this is German, but don't worry. Like in English it looks completely the same. Even if you don't have the pro, this will work. You can ask something like this. Make a picture of these. Then we include our prompt. You will see that we will get something that is really specific. Now, while this thing is creating, I want to go back on these. You don't necessarily have to include all of these. Our prompt could also be this, dog. Yes, dog. This is a prompt. And you can totally make pictures and videos out of this prompt. But do you remember what the diffusion model is and what it does? If you are not specific with your prompts, your output will also not be specific. The diffusion model is trained on a lot of dogs. If you don't tell the diffusion model, you want to have an illustration or the dog should be on the moon or wherever you want to see this, then you will not be specific. If we simply copy this word like I could also retype it, but I want to copy this because I'm lazy. Then we go back. Now you see the first thing. Of course, this is really, really specific, but if I only type in dog, I have absolutely no clue what I get. Okay? I have to admit it, I have to tell, at least make pick of this. Then I include the dog. Let's just go back on these. You see how specific this is. This is, without a doubt, an illustration of a dog. Of course, the dog is walking on the moon. You can see this. We have our neon lights. It is colorful. The dog is also energetic, like you can see how much fun this dog has and it is a full body view of this. And if we press on these, you can also see it a little bit bigger. All of these pictures, they are really specific because our prompt was specific. Don't worry, yes, these are only pictures, but I want to show you these with pictures because it is easier to show you. If you are specific, you get specific output. If you are not specific, you will not get specific output. Here you can see that, for example, make a picture of a dog. And of course, we have a picture of a dog, but it could be a photography. It could be colorful, it could be whatever. Now we have like a little dog, we have some butterflies around the dog. It is colorful. It looks something like a sticker with white background. We are not specific and if we let this prompt re run again, we will get also other pictures that are not specific. Maybe we get a photo, maybe we get something else. We simply don't know is this a problem? Yes and no. If you don't know what you want to make, you can totally run with these. Just tell the diffusion model. Make a dog and you will get a dog. But if you have something in mind, you need to take this into consideration. Most of the time, you want to create something that is specific. You have something in your head. If you have something in your head, you need to think about the object, the medium, the setting, the lightning, the color, the mood, the composition, and of course, also the movement. And I have to tell you, yes, you can make this even bigger, especially the part of the compositions. You can also include stuff like shot from slightly above. You will get a shot that comes from slightly above. You can also include the warm eye view. The warms eye view is something like this. When a warm looks up to you, you can simply play a little bit with all of this, but you need to be specific if you want to have specific output and if you don't want to have specific output, just go with a word. You will get a dog. In the next video, we will take a look at Zora because Sa is the future in video generations. And then we will try to create our own videos. But like I said, first I want to show you Sa, because Sa is enormous. 5. SORA by OpenAI: Overview The Future of AI Videos: This video, I want to introduce you to Sa, because Sa is completely awesome. Sa is the tool from Opmei, the tool that can create videos out of text. The fun part is that Sa is also a word simulator, or at least it should be a word simulator in the future. More on that, of course, later in the course. And I will also show you some concept of this research paper in this video. But first, we take a look at the videos from Sa. This right here comes directly from Sa. This is the introduction video. They simply want to show us what Sura can do without any modifications. Here you see the prompt, A cartoon kangaroo, basically nadisco. You see this thing is dancing and this looks really good and enormous coherent. Then the next prompt, the golden retriever puppies in the snow. You also the playing of this puppies. This looks enormous good. Also, stuff from the nature works of course really well. You see the camera zooms in and you can imagine how all of this will go into future because you can create stuff out of thin air. You can make stuff that is not realistic, but you can also make stuff that is realistic. You see longer prompts with a lot of descriptions. You can also see that I think this model understands physics. Yes, you can see all of the generations. They seem that the model understand physics at least a bit because all the people, they move like they should move every time you see something in nature, all of these moves really coherent. There are basically ships in black coffee and they are floating around like this model understands physics. Also the people, they are looking really nice. This thing looks also really good and also the cat is nearly as it was real. You may be also thinking that Mahmud are back in civilization because all of this looks enormous. Now enough with the videos, let's just take a look at the research paper. Because Sra has the potential to be a world simulator. They tell us right here, we explore large scale training of generative models on video data. Specifically, we train text conditional difusion models jointly on videos and images of variable durations, resolutions, and aspect ratio. We leverage a transformer architecture that operates on space time batches of video and image latent codes. Our largest model, Sa, like the stuff that you already saw, is capable of generating a minute of high fidelity videos. Our results suggest that scaling video generation models is a promising path towards building general purpose simulation of the physical world. This sounds awesome because Op Mai tells us they will scale these models and these models understand physics. This is really promising. Maybe you think why these models should understand physics or what should we do after these models can understand physics? We can do all of it because we can basically just like simulate other worlds. Maybe you saw the dog from Chim Fun. Chim Fan is a guy from Nvidia and he talks about Isaac Chim. I just want to explain it real quick. Here at Invidia, they made a model that can simulate words. In the words, they train robots to do stuff. They can train these robots really, really fast as soon as the robot can do stuff in the simulation, it can also do stuff like in the real world, you should probably see the Stat dog for yourself. I just did this really, really quick here. But they try to simulate the physical world to train robots. If Sra can do this on a really high level, we can do all of it. We can train robots in the simulations. Now, back into this paper here you have also a lot of examples. And here is how this works. Basically, yes, this is just a diffusion model. You make pictures, you make a lot of different pictures, but they also work similar than a large language model. Here you can see it. We take inspiration from LLM models with acquired generalized capabilities and training on the Internet scale data. He simply tries to tell us that these models, they work similar to a large language model. An LLM is for example ChetchPD. Chechpd makes tokens out of text, basically makes these pictures smaller, it chunks it down into patches. This is basically the same stuff as tokens. And then these models can be more efficient. At a high level, we turn videos into patches by first compressing video into lower dimensional latent space and then basically decompressing all of it into patches. Here, you can dive deeper if you want, but I don't think we need to do this. You can also see that they use this diffusion model. They start out of the noise and they start with the denoising process. Here you can also see the difference between that computing power, the first video here, this is base computing power, excuse me, for my slow internet. But I think you can see it the first video, this is the slow computing power. Then the second video here you can see four, the computing power, and the video starts to work like a lot better, a lot more coherent. Then here, the last video, this is 32, the computing power you see, they need really a lot of computing power to scale these models. Sam Altman from Pm, I even tried to raise $7,000,000,000,000 for computing power. He wants to make chips. If you see this that you need like 32 x the computing power, you also understand that this thing needs a lot of compute. You can also make different resolutions. You can make stuff for Instagram, for Facebook, for Youtube, for whatever you want. Also, the coherence under water is really good. You can also start with pictures. You can start with a picture like the Sheba, and you can animate it. This looks really good and really nice. The second animation is also good, and the cloud that tells us sa is really amazing. Here is one of my favorites. This is enormous. We start first, we start with the small little wave and then we go into this big wave. This wave is really cool, at least in my mind, extending of generated videos. Of course also possible, we can make endless loops. You can see this for yourself on the web page. We can start with input videos and we can lay prompts over it. Just imagine what these things can do. You can create whole films with these. This is completely mind boggling. You can also see you can start with a video and then make other stuff into the video. Here you see it gets into something really funny. You can even play Minecraft if you want, because all of this is generated out of text. Just look at it. This is really coherent, this is how the Game Minecraft looks like. I think this is completely mind boggling. This is Sra, and of course I know right now we can not use it. But I am relatively confident that we can use this relatively quick. And as soon as we can use this, I will make updates. Sra is more than just videos. Just think about the applications. If you want to create a whole film, you can do it with Sra, just like one or two people. If you have a narrative, you can create whole films with these. It was never possible until now. If you want to make a film without a I, you have to hire a gazillion people and you need to spend millions and millions of dollars. Right now, maybe you can do it with a strong ChPU and Sa or at least as soon as we have access. But don't worry, we will take a closer look at the tools that we can use right now in the next videos because we can really make a lot of stuff even without Sra. But I think S is the next level or is the next cool thing. Because it can also understand physics. It has the possibility to simulate whole world, at least if we believe what Omai tells us. 6. Section 2: Easy Tools for Video Generation: In this section, I want to show you the easiest tools that you can use to generate your videos. We will use stable video diffusion. Run ML B collapse, Moon Valley, AI, hyper big Sors. And you will also get a nice overview of the coolest tools, for example, Kyber and much, much more. Yes, there are a lot of different tools out there. No, you don't need to use every single tool under the sun. I just show you the best in the best tools, we will go a little bit more in detail and of course, I will update this section over time because I assume that we will get new tools that are better and better over the next days, weeks, and months and the next video. This is a really nice little video because you can nearly forget every single thing about sets, about prompts and so on. Because in the next video, I will show you stable video diffusion. And stable video diffusion is the easiest tools from all, it's open source, it runs on stable, stable diffusion is from stability AI. And you can use this completely for free. Every single thing that you have to do is to upload a picture and you get the video completely automatically. Yes, you don't have control, but you can use it and you should use it. Like I said, this is the easiest tool and then we dive into the other stuff, have fun over the section and try to make all of this work for yourself. 7. Stable Video Diffusion: Free Image to Video Tool!: In this video, I want to talk about stable video diffusion. Stable video diffusion is a project out of stability, AI. Yes, all of this is from stable diffusion. And you already know the code is open source, so we can use this totally for free. Now, you can install automatic 11, you can install comfy. Yes, all of this is relatively complex and you need a lot of time and you need a strong CPU. But in this video, I will show you how we can use this right now, completely for free, without installing anything. And we don't need any GPU power on our PC whatsoever. But sometimes we have to wait a little bit. This is the only downside. We can simply use a free hugging phase space. Yes, hugging phase gives us access to stable video diffusion. We can try this out and we can generate some videos. Now, the videos are not completely out of this world, at least not right now. We don't have a lot of control for all of these. We can simply upload a picture and stable video diffusion will create up until 25 frames, I think. But the frames will get animated. And I am 100% sure that all of this will get much, much better. Remember, this is the worst version that you will ever get access to. Just go on hugging face and play a little bit with these. Now I will show you how this right here is the hugging face space. You simply go over on hugging face and type in stable video diffusion. You can also just Google It or use the stuff from my description. Everything you have to do then is of course, just upload an image and then press Generate. I already did that. And you see right now we have 200 seconds of generation behind us. And they expect that we need to wait 264 seconds. And sometimes this even takes longer. As soon as I started to upload this image, I immediately get a request. Hey, a lot of people are using this stuff right now and you don't even have to upload your own pictures. You can also use pictures from down here. You can simply search for pictures that you like, You simply press on them. And then you can create your small little videos outside of stable video diffusion with one simple click. Like I said, you don't have a lot of control, at least not right now. But I am sure this will come and I am sure that we will be on top as soon as we get control. Just try this a little bit out until we get control, and I hope that we get access in like 5321. Let's just play this and see what we get out of these. You see also this rocket launch is awesome. We can zoom in into this video. The rocket launches completely, so the rocket goes outside of the picture. I think this looks completely awesome. Stable video diffusion is one of the coolest image to video tools out there is completely for free. The code is open source and you should totally try this out. Yes, you don't have a lot of control out of these, but the image is unbelievably coherent and you can make enormous good videos outside of stable video diffusion. Now you have even the possibility to make longer videos. First, I would recommend to download this video, and then you can simply slow it down a little bit. You go into your favorite video editor and you slow it down. Then you have an eight second video. Then you can use the last frame, re upload it right here and generate another video. You can do this over and over and over again, and you are able to generate really nice and coherent videos completely for free. Of course, you can also make videos outside of these. I think I need to try a video also of these because like this is a nice little video. Just let's see what we get out of these. You see? Yes, it's somehow coherent and the thing right here is burning. The building is burning behind. This gets a little bit deformed. Let's just take a look off this generations because I think I am also somewhat cautious to see how they look that somehow. Okay, let's just see what we get out of this video. This is a complete mess. I think, I think the rocket launch was some of the best things out of these. Not bad. Now let's just see what we get out of the. Yeah, it gets deformed and she looks like really, really bad. I think this is also a cool video. In this video, you saw stable video diffusion. This is based on stable difusion. Who would have thought it's completely for free to use because the code is open source? The company behind stable difusion is called Stability. They give us every single thing for free. Yes, you can install everything locally, but you don't need to. We don't need to use our own GPU. Just go on hugging face and try this stuff out totally for free without installing anything. I am sure we will get a lot of control outside of these tools in the future. Right now it's for playtime at least that's how I see it. We can play a little bit, we can explore a little bit, but I think this is really are nice and possible future for the whole video generation. I think we will be able to create small videos maybe in the future. Sometimes in the future you are able to look entire films that are created just from pictures and text. I am really, really cautious to see how the future unfolds before our eyes. 8. Moonvalley AI: Free Text to Video Tool on Discord: In this video, I want to talk about Moon Valley AI. Moon Valley AI is a tool and we can make video out of text. We can really start with just text and we get videos. And these videos are relatively cool. They are somehow coherent. They are two to 4 seconds long. And the coolest part is this is completely for free. At least right now, we can make videos however we like. We can make really hundreds of videos completely for free. But of course, we have also one or two downsides. First, the generations are relatively slow, and second, the interface is in this court. For everybody that doesn't know this Court, this Court is simply a jet application similar to Whatcepp Moon. Al AI is a viable in this court, but it's for free. Let's just take a look at the web page. This right here is the web page. And you see also here they are hiring. If you want to work in the sector, maybe just talk with Moon Valley. Ai. Moonvalyai is a groundbreaking new text to video generative AI model. You can see two videos of them right here. I think the tiger looks somehow good and also this fish is okay. Just remember this comes only from text. I think this is good. Of course, they have a lot of examples. You can see them for yourself. But what we will do right now is press this button. Try the beta. This pop up will appear. Yes, you need to give confirmation that you want to add Moon Valley AI into your discord. You simply press on this button. If you don't have a Discord account, yes, you need to create one. If you need help to create your Discord account, just hit me up. But I don't want to bore anybody in this video. We skip these. We simply press here that we want to accept here, everybody on this court. And then we go onto this Court into web application and we are in the right server. So here we are in Moonwalytai. Here you can also see the stuff directly from Moonwaly. Yes, it's free, even Moonwaly tells us free to use. Make your first video on MoonwalI. Navigate to as such a room, you can always press on this and you are in a room and you can make similar videos than the guys that made the videos right here. You can see a lot of people are already creating videos here. And you can also see that most of these videos, they look relatively good. First, you can always go back on the left side here. Here you can see this guide. So you can go back and how you can create video, use the create command and provide the prompt you like to utilize. And then some example prompts. So you can always see for yourself how all of this works here. You have a lot of epic use and so on, but I want to make this fast. We simply press on this new room, and then we see for ourselves what other people make and what we can make. Here you always see a lot of different prompts and a lot of different ideas. This right here is the set. You see a lot of different creations. And you can always see what these people type in to get the creations. If you like, you can always press play. And see for yourself what all of this means and what you get right here. Here, you also see this right here, negative prompt. Now you know what a prompt is and you know what you need to describe in a prompt. But I have to tell you this right here works with stable difusion in the background. Stable difusion is somehow special because in stable diffusion, you can also type in your negative prompt. In your negative prompt, you can simply type in the stuff that you don't like to see. For example, text, watermark letters, two hands, two faces, three hands, many fingers, crooked fingers, crooked feet, and so on. You also see some brackets right here. If you put stuff in brackets, this only means that the weight of these words is a little bit higher. I have to tell you, you don't need to be that specific in the negative prompt. You can also just go and copy all of this negative prompt and throw it into your generation, even if you want to make animals. Because stuff like crook tens and many fingers and crooked feet and so on. All of this works completely perfect In this models, you just need to type in a few words that you don't want to see in your negative prompt. You don't have to be that specific in the negative prompt. This is just a nice little support for the diffusion model. Now we start to create ourselves. We type in Create. Here, you can always see prompt style duration. If we press on it, we first start with the prompt. Of course, let's just make a simple prompt, but we take this into consideration. Let's just say we want to make an illustration of a ghost on Mars. Neon, like, vibrant color of let's just say calm ghost and it should be a full body view. Don't worry if we have some typos here. Typos are completely no problem because these diffusion models understand what we want. Now we press here on style, and here you see what you can use. Comic book fantasy, Anime realismus, or Red Animation. I think this should be somehow fantasy then the duration. And if you press on it, you can always use a long or short duration. But the waiting time, of course, increases. If you want something short, you wait two to 5 minutes, medium, four to eight, and long eight to 12 minutes. This is the only downside from this tool. I think we should make medium just to make this a little bit easier then then we have a lot of different options. Image camera, model, version, negative, prompt, end seat. If I press on image, I can upload images if I really want to. But right now I don't want to upload a reference image, I simply delete this. But if you want to upload these images, you can totally do this if I press on it again. So I simply delete the image. If I press on it again, I can press on camera. Should the camera zoom in? Zoom out, pan, left pan, right, up or down. I think the camera should zoom out. Then we have the model version. We can use version one or version two. Version two is the default. Version two works better. Then we have the negative prompt. You remember what we can include in the negative pront. We can also just copy the stuff from the guy above us because we simply need a few words. And all of it will work seamlessly if I press on it again. We can also use a seat if I press on it. Let's just say we want to have this set, that the video is a little bit more coherent. And then this is the last thing right here, This is the image. But like I said, I don't want to use an image, I don't want to mix this up. The only thing that is left is to send this out and then Moonwaly tells us, we're working on your video. We will notify you when it's ready. Here we have our prompt ID. We will get notified as soon as this video is ready. And there we have our video. You'll see this is the prompt. Then the model is the fantasy model, The length is medium. We have our seat. This right here is the negative prompt. We zoom out. We use version two, and we have the prompt ID. And this is what we get out of Moon Valley. I, I think this looks somehow cool. Let me just make this big. I think this looks somehow cool. We have every single thing that we have included in our prompt. This looks really spacy. This looks like a guy on Mars. I think this looks like a cool, nice little AI video. Let me know what you think. Of course, you can do a lot of different stuff in this tool, Moon Walley. Ai is relatively nice. Of course, you can simply see for yourself what other people are making. Just let yourself inspire you by the generations from other people. There are really a lot of people that make a lot of cool stuff. You should totally look at these and make cool stuff for yourself. In this video, you have seen Moon Walley, AI. Moon Valley. Ai is really cool because this is completely free forever. At least right now, you can make as many videos as you like. The only downside is that you have to work in this court. But come on, we can get over this. We have unlimited generations. The generations are relatively good. We can work with text, we can use prompts, we can use negative prompts. We can use sets. We can use different camera options like zoom in, zoom out, pan, left pan, right pan up, pan down. You can also include your own pictures if you like, as a reference. I didn't show you this because this is your nice little thing that you should do at home. Just go in the tool, upload a picture from your computer and use it as a reference, and let me know what you get. 9. RunwalML: Everything You Need to Know: Let's just talk about Runway ML. Runway ML is a nice little tool we can make videos out of text. Runway can do text to video, it can do image to video, it can do video to video, and you can also edit your videos. So Runway ML can do all of it and you can start totally for free. Yes, if you want to make really a lot of videos, you need to upgrade your plan, but you can start totally for free. And I think we should totally do this. Runway ML is a leader in this industry. So first you go onto the website. So on Runway L here, you can also see that you basically need the subscription if you want to make really a lot of stuff, but I think you don't need it to start. You can upgrade your plan if you want, but I think we should start completely for free. I want to close all of this. Bring your imagine to life with Chen two On the left side, you can see what runway ML can do. We can go on Runway Watch here. We can simply watch the stuff that is created from Runway Dot Ml. You can see all the generations that other people have made. Some of them are really good, some of them are, yeah, Maybe not that great, but I like a lot of these generations. Let's just see what these clouds have to offer. I think they look nice. Can you see the face inside of the cloud? The cloud starts to merge into like persons or whatever you want to call this. Here you see hands and this looks really, really awesome. I think some of these videos from Runway Watch, they look really, really cool. The generations are stuff that you can expect out of Runway Dotel, and this is stuff that is nice and easy to create. You can simply type in a text prompt. But of course, Runway ML has also some generations that I personally don't like that much. For example, this right here, sometimes you get a little bit like special results, I want to call it. But most of the generations, they look really, really nice. You can also go on assets. Here on assets you have favorites and you have all here are just the generations that you have already made with Run ML, you have the video editor projects. As soon as you have edited your own projects, they will appear here. You can generate videos, you can edit videos, you can generate audio. You can also make images free D, AI training and much, much more. But let's just go back on home because here on home you see a lot of different things. Video to video, text to image. You can also remove the background. You can do text to image. You can do image to image. Image to image is also relatively nice, and of course, you can also do text to speech. This is also a nice feature that Runway ML can offer. If you scroll down a little bit, you see also some tutorials. They show you like really, really quick, how Runway ML works and you can discover and remix the stuff from runway. But let's just start with video to video. This simply means that we can upload a video, we can drag and drop a file here, and then we can edit our video. We can simply lay prompts over it. We drag and drop our video inside of this nice little box, and then we are basically ready to rock. I have to admit it. This thing takes a little bit until it's uploaded here on the right side, you can see what you can do as soon as all of this is uploaded. You have you and the style reference. You can use image, a present, or a text prompt. You can type in stuff. Or you can also use pictures just like here. If you press on these pictures, you have a lot of different options, but you can also upload different pictures. Let's just assume we want to have the Cyberpunk City. As soon as I use this image as a reference for my video, we will lay prompts or we will lay this style over my video, and we can simply make myself into a Cyberpunk. We have also the style strength. The higher the style strength is, the higher of course, the cyberpunk city will be. I normally I use stuff like between 20% sometimes up to 60% But I have to admit if we use a high style, just like 55% we get nearly just Cyberpunk City and you see also the set. So the seat is always important. You already know what a seat is. We use a fixed seat. This works automatically and because of the fixed seat, we have coherent videos. This is really important. Learn how Chen one works. So we can also press on these. Then we are on the website. Here on the left, you see the original video and the image prompt. If you scroll down, you can see you can use the style to structure consistency. So you can use the style weight, style weight of one versus a style weight of 15. The higher the style weight, the stronger. Of course, is the style you can use different. Also the set is a little bit different in every single generation. You understand what the seed is. We need to use a fixed set. We have the frame consistency. The higher the frame consistency, of course, the different the videos will look like. With consistency five, you see a little bit of flickering and you can use up scales. The up scales are also really nice because you can even change resolution and you can remove watermarks. But in order to do that, I think you need to pay for this tool and you can also mask different stuff out, so you can do really a lot of different things inside of Runway ML. And here you have a nice little comparison from the original to the other. Now you can see my video is uploaded, and as soon as we have uploaded every single thing, we have like a four second video generation inside of Runway DoteML. If we kick, click, play, you can see that I move here, so all of it seems to work relatively fine. And now we can lay our picture over it, because we want to have a nice little preview. If I press preview, we can see just some pictures, how all of this looks, as soon as it will be generated. So you can simply see, yes, we are relatively strong, so we have really strong cyberpunk view right here. If I decrease the strength just a little bit, I think the human touch should be a little bit bigger. But I have to admit, if we use such a strong cyberpunk view, this is always relatively strong. But like you see, it gets a little bit better. Let's just decrease it to 23% Preview one less time, I think this is nice. Now we can see that here, there is also a human in the cyberpunk city. I press generate this video, now we need to wait a little bit. This takes like somewhat between 1 minute, sometimes also 5 minutes. You see 21% So this needs a little bit of work until it's done. But I think it's worth the wait because the video, they look cool. Now we can press play and you see I move relatively seamless. I won't play the audio because this is German, but we can also use different styles. You can see we can make myself green, for example. And if we press preview style, you see that I get changed immediately in a thing that looks somehow like damn frog. But I think this looks cool. But we can also make myself into a bad cartoon if we really want. So we can use whatever we want. If we press preview style, we can always see how I would look as a bad cartoon. I think the bad cartoon command like this, looks really nice. You can always press preview before you make your presentation. If you press some prompts, you can also describe just what you want to see. Let's just say you want to get into a real frog. You can always try different prompts of course. Or maybe you want to be king of the rock. Let's just say rock as a king. We press preview styles and then we see what a rock king, what a real rock king would look like. Come on, this is hilarious. We can make myself into a rock king, and I think this looks also relatively nice. If we want to generate this, we can totally do this. I can always transform myself in whatever I want. Now let's just go back to runway, and now we test text to image or image to video. We can use text and we can make videos out of text. We can simply also drag and drop stuff right here. So we can upload pictures if we want. And we can also animate our pictures. But first, I want to start with a nice little text prompt. Inside the eye of a storm. Calm and peaceful in the middle, but harsh on the outside. I think this is somehow okay. Maybe the prompt is not perfect. Let's just see if we get an eye, a real Y, or if we get a tornado or something. But you can also see you can use different sets. You can up scale, you can remove water marks. But in order to do that, you always need to upgrade your plan. But I think right now it's not really worth it. But we need to try this out before we pay for this stuff. You can use different resolutions, 16 by 99 by 16. You can use different styles. Should it be a three D cartoon, should it be like three D rendered, You can simply search for yourself, what you like. Let's just say Thriller. I think thriller looks okay. You can also press on camera, and here you can use different camera movements. Do you want to make the horizontal axis move a little bit more or the vertical axis? You can also pan into different directions. You can zoom. You can also role, so you can always see on the left side how all of this behaves. Even the role, I think the role looks really cool. If we want to pan is also a little bit interesting to see. You can always play with all of this. Camera motions just increase or decrease it a little bit. I have to admit, if we use really strong camera movements, the video will be a little bit of a mess. But I think it's worth trying to do a little bit. Let's just press safe and then we can run all of this. But before we do that, I also want to show you the motion brush, because you can also brush with the tool over the pictures. But just if you upload it, let's just press Generate and see what we get out of this nice little prompt. I would assume that we get maybe a real, or we get the storm that I had in mind previously. Remember the prompt engineering? We also need to make good prompts if we want to have good outputs. I think it's maybe a little bit of work, I think, yes, damn it. We have I, but I think also, this looks like relatively cool. Just remember, you need to take into consideration, if you want to have specific output, you need to be specific with your input. Yes, we have like an I, I think we have a real storm going on right here. Maybe I should use the word tornado or something else, but I think this is totally fine so that you can see that we really need to be specific with our words. If we don't be that specific like we get an eye because the diffusion model sees any in the prompt and is trained on ice and generate ice. The camera movements are strong. But we can also upload other pictures. This, for example, is a picture that I have made with another diffusion model. We can do a lot of different stuff with pictures. We can use a motion brush and paint over it and we can animate it. So in simple terms, we can make pictures move that we upload. I think moving pictures is also relatively nice feature. Let's just see what we can get out of her. You see we have a woman with like pink hair. This is like an AI influencer. We delete our prompt and we make, let's just say waves. I think waves work. And if we use the motion brush, we can paint over different stuff from this picture. We can use just the hair, for example, if we want more motion in the hair or if we want more motion in the clothes or in the waves. Just let's animate the real waves. But I have to tell you this motion brush, this is not always completely perfect. Sometimes we also animate the whole pictures. We can also increase and decrease the movement of the axis. And we can always delete the stuff if we don't like something. If you think you don't want to have something right here, you can always include it or exclude it, just like you want. You have also different brushes. You can use this brush or another brush. You can always play with all of these things. But I think we should try to animate also, like this small piece of this hair. So you can also decrease the auto direction and you can simply take it into manual. Let's just play it done. Always remember, none of this will be completely perfect, but it should work. We should be able to animate it. Let's just include wind. And then we press Generate, and we see what we get out of this. Picture. I think maybe like the whole picture, it should move, and maybe it moves also a little bit too strong, because this is always a bit tricky. Damn. And there we have it. These are really windy waves, you see. The whole picture is animated. Think it looks really, really nice, but it is of course not coherent. And now I get flagged. I don't know why I get flagged, but I want to work with another picture. Now, I uploaded this picture right here. If we press on motion brush, we can always animate our picture. I think we should use it a bit sparingly. We simply increase the strength here just a little bit, and I just use the hair and then we press done and we see if we can make this a little bit more coherent. Sometimes it can be a little bit messy. If we want to animate pictures, we decrease this things also, and then it will be a lot less strong. We press safe, so now we have just a little bit of movements and it should work a little bit better. You can always press on the tools again and repaint the things that you like. Let's just press Generate. I think the moves should be a little bit more gentle now because we have decreased the strength really by a lot. It's always important to make this not too big because if you increase it really a lot, you get the strong movements just like in the last generations, and this makes it really bad. And now you see the movements are gentle. Because we have decreased it. Yes, we have a little bit of flickering. Yes, the eyes get a little bit deformed, but degeneration is like ten times better. So you need to be a little bit gentle with the movements. If you are not gentle, it's totally okay. You get good outputs also with not gentle movements, but the outputs get a little bit like strong and they are not perfect. If you want to have gentle moves, you can be a little bit more coherent. Let's just try another picture. I uploaded this picture again. This is also a nice little picture from like a influencer. We press on the motion brush and we can also brush this stuff for ourself. If you want to have really, really gentle moves, just brush over it with your hand and tone to use the automatic selection tool. And then you can animate also really small stuff, but I think you get the idea. The less you animate, the more coherent the pictures get. The more you animate, the more camera movements you include. The stronger the animation get and the messier all of it looks. Let's just go back to the dashboard. You have seen video two video, you have seen text video, and also image two video. We can also remove background. This is also a nice little tool. If we press on it, we can simply drag and drop a video from ourself or from whoever we want, and then we can remove the background. This is also a nice feature, I want to upload this nice little dancing lady. This is a video, I think, from Ps. As soon as this video is uploaded, we can edit this video a little bit. I think now it should work. We simply press on to timeline and then we will get our video on our timeline and we can edit it. And there we have our video. And now we can mask different stuff here, let's just try to mask this thing in the background. If I use this thing, if I mask all of this thing, we should exclude everything else. But except this thing, then we simply have like a video without background. We only have this thing that I masked. I think this is a nice little tool. Like normally, yes, we mask, of course persons out. You see the person is dancing and we can of course, also include our person. So let's just include the person and the thing from the background. We press done masking. And then we will have just our person and the thing from the background that we have masked. And every single other thing will be excluded. So you can see it. It works really, really good. I think the person is dancing and we also have like this light or whatever, this is in the background. And the rest of it is a green screen. So you can edit your videos really, really nice with it if you really want. You can also include like different backgrounds because now you have a green screen. You can always press export if you want to download these pictures and you can press press play if you want to see it. Once again, you can basically include this in every single background where you want and you have a girl dancing and this light in a new background. Like I said, this is a beta feature, so none of it will be completely perfect, at least not right now. But you can also use different other things. You can include assets. You can include, for example, text. You can include whatever you want. So this is basically a video editor, but I have to admit, I don't like to edit my videos in this editor. I just use it to mask out different stuff of videos. Let's just go back on home because I think this remove background tool, you can try it but it's not perfect. And here you see text to image. I have to admit the text image works in other tools a lot better. We can use stable diffusion. We can use mid journey. We can use Ali, we can use Adobe Firefly Text to image is not perfect in run way ML. And that's why I don't use text to image at all in this tool. If we press on it, you can see basically just what we can make the pictures, they look somehow, okay. But I don't really love these pictures. And they are also cherry picked. You can simply describe with a prompt what you want. You can use different aspects, ties, resolutions, different styles. You can generate pictures, but I don't love it. In this tool, you can do the same thing basically also with image to image. You can upload an image if you want and you can make similar images. And also this doesn't work that great in runway, but you can also go to text to speech. If you go to text to speech, you can simply press on it. And then this works like somehow, okay? Because you can create a nice little voice. The voice is called Kathy, at least right now you can see it in the left corner. You can simply type in what you want to hear. And then you will get your speech pack. If you press on this right now, you can simply type in, for example, hello, I am Kathy. Because like she is called Kathy. Then we press Generate, and in no time whatsoever we will get our audio back and you can use this audio forever. You want Hello? I am Kathy. Hello, I am Kathy. You hear this looks like relatively okay. If you want to download it, you can totally download this. But I have to admit I like 11 laps more than one way ML for this text to speech generation. You can make this longer. You can also make like stuff in chat PT and paste it in. Let's just go back to dashboard because we take a closer look at 11 laps later. And 11 laps is a bit better. But you saw that you can make video to video inside of Run ML. You can make text to video and also image to video so you can make every single thing that you like. With this tool, you can remove backgrounds if you want. Like I said, I don't use this Aton in this tool. But if you want, you can do this. Text to image is not perfect. Image to image is also not perfect. The text to speech, this works somehow, okay. But I have to admit, I don't use it a lot because I like 11 labs. You have seen Runway. Mel can do all of it. You can make text into speech. You can make image into image. If you want to try it, just go over it, text to image, I would say just use another tool. The remove background is somehow okay, but the text to video, this works really good. And also the image to video works good. You can use the motion brush and much, much more. The video to video is awesome, because you can lay prompts over your videos. So this was basically runway Ml. Runway dot Ml is a nice little tool. You should totally try this out because you can start totally for free. Do it. 10. PikaLabs Text to Video, Video to Video, Image to Video, and Video Editing: Let's talk about collapse. Because collapse is a tool that works really good. You can make basically the same stuff as in Runway Dol. You can use text to video, video to video, image to video and match much more. And you can also edit your videos. Let's just take a look how all of this is in the interface. But first we go to the web page. If you go on their web page, it looks something like this. You see immediately trip. But if you scroll down a little bit, first you go outside of this window and then you can see what's on their web page. You can press Create, but you can also see the stuff that is generated by a. You can do text video, and it looks something like this. They try to always make raccoons. Don't ask me why, but I think raccoons are nice. Come on, you see, you can also make image to video, and if you press on it, your images will start to move. Also image to video works relatively nice inside the beak. I think the generations are also a bit better than in runway, and you can do video to video. You can see, you can change your videos, and this looks really, really awesome. If you scroll down just a little bit, you see also this lip sync. Yes. You can also do lip sync, and you see it right here, the lips, This looks somehow okay, but I have to tell you, this is not perfect. If you do this like with real humans, the output is not that good. I want to show you later in the course, of course, an alternative that makes better lip sync. But this is cool. Modify region. And modify region works relatively nice. You can simply modify stuff from your images, from your videos, not only from your images. And some things works awesome, other things are not perfect. But you can see these examples, they are completely awesome. You can also expand your canvas so you can change the resolution. You can make stuff in 16 by nine, in nine by 16, in one by one, in whatever you like. You can out paint different stuff. Of course, you can also extend the video length if you really want to. You can simply extend a video over and over and over again for 4 seconds until you have something that is decently long. Here you can basically see a lot of examples if you want your ideas. Ps command Tripka, let's just try this thing out. Let's just see how all of this looks in the real interface. And of course, you need to make an account. You can either sign in with Google or sign in with this Court just like you want. As soon as you are in, you can go onto the Explorer Page. This right here is the Explorer Page. You can simply see what other people have made and you can also make similar stuff. It's really easy to make similar stuff. You can always copy their prompt. If you go with the mouse over these pictures, you can see how all of this looks. This robot, for example, is moving relatively coherent. I think the videos, they look good from collapse. You can also download these videos if you like. You can totally use this video. You can copy, for example, this prompt if you want, or you can simply see again, a raccoon. These raccoons, I think they look hilarious. I think I understand why they always try to use these raccoons, because, come on, the raccoons are nice. You can also make other stuff here, like New Year from Japan. At least I think this is New Year. And you can see also some dragons. And if you press on these, you can simply copy this prompt. And you can also throw it down into the prompting box and make the stuff for yourself. As soon as you press Generate, you create the same video as the guy that made this video. Yes, all of this will be a little bit different, but I want to show you something else. We want to start with image or video. So we can simply press on these, and here we can upload videos from ourself. And I think we should upload again this nice little dancing woman, because this is a really enormous, hard video to process. You can modify region, you can also expand the canvas if you want, and you can do lip sync. But like I said, this is a really hard video. But let's just try modify a region because we want to do hard stuff. You see we have here this generation box and we can move this around, You can make it bigger, you can make it smaller. I want to try something that is really hard for this diffusion model, because this woman, she moves like really fast. I want to create a new hoodie for her. I think she should wear something that is maybe red. First, we need to try. To take the box in the right place and we throw in a nice little prompt. Let's just think, I want to make it, a model is dancing in a red hoodie. Then we need to accept that it maybe not works perfect. We need to make this a bit bigger because we need to have every single thing in the frames here. Let's just press generate. Come on. I think this will not work perfect. Like I said, I want to show you hard stuff here. We have our generation. Yes, it doesn't work perfect. But this is the stuff that I want to show you, not always perfect. Let's just take a look at this. So here you can see it. Yes, it works, but it is of course not perfect. She has a red hoodie. It works somehow. Okay. She's again, dancing, so no problems here. But the movements, they are really too harsh for this technology. You need the movements a little bit more gentle if you want to have good videos. But like I said, I also want to show you the downsides that we have right now. It would be really easy for a nice little slow motion video to generate the clothes, or to add, for example, a hat, or to add some sunglasses. So I want to show you the challenging stuff so that you can see that it doesn't work perfect all the time. We will get there eventually, but right now we are unfortunately not there. But we can try to expand the canvas. In order to do that, we delete every single thing right here, and we press Expand Canvas. And here we are. So you can see we can simply press on different resolutions. And then we can expand the canvas however we want, 16 by 94 by 34 by five, Whatever you want. You can also place it wherever you like. I want to do also here, the really hard stuff. This is a realistic video and I want to place it in the middle. If you use illustrations, this works really, really easy. You need to insert also a prompt. What maybe model is dancing in a let's just think, what is this exactly? Just a room command. So we include a room. I press Generate, and then we will get something. But I think also this will not be perfect. Like I said, if you use an illustration, this will work perfect. You have seen the videos from Pka itself, but I want to push it to the limit. Let's just see how the generation went because I want to show you the hard stuff. Yeah, let's just make a full screen. I think this looks somehow okay, but it's of course not perfect. You can see on the left hand, the right. We have out painted our pictures without any doubt. But of course it's not completely perfect. But that's the stuff that I want to show you. You need to understand this. If you want to make realistic stuff. This technology right now, it's nearly there, but it's not completely there. You see, all of this works, but it's not completely perfect. Maybe you ask yourself, why does this guy show us all the time, the stuff that doesn't work perfect? You can also make better stuff. I want to show you this because you need also to understand the downsides of these models. If I simply show you stuff on illustrations that works seamlessly, you maybe get false information. I want to be as transparent as possible. I want to push these models to the absolute limit so that you understand what you can and cannot do. You already saw the presentations from Ka. They look awesome. You can make similar stuff, Make a raccoon in the space that tries to catch something. If you want to make videos or just maybe try the painting with something that moves relatively slow, all of this will work really, really good. But if you want to go to the limits, also the tools come to the limit. I want to show you this. But of course in the future we will get to this point and then I will update the course and we can make entire films. Maybe just with some prompts. Let's just go back here because we can do lips. We can also lipsyn this video, but I think this video right here, this would be an absolute mess to Lipsynch. We need to Lipsyn something other. Let's just think, I think we should lip sync myself. This is also really, really hard for K. We have other tools that make better Lipsyncs in realistic characters, but I want to show you these. This works relatively easy. You just need to upload a video. I uploaded a video from myself, and you can also see it right here. For best results, use lips with front facing human video and clear high quality audio. So we can simply upload an audio that is high quality and we can make lip sync. But in order to get an audio, we need, of course, to make an audio. I like, for example, 11 laps to create an audio. So we want to animate this video with lip sync. If I press on these, you can see we can lip sync the audio. And we can simply see for ourselves what we want to use here. We can use Rachel, for example, or Drew or somebody else. I think we should use Rachel. And we can press Generate Video after we input our text. This is like not from 11 laps. And if we press generate voice, we will get a voice. But it is voice. This is not really great. We can totally make this voice. But like I said, the voice is not optimal. In order to get a good voice, we should go to 11 laps, just Google 11 labs. And then we can simply press on the first link. And then we are on the web page from 11 Labs. And here we can create audio that is a lot better than the audio here from K. We can go into 11 labs. And here we can type in our prompt. For example, give me more attention. I have a nice voice. Come on, this should look great. We use like a British girl, because I want to be a British girl. Come on, I just like to use funny stuff because the world needs to be funny. This is how it sounds. Give me more attention. I have a nice voice. Then we can simply download the stretch, and as soon as this trach is downloaded, we can upload it, of course, into Pka. If we upload it, we can simply drag and drop it right here, and then attach and continue, and we will get our output. Maybe you ask yourself why I'm doing always such stupid stuff? Yes, I animate myself with a voice from a girl. I think this stuff is funny. If I want to make like serious stuff, maybe nobody will remember. And I want that you remember. That's why I make always silly stuff. I also think that this doesn't work perfect. If we want something that works perfect, we need to make an illustration. And like an easy text, but I want to make a male speaking in a voice that is English. And of course that's a girl. And it later in the course, I also show you stuff that makes really enormous good lip sync. We can also use Hagen and a cop notebook. This is a lot better with lip sync, but I want to show you also this right now in, even if K is not perfect with this here, you can see it. Now we can create everything, or now we have everything. Let's just see for yourself. You see my lips. They move completely different. And this is how it looks. If I make it big. So you see my lips are moving completely different. Yes. This is now without the voice. I will show you the voice later. I just want to show you that the moves, they are moving completely different. It worked, yes. And maybe this is like not perfect right now, but I want to show you how this looks and how it sounds right now. Give me more attention. I have a nice voice. Give me more attention. I have a nice voice. So you see, yes, it worked. It's not perfect, but I want to show you all of it. I don't like to show you always the positive things. I want to show you every single thing because this is important to me. You see we can do really a lot of different stuff here with Peak, we can do modify region and modify region worked like okay. If you take other pictures, it works better. You can expand canvas and make your pictures pick. You can do Lipsy. Yes, the Lipsync is not optimal. But we have also, of course, another tool that can do Lipsys a lot better. We have Hedren and we have a Cop notebook. And I will show you this stuff of course later. Not all of it is perfect. If you press on Explorer, you can see always for yourself, what other people have made. Here are, for example, some Lip things that work, maybe a little bit better. You see always the prompt down here, so you can search for yourself for the stuff that you like. You can always delete your stuff. And you can simply re run a prompt from here. If you find something that you really like, you can simply search it. Lip sync. Works with all of this like somehow. Okay, let's just copyDse, because now I want to show you something cool. We press retry, we are simply here, and then we get our nice little video immediately as an output. Here is our video. We simply retry this here. You see the video? Yes, this video looks like really good, I think. We can also do other stuff. Let's just start with this nice little prompt. Free de animation. Acute boy is standing in a house. Spring festival, interior, Lunar New Year Holiday, wipe snow outside, land tours. You see this is the prompt. This is the same prompt as previously. And now we can also change the resolution. If you click retry, you get always exactly the same stuff. But you can also change like frames per second. If you just copy the prompt, you can go down until eight frames and up until 24 frames. Of course, the video will be better with 24 frames. You can also change here different stuff, so you can pan, you can tilt, you can rotate, you can zoom. You can simply press on these things and all of it will get included in this picture. Let's just presume rotate and also the pan. And then you will get an output that is relatively cool. I think at least you can go on the parameters. You can also include negative prompts and sets. On the negative prompt, you simply type in what you don't like to see. For example, ugly, bad, blurry out of frame, and so on. So you can include what you want if you like. You can also include a set. Remember, it will get a little bit more coherent if we include a set. So this is a nice little trick. The consistent with the text, the consistency with the text. You can increase or decrease it how you like. The prompt will get executed more precisely if you increase this, but it's possible that the output gets a little bit lower resolution. Let's just try to generate this. I think the generations work relatively fast here, so we simply wait like one to 2 minutes and then we get our whole generation. And the coolest part is we can also upload a video from ourself and lay these prompts over it. So we simply try this with our Lip Sing video. The second generation was, yes, it was okay. But like the eye is broken, It's completely broken. Maybe the consistency of the prompts was a little bit too high. But you see the camera movements, they work completely fine. Just the eye, or maybe both eyes, they need a little bit of a fix. They are of course not perfect, but I assume that this stuff will get perfect over time. Now you see what we get. If we lay prompt over my video, this works really nice. We lay this prompt over my video from previous, from the lip sync. I think this looks relatively cool. The original versus the generated. Let's just make the speak, you see the cartoon stuff. This works always a lot better. In this video, you saw collapse. Collapse is right now, an all round tool that works relatively nice. You can make videos out of text, so you can do text to video. You can make video to video. You can do in painting, you can do out painting. You can also animate pictures and make videos out of these. So you can basically do all of it. I have to tell you, yes, I showed you always the stuff that is not perfect. I want to use a girl that is dancing like really aggressively so that you see the flaws in these models. I also want to show you a lip sync that it doesn't work perfect. I assume that all of this will get perfect over time. But right now we are not there yet. But we will get there. If you want to have perfect outputs, just do stuff with animations. You can lay prompts over videos just like you saw. And this looks really good. If you want to paint stuff, do stuff where the motion is a little bit gentle and slow. If you want to outpaint stuff, just try a little bit. Of course, if you want to do a Lip sing, I will show you a technology that can do perfect Lip sys. Also, right now, K is not the tool for that, but I think BK is cool because you can try K for free, at least for a little bit. Try yourself a little bit out. Have fun with these tools. 11. Haiper AI: Supported by Google Deepmind: Hyperai is another video generating tool that works really nice. You can make videos out of text. Let's just see how all of this looks. We will make this relatively fast. I go on this side so that you can see more. You go on hyper, and here you can explore. You can see what you can make. First, I want to show you that you can create videos in HD. You can animate images in HD, but you can also repaint your videos. But the repainting, yes, I have to tell it you can't do it right now, but this will come relatively soon. You can also make normal videos and normal animations out of your images. The only difference is here HD and not HD. If you go right here, you can also extend your videos and also this comes soon. Now I'll show you this web page. Why is this web page even relevant if we don't have all the features? Because this web page is supported by the mind. I assume that this web page will get better and better over time. And the second thing is this is totally for free, at least right now, you can test it completely for free. Here are some videos. The videos are good. The videos are coherent. You see this dragon, This looks really good here, a robot with the Earth in the hand. I think this is a cool video. This is also the right time for this video here a girl in a car and you see also realistic videos work really nice. This glass pony or whatever this is, maybe not a pony, it's a horse, but also this looks cool. Here the sky on the moon with a husky or Shiba, this dragon out of bubbles. You can see we can make really, really good pictures. And I am stunned by the coherence and also by the quality of these videos. You can see the videos, they look not completely real, but the realisms in these videos is really nice. Even this right here, so these two people that kiss each other, I think this looks awesome. Also, this car looks really nice. Now I want to show you how we can make this stuff. Before we make stuff, I will also show you that we can of course, like this and then you will find it. If you press on this button, you like it and then you have it in your collections. You can also download it and you can also of course, share it on social media or wherever you like. If you press on such a picture, you are here and you can see it in full screen. And you can also see the prompt. A supercar driving in heavy rain and camera behind it while the car driving through traffic. I think this is like exactly this video. You have always the ID. You don't need to worry about the ID, but the deceit is important. If you want to recreate exactly this video, yes, of course. You can also download this, but if you want to recreate exactly this video, you can, for example, copy these prompts, you simply press on copy. Then you can also use the seat and you will get exactly the same video. But this is not what we want to do right now. If you think this is cool, you can also of course it here. Now I want to close this window and I want to show you the other stuff. If you press here on creations, you see the things that you have made. Right now, there are some generations. Here is a lion in the jungle, and here are three people on the moon. And if we press play, you can always see how the video looks like. The videos maybe not optimal. If you go on favorite, you see the stuff that you have shared, that you have liked, for example, this car. And you can also get on your profile, you get a nice little starting guide. You can get help if you press here on more, you can log out teams and conditions and the private policy. If you want to have more details about the private policy, let me know or click on this button. Of course, if you don't like the dark mode, you can also make it bright. But I hate it when it's bright, I want to work with the dark mode. Now we go on creations, and we do something for ourself. We can simply go down here. And now we can, of course, describe our video. And I want to run with a prompt that is really, really easy. We don't describe every single thing exactly. I want something like this. Iron Man climbing Mount Everest. And here we can use some genres. We use old film, watercolor, cyber punk errand, Lego, bluer, background, Chi. Ask me what this thing is, maybe I should Google this or ask ChatGPT, Steampunk and Impressionism. But I think I don't want to use something of these. I simply want to press Create. And we use HD if you press on the. You can also use like smaller models. So you can make, for example, your normal text prompt without HD, but I think we should use HD. If you press on this button here, you can either share it publicly or you can also leave it private. I want to share this publicly. This is completely okay for me. If you press on these, you can use different length of this video, 2 seconds or 4 seconds, or most likely 4 seconds are just coming soon. But I want to use a seat. I think we get a better output with a seat then. We simply press Create. There we have it. Let's just see how our Iron Man looks like. I think this looks really good. It's really, really coherent. He is on Mount Everest and you can clearly see, this is Iron Man. And now I have to tell you, copyrights can be a little bit of an issue because it's maybe not that optimal to create Iron Man. Iron Man is a creature from Marvel, and I think, yes, it's completely okay if you just make it from time to time, but don't try to make entire films or entire videos out of Iron Man, and like try to sell them or whatever. I think you should get these. So maybe it's not perfect to create Ironman and like make a movie out of it. But of course we know this. You can also rate it if you like it. I think this is great. You can share it. You can do whatever you like with the generations. In this video, I showed you hyper. I showed you this tool because I think this has great potential. This comes from deep mind and right now it's completely for free. They will add longer videos, they will add more features. Maybe as soon as you see the course, you can do a lot more in this tool. Of course, I will bring updates as soon as more as possible. Have fun trying this tool out. 12. PixVerse, Morph Studios, and LTX Studios: In this video, I want to talk about pigsworse. Pigsworse can also create videos, and right now it's completely for free. After you have seen Bigsworse, I also want to give you a nice little outlook about two tools that are really cool. Because until now, yes, you saw a lot of different tools that can make good videos. But the thing is these videos are always short. Of course, we need tools that can make longer videos, can make coherent videos, or can loop different videos together. If you can loop different videos together, you can also create standing art. This right here is Pick worse and Pick Works Works really, really similar to all the other tools. Here you can see some of the generations. I think this is a nice generation also. This cat looks, at least at the first glance, really, really good. You can always crow down a bit. Come on. This tiger is also nice, so you should totally look about all these generations. This is somehow like maybe a husky. They are cute. Come on. You can also make realistic things. I think this woman looks really good. Yes, maybe we have a little bit of flickering going on when she blinks, but also this bird looks really realistic. Yes, even with fast movements like this, you see this tool can make really good videos. And the coolest part is this is right now, at least totally for free. You can test all of this out. Of course, if you want to recreate some pictures of here or some videos of here, you can always press on them. We can press on this video and you can see it here for yourself. I think the jelly fishes. Yes, they look awesome. The prompt is Jellyfish rises. This is a nice prompt. So this is really, really easy. You can go to create, you can upscale, you can retry, You can of course, also copy the prompt, Copy the seed and the strength of the motion. Now the thing is this is not from a simple plain text prompt, but this is from this picture. He uploaded a picture and he animated these. You already know how all of this works and you can also get stuff from stable video diffusion that is similar. What I want to do right now is to simply copy these. And then we press Go Create. And we make a similar picture, but we will do it out of plain text. We don't want to include like new images. Okay, right now this image is included. You can see you can always retry this, but I want to exclude this image. I simply want to go to text and of course, just like in Runway and be aka you can always upload your own images. But I want to go to text and then I can simply insert my text. Jelevish Rises. Then we can use a negative prompt. Let's just make it sim ugly, blurry And that's it. Come on. Inspiring prompts to al clips. You can go with the mouse always over these things, and you can read it yourself. By turning on this feature, your prompt will be analyzed to find the most related and appropriate prompts. You can include this. And then you get maybe a little bit better prompts and this gets completely automatically update. I want to include this. Then we can also use a style, do we want to have a realistic style enemy three animation and CG style of aspect ratio. I think I want to use realistic aspect ratio in 16 by nine so that you can see it better in this course. Then you can play with the. If you want, this makes no big difference. But if you use, for example, this seat. And then we press, we will get our nice little generation of course. And then if we do this again, we can also make completely the exact same picture, or the exact same video if we use the same seat. But now I want to wait until we get our output. Here we have our jelly fishes. You see the two pictures or the two videos? I think at least they look really cool. We have good movements, we have a really coherent picture right here, and the same stuff also here. And we can make it be the resolution is good. We have a little bit of movement. So all of it looks really cool. I at least, so you should totally try this web page out because on this web page you can make videos, at least right now, completely for free. And now I want to show you two tools that can make longer videos, that can clip and loop different videos together. The two things that I want to show you are, first of all, and more studios is really cool. More studios works really nice, but I don't have access. You should totally enter your e mail address right here and just hope that you get access, because more studios is cool. I want to show you their introduction video right now. What's your AI Filmmaking workflow? Text to Video. Image to video. Video to video. How about idea to video? Introducing Morph Studio, an all in one AI powered film creation tool with an easy to use storyboard format. You can easily create and edit shots together based on text or images, and generate videos in different styles. More studio helps you connect your ideas, test different versions, and export to any post production software. We believe in a sharing community. Users can upload their workflows to the gallery and allow others to access them for their own projects. Welcome aboard the ship to the future. Join the waiting list today. So you see more studio is cool and you can expect similar stuff out of LTX Studios. Also at LTX Studios, I have made it into the wait list, but right now I can't get in. So you can simply press join the wait list and we together hope that we can get access more studios and LTX Studios. They are really, really nice. They are really awesome. There's a lot of demand and I can't get in. As soon as I can get in, I will make updates immediately in this course. I hope you get access. You should totally try this out. And if you don't get access, just play a little bit with Pix Worse. Of course, later in this course, I will also show you how you can edit your normal videos out of Pka, out of Moon Valley, out of wherever you want, with a nice little editing program. And then you can also string your videos together and make something that is a bit longer. See you in the next video. 13. More Tools Overview, InVideo, Fliki, Kaiber, Lalamu, and More: In this video, I want to give you just a quick little overview of some tools that are also somehow okay, somehow nice. I have to admit it, there are so many tools out there right now and I really think not all of these tools will make it. Some of these tools will not be around in one year or in two years because like there are too much. I don't think that all of these tools can run profitably over time because they always need a lot of GPUs. The best tools they will stay there, and the worst tools they will get washed out of the market. That's why I want to give you an overview of some tools. I have to say it, I don't know it exactly what tool will take off and what tools will not take off. Because this is also about marketing and about a lot of different factors. But you need to try this for yourself. You need to see what's out there. And I think like one or two tools that I show you right now, maybe they are not around forever. One tool is in video AI. This tool is okay. You can use it if you want. You can simply go on their web page and you can try it out. The second tool is Flicky. In Flicky you can make also like similar stuff. Then in the tools that I already showed you, there's also a demo. Like you can see for yourself all these demos, I don't want to waste your time. Then another tool is L Lamu and Lala Tu is especially good about Lisin technologies. But I have to tell you, I will show you something for Lisin that is better than Lalamu and that is completely free to use. The last thing that I want to show you, this is really awesome. This is Kyber. I think Kyber will stick around for a long time. Kyber was one of the first tools, was one of the best AI video generating tools. And in Kyber, we will make a deep dive later because even Linkin Park has made their music video with Kb. Of course later in the course, I will also show you how you can use stable difusion like a pro. After that, I will also show you how you can use Kyber. Because Kyber, this is really, really good. This was just a quick overview about four tools. I am relatively confident that Kibi will be around for a long time. I need to be honest, I don't know if the other tools are around forever, but maybe I'm wrong. Maybe one tool takes completely off. And of course I will show you that tool. And you know what, As soon as we get a of course, I will update the course. 14. End of Section 2: Over the last few lectures you have seen really a lot of different tools. We started with the easiest one with stable video diffusion. You can upload a picture and you get a video. It's instant and it's completely free. It's open source. You have seen Moon, well the AI and you can work also for free. But in this court then we have Runway and aka the two things are at least right now, the big dogs. But of course you need a subscription if you want to make a lot of stuff in runway. And aka Hyper AI from Deep Mind is also a nice tool. And right now it's completely for free. Big worth works good. The cool stuff would be of course, LM Studio and Morph Studios. I hope we get access really soon. You saw also an overview of some other cool tools, and especially Kyber is a tool that we will take a closer look later in the course. You have seen a lot of different tools. I have something for you that you should try. Just go in one tool, just 11, that you like and make a video. You'll know that learning is doing and you get better by doing more. Start to make your own prompts. Start to make your own videos. Just play a little bit around until we create whole videos and we got into a completely pro with all of these AI videos. That's basically it. Just create a video. And in the next section we talk about I avatars because also the things are enormous. See you in the next section. 15. Section 3: AI-Avatar: This section is all about AI avatar tools. We will look at the big players. And I have to say it right now, the biggest player, or at least the best player, is right now, Hagen. Because you can clone yourself. You can make yourself speak in other languages and you can do a lot of different stuff with Hagen. You should totally look into the tool we take. Also look at DID synthesia Eli and that's basically, these are the big players. Yes, there are a lot more tools, but these are right now, the best tools. We will take a closer look why you should take a look at these. Because like big companies, they use these tools all the time to save a lot of money. And maybe you should too in this manner have fun. In the next video, we will take a close look at Hagen because Ag is awesome. 16. Heygen!: In this video, I want to talk about Hagen. Hagen is hands down the best AI avatar tool. You can clone yourself. And you can let yourself speak whatever you want, but you can also animate pictures. You can make whole presentations for social media and much, much more. And you can even let yourself speak in other languages. All of this is really, really cool and you can at least test it for free. Yes, if you want to make more stuff inside of Jen, you need a subscription. That's the only downside from this tool, but Hagen is awesome. Let's just take a look at their website. Every single thing starts right here at their website. If you go in the left corner, you see the use cases. You can make product marketing. You can also make content marketing, learning and development, and of course, personal videos you can see. You can make a lot of stuff with this tool. This works really awesome. We have more features. We have AI avatars, AI voices. We can also video translations. We can make personalized video streaming. Avatars are also a nice little possibility. And we even have say, beer included. If somebody works in say, beer, we can press on the pricing. And that's basically the only downside you see. If you press on this link, you are on this web page, you can start totally for free, but your credits will be gone relatively fast. You can basically have one free credit, one seat, 120 public avatars, and 300 voices. Like I said, this is the only downside. But you can update your plans, so you can make a creator plan, a business plan or an enterprise plan. It depends a little bit how much stuff you want to make and if you want to pay annually or monthly. Of course, annually is a little bit cheaper. You can always see this for yourself. And I go back on the main web page because here I want to show you some cool things. We can here see in this demo how all of this looks life. This right here is the CEO of Hagen. So we can see how all of this looks in life. If we press on these, hey there, welcome to Hagen, where you can easily create fun, high quality videos using our AI advertiser invoices. In just a few clicks, you can generate custom videos for social media presentations, education and more. So you see this looks really awesome. And if you scroll down just a little bit, you also see that we can make a lot of different stuff. And we simply press get started, and then I will show you how all of this works. Of course, you need to enter your e mail address or you can also make your account with Google. You have on the left side home, then templates, avatars, voices, video and assets. And of course, you can make video translations. This are really off webinars. You have, again, the pricing, the labs, the tutorials and the news. So you can do a lot of stuff here. You can also create an instant avatar, but this is in the pat feature. And the instant avatars, they look like the stuff that you see right here, for example, Matthew or the Hagen CEO. All of these are AI avatars. These are clones from their people. If you press on Edward, it looks something like this. Welcome to the new era of video creation with Hagen. Simply type your script to get started. You see these clones, they work really, really well. You can also make them in portray or however you want. You can also, of course, create an instant avatar. This right here is prices. You need an instant subscription if you want to create your own avatar. But the thing that I love most is the video translate. Just press on it, because you can test these, at least right now for free. Here, you need to upload a video. I just want to upload some simple words and then we can translate our speech. We can simply press on what language this guy should talk. In this example, me like I normally, I speak German, I speak English. I also speak Italian. Let's just search for a voice that I want to have myself, English. You can hear my accent right now, so we search for French. Come on, I don't speak French. I have no idea if this will come good or not. Number of speakers, we speaker for example, and then we simply press translate. This video, I have to tell you right now. I personally, I don't speak French, but I have done this a lot of times, and normally we get output that is really, really good. I have tried these with German, with English, with Italian. These are all languages that I personally speak. Yes, I normally speak just German. German is my main language. English and Italian are not that great. You can probably hear it, but I have no clue how this thing works with French. I can't really tell you if the output is good or not. Maybe you can tell me if the French speak is good or not. We press and translate this video, you see it costs us a half of a credit, and that's basically half amount of the credits that you have on this tool. This gets expensive relatively fast. You can test just for a limited offer. Let's just press Submit. And then when all is done, it's look something like this. You will get an E mail in this email, you will have your video. You will get this e mail in like five or 10 minutes. This E mail is of course, from Hagen, and we can simply press on our video and see how all of this looks right now. Video. So video. So maybe you can tell me how this is. Yes, I think it's French, but I have no clue if this is right or not. We just need to believe it. You can also of course, download this video. And I would assume that you should download it because you have not unlimited things here to download. Let's just go back. You see also that you can of course, make a lot of different things right here, you can always press on photo avatar, for example. Here you can use different avatars. You can use Felix or Elliott or whoever you like. You can make whole presentations. You can also go on this templates. And here you get complete templates. And you can simply see for yourself if there is something that you really like, that you really love. Let's just press on this business report. And then I want to edit this business report a bit. You can always look for yourself how all of this sounds. Welcome to this versatile business presentation template. Use this introductory slide to present your company and individuals involved. I think this looks relatively good and of course, we can edit every single slide of these. Let's just add it to this stuff. We are here in this editor and we can animate everything. We can move him around, for example. And we can not only move him, we can also move the text around. And we can, of course, also insert other stuff. And you have just a normal video editor where you can simply delete all what you want. Let's just place him in the right corner and delete this business report. Let's assume that you want to make something in German, We'll come in. This means welcome, like nearly as in English, your company. We can also delete maybe your company and say Nadu. This means hey you. You can simply insert all the stuff that you like and you can even change the speaker. So let's just say you want to have Brianna. And then you can simply click on her and insert her. And this whole presentation will look relatively nice, at least in my mind, because we can edit every single slide that we want. We can insert everybody that we want. We can make all of this customized to our content. It all depends on what you need. You can give titles. You can give subscriptions. You can edit stuff. You can insert text. Here on the left side you see You can insert also different text. So different styles, you can insert elements, you can insert stickers, icons, images, videos. Let's just throw this dicktok thing in here. You can make it bigger or smaller or whatever you like. This right here is like a **** slide from God. You can also go back, of course, let's just assume that you don't like this at all. Then you can always see for yourself if you find other templates or maybe other avatars. Let's just assume you want to make something completely new. Let's just see if we find someone here. I think we should use Elliott AI marketing. I think this guy looks awesome. Edit avatar and we can create with him. Now this thing is inserted and we can also edit him a little bit if you want him square or maybe in a circle. You can edit every single thing how you want it. The speaker is Sarah, right here. I think this is okay. So we save this. Now we have Elliott saved, but the speaker is Sarah, of course. I think this should be funny. We press on him and make create landscape. I want to have the 16 by nine, then we get back in our studios. And here we can insert every single thing that we like. We can go to script. Here you can see the script. The script is welcome to the new era of video creation with Hagan. Simply type your script to get started of this will be set by Sarah, even if this Elliott guy is a male. But you can of course, delete all of this and make it how you want. Let's just listen to him or to her, however you want to call this. Welcome to the new era of video creation. I think this is funny. Come on Sarah, as Elliott is really good, but of course you can always delete this. You can, for example, delete the text right here and insert the text that you like. Of course, simply delete it. You can also use like not Sarah, and let a male speak. You can customize all of this however you want to customize these things. If we go back one more time, you see you have a lot of different speakers. And the cool part is you can also upload your own photos if you want. You can make everybody speak that you want. You can even use like, for example, pictures out of a diffusion model, out of mid journey stable diffusion, firefly or something else. And we have also I scripts. The AI scripts are really nice because they are enormous easy to use here. Works sketchy between the background, we type in a topic Telestar Hagen aven solte. This is completely German erstellen script. Um Hagen vendenzolte. This simply means make a script. Why I should use Hagen? And of course this is in German, but you see this LLMs, they can of course, use every single language that you want. We can simply use that language, so it's of course German. What's the tone? Should it be professional? Just okay, and press Generate. And then we get a complete German text and with seven sections. And you can use every single section, You can simply press Create Video. And then you have every single thing also in German. Even if you don't speak any word in German, If you press Create video, we will have every single thing here inserted. And the speaker speaks, of course, German. Right now, you don't have to speak German for yourself because like this tool can make all of this in German. You can always edit it how you like. You can insert, for example, like the speech view or whatever this is. Yes, I think a beach view. Let's just make it a bit bigger if you want to insert this right here maybe. And also like a Whatsapp sign or something else like this businessman, this German businessman is always a viable at Whatsapp, even if he makes vacation. And he will tell you that he is always a viable, like the Germans here. And of course, you can press on instant avatar and you can create such things like the gen coded. But in order to do that, you need, of course, to press create instant avatar. And this is a paid feature. This is the only downside. I personally, I don't pay for this feature. I had this feature, but I don't see myself using this feature a lot. But I need to show you this because this is the best AI feature. When we talk about AI avatars, Jen can create all of it and you should know about Ag. And, and you can test it completely for free if you want to make a lot of stuff inside of Ag. And of course you need to pay, by the way. Later in the course, I will also show you a free cop notebook where you can also clone yourself. Where you can insert different voices into different bodies, into different bodies. All of this can also work completely for free. But of course if you want to create such an AI avatar with yourself, you can totally do this. You press Create instant avatar. Then you need to upload at least 30 seconds of voice and video from yourself. You press Create, and then you get something that looks exactly like the Hagen CEO or like Edward and Blake and Matthew and Lea. And you can simply make yourself into such an AI avatar and you can make text. And all of this will look like you would speak, Yes, this works relatively nice. And after that, you have your own AI avatar here on the left if you really want to use this. In this video, we talked about Hagen because Hagen is the best tool if we talk about AI avatars hands down, the best. But it's not cheap. That's the only downside. But you can test it for free. And I think you should totally test it if you want to create content for social media. If you need to make like videos for your employees, if you want to make like something funny, maybe you want to tell your wife or your man happy birthday. In another language, you can do all of it inside of Hagen. You can test it for free. But like I said, it gets a little bit expensive if you want to do a lot more with this tool. So have fun trying this out. 17. Overview of the Most Important AI Avatar Tools: D ID, Synthesia, Elai: We have a lot of different AI avatar tools out on the marketplace. We can use most of them in almost every single tool. You can make something for free, at least a starting trial. You can in most tools like five to ten generations completely for free. In this video, I want to give you a quick overview. Did you see, first of all, of course, what's on the market? And second what you can and probably should test. I want to show you the three biggest tools right now. Did synthesia and maybe also Eli. These three tools are right now after Hagen, at least the best tools, if you talk about I avatars. The first tool that I want to show you is called Synthesia. At Synthesia, you can make similar stuff then in Hagen, but the only downside is you can't clone yourself. You are limited to the original AI avatars. Here I want to show you their quick introduction. Hey, I'm Alex. One of over 160 AI avatars available in Synthesia. In a few clicks, you can create a free video just like this one. You see this is like the quick introduction. It looks relatively nice. Everything you have to do is to press create an account. Then of course, it depends what you want to do. You can make a starter kit for $20 a month. The creator kit for nearly $60 a month, And an enterprise plan. But of course, you can test every single thing for free. At least if you start out here on this web page. In order to do that, you can go on to features and of course the use cases, They are similar to Hagen Learning and Development, Sales enablement, marketing, information technology, customer service, service or enterprise. If you press on all features, you will see what we can create inside of Synthesia. Right now, my credits are gone for Synthesia, but I want to show you a video that I have created a few days ago. Kinzgndoibstst videos are heightened. You can see this comes from Synthesia. The generations are completely the same as in Hagen. You simply type in the text that you like and you can use their stuff. The next tool is DID, DID. I think it works a little bit better. You simply go onto ID. And here you can see every single thing. Also here you can start a retrial. And here you can also animate like your own AI avatars. You can also make a picture like in michen or from another diffusion model, and you can animate it. Here you have a nice little small introduction video so that you can see how the stuff from DID looks. Creative reality studio saves businesses money, time and hassle, and makes videos and presentations more engaging. Making videos has always been complicated. But with creative reality studio, it's simple. And anybody, anybody, anybody, can be a presenter. Choose from one of 270 voices across 119 languages. Just upload an image, add your text, and in moments you'll have a video ready to download or share. Start your free trial and superpower your training content, presentations, sales, marketing and more. That's basically, you saw every single thing that you can do inside of ID. In this small little presentation, everything you have to do is to simply press, Start free trial. You can type in your text just like you saw in the presentation and then you have your AI avatar. Yes, I have to tell you, all of this works in Hagen. A little bit better, at least right now. Welcome to the employee assessment training. My name is Ava, your AI trainer. The last thing is li at ALI, you can do like the same stuff as in synthesia. And the DID, if we press play on this button, you'll see in a quick glance what this is all about. Welcome to L. Have you ever imagined being able to create amazing videos with real people? Now, with the help of AI, you can create engaging videos with the real presenters in minutes just from text. Say goodbye to costly studios equipment, and real actors. Just follow these three steps. Choose a presenter from the library, add some texts and visuals for slides, and here we go, Hey, my name is Cody and I am your digital video presenter. You can also create videos by simply leaving a link to your blog or website using our generate from text feature. And here is your result. Videos that don't get many views or engagement need to be promoted in different ways, such as through social network postings or e mail campaigns. Moreover, you can create your personal avatar and use it in hundreds of videos in more than 65 languages. This is how your CEO can speak multiple languages. I can now speak Spanish, binaries, and can say hello to all my French customers interested. Sign up and try it for free. Now, bye bye from Eli. This is how I sound in English. What do you think you see? Also, Li or li? Maybe I have pronounced it wrong previously. Can do such videos. These AI avatar tools are here and I think they are here to stay. I don't want to bore you, that's why I just took you to a brief overlook over all of these tools. They all work like nearly the same. You chose a character or your upload a character yourself, then you type in text and you get your nice little AI avatar. And you can always make slides. I have showed you every single thing in Hagen. I think Hagen is the real deal. I think Hagen is the best tool at least right now. And all other tools are behind Hagen and they work completely the same. That's why I don't want to steal your time. So just try one of these tools out and I think you can create standing videos. And if you want to be like really, really smart, go into Hagen DID Synthesia and Eli and make all your three generations. And you have like 20 or 30 videos completely for free. Just make an account. Type in your text and you're ready to rock. 18. Case Studies: Why You Should Use AI Avatars: Maybe you ask yourself, why I show you this AI avatar tools. Let me tell you. They are not just a joke. And the biggest companies on Earth, they use these AI avatar tools and they save money with these. Yes, you can build also a social media brand with these tools. You can create videos for yourself. Or if you have a business, they can help you scale faster and save some money. Let's just look at some of the case studies from Synthesia. And of course all of this is for all these AI avatar tools, ID, Hagen, and so on. But Synthesia has nice case studies. They have over 1,000 reviews, 4.7 out of five stars in the case, that is delucoson BS H develops 70% more efficient trainings. If we press on this link, you see the following. If we scroll down, you can see it right here, that BSH is the largest manufacturer of home applications in Europe with over 60,000 employees. One of the leading companies in the industry worldwide. If you scroll down a bit, you can see how all of this works. The challenge science SH, is a global company with knowledge spread across the world. Sabina and her team wanted to ensure that everyone in the organization had access to that knowledge. However, it was important that this was done in an efficient way to make expertise a Bible. Regardless of time and space, dedication of learning was an obvious choice. The team at BH didn't want to rely on boring PDFs or slides to click through. And then they took Synthesia. You can see it right here. They wanted to make videos. Normally, videos cost a lot to make you are not flexible, and the translation is, of course, really hard to scale. Then the solution was simply Synthesia. She created a script, then she inserted it into Synthesia, chose a navatar and basically added some slides, and that's it. If you scroll just down a little bit, they have reached over 30,000 views with the trainings. They have over 30% increased engagement in this E learning. They saved 70% in production cost. So this is really awesome. I think all of these companies will use such tools. And of course, there are like a lot more here. We have Xerox, we have Zoom, we have also U, we have Latam, we have Sim Corp. Let's just take a look at Sim Corp, for example. They have over 3,000 people and they are in the investment space here. You can always read the challenge resolution and so on. So they simply all save money with these tools. They make stuff five x faster. And that's of course, enormous, even zoom. I think everybody knows zoom, or at least most of the people know zoom. Zoom is a big company, it's even publicly traded. All of these companies, they have saved time like up to 90% You see this is really, really cool. That's why I also show you the AI avatar tools. Maybe you have a Youtube channel or you want to start a Youtube channel and you want to make education on your Youtube channel. These AI avatar tools can really help you if you have a business. They can also help you. They can help you in a lot of ways. You need to be just a little bit creative and I am relatively sure that you can create some videos that will help you out. Use Hagen use. Did use synthesia. Use Eli. I think most of these tools work fine. 19. Recap of Section 3 and a bit of Homework: Over the last videos, you have saw the best AI avatar tools. We started with Hagen because Hagen is awesome. I gave you detailed instructions how to use Hagen and then we took a brave look in DID synthesia and also in Eli because all of these tools work. Similarly, we just took an overview. You have also seen that big companies save a lot of money and so should you. And maybe you ask yourself for what you should use these tools, you can use them to boost your social media reach. You can use them for SEO. You can use them to make like product descriptions. Maybe you want to sell a watch. Just use such an AI avatar tool to describe your product. You need to be a little bit cautious. You need to think for a moment, but I am relatively sure that you will find the right place for these tools. Just think for a moment, Where could the video help? Maybe the AI avatar tools can help you generate your perfect video. Maybe you also understand what learning is learning is doing. And you get better by doing more. I would really recommend you to just go in one tool and make a nice little video. Do it just for fun. You also remember that sharing is caring and that we can learn better together. Do me a favor, just share this course with one of your friends. If this course helps you out, I am relatively sure that this guy will describe the value that he gets out from this course to you because you told him. So the status of you rises. I think this is a win win win situation for me, for you, for everybody in this community, and especially for the guy that you sent this link. Thank you for that. 20. Section 4: Ai-Video like a Pro: This section will make you in AI video pro because you will learn the most advanced features tools and much, much more how you can create standing AI art or I videos. You will learn what Github is because on Github we will borrow some code. We need this code to run it in Google Cop. You will also learn what Google Cop is because we use Google Cop to run the code to create our videos. As soon as you know what tab in Google Cop is, we start with the coolest parts, because we will also make deep fake. So you learn how to deep fake a voice. You will learn how you can use a to Lip in Google Cop. This is completely for free and you can make lip sync just like in collapse, but like I said, completely for free. Unlimited generations. We will do complete face swaps of our AI videos, of course also in Google Cop and in art because C art is completely awesome. And right now also for free to use, all of these tools work with stable difusion. So you know, all of this is open source. Then we will take a look at two of the coolest Cop notebooks, the Forum Diffusion and stable P Fusion. Also the two tools, the two extensions work with stable diffusion and you can make amazing animations. With these tools, you can change your videos so you can overlay it with prompts and you can do enormous cool stuff. You can even download your own Lauras and make enormous videos. All of this will be awesome. I think you will even learn how to create your own AI music video. And of course we will also take a small look at Kyber because also Kyber is, first of all, it's an alternative. It's easy to use but you need to pay. But I think Kyber is also cool. You learn some valuable stuff in this section. I hope you will look through this. 21. Introduction to Github: The next web page I want to show you is Github. And Github is awesome at Github. There are a lot of developers, it's something like a community. A lot of developers share their code, so Github is just like a library. We can go and borrow code from good programmers. I myself, I'm not a good programmer and for that reason we can also borrow some code from Github. There are a lot of cool people and they are doing amazing stuff. They put together entire Colab notebooks and we can work with them totally for free. More on Google Cop of course later. But in this lecture we will make an account on Github, because you will need this later. Github is an absolute amazing web page typing Github in Google and press on the first link. Most likely you will be on a web page that looks something like this. And of course you need to sign up with your e mail address or you can also sign up with Google. Please just click sign up. Then you need to follow all the instructions after you have created your whole account. The web page will look something like this here on Github. We can do a lot of different stuff. Just take a look at the whole web page in the left corner, you press on this button, You can go to Home Issue, pull, request projects, and so on. If you simply press on home, this is the web page where you are in the right corner. You can also search for specific projects if you like. You can create new projects on this button if you have some issues. This is the button for you pull requests, and here are the notifications if you have something new. If you click on this Explorer button, the web page will look something like this. And here you can search for the newest, hottest topics on Explorer. There are always some cool videos and much, much more. You can go on trending and find the trending new stuff from all the programmers on this web page. So for example, movie, web, and much, much more. You can search for the collections for topics. Just take a quick look at this web page. And if you are a little bit overwhelmed right now, don't worry. Because this is just the introduction to Github. We will use Github later and just to search code, so you don't need to worry to do anything fancy on these web pages. We simply search for the right cop notebooks later in the course. But I think right now is the right time to introduce you to Github, because this is a lot easier than it looks. We just need to try to search the code that we will use inside of Google Cop, because we also talk about Google Cop. I will show you what Google Cop exactly is. In the next video, we will combine Github and Google Cop. We simply search for Cop notebooks on Github and we can run this completely for free. That's awesome. See you in the next video. 22. Introduction to Google Colab: In this video, I want to introduce you to Google Cop. I already told you quickly what Google Cop is. Google Cop is simply the cloud solution from Google. They borrow us some GPU and CPU power, or even TPU power, to make it easy. They simply borrow us some power that we can run code, we can run different code so we can run code to make pictures. The pictures need a lot of power or a lot of GPU to render. And all of this code we can have from Google Co. If you ever wonder what code you should run in Google Cop, of course, you can either borrow your code from Github, you already know what Github is, or you can make code yourself in chat GPD. Give me the code for Snake for example. You can of course write code in chat PT. Like no, this is not a complete course on chat ChpT. I just want to give you a quick overview here. You get code from chat PT. You don't need to understand all of this code, but you can totally make code inside of chat chPT and we can run this in Google Cop. Like I said, this course will not be about installing Python or running snake inside of Google Cop or playing games with Google Cop. This is just an example because later we just borrow all of this code from Github, because on Github are such nice Cop notebooks already completely done so that you don't need to know all of this. I just want to show you what Google Cop is in this nice little lecture, so don't worry about this code. This is just an example. Now let's just take a look what Google Cop is in detail. You can simply Google Google Cop and then press on the first link. And something like this will open up. I use a lot of different Google Cop notebooks and we will take a closer look at Ats of Google Cop. But first you can see what Google Cop is exactly. For that, we simply press new notebook. Yes, this is German, but don't worry, Simply press new notebook, and then you are in an empty notebook. There we are. And here you can see you have basically nothing but in this environment you can run everything. Here works. For example, Biton code and much more. Everything you have to do is to simply include the code into the small cells and then press play. Right here we go back into Chachi PD because we already have code. We have code for Snake right now. We also need to install this Pi game. We copy the code. We go back into our Google Colab notebook. Here I include the code, and then I can simply press run on this cell. And we will run this cell, and everything will get installed into this small Google Colab notebook. So first you see we have here some Rams. And we need, of course, to connect to our Rams Right now it's connected, and requirement already satisfied. So Pi game in and so on. So everything is installed. Right now, we have installed our Pi game. We use the GPU from Google to do this. You can see right now, this is completely for free. Yes, you can update this if you want, but for now we work with the free stuff here. The second thing we can do is to simply press code in the left corner, and then we get our second cell. In the second cell, we can execute our code. First, we have installed some stuff. I click copy code, We go back into our notebook, and then we include our code right here. We can simply press play on this button and the code will get executed. And now the code did its job and we only get a linch. Why we only get a linch? Because in Google Cop you can absolutely run code. But you can't make this because you need a desktop variation of these. You can go on this link and you can see everything yourself right here, if you want to make a desktop app out of these. I just wanted to show you how to run code in Google Cop. If you want to have code that is working in Google Cop, of course, we can also do that. We can make, for example, another code, give me the code for guess the number, and there we have the code for guess the number. And this can absolutely run in Google Cop. We go back into our Google Cop notebook, we can simply delete all of this code. I think we have also no problem if we use these cells right here or we use a completely new notebook. Now for guess the number, we don't have to install anything into this Colab notebook. We can simply include this code right here and press Play. And then the code will get executed, and then we are ready to play. Of course, we need to connect our Ram first. And here we have our little game. So guess the number 1-100 Let's just do it Strategic. I use 50, okay, This is too high. So we do it. Strategic Again, we do 25 too high. 13 too high. Again, seven too low. It's 7-30 And now we just need to try. And there we have it. Congratulations. You guessed the right number. We have made our small game in Python, and we let it run in Google Cop. In this video, you have learned what Google Cop is. Google Cop is simply a cloud solution from Google. They borrow us some Apu power and we can run code in this environment. This is especially cool because we can also borrow code from other people. You have seen how to run this, so we can include code and then press Play. And then we can make new cells include other code, and press Play again. And we can do this over and over again. This right here, for example, is the Google Colab Notebook. That's called the Forum Stable Diffusion. With this cool notebook, we can make pictures first. This seems a bit overwhelming, but don't worry, because we will do this step by step. This is easier than it looks because we can't just borrow all of this, so we don't have to type in a number. We can simply go into this notebook and fill out everything, and we get this notebook from Github. And that is the synergy of all of these. So you can go to Github and simply search for Google Cop notebooks. If you press on these and you search, for example, for devorum Google Cop, you will find a nice little Google Cop notebook, and you can press on the link, and then we can run the code in a Google Cop notebook. Here we are again. This is also stable diffusion. The Devorum version, we can simply search for code on Github and we can find dire Google colab notebooks to run the right code. Again, in this video, you have learned what Google Cop Notebook is. It's from Google. You can borrow GPU and you can run code. The coolest stuff is that we can also borrow code from Github and let it run. We can simply press play on this button and we don't have to write a single line of code. That's the bower of Google Cop in combination with Github. That's the synergy and that's why I showed you both of these. 23. Stable Diffusion [Wrapfusion] in Google Colab: Stable rape fusion. That's also a cool tool that I will show you how to use in this course, step one. Of course, you know it, you need a file. We do basically the same stuff as soon as the frame comes around where we want to transform. I cut the video and then I extract the part where I want to be transformed afterwards. Of course, I go one step back, because we need that later. Now I have the file. I put it simply into a folder because we need that file later. The next step is we need to go to Patrim. We work in a Cop notebook, But the Cop notebook we will get only access if we subscribe to someone on Patri. He give always the newest versions and he gives tips and so on if you have any problems. There's also a discord server for me. I think this is really nice. Here I am into patrian and the name of the guy is X L E L A, or most likely L x S. If you read it from behind, he is really active. He gives some updates every few days. And right now we are in version 0.20 That's only a preview. We need to be a power member. If we want to have this, I'm not a power member, I just spend that $1 But we can have stable B fusion version 19.8 You need to download your version that you like. Sometimes the newest versions don't work that well. A version that I always liked using was the version 0.16 You can also simply press on this file. If we press on this file, you can download it on your computer. The next step is to click on the Cop notebook. I will link you everything. This right here is a normal notebook. You need to go to uploads and you need to find your file. Right now I use version 0.16 because this works always, really, really nice. We upload here our version and then we are ready to go. Now this cop notebook, maybe looks a bit overwhelming because it is really, really big and there's code included and so on. But this is really easy to use. We don't have to do a lot. First step, I like to go to run time. We need to make sure that we use GPU and I like to use a stronger GPU. So change run time. The version 100 U works fine. Also the 100 U works fine, this works for me. If you use four U, this will take like forever. Then we scroll down to set up. We need to press play on the first cell. Of course, we need to give permission because this will get access to our Google Drive. We give permission right here. Then we need to scroll down a bit. We need to go to basic settings. What's the batch name? We just call this test. But we've do these because I think I already have a test then. What's the width and the height? Simply the resolution. My video is in 169. This resolution works fine for me. Now we need to go to Video Input Settings, and of course, we need to give our path. We click right here as always. Then we can upload our video right here, or we can even upload it in our drive. But this right here works fine for me because I just want to make this video. I don't want to have this in my drive. This is totally fine for me if it gets deleted afterwards. As soon as this is uploaded, I called it wrap me, we press the three dots. We press copy path. We delete this path right here and we put our path here in this field. This field is the next important field, extract frame. If we put one right here, every frame will get reworked. Every frame will get extracted and reworked with our prompts afterwards. If you put two right here, the video will rendered in the half amount of the time, but I want to have every frame rendered in this video. Then we scroll down if you like, you can store every frame in your Google Drive, but you don't have to do that. But I'd like to do this if we want to work in Blender afterwards. For example, after that you can scroll down even further until you get to this code here. It's also really easy. Everything you have to do is instead of CPU, we use CPU in this field. He is Model path and control net models. These are simply the default models that we have right here. I want to make the short and easy. So we use the default models, but you can also go to Cvti.com if you need to have a specific look. You can download models for free right here on CVI. Upload them into your Google Drive, and then you can give the path to the model that you uploaded. But like I said, I want to make the short and easy. We use the default models, that's basically it. If you scroll down a bit to non GUY, it comes right here. You can also delete the prompt and use different prompts. But right now I want to use the default prompt because we can change the prompt afterwards. Everything we need to do right now is simply run and run. All now the Cop Notebook will do its stuff, but this takes a bit of time. If you scroll up to the top, you will see that every cell will get automatically activated. The Colab Notebook does simply stuff. It installs different stuff, it downloads stuff. This takes a while. This took about 5 minutes. If we scroll down, you can see that the program is doing his stuff right here. So it extracted some frames. And then at some point, you will see a preview under the cell. Do the run. This is the cell that is nearly the last cell in this Collab Notebook. If you like this output, perfect. If you don't like this output, you can press Stop right here. Then you can scroll up until you see here the prompting box. Until you can put whatever you like in the prompting box. You simply include the things that you want to see in this prompt. In the negative prompt, you want to include what you don't want to see. Text, naked, nude logo, cropped to heads and so on. This works all really, really good. I think this cyber bunk look, I really like this. I want to transform in such a cyberpunk. The only thing that I want to change is I don't want to be a woman, I want to be a man. Remember, you can always copy prompt. You can ask Chachi BD for stable diffusion prompts and so on. So now we press play again. Down here we have another prompt. Now the only thing that we need to do is press Play again, right here. And then we should transform into a Cyberpunk man, instead a cyberpunk woman. And perfect. Now this works really, really fine for me. As you can see, now we turn into the cyberpunk man. You can also see this takes a lot of time. I think this will take maybe up to an hour because this is not accurate. Sometimes you can really wait a lot. These are right now, 142 frames, just from experience. Sometimes you need to wait for an hour, you need to drink a coffee. You can see right here all the images they will get stored into my drive content drive, my drive stable. B, Fusion images out, I go into my drive, now I go into AI stable be fusion images out the test order. Here you can see the frames come. Now we have now our third or fourth frame. And the frames will be collected in this folder. But we can also make a video as soon as all of this is done. I see you. As soon as the video is rendered, just make a coffee. And now we have it. This is right now, the last frame. And now we want to create our video from our frames. All the frames are now also collected in our Google Drive. And we can also use the frames and put them in Blender, for example. But I like the easy way I want to create the video inside this cop notebook. That is really easy. All you need to put right here is the right frames per second. Remember, if you have a video with audio, it's important that you use the same frames per second in this video right now. Because if you use a higher or lower number of frames, the audio will not be synchronized. My video has 25 frames, and for that reason, we need to put 25 also right here. Then we are basically ready to render. We click Create Video right here, and then we go. But I already showed you how you can do that. Simply press play and you will get your video. In this video, I will show it in Blender. We need to go into our drive. We go into our folder where all the frames are. I want to delete the first frame, and then I download all of these frames. These are 141 frames. If we scroll down, I already downloaded them and I saved them into a folder. Remember, you don't have to do that, but I just want to show this in blender because I think maybe it's valuable for some people. Then we go into blender. We need to make sure that we go to the video sequencer and it all should look like this. Then we aim for a resolution that is perfect for us, 1,900.25080 is perfect. The frame rate, I set it to 25 frames, just like I told you, and the output should be a video. Now we press Ad, and we search for our frames. We press image or image sequence. We need to find them. So these are our images. And if we press A, all of them get selected. Simply press A on your keyboard and then press a image strip. Now they are basically all added into your blender. Then we need to make sure that we save them in appropriate way so you can leave them into this folder, or you can create a new folder right here or search for a folder on your desktop. Just do how you like it, Maybe we do it into perhaps accept it. Then we need to adjust our frames right here. Frames start, one is okay, but the end is at 141. And the last thing we want to do is into the left corner press, render and render animation here. You can also see the process, if you like, I will see you as soon as this is rendered. This is normally relatively fast and now it's rendered, everything will get automatically into the folder that we set, the folder wraps. Just remember, you can also do this directly in the Cove notebook. Just press play. But I think for some people it's may be cool also to know blender. The next thing is really easy, we go back into shortcut. We include our video right here, and then we just throw it on our time line, directly over our video. And it's basically done. You can see, as soon as I clap right here, the next frame steps in, and I will get into this cool animated stuff. Let's take one less look at this. I think this is really good. Perfect. Now we want to export this, and basically it's done. In this video, we took a look at stable web Fusion. It's a Cop notebook behind the Colab Notebook Works Stable diffusion, it's relatively easy to use. Just use a video clip where you want to overlay your prompts. As soon as you have found your video clip, you need to go into the Colab notebook, but you only get access if you download it from Axle on Patrion. As soon as the notebook is opened, just connect it, press a few play buttons. If you really like, use different models from CV Ti and then you simply press all, of course give the paths and so on. Just like I told you, this will need a bit of time. Maybe you need something like an hour or so to render these. After it's rendered, you can make a video of it directly into the Cop notebook. But in this video I wanted to show you that it's also possible in Blender, you can download all your frames and you can make your frames into a video also in blender. Then we just throw it into shortcut. And it's basically done relatively easy, but you need to do a few steps. You need to go to Patrian. It's a bit of work, but thanks to this tutorial, it's easy. And remember, as soon as it starts to render your first frame, you can always post this and adjust your prompts. You can put anything that you like into this prompt. You can also transform, for example, into a turtle or whatever you like. Just try it out. 24. Overview of Deforum Diffusion to make Ai Animation: I have to tell you right now, this is not a complete course on the forum stable diffusion, because the forum is a really big extension on stable diffusion. But I need to include the coolest stuff about the forum stable diffusion. If somebody has more questions, just hit me up. I have also a whole course on the Forum Stable Diffusion, where I go in detail on every single step in this notebook. But like I said, I need to include the coolest stuff inside of this course. I think if you just try a little bit with different parameters, you are also able to create standing AI videos. Even if you just look the next few videos. Now, what is the forum? You already know what stable diffusion is. Stable diffusion is the open source model that will create pictures. For us, the forum is similar but different. The forum is simply an extension that runs on stable difusion with the forum. We can create a lot of different frames, and if we use a lot of different frames step by step, one after another, we can create whole videos with stable difusion. I want to show you how we can create music videos with Devorum difusion, and also how we can create animations with Devorum Difusion. I have to admit that this is not so easy to use and you need a little bit of time to render all of these, but it's completely awesome. So in this video, I will give you a quick overview of this Colab notebook. Then we will create a small music video, and after that we will make a small three D animation. Like I said, you need to try with these a little bit. I would assume that this gets also easier to use in the future. Let's just see what we got right here. We have a cop notebook, and of course all of this works completely the same. You need to make sure that you connect to the right run time. It would be if you are in a three cop notebook, as always, the run time t four, just use the three version and connect with it. As soon as you have chosen the right run time, you need to press again on run time and run all. Now every single cell will get executed right here. This notebook will ask you a few times if you want to give permission to use the code from Github, and if you want to connect your Google Drive, you need to give permission. If you give permission, a folder that it's called, AI, will be automatically created in your Google Drive. If you press on this folder, there are all the models and all the stuff. All of this will get downloaded completely automatic here in this nice little folder. You can also upload stuff yourself. You can upload, for example, screenshots here, for example, a few screenshots. I use the screenshots to start my animations because later we need to give a path to our screenshots. You need to upload your screenshots right here if you want to make advanced stuff right here in this notebook. If you want to make animations out of your own frame, you always need to give the path to your screenshot. Of course, you will learn later where to give your path exactly. But first before you give your path, you need to upload your screenshot right here in such a folder. That's basically it. That's the standard thing that you need to do before we can use this coded notebook. Now let's take a brief look on the interface because I want to give you a quick overview before we start with our first generation. If we scroll down a small little bit, you see that we use a Nvidia GPU. We use the environment and we have the set up. Now, you know, every time we press play on such stuff, we simply start to install different things. The default settings work really well in this notebook, but you can do a lot more. You can press simply Play step by step in all of these cells. The Model Path is just simply the stuff that you use. Of course, you can also give permission to your Google Drive, you have to give permission to everything right here. The model set up we use Da. This right here is the inferience. We use a model checkpoint that is automatically set up. This protogon model works completely fine. It's easy to use, of course I have to tell you, you can also use other models. In this Colab notebook, I would assume that you should start with the default settings. If you scroll down a little bit here you have the stuff that you can and probably should change the animation mode. If you press on it, you can use 23d Video Input and Interpropolation. If you use two D, you make simply static animations. You animate frame by frame, just normal. If you use three D, you can activate camera movements that go in every single direction. If you use video input, you can make similar stuff as in B Fusion. But I have to admit, B Fusion works better if you use video input. And this right here, I don't use it, this doesn't work that nice. Here you can type in how many frames you want to create. The border, we always use this at replicate. Then we have the motion parameters. This right here is the angle. If you animate the angle, you can rotate your animations. If we use the zoom, we can zoom in or zoom out to our animations translation. X, yz3d is three D rotations. They are also cool because on X and Y Xs we can simply go up or right and left. The translation said, this is the zoom. If we use the three D animation and then we have the rotations. The rotations simply rotate as you can probably think if you use this, don't worry, I will give you a blueprint in the next video. Every other single thing right here. You don't really need to understand this because the default settings work really, really well. Like I told you, this course will get much too big if I include every single thing right here. But if you want more info, just hit me up and I will include more lectures. If somebody needs to learn this in more depth, then we can scroll down even further. Because here we have the prompt here, we need to type in what we want to see. For example, here you see zero double point, and then a prompt. Now if you type in here a prompt, you will get this prompt executed from frame zero until frame ten. And of course, you can also include negative prompts, as you always know in stable difusion. Now you can new stuff right here. You type in the frame number and then simply include the new prompt. For example, do we have from the frames zero to ten a beautiful lake? Then we start with a portrait of a woman, and then we start with a dog. This right here is how this works. Then you go down and we have the load settings. On. The load settings, we can work with the default settings, no need to change stuff right here. Then we have the image settings here. You need to give width and height to your resolution. 512 x 512 is of course a square image. The bit depth output, we simply leave it at eight. We can play with the set if you want the sampler. I always work with oiler. The steps work really fine at 50. You can also go down the scale at 70 is perfect. Then we scroll down even more here. You can give a batch name, stable fund works. So you can also work right here with the default settings. Then we have the set behavior. You can use it, but you can use a fixed set. If you use a fixed seat, you need to change this number. Minus one is always a default seat. You can use, for example, a set. If you use fixed set right here, you should change this number. Then if you scroll down even further, this right here, we don't need it. But this is important if you want to make an animation that starts with your own frame, you need to use in image. If you need to check this, the image strength, this is simply how strong the first frame will be replicated with the second frame. So if you put right here, 0.8 the second frame will be 80% of your first frame. Don't worry, you will get this as soon as we start to animate it. Then here you need to give the path to your input image, to your int image. Out of these, we start our animation here. You can also put in a mask if you like, but we don't need it. The other stuff will leave at the default values. Now if you are overwhelmed, you need to do this step by step. You need to try this a few times. Like I said, I can also add lectures in this course, or you just hit me up. If you need more information about this notebook, then we scroll down even further. Because all of this you don't really need the rest of these. Here you see create videos from frames. Here, your video will be created basically, that's basically the overview of this big, big Cop Notebook. In this video, you saw an overview of the Cop Notebook that can create videos with stable fusion. We can animate every single frame. Like I said, this is a big notebook. This was just the overview. Over the next two or three lectures, you will see how we can use this notebook. Because we can create music videos, three D animations, and two D animations. You need to play a little bit with this notebook, like I said. So have fun trying this stuff out because this is awesome. Like I said, if you need more information, please hit me up. By the way, I think this article will also help you if you go on stable diffusion art.com slash deforum. You have this quick guide. This quick guide is completely awesome and they show you everything that you can do with stable diffusion deforum. These are basically the animations that we can make. With these, you can simply scroll down and they will describe you every single thing. What do they do? For example, the zoom, the angle translation, x translation y much, much more. First, of course, what is the forum? Then they tell you also how to use it. So you can use it of course as a deforum extension on automatic 11, 11 on the stable diffusion webby. Or you can use also other interfaces like Windows Mac or Google Cop. Like I said, we Google Cop because this is the easiest tool and we don't need any cheap U on our computer. They simply tell you every single thing how this works. And if you scroll down, they tell you also how the parameters work. Here you see this is the stuff that we can create. This is, for example, two animation. This is relatively easy to create, but of course, if you want to make music videos and three animations, this gets a bit more advanced. Here you see the two demotion settings always with an example. For example, the zoom. If you put in values like this, you get a negative zoom. And if you want to zoom in, you need to put in values like that. Then of course you have the angle. The angle rotates, translation X, translation y. As you can see here, the perspective flip. So you can see we can do really amazing and wild stuff in the notebooks. Then the three D motion settings. This is the most advanced stuff, but this is really, really cool because we can rotate with our camera in every single direction however we want, we can combine all of these. Rotate, zoom out, and much more. Here you see the three D rotation x 3y3d. I think this quick guide will totally help you and show you nearly every single thing that is important for this notebook, of course. Also the next few videos will show you exactly how you can create this stuff. You can see also in this video how we can zoom in and transform and much more. The notebooks are completely awesome because this AI animation is stuff that not everybody is able to do, but you will be able after the next few videos then hey, they have some tips and much, much more. So you see this is a really, really cool guide on this super cool notebook on the extension of stable diffusion that's named De Forum. 25. Mkae AI Music Videos with Deforum Diffusion: Suicide or audio to keyframe string generator. This thing blew my mind. As I explored it the first time, there is the possibility to use a specific music or a specific soundtrack and to extract the key frames from specific parts of this music. For example, just download the music, which of course you will need to have the permission to use. You need to use copy free music, maybe from the Youtube audio for example. You can download it and you can upload it in a tool where you extract a specific part of that song. For example, I already downloaded a song and you can go to I, lala, I is my favorite tool for extracting these things. And it works really easy. But you have to make a subscription if you want to do more. Here there is a possibility to test this thing. And I think you can test it for about 10 minutes. After 10 minutes, you need to make a subscription. If you don't want to make any subscriptions at all, you can use phonic mind. On phonic mind, it's basically the same thing. You just upload your music and then you can extract what you want to hear. Let's just make an example here on Lala AI, but it works in every tool nearly the same. You can also run local things on your computer. The first thing you need to do is of course upload your song. So we simply press Select New File. We can drag and drop our file right here. As soon as this is uploaded, it starts. Then you need to select what you want to extract. And in most cases, the drums work best. Let's extract the drums for this specific example, and we make it only 30 seconds so that it is not too big. As soon as everything is done, we need to go to our audio decree frame string generator here, of course, we need to upload our file. We can drag and drop our files of only the drums right here in it. And you see it, we have 34 seconds of our song, and we have here our key frames. All key frames are here listed. The first thing you need to check is the frames per second. You need to know how many frames you want to create. For example, if you take 24 frames per second right here, you will get an animation that has 800 frames. If you take 12 frames per second, your animation is of course, shorter. It is only 408 frames because the soundtrack is 34 seconds. And if you time this 34 by 12, you get the 408 frames. If you time it per 24, of course you got a lot more frames. The next thing right here, that's the function we work in our example with the zoom. And the default settings work perfectly well for the zoom. But what does this formula mean? Let's just make a quick example. We have one plus x to the power of four. What does this mean? For example, if we go into the forum, and let's take the zoom for an example. The zoom is a multiplier. Of course, this right here would be a static zoom. No zoom at all. And you can multiply it by this value. And you will zoom slowly in if you work on other things, like maybe the y axis or the X axis. But I like the zoom the most you need to change this. You could change this to only x, to the power of four, because x is always the rump and to the power of four. But because zoom is our multiplier, we need to work with this one plus x to the power of four. We simply multiply this, and every time the Ram hits, we will get a zoom. You can see it right here. For example, at frame 403, we don't have any zoom. 404, we don't have any zoom, but at 405 we have a zoom that will be exactly when the ramp hits. If we work with our zoom, of course, you can also change this x to the power of two. And you may be think that x to the power of two makes smaller adjustments in the zoom. But that is not right in this formula. Smaller is bigger. And I tried to explain to you why in our example, x is smaller than one. If you multiply x with something and x is smaller than one, a higher value to multiply means a lower number. Let's just make a quick example. So that right here is our example. If the drum doesn't hit at all, we have zero. If the drum hits really strong, we have 0.3 If the drum hits just a little bit, we have 0.1 And our formula will go like this, one plus 0.3 to the power of four equals 1.081 And if you take the same formula, but with a lower multiplicator, we get the following one plus 0.3 to the power of two equals to 1.09 As you can see, the value of the zoom is higher if the multiplicators or this thing right here is lower. So you just need to remember, if you take the power of two, your output will be stronger than to the power of four because we work with values that are smaller than one. That was basically the explanation. Let's assume we want to have a strong hit. So we work with the power of two. Because we work in the zoom, we can leave the rest the same, just send it up. And you can see it also right here. For example, on frame 405, we have a zoom of 1.15 If we take it to the power of four, our frame 405 is 1.02 so a smaller value. If we take to the power of three, of course it will get right in between. I think I want to work with the power of three in this basic example, because I think that zoom makes the most sense for me. And now we are ready. We have uploaded our zone. We choose a frame rate of 12 frames per second. We choose this formula because we work in the zoom. The zoom is a multiplier and we took it to the power of four. We have our 408 frames and we simply press copy string. Now we go into our collapse. We put it right here into the zoom, and you can see we are basically ready to go. I want to work with a two D animation right here. So we simply choose two D. We take our frames. Our max frames are 408 frames, as you know, border replicate works good for me, the angle we put at zero. And of course we put also the other translations at zero because I don't want to have any movement at all for the translation x and the translation y. Let's make something funny. Right here, we want to rotate to the whole video. But the zoom hits only when also the drum hits, the other Xs will get ignored because we use to the animation noise and strength is right with this. The next thing we need are the prompts. I want to simply delete this prompt or this activated. And I want to have the same prompt for every frame of our animation. Just the rump should animate this thing. Our prompt is an epic battle ground city, basically in a cyber bank style. Of course, I just copy this from Lexica. Our image settings are fine because in the course I think this resolution works fine. Scale is okay. The steps are okay. We don't use, of course, any It image, so we don't need. The next thing that is important to me is this right here, the frames per second for our video. We need to choose the same rate for frames that we had in our key frame generator. If you take here 12, you also need to make the animation with 12 frames, or it doesn't work at all. If you just take one frame more right here, the whole animation will not work. It will work, but it is not at the exact same time as the rump hits. So now the only thing we have to do is to render all the max frames at 500. Right here work fine for me because the other frames don't get animated run all and we are ready to go. Everything worked well. And I just downloaded the file and uploaded it into shortcut. I took also the audiophile, and let's just see what we get. I want to play the whole song right now and really nice. Our drums and our animation are perfect. It works really, really good. And it was relatively simple to create. We just took our audiophile, we traded the drums, then we had to upload it into our key frame string generator. We need to choose the right parameters for the zoom. It's a multiplier, and we need to take one plus x to the power of 23 or four. It depends a bit how strong you want to have your zoom. You need to take a lower number, If you want to have a higher value at the zoom, we need to take the right frame rate. And then we are ready to go, just download your animation, upload it into a video editor, and just synchronize it with your audio file. And you can see it also right here, the audiophile is exactly the same length as our animation. So it works really, really well. No need to tweak this thing because I am happy. I'm sure you can also make really good animations that are synchronized with your audio files. So good luck with that. Super. 26. Create 3D Animation in Deforum Stable Diffusion: We've seen a lot of different stuff. Let's just use this stuff, let's just use this knowledge to create a full project. I have something in mind and I already started something. Just take a look with me. I sit here in my kitchen and I just try to drink a coffee. No, normally I don't wear glasses. That is just for the video. I nearly didn't see anything with the glasses. I put on my glasses. I drink the coffee, I see something in the papers and I do write this, this right here, That is our last frame. I want to start our animation from this frame on. Maybe I could include something right here. Something like AI is going to take ba, the water or something, and as soon as I see it, I spit out my coffee. From here on, we want to start my animation. We take a screenshot of this picture and we'll load it up in Google in our F. You already know how to do that, so let's go in Google Colab and we want to make our animation. Our animation should be in three D. I think the right frames are something like 300. I think that is more than enough for a video like this. And maybe I even can take this video for the intro or promo for this course. We will see how this goes. The next thing we want to animate is all of this angle. Of course, I put a zero because we don't want to rotate at all. The zoom, we don't need to zoom. You already know if we want to zoom in three D, we need to animate that translation. Set the first xs that I want to animate is the translation x. And now we need to think what we want to see. You know, we start with our key frame like this. We need to start to animate our prompts. I think I want to get into a zombias always and maybe after that we tweak our prompts. We need to take a look what we want to see, the first thing is of course, our axis. How they should behave. I think I want to animate my axis like this. This right here is our object. I want to start zooming translation x and translation y does nothing but we start to zoom in. And then the three D rotations rotate up to the right. I think we should change prompts as soon as we start to rotate up. I think that should look relatively nice. And I want to take keys, string generators, to make my animations faster. Just for the C of this tutorial. I will also animate the x x, but I think that wouldn't be necessary in our example. But just let's do it. We can always tweak if we don't like it. So the first thing we want to animate is the x x. I need my 300 frames from minus ten until plus ten, and let's go. But the x x, I think they should be relatively even until frame 150, at 150 starts our magic, and the x x should go to the right. Maybe just a little bit. Something like that. I think that is fine. So just copy string And we put this at our x axis translation y. I don't need that at all here. I want to start exactly at zero. The translation said, that's something we need. The next thing I want to animate is the zoom. I think I want to have a zoom that starts really slowly and increases a bit. We start at frame zero. We go, I think, until frame 150. After frame 150, I want to increase the zoom drastically. Maybe 23, and it should happen until frame 160. After that, the zoom should slowly be much slower until the end of our animation. That right here is our zoom. Copy the string, and under translation set. The next thing are the three D rotations. I want to rotate to the right at frame 150, we go back right here. This time we need lower values minus three until plus three, because, you know, this axis behaves much, much more. I want to also have, of course, the 300 frames, we start at zero, and we should go maybe until frame 150. With zero rotation right here. But after that we start to go upwards and maybe until a value of 1.55 in. That is fair. Or let's just play with one right here. Copy string. And at our x axis, same thing I want to do with our y axis. And now the only thing that we need is our rotation. I think I want to have a rotation. Or we can even let this deal and maybe re run it if we think we need a rotation afterwards. And the next thing we need to take is the noise. I think we can start with a noise of 0.2 But at frame 150, some magic happens. We want to increase this after frame 150 a bit, maybe to 0.04 Then I think that strength is right for us. We could work with HSV in this example. I think that makes sometimes the better outputs. We need, no video input at all. The rest of this is fine. And now let's go to the prompts in what we want to transform. I found a prompt on Leonardo and I want to transform in this. I hope that works. Just copy prompt. We go back into coap, we put it right here at frame zero. We also have a negative prompt, minus, minus neck. We need to copy our negative prompt and we put it right here. That is our positive prompt. And we want to change prompt as soon as we rotate. Remember we start to rotate upwards and to the right at frame 150, and also the zoom starts to be more quick. I want to change prompts maybe at frame 160 or 170, so that maybe new things appear as soon as we zoom out of our frames. So I want to activate this prompt right here, maybe from frame 100, let's say 165 to start with, we need to tweak if it's not good. What we want to see right here in the background of us, in the background, something like should appear, copy prompt at frame 165, we want this writing at my first prompt. I want to add right here a plate girl on a dark throne drinking coffee because I spitted out my coffee, and basically that's it. Load settings are fine, our resolution works well. The seat doesn't really matter, but we have a seat. Oiler, 50 steps. We need to lean a bit heavier into our prompts. Maybe eight prompt settings are fine. We want to name this. Spitting my Coffee. We use the seat. Here it seat I think works also. Well, we need to use an Init, so we activate the in. The strength is 0.9 That is fine. And now we need to connect our Google Drive and give the path to our Init image. We do that as always. We click right here. We connect our Google Play this cell. Give permission if it doesn't appear. We need to refresh. We go to my drive AI. Now we need to take the right screenshot, start keyframe, copy path. We need to put the path right here and we can go on. What's next? We don't use any mask, we don't use these things at all. But I want to have 24 frames because my camera also shoots with 24 or 25 frames. I think we are ready to go. The next frames down here are to 500. I think all the rest is relatively fine. Yes, the prompts change right here. The only thing that we need to do right now is go render. Simply run all or just run here, the animations After that run the prompts and now run the load settings and we are ready to go. I see you as soon as everything is rendered. And here we have our video. Let's just take a look. I transform. The transformation is good and smooth. We start to zoom. Think we should rotate right now to the right. Also the prompts right here. They are really nice, I think. Now we change prompts and rotate to the right. Yes, change prompts to the right. I think more noise. Should start kicking. Yes, more noise so we can create the stuff faster. I think the whole animation is relatively good. Of course we can and probably should tweak that if we want. Maybe the camera rotations are a bit too harsh at the ending and a bit too slow at the beginning. We could simply tweak that if we like, but I don't want to tweak this. I think that is relatively good. I downloaded it. We uploaded in our shortcut. I tried to put it right here on my video. I think that transformation should also be relatively smooth. Let's just take a look. Now I spit my coffee. And now I should transform. And you see the transformation is really smooth. And then I'm relatively quickly, this lady. Yes, I think the video is relative good. Let's take a look at the whole video first with my stupid glasses, and I can't see anything, but that's okay. Start to spit and now I transform. The lady is really good. The zoom is also okay. It's not too fast. That's relatively good. Also the rotation, I think I like that it's maybe a bit too fast at the end, but I want to simply leave this as it is. We can always tweak it if we want, but I want to leave it as it is. I think that's really, really nice. And you can see, thanks to the increased noise at the ending, we can create the stuff faster. Stable division keeps up with creating our new stuff. We can also make this longer, or animate this further. We can do whatever we like with this. Now we got to a full project. I think you know what you have to do now. Exactly, start your own projects. It's relatively simple. Just make a video and it doesn't matter what video. Start with animation, something like a clap or something like that. And then start your animation. Take your last keyframe. Upload it in your Google. Give the path and use the tweak the prompts. Tweak the camera access. Use keyframe generators and you can and will create relatively good stuff. We did exactly what we want to do. Our object was right in front of us. We zoomed. After that, we rotate upwards and into the right corner. And the zoom was faster and we rotate even faster. We did exactly what we want to do, thanks to this course, you can also do exactly what you want to do. If we want to tweak this, maybe we could decrease the speed a little bit of the last frames. We could also change the prompts a little tiny bit later, and the outputs are really amazing. But I think this output is relatively good. We should probably just overlay a little bit of music over this and we are ready to go. So let's go create your own animation. 27. Kaibe: A solied and easy alternative to Deforum Stable Diffusion and Wrapfusion: Let's talk about Kyber. Kyber is really easy to use and I just want to give you a quick overview here. You see as soon as you go to Kaib, of course you need to make an account and login. Then you can press on Products Gallery, Kyber Studios, my video and about. If we press on products, you can see what you can make in Kyber. You can do audio recivity. And you know we already done this with Devorum Diffusion, I think in the forum, this works a bit better. And of course Devorum is free. But you can totally just press try now and you will make awesome animations also with Kyber by the way, linking bark mate the whole music video with Kyber with this tool. If you scroll down a little bit more, you see that you can also make animations where your words take shape and still images come to life. Of course, you can simply animate pictures. If you scroll down just a little bit more, you can also transform. This is basically just like the to the table divorum diffusion. This is basically the animation with devorum diffusion and with the combination of the music just like we saw this right here. This is basically just like stable red fusion. Change the look of your videos with just a few clicks. You can always press and try now and make all of this happen. Like I said, all of this is nice in Kyber, but you have to pay relatively fast. And now I want to show you some stuff that I have made with Kyber because we can test these completely for free. But in my mind, it's better to learn to use stable diffusion because open source and free is always nice to have. If you go to Kyber, you can go to the gallery and see what other people did with Kyber. But I want to show you what I did with Kyber. So we go to my videos. I've simply created two videos. Because I must admit it, that was for free. I don't want to pay for this because the control is not that good, the interface is really easy to use. You simply upload a picture of you and you can type in these prompts and you can type what you want to see. Camera, for example, zoom in. And you can also include a rotation if you want and you see the output. So the output is relatively good, but you don't have a lot of control and you can only do this in two D in my mind. It's really not worth it right now. But the other feature is relatively good. You can also overlay your pictures with prompts. As you can see right here, I simply uploaded a video from me where I am basically just handling around in my room. And you see what Kyber did to it with overlaid prompt, amend dancing in the style of illustration, highly detailed simple move and so on. And it was really, really easy and it is intuitive. You can't just upload it. The prompts are nearly created for themselves because you just have to type in two or three words. The rest is the prompt magic from Kyber and you get your video at the end. And you can also animate whole videos. I think Linking Park did their music video completely with Kyber. In Kyber, you can work. It can be a bit expensive if you need to do a lot of it, but you have not the best control in my mind, you should work with Devorum diffusion and there's also P diffusion. P diffusion is another Collab notebook. And in my experience, P difusion makes good outputs, but it can only overlay prompts over your videos, but you can use control net and stuff. So P diffusion is relatively nice, but I won't dive into it because I'm not an expert in P diffusion wrap diffusion is easier to use than devorum, but I think also you have more control in devorum and you can make cooler stuff with devorum. In no other thing, you can make such good three D creations. 28. Animate Images with D-ID: Let's assume that you have already made a few pictures in micherny or in another diffusion model. These are really easy to use because you know how to write prompts. This right here for example, is the Mitchy web page. And if you simply type in some prompts, for example, a woman on the beach, you will get some pictures. As soon as you get these pictures, you can animate them with the ID. You already saw the ID in the beginning of AI avatars from this course. And now I want to show you how we can use the ID to animate some pictures. Right now, I want to wait until I have my picture and then we will create something from DID. I think I will make a few more because I am not satisfied with these pictures. So you will see what we get as soon as we animate it. If you ever have questions about how to make normal pictures in mid journey or other AI tools, just let me know. I can totally include some lectures about these. First of all, you can also use the Colab Notebook, a too Lep. But I must admit, this doesn't work that well. It works sometimes, but sometimes not. But ID is a cool alternative. Did is a tool, an online tool. And you can use it, first of all, for free, but just for 20 credits. But you can play a bit with this. I think you should try this. If you really, really love it, you can make a subscription. But the free trials are for everyone. So first of all, you go to Studio.d.com and then you basically land on this page. If you choose to register with your Google account, then you press Create Video Here, you have some nice options. We can of course, also use the pre designed speakers right here. But we want to let our mid journey picture speak because we have made it. So we press at, we simply include our picture right here, and there we have it. Here is our picture. And now we can do really a lot of different stuff. We can also just type in the text that we want to have here, so we can type in the text. We search, for example, for Rachel or another speaker that is for free because she is like for premium. So we could use Jenny. But what I really, really like is to use 11 labs. Now we go into 11 labs and we make our audio. Now we are back into 11 labs. And I think Elon is not the optimal voice for our girl. Okay, now we do something really funny. We use the deep mail voice, but we want to make two videos. First of all, we want to have the deep mail voice. I want to have this text. Hello. I am really glad that I've hit puberty and my voice finally sounds the way I feel generate this. I played a bit with points like always and then we simply download our voice. Next step we go back into ID and we simply upload audio and we include our voice from 11 laps if you like. You can also record your audio right here, or you click here or do your simple drag and drop and include your audio right here. This is really fast and everything that we need to do right now is generate video. This works also really fast. You can see this costs only one credit. We have 20 credits left right now. In something like 20 seconds, we have our video and I want to see how this all goes. Hello. I'm really glad that I finally hit puberty. My voice finally sounds the way I feel. I think this is really nice and now I make a second one with a girl voice. We search for a girl right here. I don't want to have the British girl. I think I want to have Bella. I'm American Soft and that's basically cool, I think. Or even Charlotte Cheese. For video games, we use Charlotte. And then we delete this right here and we include this text. Hello, I'm Alice. I'm really glad that Arnie brought me to life and 11 laps recommends to switching to multilingual version one. We do that and now we press Generate. Now I download also this, we create a video again, here in ID, we simply upload again our voice generate video. Now we are basically done. This only lasts for a few seconds. This is like really fast. Let's just see how this looks. Hello, I'm Alice. I'm really glad that Annie brought me to life. Okay, I think this maybe is a little bit better than the first one, if you like some of these, just press the three dots and download them. You can do this, of course, with both. In this video, we took a look at the ID. The ID will animate every picture that you like. It could also be a photo of you, a photo of some other guys. Or like we did a picture for Miurny. I think with mure pictures this is like really, really nice. Because they are really good looking and you can create everything that you like. You can also make cool cyber punk men's or cyberpunk women's and animate this. You don't have to do realistic stuff. You can even try to animate a dinosaur with this, but I think it's not perfect. Maybe it's better to stick with humans, because I tried it a few times and the outputs are not that good if they work and they don't work always, but you can try it. Ai is here for trying, so you should try different stuff. Just as a quick tip, some cyber bank woman's worked good for me. Here is a German one. Hey, Fred Blangbert. I hope you can create such cool things in the ID and 11 labs and mid journey, you can combine it all. 29. Faceswap in Seaart ist faster as Google Colab and Midjourney: We have good news. Art has added the feature of face swap. That means you can do your face swaps also inside of art. And maybe you can even skip either the Google Co Notebook and also of course the face swap both inside of journey. So stay tuned because I think this works really, really good. This right here is Spider Man or it's Arnie Man. It depends a bit on what you see. I have made this little video, you can see this is my face swapped on Spider Man and it works really good. And now I will download these. I press download, and then this video is mine. If you wonder how you can create this, this is really easy. You simply have over to art tools face swap. Then you need to search for the right video or photo that you want to swap yourself. Here you have a lot of different videos. You can simply press on this video and try to swap your face. Right now, I will delete this video. For example, I press here on this button and I confirm it. And now we can simply choose a video that we like. For example this. You can either make a video swap or an image swap. You can search either for pictures where you want to do your face swap or you can also search of videos where you can make a face swap. Let's just say you want to swap this video, you simply press on, Oh yes, here. And let's assume you really like this video. And now you can simply swap this with your face. You press on this button and then you upload a picture. I have uploaded the terminator and me, because I think essentially this is the same person. Then you simply press on the thing that you want to swap, and then you press Create, and then you are ready to go. This is really the process, and in 2 minutes, you have your video. And the same thing is, of course, also true with pictures. You can also do this with pictures. You go on image, you use the image that you like. For example here the Superman I use here my face. And I press Create and then my face will get swapped with Superman. And there we have it. This is Superman with my face. And I also want to download this picture and then you are ready. You can do everything that you like. So have fun, try to do a little bit of face swap because the art has made the entire process seamless and even easier than it was already before I forget it. You can of course, also make your own custom template. You press some custom template, you can upload your video, and then you can face swap your uploaded video with the stuff that you upload right here. And the same thing is true for image. You can upload images and videos as the target video and target image, and then of course your source image in this corner. 30. Conclusions of Section 4: Over the last video, you have learned really a lot of stuff. And most of it is really, really cool and awesome. You have learned how to make deep fakes. First of course is to clone our AI voice in 11 laps. Then we use the cop notebook A to Lip to make our deep fake video. That's basically our first project. Then you have seen that you can also do Facewaps in Google Cop. Yes, all of this works also of course, in C art, and you have already seen a lot of cool face swaps over this course. Face swap videos work in different kinds of manner. You have seen that we can make a picture talk with DID Rep. Fusion is a completely awesome tool. And we can animate our videos in Google Colab. This works really, really fine. And then we have also the forum Stable diffusion, and the Deforum is a bit more technical, but I gave you the blueprint, a quick overview, how to make music videos and how to make three D videos. Like I said, this is not a complete course on stable difusion devorum because this is a really, really big tool. But you can do most of the cool stuff if you try it a little bit out for yourself. You can basically make whole music videos and three D animations with these if you don't want to do this right now. We have also Kyber. Kyber is an easy alternative and you can try this out for free. I think it's not really worth the subscription unless you are not willing to learn Devorum Diffusion and Rep Fusion. Basically, my little homework for you would be just to try some of these out. If you don't want to go into the technical details, just go into Kyber and make some cool animations. Or you can also try to make a nice little deep fake. Because I think deep fakes are awesome and they are really, really easy to create. So have fun creating your own deep fakes and AI animations with AI voices, AI videos and much, much more. All of this is completely awesome. I love AI. I hope also you love AI. If you do it, just share this course or leave me a nice review. This would mean the world to me. And maybe it will also help you because the right people come in this course because of your rating. And if you share it, people will always assume the value that they get out of this course with you. So a win win situation for all of us. 31. Section 5: Your Project and more Tools: In this section, I will show you some examples and stuff that you can do. You will learn that you can use different tools and you will also see how you can make your own whole project. Because we start first with the terminator example. You will see a cool video. I think the video is awesome. And then you learn how you can make such videos for yourself. If you combine a few tools and you edit the videos just a little bit. You also see things like Victory Box or Cap Cut, Adobe Firefly with Adobe Express Canva and also Opus Pro. So this section is nice because after this section, you know how to make and edit some videos. Have fun. 32. Your Project: Create, Cut & Edit Videos with Pika, Shotcut and CapCut: In this video, I will show you how we can make such videos that you have shown in the previous video. Yes, I will make this a lot quicker and a lot shorter. But you can of course, make really a deep dive. I'll just show you how the process works and then your creativity is asked. Of course, you can do this however you want. You can make entire films with this process. I just beg for mercy that you forgive me because I don't want to show you how we our long video. Because like we are a week in this lecture, we do this really fast. Step one of course, we go into a program where we can create videos. For example, hyper AIP collapse, Runway ML, whatever you want. Let's just assume we want to make something like this car. Or we can of course, also make something on the Moon or on Mars. I think stuff on Mars should work relatively nice. I want to make something that is similar than this. Let me just see if we use hyper or we use collapse. I think I should just try hyper. And if it works good for us, if it doesn't work like we take k lapse. So we can insert here a prompt or we can of course, also copy this thing. But I think I want to make something in Ka. You already know how Ka works, so I don't want to bore you. The next step is of course to go into an editing program just like this. This is shortcut and we give names. Project X should work. We press Create and then we insert our videos. We simply drag and drop every single video right here in. I have made three videos inside of collapse because I think this works like relatively cool. You see we have this astronaut and of course every single video editing program works. You can use shortcut, you can use Da Vinci, you can use whatever you like. The process is always the same. You can go even on Cap Cut, because P Cut is a video editing program that works really good for a lot of people, this is completely free to use. So you can simply search for P cut on the Internet and you can download these, install it, or even use it on the Web. You find also a lot of tutorials about Cap Cut and if you need more, just Hymup. I can always include a lecturer to how you can use all of these. But like I said, all of these works normally completely the same. You always insert your videos. Then of course, you need to go to your videos and you simply right and drop them down to your time line as soon as the videos are on the timeline. I like to make them bigger by pressing the plus sign. And then we can play these videos, or we can insert our other videos, we r and drop the second video also on the timeline. And then it depends a bit because we can use the magnet and we can simply play them together. And then the third video. And then we have like three videos that tell a nice little story. First, we are in this space shuttle, then the guy goes outside of the space shuttle. Last but not least, he is satisfied that he is on Mars and he can explore a nice little new planet. I think the story line is decent. I will exclude my face so you can see every single bit a little bit better. We can always make like stuff black in the middle. We deactivate this nice little magnet and then we can always have the black overview. So you can see we have separate scenes and this works similar in every single video editor. Right now we have a video, but I think it's better to make this seamless. So I think this works a lot better. We can always play a little bit how all of this should look like. This is always a creative process. Video editing like, you can spend a lot of time on these and you should always look if this looks right for you or not. I can't tell you what's optimal, because everybody likes different videos of course. But the process is always the same. We have a 12 second video right now, but you can make these videos, of course, as big as you want. Now we need a story line. I like to make story lines in Chachi Pt because in Che Chi PD it's really awesome. I just start with something in German so that you can see we can always also translated in English. This is basically the stuff that we need from ChechiPD. Chechipd should basically make her script about an astronaut in a space shuttle. Then the guy goes outside of the space shuttle. And last but not least, he goes into the golden hour. So the sun goes under and he conquered Mars. That's basically what I told ChechipT here in German. But of course, I will translate these in English so that everybody can understand this. Like German is not so nice for most of the people in this course. I think we simply translate these in English. And this is also really easy. You can also do this vice versa, if you have, for example, a script in English, but you'd like to make something in German, just do this, vice versa, and you will get the right answers. Right now, I want to have this in English. We go into 11 labs and we create our text. We go to text to speech. Then we simply use the right stuff. Multilingual should work, The British girl is not optimal. We have here Adam. Adam is a voice that everybody loves on social media. And we simply delete this stuff right here, we go back into ChachiPT, and now we insert our speech control center shuts is ready for Mars. Landing is our first sentences. I simply include this into 11 laps, make some points because this is a nice little past. And then we use incredible to be standing here. I think this is a nice second phrase here. Also the I need to include some points for a pause at the end. And then we insert the last thing, a moment for entity. Then we insert also this. I think this should work. So we simply press Generate here in 11 laps. And we get something like this. Control center shuttle is ready for Mars landing. Incredible. To be standing here a moment for eternity. I think this sounds relatively nice. I want to download these by pressing here on the right button. We download it and then of course we can insert it in our video editing program. We simply go back in Shortcut or Da Vinci or Cap Cut. We insert our nice little audio, then we need to insert also an audio file here. And we simply drag and drop it down. And this is a little bit short, I can see it right here. We need to edit these a little bit better. But first, let's just see how this looks right now. Control center shuttle is ready for Mars and Incredible control center shuttle is ready for Mars landing. Incredible, to be standing here a moment for eternity. So generally speaking, this should work, but we need to make this a little bit better. So we need to cut the audio, cut some parts, and we listen from time to time. Like I said, this is a process that needs a bit of creativity. Let's just see how this looks right now. Control shuttle is ready for Mars landing. Incredible. To be standing here a moment for eternity. I think this looks decent. Of course. You can also include music, You can include different stuff. You can make this longer, shorter, a lot better. Like I said, this is just a quick quick tutorial that you can see and understand how the process works. Let's just see how this looks. If we add a little bit of music or something, control center shuttle is ready for Mars landing. Incredible to be standing here a moment for eternity. Like you see, this is a starting point and you know how to do this. This is of course, not a complete project. This is just how you should do it. But I think, yeah, maybe it's a complete project, so it makes all sense. We can totally post these on social media. This is a complete video with a nice little story line, if you want to call it that way, this should work. Maybe you get some clicks with these. You also saw that you can make entire films out of these, just like the guys from the Terminator remake. But I have to tell you, they made a whole film. You can watch this in the cinema. The stuff that we saw was just the introduction. Yes, you need to work a lot on these if you want to make really awesome videos. But you also saw like stable diffusion and we can make P fusion and we can make music videos. We can do a lot of stuff. If you want to make something like we did here, just try it out. Make some videos with labs with run way, Emil with hyper, with whatever tool you like. Then we insert these into a video editing program. We make something in 11 labs, like a nice little story line. We insert also some music. If you don't know where you can find music, you can find music totally for free on the Youtube library. You can simply search for music and you can insert this music. Like I said, this is a creative process. You need to play a little bit with all of these and you can really spend a lot of time on these if you want to make this perfect. If you don't want to spend a lot of time on these, I want to show you in the next video something that is also really awesome. In the next video, I will show you a tool where you can simply type in some text and you get your whole videos back from the tool. So stay tuned for the next video. But I think this is also like the real deal, because you can be really, really creative and you can die down on every single section, just like you want. 33. Pictory & WOXO for the Lazy ppl, Complete Videos with One Click: In this video, I will show you two tools that make videos completely automatically. You just have to type in some text or a story line and you get your full video. We talk about Victory and Oxa. I think these two tools, they are completely awesome. Victory is more like for bigger videos and Vox is for shorts. You can make Youtube shorts and much, much more With these two tools, let's just take a look how these things look like. I think I should blend myself out that you can see a bit more. Hello Arnold. Which content would you like to repose into videos? This is the first thing that you see after you have created an account into victory. And you can see that you can make script to video. You can start typing or copy paste your script and then you can make your video. You can always see for what this stuff is recommended, recommended for educational videos, listical videos, coaching videos, and step by step guides. Then we have also article two videos. This is really practical because you can also just insert a link. Maybe you have a block or something and you will get a video out of these. This is recommended for blocks, press releases, or any HDML article. You have edit videos using text. You can, of course, input video from various sources and then you can edit these videos also. This works like okay, I would assume. I think you should totally try all of this for yourself so that you can see this. This is recommended for add subtitles, automatically cut portions of video, create video highlights, and add logo intro and outro. So you can make, for example, a Youtube video and you can add your Intro ultra, and a logo. You can also insert subtitles. You can cut this video and you can create video highlights. Now, I have to tell you if you want to cut your videos and make a highlight, it's not always perfect, but you can try it. Like I said, this is always the worst version that you will ever use. It's always worth taking a look. You have also visual two videos, drag and drop files or Browse Compute. You can simply upload a visual and make it into a video. You can create a slide show video using images or short video clips. I think I just want to show you one simple thing inside of these tools, because you should totally try this out for yourself. We simply start typing or copy based your script. I want to show you this right here. Of course you can do all of these. We simply press proceed here on script to video, and then we can enter a video name. Let's just make this funny. We go into Che ChiBT because I am not smart enough to make something for myself. Make a short, funny script for a 32nd video about AI. For example, we sent this out and I think we will get something. We even have a title, so come on, this title is of course better than my title. Then we have the scenes and we can always just copy these and see what we get out of victory. So we copy, of course first this title. We go right here in and I want to insert it. Start typing, We don't want to type, we simply want to insert our new script. I think we should just copy all of this. We can also make this like in a story line so we don't have to make all these scenes or something else. We can also just tell chipped, make a quick story out of these if you like. You don't have to tell like every single scene what you want to see. But you can do it. We just want to do it like this without. I just want to see if Victory finds the scenes themselves because like, yeah, this is prescripted. And I want to test also the creativity of this tool. Just a little bit. We have this nice little script right now in the non distant future robo, a state of the art I kitchen assistant decides to tackle the challenge of cooking spaghetti carbonara, a favorite among humans with confidence and so on. So I simply want to copy these. You see, now it gets, of course, a lot harder for victory to find the right things here. For this video, we simply want to insert this and then I think we are ready to rock. We simply press Proceed right here. First we prepare our storyboard. This works also relatively fast, So you can see creating scenes, we get 12 scenes. We have right now 789, You can always see what we do right here. We find relevant visuals, we add I voice narrotation. We can record your own voice, we can add our logo, and we can download all of these. Right now, I think we have just the standard stuff you see right here, the scene duration. We have 14 second long scenes. The video is 2 minutes and 35 seconds long. And we can take a look at this video. Of course, we can take just a preview. So if we press on preview, we can see all of this. And if you press right here, we can do it. I think we should download this video because like safe, safe. As soon as we have it, we can always look at this video. And now I want to make a nice little preview. So we simply press on preview. You saw the video and Pigtory has not nailed it. I have tried this quite a few times, Most of the times Pi story has nailed the pictures a little bit better. Of course, we can also include the voices and much more. You saw that we can include nearly every single thing that we like, but of not every single thing is always perfect. We have more like the cooking stuff and less like the AI stuff. But like I said, this will be for always the worst version that you will ever see. Now I want to show you Warsaw, because Warsaw works really, really awesome, especially if you want to make shorts. So we can always download this videos. We simply press and download. And you can also edit all of these, so you can insert other text. And to do much, much more, you can go on visuals, audios, styles, text, branding elements, and format. You need to play a little bit with these videos, but you can make videos really, really fast. And you can start for free. And now I want to show you Bokaw, just like I said, also here on walks you have, I already have an account. Simply press login. You can always make your account with Google or with your e mail and a password. I want to continue with Google because like I said, I already have an account. And this right here, this looks nearly as Adobe Express. If you are familiar with Adobe Express, also in Adobe Express, you can make similar stuff, but I think Voca works a little bit better. You can make mobile videos, Tiktok videos, Youtube shorts in Star reels, Facebook reels, and videos for X. You have different topics, languages, voice over text styles, and background music. I want to make this really, really fast. We make Youtube short. Right now we press on Youtube short, and now you see this three prompts. Of course, you can always enter your prompt or new RL for yourself, but I think I just want to show you something from these. Produce two videos highlighting innovations from the Industrial Revolution. Integrate our line art background. We press on this prompt, we simply make generate. So you see the prompt is inserted right there. Then we press Generate and we will get our video. This is also relatively fast, You see this are the two videos. This will amount to roughly 1 minute. After 1 minute, we have our two videos. There we have it. I think this took like 30 seconds. Let's just see what we get. I press on the first one. Welcome the fun facts about the Industrial Revolution in the late 18th century. The spinning jenny revolutionized textile production. But steam engine powered factories and transportation reshaping industry. The cotton gin mechanized cotton processing, boosting production. The telegraph transformed communication, connecting people across distances. To wrap up the Industrial Revolution was a time of innovation and progress. Welcome the fun facts. In the late 18th, the first video looks awesome. Let's just take a look at the second video. Welcome to the impact of the Industrial Revolution on transportation. The invention of the steam engine revolutionized travel, powering trains and boats. The creation of iron and steel allowed for the construction of stronger, faster locomotives and ships. The telegraph facilitated communication, improving coordination and safety for transportation. To conclude, the Industrial Revolution transformed transportation shaping the modern world. The second video is awesome. I think you should totally try these two tools, Victory and Vox. Like you can start to deliver and you can do a lot of stuff on both tools. I just showed you a glimpse of these, but these are really, really intuitive to use. You just press on stuff that you like and you get stuff that you like. Maybe the stuff that you get is not always the stuff that you like, but that's just the things that we need to deal right now with I. But we would assume that all of this will get better. Remember, the version that you try right now will be the worst version forever. All of this just started, all of this gets better day by day. In no time whatsoever. We will can create entire short videos and we can simply post them on social media all day long. At least I hope so. 34. Opus Pro for YouTube Shorts, Instagram or TikTok: In this video, I want to talk about Opus pro. Opus Pro is really easy and nice to use. You can simply upload a long video and you get shorts out of it. If you have, for example, a Youtube channel, or you make podcasts or something like this, Opus will help you make some shorts out of these. Now I will show you how this works. First thing you go on Opus Clips, you can simply type it into Google. Or of course, you can use the links that I gave you. One long video. Ten viral clips create ten x faster. This is the stuff that they tell you. Every single thing you have to do is either drop a video link or upload files. That's really every single thing that you have to do. If you press on upload files, I will upload something from this course, maybe like the last video, and then we will see what we get out of these. Yes, I have a Youtube channel, but the channel is in German. And I think you will understand more if I upload an English video. And by the way, if this is your first trial, of course, you need to enter an e mail address or you need to login with Google. I already have an account, so I go with Google. Here you see I already made a few videos, but right now I want to show you these. You can simply upload a video like I told you previous or you can insert a link. Now you see that we fetching our video, we have right now, 20% analyzed. We can also say that we use this for personal preferences or for whatever we want. Let's just see what we get out of our video. And there we have it. So you see Opus pro dido grade job we have here, our first trailer. Then the second one, the third we have all in total, something like ten videos. Let's just see how this one looks. Yes, this is a whole Kino presentation. It's named Terminator Two remake. You see we have the shorts. Yes, I have to admit, it works. Not perfect, but also this will get perfect over time. You can always insert your own text, the subtitles, they are really, really nailed down. And sometimes also the camera movements work. That's just a tool that I want to show you, because sometimes it works relatively good. Try this out. 35. Canva: In this video, I want to talk about Canva. We do this really quick because Canva has also included some cool AI stuff. Runway ML, for example, is included in Canva. This right here is the web page from Canva. I have to tell you I am not a pro in Canva, but I do make my thumbnails for Youtube, for example, in Canva, and I do some stuff right here, let's just put it that way. You can always press on magic studio. And here you find all the coolest AI tools. I think you should try this out. And here you see you have some AI features. For some of them you need to pay, but others are free. I want to show you text to video because this is about AI videos. We simply press on text to video, and then here we can simply test all of this completely for free. We want to try this out. Here we are. Yes, there are a few words in German, but don't worry like we can make them totally in English. First, you can always edit all of this stuff, you just click on these. And then you can insert also your English text, for example. Test this. I think this is good right now. Now we want to make a video. If we simply press on these, you can insert five words and then you will get a nice little video. Let's just say teddy bear in the pool. Having a good time. Now we press Generate. You see this is powered by a runway. In one or 2 minutes, we'll get our nice little video where a teddy bear is chilling in a pool. That video we can insert right here. We can include videos that look something like this. You already know the videos from Runway. Yes, they are relatively nice and we can totally do this all inside of Canva. Of course, we would make this in English. First of all, let's just add something like we get a teddy. I think we should do something like this. I like to chill. What do you like if you also like chilling? Join me. I think I don't need to that letter. I think it should look okay. Something like this. I think this looks good. Now, we wait until we have our video, because I want to delete this video and I want to insert our teddy. Of course, you can make your second file also right here. So you can make here, for example, stories for Instagram and much more. There we have our video. I want to insert our video right here, and we will have our teddy in the nice little story right down here. I think this looks okay. So this is a teddy that tries to chill. I think this is nice. So we have our first thing right here and we can do this over and over again, and then we have Instagram stories. In this video, I simply wanted to show you that Canva has AI integrated. You can create your AI videos inside of Canva because Runway ML is included in Canva. If you make stuff, for example, for Instagram and you want to make AI videos also for Instagram, just do it in Canva because you can do it in one single tool. And I think that's nice. 36. Adobe Express: In this video, I want to talk about Adobe Express. Because Adobe Express is also a nice little tool that helps us edit our videos. It's especially practical if you already have videos. You want to just edit them a little bit, maybe for an Instagram post. Let's just take a look at the platform. And by the way here, we can also work with Adobe Firefly. We can include text effects and much, much more. We use the diffusion model from Adobe to help us a little bit. First of all, we go at Adobe Express. Here you can scroll down U C. You can use different text effects and much, much more, but you can also press on video. Here on video you will get a lot of different things. You can make Instagram reels, for example. You have always stuff that you can use. You can make the normal Instagram reel. And I think we should totally start with these as soon as you are here. Of course, we can include stuff because this is right now completely empty. I will simply upload I think a small Youtube video from me. This is about stable Brad Fusion. You already saw stable Brad Fusion. We can make a small little video that works relatively nice on Instagram. I think at least maybe I will also go post this video. You see, we can simply insert this video. Yes, the resolution is, of course not right here. You see, this is how it looks. I transform myself. You already know how stable re, fusion works. We can make this, for example, a bit bigger. We can also make it here a bit bigger, that all of it is a bit more in the center. I think this looks somehow, right? Yes, come on. Of course we can. And probably should include also other stuff. As soon as I am in the center, I think this is okay. We can also include text and a background and much, much more. We go on the left side, we press on elements, and here on elements, you can also search for the stuff. I think we should type in technology. And the first one, this is okay for me. We simply press on these and it gets inserted. You see, I think this looks somehow okay. And now we can go on and edit this a little bit further because I think we should also make some text right here or something after the text. I think this looks a lot better. We simply press on text on the left side, and now we need to search the right text. I think we should use the ifilight text in down here now we can search for the stuff that we like, maybe the bubbles. No, I think this right here is cool or something else. Let's just see. Yeah, I don't like this stuff. Let's just use this one. Of course we need to accept. This comes from Adobe Firefly. We will generate our text via Fusion model. Now we can simply delete this old text and insert the new text. This works like relatively good. As soon as we type our stuff, we get it P fusion. That's the technology behind the tool that I show in this video. You see the text also looks like relatively cool. We can make this bigger, we can make this smaller, and we can place it right here. I think here it looks relatively cool. We simply can also insert something down here, maybe like a tube blinker, something, because it's a little bit empty without these. Let's just edit it a little bit more. Maybe this right here, down a bit. Yeah, the center is somehow okay. The Rep Fusion, also a little, a little bit bigger. I think we can also insert an animation if we like. Let's just make the text a little bit cooler. We use something like this right here. I think this is cool. Come on, The rep Fusion needs to slide in. We slide in our Rep Fusion and then we have something really cool, at least in my mind. Then we can also download this and we can also post it on social media if we want. If we press on the, we can really post it all over social media. We can time our posts, we can share it on Instagram always automatically. We can either edit it a little bit more or we'll add it how it is. I think this looks also okay, right here, how it is. Let me know what you think. I will play these right now. Stable B fusion and cool technology in a mal sin wold test. Of course, this is German, but it looks okay down here. We probably should add something. It's a little bit empty down here. Maybe we can insert a Youtube link or something else. I think we should also make a new one with a deep fake video. This is most likely the best deep fake video you will ever see. Arnie is incredible. So we did basically the same stuff with this video. So deep fake tutorial. This right here is of course Zealand mask. You already know how to make these and here in Adobe Fire Flight, we can always edit all this stuff. I think also this looks like really cool. So first we start with the Tom Cruise deep fake. This is a deep fake profile from **** Doc with a lot of views and we simply include the stuff from my video. Let's just take a look. Deep, deep fakes. This is most likely the best deep fake video you will ever see. Arnie is incredible. So you have seen, it's really easy to create such stuff for Instagram. You can just go into Adobe Firefly. You can press Create Instagram real. And you can include your videos, a background that maybe some text effects with Ad with the Difusion model from Adobe. And if you want, you can also automatically post these on social media. I think you should know that all of this is possible. Of course, if you make your videos like with walk or puss or other stuff, you can always go into Adobe A Fly, into Adobe Express. You can edit it a little bit and then you can post it. Have fun trying all of this out. 37. Recap of Section 5 and a bit of Homework: Over the last videos, you have sewn really a lot. We started with a nice little example, and you see that people making real films out of these. You have sewn the introduction of the Terminator to remake. And I think if you are really, really ambitious, you can make either the trailer or also like the whole film if you like. And I have to tell you, the guys that make this film, they did it over a long time. I think they took a few months for this film and this where a whole team, they have a behind the scenes and much, much more. So, if you want to make something really cool right now at this minute, it will take some time. As soon as Sa. Comes around the corner, I think this will change dramatically because with Sa, we can make enormous videos up to 1 minute and it should be really, really coherent. Sa will make every single thing like a lot easier. You have also seen how you can make these videos. You simply go into pica or in the tool that you like. Maybe as soon as you see this course, of course, updates will follow as soon as we have access. There are like some guys that say we have Sra in a few months. Let's just hope that we get access. But you have seen how the process work. You make a small video, then you make another small video that is similar. And you do this over and over and over again. Then you go into your favorite video editing software. You can use Shortcut, Da Vinci, Resolve the programs from Adobe, but you can also use things like Cap Cut. Cap Cut is completely for free. You can download it and you can also use it in the Web and the applications. They work nearly every single time, completely the same. You can just cut some parts of the video. You can insert a little bit of audio, for example, from 11 laps, and then you are ready to rock. Maybe some music from the Youtube library. And I think you get what all of this means. That means that you can make videos. And of course you also saw like stable difusion, kyber and all the other tools. So you are nearly unlimited. The only limitation is your imagination. I think you can make really cool stuff. You also saw that you can use victory or walk Saw. These two tools are easy to use and they are Fast Opus Pro, you can make long videos into shorts. You also learned that Canva has a runway integrated, and you can make stuff for social media with Canva. And of course, you can also use Adobe Express for your videos on social media. I think there's something for everybody. Maybe you remember what learning is, learning is doing, and you get better by doing more. If you do something, you will learn. Learning is same circumstances but new behavior. You know that you can make videos with a I. So you should totally do this. Just do something with a I, just make a video with a I. I think this will completely blow your mind as soon as you see that this gets better, like nearly every single day. Do me a favor, go in a program that you like, maybe it's Axa or whatever you want, and just make one small little video and I think you will do great. You may also remember how we can learn better. We learn better together, and we can accomplish these if you share discourse. Thank you. 38. Section 6: The Future of Ai-Video and Copyrigths: In this section, we need to have a serious talk because first we start that AI videos are not just a joke. I think you will be surprised what's possible or how the view chair cook eventually unfold these AI models. They can simulate the world. And these simulations, we can use them to train robots. And these robots can come into the real world. The next video will be cool, but then of course, we also need to make a reality check and see what we can do right now. We need to see if copyrights are important. Of course they are important. But how and where, we need to look what we can create, what we can not create, what happens with our data. I want to make a broader look because yes, this is about AI videos, but I want to show you also this like in ChechePT and all these AI tools, because this is important and I get a lot of questions around this topic. Then of course, we make also some speculation. Because first we learn how these AI videos can make something that can train robots. And that eventually comes AGI in the future. Because we talk of this, we also need to talk about AGI in general. What happens if AI becomes better or smarter than humans. And then of course, also ethics and downsides of this topic. This section, this is a little bit broader. We need to look at all of this on a broader field because it's not just about AI videos, it's about AI in general. It's about copyrights, AGI and the future. So have fun in this section, I think you will really enjoy it. And you will understand on a deeper level what's going on. Why is every single really, really big company all over? I just think about it. 39. More Than Just Videos, The Future of AI Videos and Robotics!: In this video, I want to talk about the future of AI videos. Because I think the future is much, much, really much bigger than a lot of people think. As soon as sa got introduced, a lot of people had a little bit of fear. People on social media said that this is most likely the worst thing that can happen to humanity. And they simply think of these because they are maybe scared that people can make videos with just text and they will be obsolete. But we need to see the bigger picture. First of all, of course, you can learn how to use these tools and take this as an advantage. I think that is the smartest move that you can make. You already took the right decision because you are staying in the loop. You see how these technologies work and I think you are in great favor. But we also talked about that all of these models, all of these AI video tools, they can somehow like simulate reality. In this video, I want to bring a whole holistic picture. Then you maybe understand why this is the real deal. A lot of people talk about the AI video tools as world simulators. Opi has the goal to simulate the world with their video generation tools with Sa. The goal is really to simulate the world. I already told you this at the beginning of the course, and now you will grasp why all of this will be so important. Before we make the picture full, I will also show you the right here. This is research from Google. Google introduced us to a paper named Genie Generative Interactive Environments. Here is the Genie team, and here you can basically see what they are doing. You have a still picture and then you can simply type in a prompt and you can play a whole game. That's basically how this works. So we can also simulate like video games with simple text prompts. I hope you understand that we can do a lot of things. I think the gaming development world will be a lot better, a lot faster with tools like these. If we can just prompt our games, of course, we will make these games enormous fast. Here you can see a lot of different environments. We always start just with a picture, then we throw a prompt in and we get our output. This is basically how maybe the video generation will look in the future. You see a lot of different pictures and then they start to move. As soon as we inject a prompt. This is really bigger than you think. They have a lot of examples. And later you also see this right here. The future of generative virtual words. Finally, while we have focused on results for platforms on this website, Gini is a general method and can be applied to multiple of domains without requiring any additional domain knowledge. Basically, they trained a smaller model and they tried to simulate the world so that RT one can learn. Rt one is the robot that comes from deep mind. They basically try to simulate a world in this simulated world. This robot can learn to do tasks. We can not only make video games, but we can also like simulate words. In this simulations, we can train robots. As soon as these robots are trained in simulation, they can also do stuff in the real world. This is the genie paper, this comes from Google Deepmind, and RTX is the robot from Google Deepmind. They have good results right now with this technique. Of course, we have right now, stuff like Unreal Engine. Unreal Engine. We can also make videos, and it's relatively easy to create these videos. These videos are, of course, also for games. None of this is completely new because in Unreal Engine, we can also like simulate words with games. Sa is probably maybe also trained on such data. We don't have to take necessarily data from the real world, but we can also take data, for example, from Unreal Engine. If we train such video models on Unreal Engine stuff, this is just from the video creation. Then we can really simulate all of these words without ever going into the real world. Now I will show you the most promising things from Envidia. Vida has something named Isaac Sim and they have also Isaac Jim here. This is the web page from Vida. This is the video and I think the best thing is we see the video later. They have simply Isaac Bac Jim, Isaac Sim and much, much more. They always post the newest updates. You can see they have a lot of research here and they also show these like we have some Nvidia days from time to time and they always bring us the newest. Explanations of why they do this and why all of this is important. I think we should look just at this video. And after that video, we understand why all of this is much, much bigger than you think right now. Let's just play this video. Successful development, training, and testing of complex robots for real world applications demand high fidelity simulation and accurate physics. Built on in Video's Omniverse platform, Isaac Sim combines immersive, physically accurate photorealistic environments with complex virtual robots. Let's look at three very different AI based robots being developed by our partners using Isaac Sim. Fraunhofer IML, a technology leader in logistics uses in vida. Isaac Sim for the virtual development of Obelex, a highly dynamic indoor outdoor autonomous mobile robot or AMR. After importing over 5,400 parts from D and rigging with omniverse physics, the virtual robot moves just as deafly in simulation as it does in the real world. This not only accelerates virtual development, but also enables scaling to larger scenarios. Next, Festo, well known for industrial automation, uses Isaac Sim to develop intelligence skills for collaborative robots or cobots requiring acute awareness of their environment, human partners, and tasks. Festo uses Cortex and Isaac Sim tool that dramatically simplifies programming Cobot skills for perception. Ai models used in this task were trained using only synthetic data generated by Isaac Replicator. Finally, there's animal, a robot dog, developed by a leading robotics research group from ETH Zurich and Swiss Mile using to GPU Accelerated Reinforcement learning animal whose feet were replaced with wheels, learned to walk over urban terrain within minutes rather than weeks. Using Invidias Isaac Jim training tool, the locomotion policy was verified in Isaac's Sim and deployed on a real animal. This is a compelling demonstration of simulator training for real world deployment from training, perception and policies to hardware in loop. Isaac Sim is the tool to build AI based robots that are born in simulation to work and play in the real world. You see this is really bigger than we initially thought. Let's just put this in perspective. Vida make something makes a simulation. In this simulation, we can train robots. This works really well. All of it started with a project that was named Voyager Voycher is a project that can play Minecraft. With this project, they basically showed that we can have a computer that plays Minecraft just like a normal person. There's also a really strong dog out there from Jim Fan. This is a senior developer from Nvidia. He simply tells us the whole story. We basically started with Votre Digital. This is, this is the computer that can play Minecraft. And how does this guy learn to play Minecraft? They simply give him an explanation what he should do, and then he uses reinforcement learning. He writes code for himself and then he gets rewarded. It turns out that this code comes from ChechPT. It's a paper out there that's called Eureka. And with this paper, they also show that Chet CPT can reward these machines better than humans can. By the way, reinforcement learning from human feedback is normally the thing that we have to do to reward these robots. They learn by doing. If they do something right, they get rewarded by humans. And then they know, hey, this was right, I need to do more of these. It turns out that a ChechPlso, I can write this reward code basically. They have shown that a computer can learn to play Minecraft. And then they took it a step further and they tried to simulate the real world. And here they try to train robots. They do basically the same stuff and as soon as the robots can be good in this simulation, can do stuff in the simulation, they can also do the real deal. So they can go out of the simulation and they can do stuff in the real world. This is enormous and that's why I video is so big. If we have models that can simulate the world, we can do stuff that is enormous. We can train every single thing in a simulation. And then we go outside of the simulation and have the stuff in the real world. And that's why AI videos are so big, and that's why a lot of people fear AI videos. But we learn how this technology works, so at least we understand why we have no chance against machines. Come on, it's funny. 40. AI is here, what can we do rigth now: You saw the applications. They are really gigantic. Of course, we need to ask ourselves, is this all the right direction? What happens if we go to like the AGI? What if machines get smarter and better than humans? Will there be stuff left to do for us humans right now? I have to tell you, nobody really knows that questions. So we focus on the stuff that we know right now. I want to talk also about copyrights and all of this stuff. What happens if you create something with these models? Can you really do whatever you want? Can you upload it on Youtube? Can you like sell it? Can you do whatever you want with all of this stuff? And that's what we will look in the next video because you have seen that we can create a lot of stuff. We have used like 11 labs, we have used Chat GPT. We have used all these AI video generation tools. So we use also diffusion models. And that's why we also need to look at copyrights. In the next video, we'll look at copyrights. What happens if you create stuff with artificial intelligence? I want to make this a little bit broader. We look also on the stuff that you can create with Chat PT, on the stuff that you can create with any diffusion model. Because this is all the same. As soon as you start to create AI videos, you have stuff from a diffusion model. This make the pictures and videos. Of course you have stuff from maybe ChachPT because Chat ChPT writes your story. You have also stuff from 11 labs. This makes your voice and that's why we need to look at copyrights. Can you create every single thing that you want with AI? And of course, what happens if you upload some pictures? That's the stuff that we will look next later. Of course, we will also see how this can unfold with this AGI. But I have to tell you right now, we don't really know. All we know is that the progress is there and that we should use AI. 41. Copyrights: What can you make?: In this video, we talk about copyrights. What happens if we create stuff with I? Can we sell it? Can we do whatever we like with all this AI stuff? Of course. In the next video, we will also see how all of this works with our stuff. If we upload, for example, a picture of our face, what happens with this data if we upload like text from our business? Is this safe? But like I said, first we need to start. What happens if we generate output with I. Can we really use it? And the best article is here from AI. So we start with text. And don't worry, we will do the same thing with pictures also in this video. If we scroll down to the bottom, you see the copyright shields and Op tells us this Openai is committed to protecting our customers with built in copyright safeguards in our system. Today we're going one step further and introducing copyright shield. We will now step in and defend our customers and pay the costs incurred if you face legal claims around copyright infringements. This applies to generally available features of ChechiPT Enterprise and our developer platform. In simple terms, OpmayI tells us, don't worry, we will cover these. You can use our text wherever you want. And if somebody claims this is copyright protected, we will protect you. This is really nice. But keep in mind what they said exactly. Just if you are a user of ChechiPT enterprise or if you work via the API. If you work in the standard ChechiPT interface, most likely they won't step in if you get legal claims. But the possibility that you get legal claims is also relatively low because they are not interested to pay legal claims like every other day. They must seem really confident that this is no problem. And here comes the but also Openai has legal claims. For example, the New York Times, The New York Times made legal claims against Openai, that Opi trained on their data. They trained on data that was behind the paywall. Now they say some things sometimes and in the end of the day, opemyy got natus. But like I said, we need to be a bit cautious. Now op Maya tells us this is no problem. I am not a lawyer. And for that reason, we need to assume that this is correct. But what is with pictures? What is with pictures out of stable diffusion, out of mid journey, out of Adobe Firefly. These are the big players. Let's just take a look. This right here is the fastest thing from Micherney, and this comes from the Michurney Help Center. Michurney subscribers own all the images they've created, even if their subscription has expired, and they are free to use those images however they like. There are two small exceptions. If you upscale an image of another user that upscale is owned by the original creator, not by you. It will be appear in their gallery on the website instead of yours and you need their permission to use it. This is number one And number two, if you're a business grossing more than $1,000,000 a year, you need a pro plan. That's basically it. If you're grossing more than 1 million a year, you should take the biggest plan and then you are ready to rock. Marnie, also tell you more stuff right here. If you go on content and writes. Also mine tells you this right here, Please consult your own lawyer if you want more information about the state of current intellectual property law in your era, Basically, yes, They say you can do whatever you like with the pictures, except with these two rules. If you do more than 1 million a year, just use the big plan and don't steal stuff from other guys. That's basically it right here in mine. But we have one small exception. Let's just assume we make a picture of Donald Trump that eats Nutella in a BMW and Mickey Mouse is also in this car. This is not the smartest move to do if you want to sell this picture because there are a lot of companies, and also companies have the copyrights of disguise. Mickey Mouse is owned by, I think at least by Disney. Nutella is owned by Ferrero. Donald Trump. I don't think that Donald Trump would like these. Bmw is also a company, and you can't use the company names in your content, Just do any stupid stuff. Come on, we don't have to talk about this, but we need to take this into consideration. Yes, you can make pictures of Donald Trump, of Elan Mask, of BMW, if you can of the Mickey Mouse. But don't try to sell this, So don't go out there and print like Mickey Mouse on one gazillion shirts. This is not a smart move to do. I don't think that Disney will love you if you do that. And that's also basically it about Journey. And here comes Adobe Firefly. Because Adobe Firefly is a little bit special. Yes, you can basically use it commercially. But this article, this here, is the special thing. If we scroll down a bit, this right here, I want to show you generative AI. As with any, AI is only as good as the data on which it's trained. Mitigating harmful outputs starts with building and training on safe and inclusive datasets. For example, Adobe's first model is trained on Adobe stock images, openly licensed content, and public domain content where copyrights has expired. So Adobe tells you you can sell everything whether you make with Adobe Firefly, for example. You don't need to worry about copyrights because like you can't make pictures of Mickey Mouse. Because like Mickey Mouse is protected by the copyrights. And most likely the Adobe models are not trained on this stuff. This is the coolest thing about Adobe Firefly, so you don't need to worry about this. And then you have stable diffusion. And you know, stable difusion is completely open source. That means you can simply do whatever you like, but always take into consideration to don't do stupid stuff. So think of Mickey Mouse. Don't try to do these things, but you can create whatever you like. Then you have also the tools like 11 Labs, the video creation we talked of about a lot of this stuff. And always do the same things theoretically. You can use every single thing that comes out from AI and you can make money with these. You can use it commercially, But don't do stupid stuff. So don't make like deep fake voices and deep fake videos of Donald Trump, how he does like stuff that is not correct and try to sell these videos. So just work normally and then we will not get in trouble. And like I said, I am not a lawyer. This is the stuff that the company tells us. And that's basically also what I do, and that's basically also what I can recommend from my position. In the next video, we need to take a closer look what happens with our stuff. What happens if we upload data? Will this data be spreaded all over the world? 42. My own safety: will my data be in the Web?: In this video, we will talk about your data because also this is really, really important and here are a lot of nuances and just like I said, I am not a lawyer and I am relatively sure that this stuff will change over time. But as you know, I will include updates. If this happens first, let's just start what happens if you are in this normal GBT interface. Let's just assume you are a business, or you have a business, you have employees and much, much more. But you need to take into consideration, if you type in stuff right here, I can and probably will train on this data. This means if you type in here stuff from your business and you don't want that, pay trains on this data and you don't want that. For example, this data shows up somehow somewhere in a feet of other people. Then don't type it in. This is really, really important to me, so you need to take this completely in consideration. But, but, but we can do a lot of stuff to make this safe. Number one, you simply press on your profile, you go to my plan. And this right here, this is the team plan. If you use the team plan, you are safe. You can simply make an account. Yes, this is a bit more expensive, but you can also bring in other people from your company. And I will not train your data, but I want to make this crystal clear. If you have the normal GPT plan, pay can and probably will train on your data. If you have the team plan Pi will not train your data, but you have more options, you have this other option and you can simply press again on Settings and Data. This time here you can press on data controls. You can deactivate, chat history and training. Now in theory, hope I will no longer train on your data, but this is not 100% confirmed and you should always press and learn more because this stuff can and probably will change. But we have another solution to make this completely private. We have this open I playground and every single thing that you do in this playground will be completely private. Openai will not train on your Data API and playground requests will not be used to train models and always press learn more because like I said, this thing can and probably will change here. You make API calls and you need to pay per do you need a billing account? And of course you can make the billing account right here on settings and usage and you can and probably should also give rate limits, keep that in consideration. You already learned this, but I want to make a summary so that we understand how all of this works. By the way, every single thing that you do over there API, so if you train a chat bot and much, much more, your data will be completely safe. And if you want to build a GPT that is private, do it here on the assistant API, press on Assistant. And here you can, of course, also upload your files. Also upload your knowledge and Pi will not train on these and that's basically how you can work completely safe with Chat GPT. There's also the possibility to use the Che GPD Enterprise Plan. Yes, they have a few plans, Che GPT teams and Che GBD enterprise if you want to work in a normal GPD interface or if you don't want to do this work over their ABI. Now let's talk about all this other stuff. So not large language models, even if large language models is maybe the thing that can do the most damage. But let's just talk about pictures. What happens on mid journey? What happens if you create pictures or if you upload pictures from your own face? This gets interesting because every single picture that you create can and probably will be in this Explorer page. And people can and probably will find these. People can and probably will take your prompts and run them themselves and then they have your pictures. Yes. That's simply how it works now. Yeah. You can also make a strong strong plan in mid Tourney. If you don't want to do this, just use the strongest plan and you are protected. You can physically the only thing that you can do inside of mid journey. If you want to be private in mid journey, use the strongest subscription. And always make your own discord server or work on their web page. If you work on their webpage or on your own server with the strongest plan in the private mode, you are safe. That's at least what Micherney tells us. And like I told you, please read for yourself. Just go on the articles on mine. On the Q and A, I am relatively sure you will read the same stuff. But we have also other tools like for example, stable difusion. Of course, stable difusion is open source here, you have absolutely no problems. We have worked most of the time in a Google Cop notebook. Every single thing that you upload in a Google Cop notebook will get deleted as soon as your session is over. So don't worry about privacy in Google co lap And with stable diffusion, this is basically the only thing that I can tell you. Stable diffusion is safe. That's just something that you need to keep in consideration. We work here like on the Internet, and if you post something, if you throw something into the Internet, into the cloud, the possibilities are high that other people see this. Now, what's my solution for all of these? I only use stuff in I tools that is okay for me. I type in only stuff that is no problem for me even like to post on Facebook. I only use pictures that are also no problem for me. If I want to upload a picture into Che PT Journey or another tool, I always think about it. Would this be okay for me to post on Instagram, to post on Facebook, to post on Linkedin? If yes, no problem. If no, I don't do it. These are simply the rules for me. You need to search the rules that work for you because copyrights are a little bit crucial and they change all the time, and I am not a lawyer. You need to keep this in mind. Now let's forget all the copyright stuff. Just have someone using I tools and don't think all the time about lawyers and copyrights. At least I don't do this all day long. 43. Will AI take all Jobs?: In this video, we will talk simply a little bit about AGI. So artificial general intelligence. The point as soon as AI is better in most tasks as humans about maybe a super AGI. So the point where AI is better than all humans combined what happens with chop outlook into the future and what all of this means. And of course I don't have a crystal ball now. First the brightest heads of the entire world do not know what happens. All the CEO's of these companies like Sam Altman, El, and Mask, the Microsoft CEO Nadella, Mark Suckerburg from Meta. All of these heads, they don't know what exactly happens in the future. But all of them think AI is the next big thing. All of them think AI will change the world. All of them say AI will change every single job on this planet. And not everybody says the same stuff Now, Some of them think AI will change everything radically, and others just think that AI will simply change the job a little bit. We simply don't know it exactly. Also, humans are not good to see in the future. If you asked the people a few years back, what jobs will AI replace, everybody would have said blue collar jobs. Right now, the completely opposite is happening because AI takes white collar jobs. The creative stuff like writing, making illustrations, painting, making photos, and so on. In all of these, the cases, AI gets better and better and better, and is also better than like 99% of the people. At least. Ai is a lot better than me in writing, in painting, in making pictures. We simply don't know what happens exactly. But we will know that AI will change a lot of jobs. And our job is to adapt to these. We simply need to learn to use. I, Sam Altman, the CEO of Pei, just told us, yes, jobs will change, but jobs will absolutely not go away. A programmer can now program like roughly three times as fast, but the demand for code also goes up. This is a good thing because we need this code, and all these programmers simply can become more efficient. Of course, we need to learn how to use this. The same thing may be also with FSD, With full self driving, maybe I don't know when, but maybe as soon as this get online, people will just be more productive. They can go at work and as soon as they are in their car they can be productive. They can do stuff on their phone, on their laptop, or even like chill and the watch some Netflix or whatever. You get the point. Yes, AI will change a lot, but we don't know exactly what happens as soon as AI will hit. Yes, jobs will be replaced. Jobs will be different, but we don't know exactly how and when this happens. Yeah, maybe some jobs, they go away completely. I think maybe of a voice over artist. I don't really think, to be honest, that we need him right now because a lot of tools like 11 labs, they are nearly just at the edge at matching them. Yes. Right now they are maybe not better, but they are, of course, faster and cheaper. If we use stuff like Adobe to enhance the quality. Also, the quality is pretty damn good. I think like voice over artists, they are not needed like forever. But this is no problem if you are a voice over artist, just hit me up and let me know what you think. Now, I don't want to say anything negative about your job, if you are one. I'm just be cautious to understand what you think about the situation because everybody thinks different. Of course, basically the same thing with programming. Programming will change radically, especially if we reach a point where AI can program and improve on itself. Also, right now, there are studies out there that show that GPT four can make good rewards for GPT. Basically, this is like the basic concept that we learn, so the reinforcement learning from human feedback. And we have basically studies that show that GPT can reward these AI models better than humans can reward them. They learn better. If we let GPT to decide when and how to reward them, just go a little bit further. This machines will do this completely on their own. In the future, this stuff starts to get better. What happens as soon as we reach like super AGI GI that is smarter as all humans combined. Now first of all, I don't know if this happens, when this happens. And nobody can tell us these also, nobody knows what happens. We can't really imagine that. We can't figure it out. Because like this, AI would be smarter than all humans. This AI would think about stuff that we are not able to think. Yes, this is scary, but maybe everything will turn out so damn good that we can't even imagine it right now. All of this is pure speculation. Don't let anybody tell you that he knows what comes in the future. You simply can't know if this stuff all plays out. If we reach a super intelligence, nobody ever on this planet will know what happens. Because the superintelligence will be smarter than every single one of the humans that try to tell you what happens. So you can't know it. It's just physically and mentally, it's not possible. We have also like Algos also right now, this star thing. The star thing was around in 2023. I think there was basically an algorythmus that was able to understand math like on a deep level and it was all just speculation, to be honest. But what if Algo comes around the corner that understands math? Like really, really, damn good. Maybe that understands math on a whole other level than humans. This Algo can simply break the Internet, break, block, chain, break, crypto, break the stock market, like break every single thing. So we simply don't know what happens. But until up on this point, every single thing on Earth got better, Everybody has to eat at least most of the people on the world. So let's just talk general. So more people in this world have to eat than ever before, I think, I hope. And basically we need to believe this because we are optimists that all of this gets better and better. People like El and Mask also talk a lot about a basic universal income. So AI can work on its own. Ai will create value and so humans can work less. And there are still resources for all of us. Maybe we go back just a little bit, like a few hundred years ago, people just did stuff to have something to eat. The rest of the day they simply laid around, had some fun, and so on. Now people are working like 2047, Maybe we can slow down a little bit also. This would be nice, wouldn't it be? Let's just see with bright eyes what the future brings. At least I am optimistic. And let me know, what do you think? 44. Ethical AI use and the Downside of AI: In this video, we talk about ethics risks and also the downside of AI. I will simply show you some bullet points and you can also let me know what you think of all of these. First of all, what's an ethical AI? So guidelines for ethical AI. Of course, all the companies that develop AI, they try to do it ethically. Openai, Meta, Google and so on. They all try to be ethic, they try to use training data, that is okay. They try to make the models in such a way that they are not biased or at least not that biased. We have learned that models are always biased, but they try to do their best. But what's especially important, at least for me, transparency, fairness, non harm, accountability and privacy. And what you can basically do so that it's more or less what these companies can do. But what you can do, just work with common sense. You already heard it. Just make stuff that doesn't harm people. That is my advice for ethical AI. We don't have to overthink this, the companies will do a lot and we just work with common sense and don't do any harm. Let's just talk one more time about data protection because I got a lot of questions about these, even if we just covered these. Data encryption and anonymization is of course really, really important. This is also done by the companies. Yes, it's possible that ChechiPT trains on your data, but of course they try to do this private and not public, your content. I just want to tell you you need to be cautious because like you'll never know. Pay attention to security when creating. I like if you make a chat board, if you do something, always try to use passwords wherever possible. This is also like just common sense if possible and if you don't want to do stuff on the Internet in the Cloud, whatever, install things locally. If it's possible, no connection to the Internet. Don't no worries. Use things like ChachiPT Enterprise or ChchPT Teams or at least disable the chat history. This is also some stuff that you can do if you don't want to share your data and of course you can work through their API. We already covered these, but like I said, I got a lot of questions, so we need to address this. Again, before we talk about straight the risks and the downsides of AI, I want to talk about something that is like in between autonomous vehicles. This is an enormous potential. We can not only save time, but also save human lives. If we get to a point where an autonomous vehicle drives better than all humans or than most humans, we will save human lives. But of course, it's not easy and it's challenging. And we have more challenges than just the challenge to get there. Because legal framework and ethical use cases in decision making. This is, I can't describe how hard this is, just think about this yourself. Just think, for example, if an AI needs to decide what it should do, maybe you are in your car, maybe you with another guy. So there are two guys in the car. And then there's another guy on the street. And the AI needs to decide either to basically like kill the guy and smash into this guy. Or go into the ocean or wherever, or into another car. And like risk the life of you and the guy next to you. Now what the AI should do in this use cases, let's just assume that the AI knows exactly either you need to die or another one needs to die. Or let's assume the AI needs to decide if a, a little child has to die or two older people. This is stuff that nobody can really tell. I think this is enormous difficult to get. This is really enormous hard to solve. All this decision making is really hard to solve. But we can also make decisions that are a bit easier, or at least that are not deciding life and death, for example, in the business world also. Here we have really big downsides. Let's just say we use AI for data analyzing and to support decisions such as risk management or optimizing supply chains. What if the AI gets it wrong? Yes, this is possible. Let's just say you analyze really large amounts of data, you don't control it, and then you are in complete trouble. Yes, even if the AI will be better than most humans. As soon as this stuff happens, people will hate AI because like nobody wants to pay with their money. Machine learning assists also in predicting customer behaviors, aiding in understanding market strategies. And the same thing right here, of course this can help. This will help basically this helps also right now, just think about Netflix. They know exactly what you consume, and then they will give you similar stuff. Think about Youtube, but what if these machines make the wrong decisions? All this is also a downside of the AI. Now let's go straight to the risks and downsides of AI, amplification of existing discrimination when trained with biased data. This is huge speak and you also saw it. We can make a choke of men but we can't make a joke of woman. I personally, I have no problem with like jokes about men, of course also harmful purposes. For example, deep fakes. You know how to make them. Police do this with common sense cyber attacks or for surveillance, for example. Incorrect decisions that could lead to financial or physical harm. Wrong investments or medical instructions, car causing accidents and so on. We already heard this. Just think about if an AI starts to be a surgeon. If I like tries to fix your hand and an AI makes mistakes, all of these are enormous problems. These problems are not easy to solve. This right here, this is for me, like this is big. People get lazy. Nobody talks about this. I think of myself here because as soon as I have to drive with my car, I am always lazy. I just used the navigation from Google. I don't know where I need to go, I simply used my Nave. If I think back a few years when my dad, like he was used to drive, he always knew where to get. He doesn't need it a navi, like because there was no navi in that time. People just knew where to drive, when to drive, and how to drive. I myself, I am completely lazy. I'm not interested to learn this stuff because like the navido does everything, we get completely lazy. And if such a technology doesn't work, like I don't know where to go. Yes, this is also big, this is also a downside. Same thing basically also with all these AI tools, like if you never ever write an e mail again, you don't know how to do it, you can't do it without these AI tools. Laziness is also a big thing. And that was basically all the stuff that I wanted to tell you in this video. Yes, maybe a little bit boring. Yes. Maybe a little bit different. But I think we need to address this also. The downsides of AI there are there and we need to take all of this in consideration if we use AI. To summarize the use AI just ethically. So just use common sense, don't do any harm. And keep in mind that there are downsides. If you use AI for every single thing, you can do it without AI. Ai, of course, is biased and can make failures. Not every single thing of AI is completely true, and every single AI tells you that even chachiPT like down here, they tell you, hey, our output, it's totally possible that all of this is wrong. Take this in consideration. 45. Recap and a Work Assignment!: So this section was a little bit special. First you saw how AI videos can eventually simulate words. And we can use the simulation to train robots, and that can eventually lead to AI that is really, really smart. Then we took a look at the copyrights, what you can create, what you can not create, and is your data safe. We also took a look at AI, or AGI in general, and the downsides of AI. This section was a little bit broader and of course, you also know what learning is, learning is doing, and you get better by doing more. Learning is same circumstances and different behavior. And what you can learn from this section, first of all, of course, just take a quick look at the copyrights. Don't do any stupid stuff. You know that you should not create stuff that is copyright protected, like Mickey Mouse or even people don't do that. You know it, you don't do it. You have learned, you have also learned that AI and AGI is potentially big in the future and you know what you can do. To stay in the loop, just play a bit with these tools. You are on the edge here because you understand how all of this works. That's the best thing, and basically that's the only thing that we can do, and that's my recommendation. Please stay in the loop, play a little bit with these tools and just take a look at the copyrights. Don't do any stupid stuff and you also know how you can learn better. You learn better together. Please share this course with someone where you think he would benefit from this course. This is really cool, because not only I have a benefit, so you have a benefit and the guy that comes in. Because if you tell this guy to come in and he gets a value out of this course, he will describe the value to you, Your status rises. Thank you for that. I hope we will learn better together. 46. My Thanks and What Comes Next: Congratulations, you did it. You gave me your most valuable asset on this planet, Your time. Thank you from the bottom of my heart for this. This is the last lecture and you have done it all. We have started with the basics of AI videos. What's a diffusion model? How all of this works? And much more. We even covered what a video is, it's simply a lot of frames. Then we took a look at the easiest tools. We started to create our AI videos in collapse, in runwaML, in hyperim. After that, we took a look at AI avatars. They are cool and they can scale your business. We started to get a pro because we used Google Colab, stable difusion, and much, much more to make deep fakes, to use stable P fusion deform diffusion. And this is really the hard stuff, but we can do enormous things with these. Animations are cool. We took a look at how you can use all of these. So you can even make your own projects just like determinator video. Just combine these tools like Ka and edit it a little bit in shortcut. Cap cut, Davinci resolve or whatever you like. Then we even took a look in the future, you have seen these models. They can really simulate the world. And these simulations, we can use them to train robots, and this is maybe the future. We can also make video games like this. This is cool research and I think we should totally look at these. You have really seen it all. You also saw what you can make and what you cannot make and what happens with your data. You know every single thing that you need to know about AI videos. I have a little bit of homework for you. Please just stay in the AI is here to stay and we need to learn about it. If we learn about it, we can really take advantage and not just cry that it's there, but you took the right decision. Because you are in discourse, you are a learner. You also know what learners do. Learners use the same circumstances but use different behavior. They use AI tools. They also know that we can learn better together. I have two things that you can do for me, for you, and for all the guys out there. Leave a nice five star review and we are all happy. Thank you for that. And if you share the course, more love goes out in the world. Thank you.