Find The Next Insane AI Tools BEFORE Everyone Else (Pt. 2)
Table of Contents
There's not really an area for text prompts yet, but you can put a music prompt in here so you can upload your own music into the audio file box here. And then there are a ton of styles that you can play with. Let's see what lighter light igniter igniter does. And then pulse react to it was reacting to the percussion. This time let's have it react to harmonic. The motion react was reacting to harmonic. This time let's switch that to percussive. So those two are flipped. You can mess with some of these numbers here. I don't totally know what all of them do yet. Some of these tools, today's the first time I've explored them. This is all very, very early stuff that not many people have played with yet. You can see this one's had 280 runs. So let's go ahead and generate this.
sort of gets close to looking like a real image, but then before you can kind of make out what it is, it shifts into another image and it never actually looks like the real thing that it's supposed to look like. But it's pretty cool. It makes these like really interesting psychedelic images. And again, there is just so many styles here. I mean, I would say there's hundreds of styles that you could test and keep in mind though, that this did take a while to generate from the time I went and clicked the submit button down at the bottom to the time the generation was finished. It was probably a good 15 minutes until the generation finished. Just keep that in mind when you are playing around with it, that this one is a bit time consuming to get this final product here. All right. So the next one I want to show you is called motion diffusion model, and it's actually a text prompt to a little animation. You can see that the demo here is the person walked forward and is picking up his toolbox and it generated this little animation here. This person walked forward and is picking up his toolbox. If you look down here, they've got like these little models. These ones up here are stick figures, which you know, you might be able to use eventually with the control net pose feature and some of the future stable diffusion animation stuff that's in the works. I'm not totally sure how to get them to generate in this style here, but let's go ahead and try one. Let's say the person jumped with their hands in the air. Let's go ahead and submit that and see what happens. All right. So on this one, it didn't actually generate the video output here. We got to download it and then take a peek. And here's what it did. The person jumping with their hands in the air kind of makes it look like they're doing jumping jacks or something there, but pretty cool stuff. We're getting to a point where you can literally enter a prompt and it makes a video animation out of that prompt. Obviously it's still very, very early in what we're generating our stick figures essentially. But if this is where we are now, imagine where we're going to be a few months from now. So the next one I want to show you on replicate is this one called tile morph. This one, you set a starting prompt here, and then you set an ending prompt and it will actually create a video that merges from the first prompt to the second prompt. But because this one's called tile morph, all of these videos are tileable. So you'll notice that this left edge over here lines up with this right edge and this left side over here lines up with this, right? So if you're to take this video and tile them all together and duplicate this video four times over, they would all line up and look like one giant image. This is the prompt, a portal into another world space explosion of color, unreal engine ending in this giant prompt here, a spaceship taking off from a distant planet, epic scene displaying in the background to the moon, blah, blah, blah. Let's just see what it does here. You can see that it will slowly start to morph into the other one. But if you look around the edges here, you'll notice that the morphing is always done in a way where it would be tileable. Now let's do something a little bit more random. Let's do a wolf howling at the moon, ending with a robot in a cyber punk city. This should be interesting. Scroll down here. Let's click submit. All right. So this is what it came up with. You can kind of see like maybe a outline of a wolf on the edge here, right? And then it sort of merges and moves into, maybe there's like a wolf here in the center if you're looking closely and then it shifts into this city here. And really the big important thing to note about this is that all of this tiles. So again, if you lay all of this out edge to edge to edge, it should create a seamless pattern on the entire video. Again, very, very early. Now I got two more cool tools that I want to show you here on replicate, but before I do, I want to tell you quickly about our sponsor. The sponsor of today's video is open CV and they've got an awesome Kickstarter going on right now called mastering AI art generation. And you can find it over at future tools.io/openCV. And this course is a complete deep dive into everything AI art generation.
Fundamentals of Generative Al and Art Generation
They cover the fundamentals of generative AI and art generation.
Al Art Generation using OpenAI DALL-E-2
They cover AI art generation using open AI and Dolly too. They cover mid journey. They cover stability AI and dream studio.
Generating Infinite Images with Stable Diffusion WebUI
They cover automatic 1111 and the web UI for stable diffusion.
Diving Deeper into Prompt Engineering with Stable Diffusion
They go into some deeper dive prompt engineering with stable diffusion, some controlled editing of images where you can edit out specific things in your image using AI, inserting objects into your image using masks in painting instruct picks to picks virtual try ons, virtual hairstyles, virtual makeup. All of this stuff is covered inside of their deep dive course. And if you're not familiar with open CV, they're actually the people behind a lot of the computer vision tech that is used today. They've also worked on things like AI smart cameras, and there's some of the people behind people detection in a lot of the computer vision tools that are used today. So this isn't just a random company making a course on AI. These are pioneers in the AI space teaching you how it's done. So check out their Kickstarter campaign. Again, you can find it over at future tools.io/openCV and they have various levels that you can pledge at to get involved in their course. So check out their Kickstarter, which is raised over $100,000 already and find the backing that works for you. Again, you can find out more at future tools.io/openCV. All right, let's get back over to the tools. So the next one I want to show you is called 3d photo in painting. And you can see the example, they have this kind of psychedelic looking image here, and it generated various styles of video off of one image. It created this one where it's kind of rotating weirdly here, this one where it's kind of leaning in and out another one where it's sort of swaying side to side here, and this one where it's kind of zooming in and out. And all you do with this one is upload an image and that one image will generate all of these various videos. So let me grab my little AI generated wolf here. Let's drag this guy in. And let's go ahead and generate this and see what it comes up with. All right, and here's what we got. We got our wolf kind of zooming in and out and rotating, you can kind of see a little bit of weirdness up at the top during one point right there. Another one where the background zoom it out and the snout gets real long. That looks kind of cool. Actually got a little sort of rotation right here going on. It kind of reminds me of that Leia pics app that we messed with in a previous video, but with various styles of motion and some funky zooms and stuff like that. And the last one I want to look at is called tuna video.
And again, this one has not been run by a lot of people yet here. And you can see what it is, is you upload an input video here, and then you give it a prompt, you'd say what the video is, a man surfing, and then you give an output prompt. And in this case, they gave two outputs, a panda surfing and a cartoon sloth surfing. And you can see the two outputs that it gave right here. Here's our panda surfing and our sloth surfing. Let's say we want to change this to a green alien surfing. We have a handful of other settings that we can mess with here, but I'm just going to go ahead and leave them all on the default and click submit. And let's see what it looks like if we have a green alien surfing. And there we go. There's what it thinks a green alien surfing looks like. So again, some of these have a little ways to go before we're like really ultra impressed with what they're putting out. But again, this thing's had 378 people run it on replicate right now. These are still very, very, very early sort of demos of what's coming. So pretty cool stuff. Head on over to replicate.com, create a free account, click on the explore section and click around. See what you can find all sorts of new AI tools and technologies are popping up that you can play with right now and get a jumpstart on the rest of the world. Right now. AI is really truly in its infancy. We are at the very, very beginning. You feel like there's been a big wave of AI stuff now, just wait, it is going to explode as some of these technologies get bigger and bigger. And if you're hanging out on places like hugging face and replicate.com, where people are actually experimenting with these new tools and these new technologies, you're going to be so far ahead of the curve of everybody else. And that's what I want to show you with this channel. So hopefully you learn something and are ready to go experiment with some cool new technologies that nobody else has really played with before. And if you like all this cool tech stuff, head on over to future tools.io. This is the site where I curate all the cool tools that I come across just cracked a thousand tools at the time of this recording, but don't worry if a thousand tools is too much for you. Click on this little button here to join the free newsletter. And every single Friday, I'll send you the TLDR of the week. I'll show you just the five coolest tools that I came across a handful of news articles, a handful of YouTube videos, and one cool way to make money with AI. And I said it every Friday, all you got to do is go to future tools.io and sign up. Thanks again for hanging out with me. I really, really appreciate it. If you haven't already make sure you click the subscribe button. If you enjoy videos like this and give it a thumbs up, if you give it that thumbs up, it'll make sure you see more cool AI videos like this one, right in your YouTube feed. So thanks again for tuning in and thanks again to OpenCV for sponsoring this one. Appreciate you guys. See you in the next one. Bye.