 Our world is getting horrifyingly advanced as far as technology goes and as far specifically as AI goes. Everybody obviously knows that things can be just like created things that aren't necessarily real, whether it be like VFX or just like basic photo shops. Now with deep fakes, which are becoming not only increasingly better but increasingly easier and more convincing. Even stuff like live facial tracking with Snapchat filters and the Instagram filters and stuff, it's all getting crazy. And something that's getting really, really crazy is AI produced images. The biggest one right now that's like the most popular amongst the general public is probably this one, which is Dolly Mini. So basically what you do is you, it's an AI model generating images from any prompt. So we just give it a prompt and then it uses the data that it has to create an image on its own. So let's see, let's do something very basic. Red barn with a horse standing beside it. Okay, let's click run. Let's see if this actually works because most of the time, Dolly Mini has too many users on it. Okay, here we go. So now we have different images that this made of a red barn with a horse standing beside it. I mean, this honestly already is extremely impressive to me that it can just take these words that I wrote and make its own image of like, it's not perfect. It's fucked up looking obviously but you can very clearly tell that this is a red barn with a horse standing beside it. It's crazy. It's wacky. It's kooky crazy, bud. Let's see what happens if we just do Crank Game Police. Dolly Mini is not the only AI image generating platform out there. There's a bunch of them that are coming out. The most advanced one, which you have to join a wait list and very few people are getting in, is Dolly E2, which apparently is like horrifying, like just what you can make. And the reason why it's really scary is that one, anybody can just sort of make any image without having like any skills at all. And so things can obviously very easily be faked. But two, it's obviously very, it's very scary for different artists out there. Like specifically when I am thinking of this kind of stuff, I'm thinking of storyboard artists and people who start the process of like a creative collaboration. You know, there's so many people out there that make concept art and things like that. And so this I feel like is the most scary for those kind of people in that field where it's like, okay, we're just trying to come up with a loose concept. You know, why hire somebody when we can just use this AI thing and it spits out an image. This is awful. Oh my God. Look, it's me guys. So bad. But yeah, I feel like it's really scary for like concept artists and storyboard artists and stuff like that. But today we're actually not gonna be looking at Dolly Mini. We are going to be looking at something that is a bit more advanced than Dolly Mini, which is mid-journey. So mid-journey is obviously another AI image generation platform, I guess you could call it. So these are some of the images that I have given it prompts for and it has spat out images. So those were all loose prompts that I gave for all of mine. These are all images that are made with the same thing, but people are just giving it different prompts and you can actually see what they use. So like for this, Japanese high school girl, cool beauty, short hair, sailor suit, anime, clear symmetrical face portrait, dark atmospheric lighting, there's a bunch more tags. This, ultra-realistic photography, close-up shot of a mid-century, myth-green dragon eyes within, with thin yellow pupil. Like these, oh my God, Mr. Bean anime, dark atmospheric lighting, volume metrics, manga style, art station, black and white line art. That is insane. Like this is a little messed up down there where the neck is, but like that's very clearly, Rowan Atkinson, that's Mr. Bean. It's insane that AI made this. Oh, it's incredible. Like the detail is insane. It's so damn cool. Oh my God, Kermit the Frog on the front lines of Normandy on D-Day. And so AI-generated images are something that is a gray area in a lot of different ways because basically the way that it works on a very basic level is that the AI is being fed a ton of information, a ton of information as far as like different photos, different drawings, different blah, blah, blah, blah, blah. So there's a ton of gray area when it comes to this as far as like how these images can be used. Mimika and Mark and Amy went out to dinner last night and we were talking about this specific thing and Amy made a really, really good point, which is like there will be a time where AI makes something that is too similar to something else. We're like, for example, like you give it a prompt and it just spits out like the Mona Lisa because you're just giving it a ton of information. Like that could happen. And so, you know, it's kind of a gray area where it's like, you know, it's taking reference and it's taking images or information from other, so many different images that at some point it could just completely remake something or make something that's close enough to something where it's weird as far as like copyright law goes. But anyway, I'm getting a little bit ahead of myself here. I tweeted shooting prompts at AI to make horrifying slash funny art, gimme prompts. So basically how it works is mid-journey, you access it through Discord and stuff and then you just basically have DMs with the bot. So let's try this Ethan Nester trail cam footage. So we just do slash imagine Ethan Nester trail cam footage. And so what will happen? It'll start generating the image and you can see it sort of update live. So it's only 12% and then the image will get clearer and clearer and clearer. And so as you can see from up here, it'll give you four different variations and then you can make more variations on a specific variation. So it's like, okay, I like this third one a lot, make more variations that are based off of that third image or you can up res or upscale each of these and just get more and more detail. I think it would probably have been more like me if I did crank gameplays because I feel like crank gameplays will have more images than Ethan Nester but we'll see what happens. Okay, so these are the images, which is like weird, like this isn't a real person. These aren't real people, you know? It's so weird. Okay, so V3, let's make it, let's make more of V3. Like, isn't that odd? It's like a weird thing where it's like, I recognize that these are people but these aren't real people. This computer just made these. It's wild dude. It kind of looks like super mega's editor, Justin, nothing but lag. All right, so now we'll take that, we'll up res U2 and it'll make it even bigger. And again, the prompt was only Ethan Nester trail cam footage which is giving very little detail into what it is. And so like the AI does not have a lot to work with but let's go back to some other prompts that you guys gave me and we will get a little bit more stuff. So here's that, gosh. Let's see what we have here. A lot of trail cam footage. I want one that is more detailed. Let's see what this does. It may or may not let me because there's knife and there's certain words that are blocked. So like if I were to do like blood or something like that I think knife will work. Yeah, so minion chasing after you with a knife through the back rooms. So you can't like do blood. You can't do like any like nudity or anything like that. Okay, well, God, just the preview of this is horrifying. I don't like how this one is in the back. Like how the back is facing us. It's so ominous. Okay, here it is. Oh, oh, oh, see that first one. Oh, there's like a tiny man. Damn, that's cool. Okay, let's do, let's do V1. We're gonna let it do another variation and let's look through another prompt to see again. What are everybody saying in the back rooms? What are back rooms? Ooh, this one's pretty good. Dwayne Johnson playing chess with the cookie monster at the DMV. That is fairly detailed. God, here we go. Oh, okay. Let's do that. I'm gonna add some more stuff here. Atmospheric lighting, white tile floors. Cookie crumbs on table. Let's do realism. Yeah, let's see that. Let's see that. That gives it a fair amount of stuff to work with. This is gonna be awful. Look at this though. Yucky. Wait, it looks like he's like on a zoom call with him. What is going on? Oh man. That's the other thing too is I think with specific people probably has a harder time making them. Cause if you just give it like person playing chess with the cookie monster, then it can just kind of make whoever. That's pretty good though, honestly. God. Oh, that first one. Hold on, let's do a V1. God, this is, oh, that bottom one is horrifying. Let's up res the third one. There it is, baby. What else do you got? You got, your prompts are too loose. Give me something. Give me something real detailed. What is the lighting like? What are the textures like? What art style is it in? Gimme deets. Oh yeah. People are giving deets. Tell us, I'll be ascending to heaven holding hands with the devil with lighting from the TV. Wow. That's good. That's good. Let's see what we got. God. What are some other prompts we got? This one is also really good. Pop art Jesus. That's also good. What the fuck? What is that last one? There's text. Tell us some driven. What the fuck? That's horrifying. God. God, that is so fucking horrifying. What does that look like, Tyler? Hold on. Hold on. Up res you too. We're going to up res version two. Why does that literally look like fucking Tyler? What the hell is going on? This is horrifying. God, and like the upside-down cross candle? What the fuck? When it was all blurry, it kind of looked like Tyler. This is a dope album cover. Look at the detail though. The detail and the fuzz. It's fucking horrifying. Dude, what the fuck? AI made this. That is goddamn wild. OK, we're doing another one. This one could be really good. Gothic Godzilla with blue lightning in the background, but he's wearing a black cowboy hat and in the middle of a stormy desert. Ooh. Ooh, this is cool. This is cool. Let's do the third. Let's do the third. God damn, that's pretty cool though. I like the third the most again. So let's up res number three. Damn, dude. Gothic style. Godzilla with a cowboy hat. This genius. I want to just see how good it can make me. Crank gameplays, staring, white hair, even lighting, small, smile, green sweatshirt. Let's see what it does. I should have given it a style. It's just giving me a sweatshirt. It's just giving me a green sweatshirt. God. This looks like Matt Watson. That looks like Ninja. That's so odd. Let's do V1. Let's see if it'll do no background. It's so weird looking. And V2 actually as well. It's just unsettling. Oh. Oh. Oh. It just looks like Matt Watson. OK, I think of all of these, either one or three resemble me most. Oh, no. Oh, no. These last ones actually could have potential. I think this one is the most like me. Let's up press four. This is weird. It's going to be really weird. Oh. I don't like it. It's fucking creepy, dude. It's just it's so creepy. Like it looks just enough like me that it's extremely unsettling. You know, where it's like this AI just made a photo of me. Like, ugh, it's super creepy. Super creepy. OK, well, we have to do Pop Art Jesus riding a bicycle at midnight. We just have to. I know I'm doing a bunch of different versions of myself now to just see what it spits out. And it's not fun. It's like it's too good. It's fucked. It's fucked. Dude, they made Jared Leto. This is sick. That's second one. Hold on, let's do a version two. How does it do it? It's cool and fucking horrifying. It's so weird. While Pop Art Jesus gets ready, Justin, the wonderful editor, I don't know who's editing this video, but submitted this. So let's see what mid-journey has for this one. This will be the last prompt, giving it a lot of data. Let's see what happens. God, I'm afraid. There's too much. Oh, I didn't even see drawn by Junji Ito. That's going to be awful. Oh, it might just be Joe Biden. Damn, kind of cool. Kind of dope album cover. God. OK, well, we have to do this, the one on the bottom. Let's see what we get from that, because that one's easily the best. He combined. It kind of looks like JFK, who was turned into a bicycle. Uh-huh, uh-huh. Well, it's number two. We're going to up res, for sure. This is horrifying. Here we go. Yeah. Yeah, it seems about right. All right, friends. Well, I am going to end this here. AI can make some really cool shit, and some really weird things, and some stuff that makes me extremely uncomfortable. But that is the future. Again, mid-journey is private currently. You have to sort of be invited in. And then there's Dolly 2, which is coming out, which is also you have to hop on a waitlist. Dolly Mini, though, the one that I showed you at the beginning, it is free. And it is just on the internet, and anybody can use it as long as there's not too much traffic. But it is horrifying, a little crazy. And I'm not sure if I like it or hate it. So let me know if you like it or hate it. So thank you guys so much for watching. Hope you enjoyed. If you did, make sure to slap that like button right in the face. And I will see you guys in the next video. Love you all. Stay cranky. Bye. What is up my cranky crew?