Open transcript
This footage was thought to be lost forever. The production was plagued by strange equipment failures. Everything came to a standstill when a mischievous cat accidentally muted the speaker. The video was irrecoverable. Only the audio remained. What you are about to see are real events. Actual audio recording of what took place that day. The visuals have been restored by the magic of artificial intelligence. Hey, this is Wes Roth. I'm joined today by Dylan Curious. Today in this two-part podcast, we're going to be talking to Tim from Theoretically Media. We'll be talking about the future in which AI content is going to be indistinguishable from real life and what that means for all of us. I had a quick question. Actually, Tim, I want to get your opinion on this because a lot of artists are mad. They're very mad at what's happening with AI. It seems like a whole movement. Some of them, I've even seen people call for violence. I don't know if they're 100% serious or not, but bomb the data centers. It's the internet. Yeah, exactly. I can say anything. There does seem to be very strong emotions there. As somebody who's in that artist community, knows a lot more about this. I'm sure a lot of people watching your channel, a lot of them consider themselves artists. They produce stuff. What's your take on the temperature? Are people seeing this as something that's going to be good for artists or just absolutely horrible? I've watched the pendulum swing. It's always hard to tell because, again, it's the internet where you're going to get polarizing, bombastic opinions on the right and the left side of an issue, whereas I think that most people are honestly right in the middle. I've always kind of said that it's not going to take jobs away. If anything, it's probably going to allow for the ability for more jobs. I always talk about there was this guy that during my Hollywood days, there was this guy named Johnny. Johnny was a stoner. He would come into work every day, stoned. Johnny's job was to do rotoscoping. That's all he did. He cut out people from photos, removed the background, and that was his job. If you asked him to do anything else, nope. Johnny was a cool guy, too. Don't get me wrong. He was super into snowboarding. That was his passion. This was his job. Johnny is not going... That job is not going to exist anymore, but that's okay because that's not really what his passion was for. Anybody that rotoscopes for eight hours a day, if that's your passion, man, I don't know what world you're living in. A very simple one, I guess. But I think that a lot of those automated... The automated aspect of it, sure, that's going to be problematic for some entry-level positions. The opportunity is that... Sure, there isn't going to be... I was in Sony Animation Studios very early... Well, not early in, but during the making of Cloudy with a chance of meatballs, too. And they have this... The actual workhorse part of it is almost as big as a football stadium, and it's just lines and lines of standing desks. It's like a football field of standing desks that go all the way down of animators working, and it kind of looks a little bit Orwellian. A lot of that is not necessarily going to exist in four, five, six years. What I do think is going to exist is many more production companies at smaller scales. So you're just going to have smaller, like 10 to 15 people and or, you know, yeah, staffed animation studios creating more material, more clients, more content. I mean, is that good? Is that bad? I, you know, that's to be determined, I guess. But I think that when you bring in all of the available options to us, you know, be it on social media platforms, streaming platforms, I don't know, I guess those are the two. But, you know, that's a lot of space. And that's all there's a lot of content to be created in there. So I think that those jobs will still be available. And I think that the ability for three or four person teams, or even one person to just create content, movies and films on their own and make their own mark, that is something that we have not seen before. And that's going to be a possibility as well. How about your personal feelings on creating content? Like if you compare the thought of going out and filming with your iPhone or your camera, a few clips and putting it together versus prompting it into existence, do you feel the same kind of authorship? Or is something missing? I always push for that. That's, that's one thing that that's the fun part too, is like, I don't get a chance to do it as much because I'm stuck here shooting videos about the latest thing that's happening. But whenever I get a chance to, the what you can do with AI and your phone is kind of remarkable. There was a company called Sky Glass, they unfortunately just went under. They were doing, I don't know, one of the coolest things where you could shoot, and it would instantly rotoscope you out and put in like a 3d Unreal background behind you. So you can either prompt for it, or they had like sort of templates that you could use. And it would be 3d. So you could kind of like move it around, you would be in this environment. So you know, if you were to dress up in say, a period costume and be in a 1930s, like 40s, noir detective office, like you could shoot on a virtual set, it's basically like the cheap version of Disney's, the volume, the thing where they do the Mandalorian and all of that stuff. And the results are a little bit janky. But then you take that and run it through another AI video to video filter. I mean, like, you've got something looking pretty cool. Is it does it look as good as, you know, a $25 million feature film? No, but it's cool. And you got your story out. Yeah, I mean, the Mandalorian stage, that's like hundreds of thousands of dollars per day to use. And if you can recreate that with a phone, that'd be absolutely incredible. I'm wondering if you guys have what do you guys think about Rick Rubin making the appearance in as the champion of vibe coding? Have you been following that at all? I yeah, tangentially, like I've seen him kind of bouncing off the side. Yeah, that that's the most Rick Rubin thing in the world. Yeah. I yeah, I'm very, very fascinated with him. It's so interesting, because I briefly caught what was some of his interview. And for people that are not familiar, Rick Rubin is he's behind like, whatever your favorite band is, he's probably had some role to play. Metallica, Eminem, Def Jam Records, Slayer, like, I can go on and like, like the biggest things people have heard, you know, System of a Down recall, you know, Serge Tankian recalls this story where they had a hard time finishing Chop Suey, which was kind of one of the songs that really kind of propelled them to fame early on one of their like, big hits early on. And they were over at Rick Rubin's house. Rick says, Well, why don't we just why don't you go to my library? He had a lot of books. Have you heard this story? I have? Yeah, yeah. Just pull out a random book. And yeah, and find a line. It's amazing. It's so mind blowing. So yeah, so he tells Serge, go find a book, pick a book, he pulls it off, starts reading. And I guess it's either it's something biblical, like why have thou forsaken me in your heart? And that's the big kind of melodic line for and I was like, wow, I can't believe that that's. But my point is, like, Rick Rubin talks about saying like, he doesn't have any technical skill, he doesn't have talent, what he has is great taste. And so he's able to kind of like create this amazing stuff and really draw out of the taste for from these amazing artists. So in a way, he's kind of like what people do, I think, with AI models, they're able to get they're able to create that stuff, but you got to be able to draw it out. And you're able to you got to be able to hear the right thing have the taste to really like, what do you think about that? Would you consider that art number one? I mean, that's always a tricky thing. And I know the artists like to jump on that. I, I don't necessarily, I don't. Well, first of all, I don't really ascribe to the, the title of artist, like, you know, you are, I actually prefer to think of it as like you are producing something, you are producing something with AI. And if you think about it that way, yeah, the artist tag, I get it. Like art is, it's the whole thing of like art, what quantifies being an artist? Did you get paid? Does that make you an artist? Does that make you a commercial artist? At what point are you an artist when you pick something up and start drawing with crayons? Are you an artist? Are you not an artist? You know what I mean? Like is, is a tree with, with some sunlight dappling through it that you happen to catch in the corner of your eye? Was that art? Like, you know what I mean? Like, yeah. I mean, so there's, there's a lot of like, um, art, the word artist can be a very loaded term, right? Um, so I, I, yeah, I just choose to think of it as like, you are creating something, you are, you are creating something. Um, you are producing something. Um, you are making something. Yes. Sometimes it, sometimes it kind of even feels like you just have to go against the grain. You're like, it was art and now everyone else is doing it. And it's like, it's not art anymore. Now you got to do something else. Oh, I mean, I think that you can still make uninteresting art. Like it's very easy to make boring art or, you know, um, or just, I mean, I do hate the word content, but, um, that is, you know, the, that is the shorthand, uh, for all of this stuff at this point, no matter what you create, um, it's content. Um, it is what it is. Um, realist enough to know that. Um, yeah, if, I think that if you are like, I remember the days of prompt engineering, like remember prompt engineering courses, take my prompt engineering course. Um, those guys are all gone. Um, but I do think that there is like the Rick Rubin side of things, uh, the creativity in, um, with using say mid journey, um, is to take, um, like desperate ideas, um, visual ideas and know what to call them in order to prompt them to create something new. It's just like remixing. You know what I mean? Like, let me take a little bit of Indian culture from, uh, a little bit of, uh, Western, like literally like cowboy culture. Um, and then throw a dash of, uh, Vikings, let's just say in there and like, see what comes out of that, that becomes something new and interesting. Do you think all the genres will go down at the same time? I mean, now that we have these tools, do you think just like romance and action and Westerns will all kind of percolate at the same pace or do you feel like one will sort of kind of lead the way and a couple of them will sort of be the last? Well, that's a, Wes, you want to check that one? Think about that one. Yeah. I mean, for me, I actually had, I actually was thinking about this. It's funny within the last 48 hours, I think, I mean, to me, I would, if I had to bet on two genres that are going to be just killer in AI in the beginning, it's going to be horror and comedy. Like, I think that's just going to kill it because it's already, I've seen some very scary things that, um, like those, what are they called? Subliminal spaces or liminal spaces. Oh my God. AI generates those better. I think it naturally does it because the whole idea of a liminal space is something that, you know, we tend to assign meaning to. Like if you went to a classroom, you sort of like, this is the classroom where this is where we sit. You visit that same space at night and there's this eerie feeling because it's like out of context for your brain. And I feel like most of the stuff that AI generates, it's easy for it to generate those liminal spaces because it's almost like there's no context, um, for it, so to speak. Um, so I think that's one genre that's going to just be really, really good. Um, and comedy just again, because I just recently saw, Tim, I think you retweeted it. So it was, uh, old school, like 1980s, 1990s, um, um, news channel. Remember the, yeah. And it's like, it's like, it's like, what's happening? The guys on the scene is like, yeah, they're just, you know, there's a fire and they're like, Hey, you with the microphone, you know, help me. And it is. Yeah. So, I mean, it's easy to make silly, funny things like that with AI. So I would bet on those two as the main ones actions hard. I mean, we've seen action scenes. It's it's, that's the hardest thing. So that might take a more, more time. I think that, um, if I'll go the opposite route of what I think will be dying, um, is, uh, cyberpunk and sci-fi. Cause I think that those are two that they're the two very obvious ones, um, to generate. And I think we've all seen so many of them at this point that people are getting a little uninterested in rain, neon soaked streets and, you know, people in black leather, uh, with, you know, robot eyes walking around. So that's one that I think is that we will see slowly start to phase out. How about historical accuracies being brought to life? Do you look forward to a really smart system trying to recreate Rome or? Yeah. I constantly get this on the channel because it's not great at that. Like there's a lot of people that, um, like the world war two people who are super into world war two, like, I mean, are like, like they are like detail orientated, you know, it's like, that is not the correct uniform. That is not the correct airplane. That is not the correct wing for that airplane. You know what I mean? That's, and again, that's it's history. Like, um, I think that you can do like, you could be as accurate as like, say, any kind of like, um, Roman, uh, documentary that you might see on like the history channel or whatever, you know, those like recreationary, recreationary documentaries that you see, uh, world war two, which is funny because you have so much reference material from that, um, some film stuff, but you know, like the details are always wrong. So I think that it's fine for broad overview stuff. I don't think that if you're trying to like maintain historical accuracy and those are the people that are super into generating that type of material, like they want, they want the accuracy, which I totally understand too. Or the same, what about the same movie with a new actor? Do you see that kind of playing out? Like people pull up like happy Gilmore, but now it's starring Jennifer Aniston or something. I think there will be a novelty in that at first. Like that was always part of the whole, like you will star in star Wars. You will be the star of John wick. Like, and I think that that you'll see a little bit of that pop up. Um, I just don't think it lasts very long because it's going to cost a lot to do for any company that wants to try to generate that material. Like, and, you know, partnering with say a Netflix, or even if a Netflix is doing it, the amount of, uh, of the budget it would take to do that versus like a couple of people going like, Oh, that was cool. That didn't work. Turn it off. Let's watch the original. Like, I just, I think it'll be a flash in a pan. Interesting. I'm wondering if so with vibe coding, when that whole thing kind of blew up, uh, some people have suggested that, you know, there was this airplane game fly Peter by Peter levels that was doing like, I think 80,000 a month on advertising. And it was just a simple simulation that he did vibe coding. And some of the people complain, they're saying, well, the reason that's doing so well, it's not because it's such a advanced game. It's because of the name recognition of Peter levels. So one thing that I'm curious, like, would this open up a whole new genre or not even show what to call it, but of people that have a lot of name recognition, maybe people that are great thinkers, or maybe they're famous in some other, whether that's a sport or some other domain that are like, I will produce this thing. Cause I think Kobe Bryant was trying to get into some sort of filmmaking, maybe, um, he was talking about it. Um, and so like somebody that before could not produce movies or produce content, whatever you want to, yeah, content is don't love that word, but that's, that's what the word we use. It is. It is. Yeah. So do you imagine something like that happening where it's somebody that's never produced any sort of movies like that before anything else? And then just decides like Tim Ferriss is, if you guys know who that is, is create, I don't even want to say the name of the thing that he's creating, but, um, it sounds interesting. He created a great world and, uh, it's, it sounds like it's going to be an amazing storyline. So he's building this whole world and he's, I guess you could call it sci-fi slash fantasy. He never wrote anything like that before, but I guarantee I'm pretty sure it's going to do well just because of his name is his name. And that's his creation. He's also writing it in four hours, right? Four hours. Yeah. Four hours. Yeah. Um, yeah, I think that, I mean, anybody that, um, like in content as, as much as we hate to say it, like, um, it was funny. Cause like the one thing that immediately popped in my head with that, um, with the vibe coded game was like, there was also flappy bird and that came out of nowhere. Like there was not just like one guy that just made flappy bird. And for some reason, the internet was like, we love this. And then we drove him insane. And he was like, I have to quit the internet now. Same with schedule one. That was one person. That's a recent example. That was absolutely amazing. Amazing. Yep. Um, so I think that, I mean, if anybody's got the muscle power, like if, you know, Tim Ferriss has an audience of, I don't even know. I mean, like he's, he's in the hundreds of millions of daily listeners. So he's already got like a built-in audience that are, that's going to, you know, go towards, uh, his, it's safe. Like LeBron James decided, like, I want to make an AI movie. Like, of course, like people are like, that's going to immediately get a flood effect because of his name. Whereas if Dylan decides to make one, like, um, you know, like you, you have an audience Dylan, but you know, it's like, you too. I'd watch it. I'd watch it. I'd watch it. Yeah. Um, but obviously LeBron's movie is going to move the needle a little bit more than yours. Um, so yeah, I, I, you know, I think that there are, and there's definitely where I'm starting to see it already. Like there's name, names that are circling around. They're still like percolating because I think that there's still a little bit of like a little bit of a no, no in Hollywood right now. Um, but their work, I mean, there are people that are working on their AI generated films and whatnot that are, you know, name, name brand people. So. Do you know about the 10 to problem, the infamous watch problem that's always plagued generators? Oh, as in the, the smiley face, uh, um, on a watch face, uh, hands at 10 o'clock and two o'clock. Yeah. So I've learned about this and I thought it was super fascinating and Wes, I'm not sure how much you know about this either, but because these models were trained on so many images of watches that were always 10 to, which is how you advertise a watch, but it's not really how watches work. No matter what time you ask for, it's very hard to generate an image. I was wondering if this amazing VO3 Google model can actually solve that. So do you mind seeing if you can throw a watch in there and see what we can generate? That's a great test. Yeah. Um, let me see if I can do that. Um, we'll use, um, Google's flow for that, uh, as well, as opposed to, um, the, uh, the, uh, Google Gemini version of it. Um, we, I, you guys are, I don't know if anybody's paying for the $125 slash $200 a month, uh, plan on Gemini. Yeah. Yeah. You get the, uh, um, you get five free generations on, uh, Gemini now just text to video. Like it doesn't have all the fancy stuff going on, but, um, that's a little bit, just get squeeze as much money out of that subscription as you possibly can. Like, cause it's expensive. It's expensive. And then you got to buy credits too. I was surprised. I thought if you get the ultra plan, you're paying so much. I figured you get unlimited. Sora was unlimited. A lot of the other ones are unlimited. This is like, no, it's not that many. I think you get like, I forget how many, but it's like 20 generate, maybe more. Yeah. They just, uh, they, on the ultra plan, if you're on that one, um, you can get up to, um, they just reduced the credits. It was at 150 credits per generation. They just reduced it to 100 credits. So I think it's somewhere in the neighborhood of like 128, which is not enough for developing anything seriously long form. So, uh, let's try. Um, all right. So I'm just going to do a prompt here of it and do my, my, my channel classic, a man in the blue business suit. Yeah. Um, all right. And, and so while you're generating that Wes, did you feel like this VO3 was a chat GPT moment or did you feel like it was an incremental upgrade? No, I absolutely felt that this is a chat GPT moment and I'm actually really excited about it because one of the questions that I always get on the channel is people are like, what are, what should we be building with this? What is like the, cause some portion of the audience wants to build a business, make some money, do something like incredible with it, kind of utilize it. And I always had a hard time kind of saying, okay, this is a good time to go in pulling the trigger. When the GPT store was announced, I was at first, I was like, oh, this could be interesting. It's like the app store from the Apple from back in the days. And wow, that was kind of a flop. I mean, it didn't really go anywhere. Um, you know, with coding and stuff like that, everything's changing so fast that that's also kind of hard to predict where it's going to be. But man, with the video specifically, now that they've combined audio and video with VO3, the stuff that I've been seeing is absolutely out of this world that, um, uh, PJ, the guy that did on Twitter, the guy that did the pop print commercial, he did the, the prompting, the, the standups. Oh my God. Absolutely out of this world getting incredible views. And I've been saying for a while that I think it's going to affect Hollywood in a sense that you're taking away the barriers. You're taking away the moat, right? So you're able to now create great movies. You don't need the connections. You don't need as much technical knowledge. You don't need the money. And I feel like we're going to like my bet, I haven't really talked about this yet, but I feel like I'm going to make some sort of prediction because I feel like within definitely within five years at the most, we're going to see companies started by people that are not necessarily in the media that are going to be sold for millions or they're going to generate millions. Um, and yeah, I, I mean, I have a lot of ideas. That's why I wanted to talk to Tim and kind of get his, am I like, am I just going too crazy here or is this kind of where, where it's going? No, I, I, I think that you will definitely see a rise in smaller independent studios, being able to produce material. I think currently where it's at is, um, for folks that are in sort of, I guess, advertising, um, marketing ish, like that's kind of where you're going to see a lot of this really blow up over the next few months to a year or so is, um, VO, uh, is, is interesting because at least thus far has not, uh, only generates in 16.9. So widescreen, not vertical. Uh, and obviously in a lot of like marketing and, uh, promotional stuff, they want vertical video. So I think once that hurdle is over and a few other things like, um, in the IO presentation, they had a references feature that they showcased where you could put a product, um, in, say, I could pick, take this coffee cup and, uh, and upload a photo of that and then have the character interact with that, uh, with this coffee cup, which is actually from the Lion King of stage show. So free plug for Disney who does not need it. And we might just get sued for me having this coffee cup. But, um, uh, once, once that feature is implemented along with via where, you know, you can have, you know, your supermodel, uh, character or whoever with your product, um, you know, giving a testimonial speech about it. Yeah. I mean, for sure, you know, TikTok, Instagram, all of the, I mean, it's going to be all, it's going to go like wildfire all over there. Like, are you watching videos now though, that are AI generated for entertainment? Because I swear before I always was just like, that looks like AI and it's interesting. But like for the first time, I'm, I'm just entertained. It just comes up in my feed and I watch it and I move on and it blends in seamlessly. I mean, I, I am, I, I have been, but only because I follow the space so closely and I'm so, you know, sort of intertwined with it. Um, but now with, yeah, a lot of the compilation reels that I see on YouTube of like VO, like, you know, uh, especially like the prompt theory stuff. Uh, although I feel like that's getting a little played out cause there's so many copycats on copycats on copycats at this point. Um, but yeah, every once in a while, I'll check in with that. One of the things that I really enjoy in old school janky AI video was like that kind of America's funniest home videos sort of vibe to it of just watching cursed videos and just like the physics are all wrong and people are melting all over the place. I love that stuff. Uh, and then, so this is the VO, I think is the first time that I've watched a compilation of those types of videos that are correct and they're funny and they're, you know, and they're engaging and they're holding my attention for at least like two and a half minutes. You know what I mean? And pushing on into like the 10 to 15 minutes of VO clips, like that's, that gets to be a bit much, but, um, but yeah, no, it's, uh, I am definitely now watching for entertainment value, which is what about the standup comedy stuff? Have you guys seen that? Yeah, I have. Yeah. That stuff is pretty good. The running theory on that is that that those are Reddit jokes because Reddit, um, uh, had their, they, they made a deal with Google that Google could train on that material. So, um, it's, it's possible that many of those jokes are actually just really, really snarky. Yeah. R slash dad jokes. That makes sense. That's interesting. We did a man in a blue business suit looking at his analog watch, which reads 730. Um, the watch is upside down. So we can't really tell what it looks like. It's more like 725. So, which is funny because it's got, it still has that, that, that tend to smiley face aspect to it. So it kind of figured out a creative way to, to flip the, the, the, to hide it. Yeah. So, um, bravo, five fingers on each hand though. So there's that. There's that. Can we try something? I saw this generated in cling. I think it was this, uh, monster with the face of a clock and actually he had multiple clocks all over his body. So you can't, you can't hide that. So maybe something like, uh, uh, a monster that appears to be made of clocks or the, has the face of a clock. I don't know what the prompt was, but I saw that video. I was like, that's terrifying. Yeah. With clocks everywhere at different times. Yeah. Um, you know, okay. Uh, how about we'll do this? Uh, uh, we'll do, uh, a cinematic horror film featuring a monster with the face of a clock in a museum of clocks. Oh, that's good. That's good. As many clocks as you can get in there. Yeah. The horror genre is going to be insane with AI video. I feel like, cause it, it just sometimes looks scary without meaning too. And so I think they're going to do some crazy stuff with that. Yeah. Well, I'm just so interested in the clock thing because it's, it, to me, I wonder often, is it actually learning the physics of the world from these two dimensional videos? Or is it, you know, it's kind of like, is it next work word prediction or next token prediction, or is it actually semantically understanding meaning? And if it actually understands that clocks in the real world move and that there's like this aspect of them that people use clocks for, it's very different than just like generating. And I wonder if VO3 is naturally just picked up any of that, or if we're almost over that uncanny valley. Yeah. Valid question. I, I don't, uh, know. Um, I think that VOs, I mean, it's like, there's no, I mean, it's, we're still speculating, but I think it's pretty obvious that it was trained pretty heavily on YouTube data. Um, so, you know, it's definitely still next frame prediction that's going on there. I have seen some papers recently that are doing some interesting stuff with, um, uh, instead of generating, uh, next frame, what it's doing is this generating last frame and then going almost in reverse. Like, it's like two different diffusion streams that are happening that kind of meet in the middle that are kind of cross-checking each other. Um, but that is more about sort of solving the morph issue. Um, the world model stuff, that's what I'm really interested in. Definitely Google has flexed that that's what they're doing. So I do think that the next iteration or iterations that we'll see off of this is not only going to be YouTube data, but like Google earth data and, um, I don't, whatever else that, I mean, they have so many different modalities that they can pull from. Like if they can bring that all under the banner of VO, that's, or yeah, that's wild. That will be wild. Yeah. Yeah. And I wouldn't be surprised if they solve a lot of the physics stuff before they solve stuff like that, because unlike the video language models or excuse me, the image language models, where they're able to recognize images, like the speed gauges, that's like the notoriously, like the worst thing that they're really bad at that, for example. So stuff like that might be difficult to solve. I feel like, can you guys describe some of the stuff you've been watching this week? Like what's stuck in your mind that was AI generated? Yeah. So some of the stuff that I've been seeing, let me think, well, a lot of the stuff that I feel like that I saw was by, PJ and maybe we can put his whole handle. So people, I'm sure if you're on, on Twitter X, you've seen some of the stuff. So the Puprin commercial where everybody gets a puppy to battle their depression, that's his, a lot of the prompt theory stuff. Yeah. A lot of the prompt theory stuff was his. Have you seen the plastic one, the plastic baby? Oh, plastic baby. No, type in it now. Plastic was amazing. Yeah. That was incredible. Yeah. If I had to pick one. Yeah. That really should be. No, that was a meta puppet. Yeah. That was, it's fantastic. Oh yeah. If I had to just select one that kind of captured my intention and just for whatever reason I had to keep watching, that was plastic. And so yeah, what you guys were talking about earlier, like, yeah, I probably wouldn't watch a full motion feature with AI video, but as far as little TikTok things, a minute or two, I feel like it's going to have incredible engagement and stickiness for people. Like if you just watch plastic that it's like, cause every scene, it just takes it to the next level and next level. And it's like, are we watching a drama action or I don't even know, but it's, it was, it was good. Have you seen the little hand drawings turned into real animals? Like it'll be like a kid's version of like a cow and then it'll be a real cow, but it's all sort of misproportioned, but it like moves like a real cow and sort of seems like a real cow. It's so, they're, they're so crazy. I do think that it's funny cause the, uh, the plastic one that, uh, that was Meta Puppet, like that really reminded me of old school, uh, adult swim, um, content that was on, um, well adult swim, like at like 4am, a bunch of stoners sitting around watching too many cooks or, uh, robot chicken, unedited feature of a, or unedited, uh, footage of a bear, like just really insane experimental like films. But you know, like they, they, they, they had like this kind of manic energy to them all of, you know, like, and I think that's what VO2 in the hands of somebody that knows what they're doing and knows how to edit and knows as a filmmaker, that's where it's really gonna, um, yeah, it's going to make some incredible content there. Cause actually even with plastic, if you watch that, just like you were saying, Wes, it's like, it begins with, um, this scenario. It's about, you know, I would say like 45 seconds to a minute long before it switches to kind of another, it kind of blends seamlessly into another scenario. Um, and then each subsequent scenario becomes more and more absurd, but you know, the quality of the video output looks the same. So that's a, that's a way of like really keeping ADHD brains, uh, engaged with what they're watching. So, and that's something that I think that most people that are just doing the viral YouTuber prompt theory videos that they don't, they don't, they haven't figured that part out. They're just kind of stringing together a bunch of like eight second clips of people talking. And what do you recommend the best way to string it together is right now? Like what tools out there do you recommend? Like what's the best way to build a long form narrative? Just old school editing. Like, I mean, Da Vinci resolve a premiere, like, I mean, you still need to, you can just prompt every scene. Yeah. The way that I've always approached all of this stuff is that like it, you're, it's really almost more of an editing exercise. Um, because things like consistency and, um, physics and just all of the other random stuff that we run into with AI generated video. Um, it's, it's, you're just generating a lot of raw material. And at that point you have to kind of filter it down. It's, it's really more of an, um, a exercise in editing, I think more than anything. So that's really where the real skill of like AI video comes into play. Is the speed of change what you expected? I mean, were you kind of guessing that sort of in June of 2025, we'd be at the VO3 level and like, where do you see it going? I would say if you would ask me that in December of 22, I would have put where we are now four years ahead of where we are now. So every year it becomes like, I didn't, I wouldn't say I in 22, I would have said like, what is VO3 would probably still be two years away. Um, so it just, it, and every time the next thing comes out, like the timeline to whatever, whatever it is that, and that's, that's another thing that I fight back a lot with in the comments is that like, there's always, we went from, it's just not there yet. Uh, well we went from Will Smith eating spaghetti too. It's just not there yet. Um, and then it was like, well, the fingers are messed up and like the eyeball and blah, blah, blah. They can't talk and this and that. And there's, so the, you know, the goalpost is constantly moving backwards. And a lot of times I do think to myself, like, what is the end goal? Like, you know what I mean? What is, what is the terminus point for AI video? Is it when it looks exactly like a low budget Netflix show? Is it exactly when it looks like, uh, like mission, the latest mission impossible movie? Like where is the, where's the point that it's, it's there now for everyone? Um, and I don't necessarily, I'm not really particularly interested in prompting the latest mission impossible movie, which I have not seen. I have heard that it's not that good, but, um, um, I'm just using that as an example. Um, but, um, yeah, like where, I don't know for you. What about you guys? Where, where do you think, what's the terminus point for, for AI video? Where, where is it? At what point will you be like, that's it that we're done? Like, it's hard to describe because it's not, it's kind of a fuzzy thing, but it's when I feel like it has soul or something. I don't know. Like sometimes VO3 can give me a clip where I'm like, Oh, that felt like real. But when it feels like almost an emotion, I hate it. Like it just, like, I'm like, you try, like it's pretty, but it's not beautiful and it's not complete in a certain way, but it's hard for me to tell when it has soul. I just, there's something about the reactions and the subtleties that I would know. Yeah. Um, well, I think Tim, you nailed it in the fact that I think the advertising aspect of it is going to probably blow up first because that's kind of like some of the first ideas that I've had. I mean, in e-commerce videos, what really helped me get started with video sales letters. And some people charge $50,000 to write and record it like a video sales letter that you're going to use to send paid traffic to, and it converts for, for, for you to get the product. So I feel like now, and I remember it at some point, 10 years ago, even thinking about starting a company that would provide those services to creating video sales letter that would convert from, you know, they call it cold traffic. So you pay traffic, you shoot them over to a video and it gets the person to buy. And if you're spending less on the clicks from Google or Facebook or wherever, then you're spending less on the clicks, then, you know, they're buying more, you're making more than that's an infinite sort of money. So you can just keep, keep scaling that, um, with AI video, that's going to be just taken to a whole new level. It's going to be, you know, chat GPT writing sales copy. It's still not quite there. We can do, it doesn't do a world-class job, but it's pretty good. So, I mean, you can definitely create stuff like that. So I think we're going to see it applied in, um, marketing first and advertising. And then once that's really, I think they're going to figure out all the nukes and crannies and how to, all the hacks and all the tricks and all that stuff. And then they're going to make it usable for the, the filmmakers or whatever you want to call that new genre, the AI video movie generation, whatever. Um, actually I liked your word producer. I'm going to start using that because yeah, artist is such a loaded term, but I feel like if we're going to refer to people that create that sort of content, artists might be getting ruffled a few feathers. Producer, I think really just nails that. So people that are producing the stuff, um, I think they're going to be advertising and set the scene, figuring everything out. Then the more artistic creative stuff is going to come in. That's, that's how I see it. So the thing is, Wes, it just gave me this idea that I hadn't thought of before is like, is this now the golden era to create a new IP? Because like, should we be putting out like hundreds of different like male, female characters, dogs, ants, cartoons, and like some of those will evolve into things that make people rich in this era? That is one of the most frustrating things to me about where we are right now is, um, the people that are continually chasing IP, like making like AI videos with Wolverine and with Star Wars and with, you know, like stuff that it's going to get pulled Batman, whatever it is, it's like, like you have infinite imagination that you can go play with now. Like you can create your own thing. Like, why are you still chasing that? Like what? Like that's all of that's like, now is the time that you can create anything that you want to create, uh, anything that you can dream up. Stop dreaming of Batman, just stop. Like Batman's good. We've got tons of Batman movies. Like just like, we got more Batman movies coming, make your own thing. So yeah, that's my rant on that. Stop making, stop playing with that. I'm 100% with that. I think the studios, because they're so concentrated, they want to make their money. It's like, if, if, if they start with some rehashed idea, they, it's more like, okay, we know we can make a return on this, a brand new one. It's a coin toss. So from that perspective, I understand it, but it's like, again, like the bean counters versus the artists. In that case, it's more like the, the bean counters win because they own the studios. This kind of turns it all in its head because now that everybody can be a producer, you know, you're paying for credits and yeah, that's going to be, that's going to cost a little bit here and there, but not millions of dollars. And I feel like all the great IP, like, like Dylan, what you mentioned, like I great, I, all the great IP happened like in the nineties, a lot of it, like all the video games, like we're still playing the same, you know, Baldur's Gate six, doom seven, whatever, you know what I mean? Back from, from those days, like why like people, like do we lose creativity? It's not that people can't come up with it. Yeah. You know, the, the people who were creating comics, like in, I guess, I don't know what era, but the fifties or sixties or whatever it was, they, they, I guess they just, we don't know it now, but they created billions of dollars for these companies because they had the freedom to throw out all these little side characters that Batman would interact with on one frame and then another frame or like the, how the Simpsons built up this gigantic like world in there because they just kind of pushed it and then they all branch out and it becomes this gnarly, super profitable, you know, universe at one. And I can say on the, on the, on the count of the comics folks is that, I mean, all those guys got screwed over by the corporations. Like, I mean, they didn't see a dime off of like, I mean, there's like stories about, um, uh, it was Len Wein, I think that, that created Wolverine. I think they cut him a check for like a hundred, like, I think it was like a hundred or it was between $250,000 and a hundred thousand dollars for, you know, for Wolverine, for this character that is worth billions of dollars. Um, you know what I mean? Like that it's tragic and it sucks. Like, um, so there is a part of me that's always like, stop trying to suck up to giant corporate, you know, IP and companies because they're ultimately not, it's not in their best interest to reward you. Uh, instead create your own thing. And if it gets big enough, they will come to you. Like they will, you know, like if you create something that is a mega viral hit, um, guaranteed at some point or another, some studio is going to come calling and then at least you can negotiate at that point and you call the shots. So. And especially if you could build your own audience on a platform, YouTube, TikTok or whatever it is. Cause you know, I, cause I still distinctly remember when I saw the trailer for Ant-Man and I was, I, I'm not a very big comic book person. And I was like, Ant-Man, that sounds so stupid. Like what a dumb name. But then I was like, but it's literally Batman, which sounds so cool. You know what I mean? But after enough commercials come out and it becomes a thing and I've seen it now, all of a sudden I've digested it, but there's just this little hump that you have to get over where you're just like, it doesn't make sense until you absorb enough of it. And maybe the influencers of the future can kind of bypass the big systems to get in that way. Well, I think the other side of it too, is that like, there is, um, there's a series that my kids like watching the anime, traditional, um, animated series called Murder Drones, um, that they, in that, that was released YouTube, et cetera, et cetera. Um, and I believe that that will be, I did see that they will be releasing all of those episodes on Amazon Prime. So, uh, you know, to me, that's really the pipeline these days is make your own thing, release it on YouTube. And then, you know, like just for a distribution, they will either come to you if you've got a big enough hit, or maybe you can go pitch to them. I do know that Netflix is like, I don't think it's a secret, like, and I'm not under any NDA. They've never talked to me, but I do know that they have their own like little AI animation studio that's, that's starting to brew up. I don't think that they're looking to do giant shows or anything like that. I think you're, you're looking at something a little more in the lines of, uh, love, death and robots and a secret level, which, uh, the, the two CGI shows that have come out recently, I think it's something more along those lines. So what's interesting there though, is that essentially like three or four artists at this point, that's, that's a level playing field. You know what I mean? Like three or four artists that know what they're doing with AI tools are using the same tool set that like Netflix is using. What do you know about fable studios and showrunner, the Matt stone and Trey Parker studio, and it's been kind of pioneering this for, Oh no, that's, that's, uh, that is, um, um, that, yeah, they're, they're actually not related to, uh, Matt and Trey. Um, they're kind of their own separate entity. Um, so they were trying to, what they were trying to do was kind of create more of a, it was sort of infinite storytelling. They just happened to choose that South park visual style, I think, because it's very easy to replicate like that 2d animation thing. It was very easy for, uh, at the time there, actually their, their tech was really fascinating at the time because it was all stable diffusion. Um, and they figured out sort of layers in how to do back, generate a background, generate a character, um, a few animations, and then a chat GPT engine. I want to say they were using 11 labs in there for the voices too. Uh, so yeah, yeah, it was pretty fascinating. The, the thing that they were working on that I thought was actually cooler than that, but I don't know where they are in development with it anymore was this project they called thistle gulch, which was much more like 3d. It was like a game engine about a Western. Um, and you could sort of, um, you had freedom of camera throughout. So basically like each of the characters would have their own motivations and you could sort of, it was a Western town, um, a mysterious murder. There's a sheriff, but there's no main character. You're just sort of following these agents around that are bumping into one another and you can kind of influence where the agents were going to go, but it was a free roaming camera, uh, in 3d space. I thought that was really fascinating. And that's the kind of thing that I'm hoping that they start kind of leaning into a little bit more. It's like essentially like that's the, that's when we get into the holodeck idea. Like you put, you pop on your VR glasses and there's AI agents that are working on a murder mystery around you and you get to walk around and, and, you know, talk to each one of them and all of that stuff. That's this iPad. I can't wait for that. That's going to be amazing. Yeah. That's that whole project they're doing. Um, and it's been a while since I looked at it. So, and I believe, yeah, thistle gulch, I don't remember, but they had a number, a number of different themes, right? So they have South Park, they have the, the Western one. They had like a few other ones. I'm forgetting they're all based on kind of the same underlying technology, right? Cause they're trying to like replicate like in South Park, the entire town, right? Yeah. Show runner. That's what it was called. Yes. Yeah. Yeah. Um, so it sounds like for people that are not familiar, they're like recreating the entire town and Stan, Kyle, they're running around that you can see him on the map. And when they meet somebody else in the map, Chad GPT, whatever GPT four that they use at the time generates the dialogue, 11 labs voices it, then they put it into the animation thing. So it's not like the, and here's the thing. So my, my kind of theory, cause after seeing like alpha evolve and all this stuff, this is kind of where everything's going. So where you create LMs plus some sort of scaffolding, some sort of a collection of tools that they use and that system, um, cause what they're really good at is, you know, creating brainstorming, basically creating like a hundred different ideas. So if you're able to have it create a hundred different scripts and find some auto automated way of evaluating if it's good or not, like, is this funny? Right. And maybe have some sort of specific thing that judges it, then strings it together into a story. I mean, that's kind of the idea behind show runner. If I understand correctly, I feel like that's going to be very effective. So you're telling me in the future, I'll go to like simpsons.com and it'll just be something. I just tune in for like a couple of hours and tune out. Yeah. So is that what you're imagining? Um, and the, and the whole, and then I choose like, do I want to talk to mill house or Bart or Lisa? And then that zooms me into that part of the city. That could be a variation of it. And, but I think their idea is more imagine the whole thing is playing out in real time across the entire town. So multiple people can have different conversations. The storyline is progressing and the camera only picks up the interesting parts and then it strings them together into a movie. So in other words, you're simulating an entire world. You're simulating the existence of all the characters as they're walking around doing their thing in order to then just like tape them in those moments, stringing together into a story. So if I'm at Moe's Tavern, am I going to miss something funny happening at the school? Or will the AI, do you see the AI like grabbing my screen and throwing me over to the school so that I see the funniest thing that it can predict? Well, your choice. Yeah. I think how they were thinking about it is at the end, you use all of that to create like a movie that the director, the person that's actually selects. So, so the, the, the final product is a movie or, you know, where you're just watching it, but behind the scenes is you have this entire simulation that's running, but obviously you can change that to be anything. I think like at that point, everything's open. Yeah. Yeah. Yeah. That's like how they film these reality shows. They just record everything and then put the storyline together later. It's going to be like that. That's going to be so different. I, you know, I, I do still wonder in terms of personalized content, how big a deal that's going to be. Like, I, I still feel that like in terms of mass entertainment, like something like Netflix's Bandersnatch and they did a couple of other experiments with interactive endings and stuff. And, you know, clearly they've taken those offline for a reason. They were expensive and nobody, nobody used them. Um, because I, I still do think that like, although we think that this is all really cool and there's a large amount of us that do, I always think of like Joe six pack coming home from work after a day and sitting down and just being like, I just want to watch TV, man. Like, um, there's a part, I mean, I'm sure that you guys, uh, go through it as well. How many times do you sit down and you're just like, I just want to zone out and watch TV. And then you see all the stupid tiles and you're like, I don't, I got to make a choice now. Like, this is like, uh, this sucks. I just really wanted to sit down. And that's a lot of people. So if it turns into a thing where it's like, I want to sit down to watch, you know, my favorite show and now I have to choose which locations I want to be in and which characters I want to follow. I think most people are just going to be like too much. I don't. You're right. Everybody's just going to watch friends on DVD, even though there's going to be the craziest stuff in the world. They'll be like, I just want to see it for the 80th time in a row, please. Just, I mean, my wife sits there. She doesn't like, I'm sure that many, um, guys my age have wives that are like this as well. This is watch TV. She listens to TV. You know what I mean? She's like, she's, she's on her phone or she's like doing like some kind of craft thing or something like that. And we'll look up when like something happens, like a gunshot look up. Oh, he's dead. And then back to, you know what I mean? Like, it's like, it's more like radio theater for a lot of people. Um, then, and I think that if it's one of those things where if we are moving into like this agentic TV, uh, where, you know, the character is stops and it's just like, so what should I do next? And is looking directly at you and like just waiting for an answer. Like for a lot of people, there, a lot of those characters are going to be waiting a long time. Right. Yeah. And so more, more passive future. Interesting. Cause I think two branches, I think. Yeah. Yeah. I, sometimes I want to watch something passively to where I don't have to pay that much attention to it. Cause I might be on my phone or something. You just like veg out. And sometimes I'm like, no, I'm like, I'm here. I'm present. Let's turn on something that has a storyline. That's going to be very involving, like, uh, severed was recently something that I watched. Yeah. Uh, interesting. The stuff we talk about at the studio on Apple TV, um, is, is Seth Rogen is in there and that's kind of what they're talking about. Cause he wants to be, he becomes the studio head. Um, but he's kind of wants to be an artist friend to the filmmakers and stuff like that, but he's a lot of pressure just to make money. So they do the rehashed versions. They bring in the Kool-Aid Kool-Aid man because the guy that owns the studio is like, Oh, Barbie did well. So why don't we do the Kool-Aid man as like the new billion dollar franchise? It's, it's pretty hilarious. I really, I was surprisingly really enjoyed that. Um, but yeah, but that whole thing that we're talking about with the simulation stuff like that. Yeah. Maybe it's overkill for video, for movies, for TV series, whatever. But I mean, for video games, for VR, like you said, for more stuff like that, I think it's going to be absolutely incredible. I I've honestly said that I think that a lot of that stuff, the holodeck aspect of it just becomes a new branch of entertainment. I don't think it's going to take over filmmaking. I think it's going to take over, uh, you know, video games. It's just going to become a different experience that we enjoy doing for, I mean, like, well, my kids, um, they, you know, last, uh, holiday season, um, like every kid that's around my kid's age, they all got met at quests. Like that was like the gift that every, every kid in school got. That was the, that was the cool gift. Um, so, and they're playing all like the junkie, like, you know, freemium games, of course. Um, and, and whatnot, um, gorilla tag, uh, job simulator, which I find hilarious, like, um, you know, um, and so they'll play that for a while, but you know, at some point the quest comes off and switch comes out, you know, or they'll, they'll go on and watch YouTube videos or, um, whatever, you know, whatever show on the watch Wednesday on Netflix. You know what I mean? Like that's, it's, it's just like that, the VR aspect of it is just another form of entertainment for them. So, um, I don't think that it's, I don't think that it's ever going to be wholly, you know, that, uh, William Gibson, me kind of dystopian future where we're all walking around with our goggles on and not interacting with each other, except in virtual space. I think that it's something that maybe you and your buddies get together and do on a Friday night for a few hours or whatever, you know, and then go off and do something else. All right. So we're just warming up here. The rest of the interview, the part two is going to be on Dylan's channel. His channel is Dylan curious, and we're going to have a links here on the end card. We're going to have it down in the pinned comment as well as in the description. So make sure you check it out there because we're about to go much, much deeper. So click on over and I'll see you there.
