Transcript · June 7, 2026 · 14.9K words

Ep. 70: The Weird AI Video Formula Getting Millions of Views

Captions uploaded with the episode. Expect the odd misheard word. Tap any timestamp to hear it on YouTube, or go back to the episode for the summary, receipts and chapters.

0:00:00Fast Hours welcomes Tyler aka jboogxcreative

Play from 0:00:01 What's up, everybody? Fast Hours Podcast, episode 70...7...0. And we've got a special guest on today, too. So we've got Rory Flynn, we've got Drew Brucker, but we've also got Tyler Bernabe generative AI creator and strategist, here with us today. And he's doing some crazy cool shit as well. So we're gonna dig into that with him. Um, but before we kick things off, let's let him sort of introduce himself.

Play from 0:00:29 uh to to the audience here. So Tyler, take it away, man. Tell us a little bit about who you are and and what you're doing. Yeah, what's going on, everybody? My name is Tyler, better known across the greater AI sphere and the creative community as jboogxcreative. Um, I have been a full time AI creator across social, working with brands. I've now been with multiple A16Z startups on the marketing and social side, as well as just growing the jboogxcreative social network over the past couple of years and educating people and trying to um.

Play from 0:01:03 Help people figure out how to do the stuff they want to do with AI, what in whatever way makes sense for them. But um yeah, no, I'm really stoked to be here. I've been full time with this for a little over four years now. And uh deciding to get curious about AI four years ago literally changed the entire course of my professional life. And it's been a crazy roller coaster ever since.

Play from 0:01:30 And also for everyone, Ty Tyler is also a listener. So we we we know we keep things in the family here. We also expand the family here. Right? Like that's how this stuff works. So you know and if you've if you've opened a social media app in the last three years, uh you're you're probably guaranteed to have seen one of his viral videos. There's been not one, not ten, maybe like I don't know, like a hundred. So maybe more. the the the cool little um pin I like to wear in my hat and the notch I like to put on my belt is that with every single let's call it AI generation of tools, every like every new whatever whatever you call it as AI has evolved over the past four years.

0:01:51Going viral through every AI era

Play from 0:02:14 I started with warp fusion back when we had to work that out with the Google Collab notebooks. and Python and it was a pain in the ass and ten seconds took hours to get the frames for. But with every single iteration of where AI has gone since then, I've had s I've been able to go viral with every single tool, no matter how it's evolved. So that's something that I'm very proud of. Did you have a you have like a very very good eye for like number one stopping the scroll?

0:02:35The anatomy of scroll-stopping AI

Play from 0:02:42 Number two, having a little bit of like I don't want to say it's not necessarily shock value, but it's not something I've seen today. Let's put it that way. And it's not it's not repurposed, it's always new, it's always original for whatever theme. You know, like I remember that one of the first ones I saw was the probably the seductive, you know, food videos, which were I when I saw that I was like my god, it's a cheat code. Like this is the best it might be one of the best social formats I've ever seen in my entire life.

0:03:07The seductive food video origin story

Play from 0:03:09 funny story, funny story about the seductive food ones. And if you and if you want to show one of those, if you go to the Reels page, I have one of them pinned at the very top. It got like tw twenty, twenty-four million views and the the donut, the very first one right there. Um, I actually made that live on stream with everybody as they were watching. And these videos were a response to me being so burnt out with making content and posting every day that like I needed to step back from taking myself and taking everything so serious.

Play from 0:03:41 And I just needed to do something to make myself laugh. And that one, just like that one thing, completely turned into like a six-month detour of degenerate food craziness. They're so good though. Like I whenever I would put whenever I'd pull one up on my screen and I'd see it, I was just I was just like, This is incredible. You know, like it's it's so creative, number one, but also like just just a great way to to probably unearth some of the dirtiest minds on the internet.

Play from 0:04:14 To be like it. Dude, I I I came up with it because like so even those back in the Animate Diff Comfy UI era, I was making those like I had a consistent community of like girls from the pole dance community I was working with, and I would use videos of them and I would turn them into dancing food with animate diff and through a bunch of like masking tricks and after effects. And what I started to realize was like, what does everybody actually want underneath everything?

Play from 0:04:41 Everyone loves food and has some sort of emotional connection to food, whether that's good or bad. And every everybody ev everybody wants sex. It's like it's those two things and you mash them together and the v velocity on that content is crazy. It's a it's a shock factor, you know. Like you this is I think you may actually know her uh on on X though, Lola. We've talked about her stuff before. But there, but there's that same shock value.

Play from 0:05:08 It's different, right? But the the the commonality or the string there is this shock value. It's something that you're just like not expecting to see, and in some ways, kind of goes into this slight bit of uncomfortableness that is just so unique and um really. I I think like differentiates itself from the rest of you know AI content across the board that we see. Part of what I always kind of ask myself as I'm doing all these things is like I I scan kind of what's going on in the greater AI creative sphere and whatever the meta currently is, I try to run as fast as I can in the opposite direction.

0:05:32Running opposite the AI meta

Play from 0:05:49 And even right now, like now I s I'm starting to see a lot more AI anime stuff and it's making me want to run in the other direction. But um I feel like now I actually have some stories to tell. So I'm gonna I'm gonna keep exploring this a bit more. I'm gonna try and pick the right photos so we're able to not get piss off YouTube in this process at the same time. So But dude, it's it's like so cra it's like it it feels every time I see I'm just like, this feels exactly like what I open social media for is to find something that's new, not the 'cause that that whole trend, which I, you know, it's only been sort of like amplified by AI, which is like, you see what works with one reel, take the exact format and do it your

Play from 0:06:36 way, right? Like, and that's how what how people go viral is like, here's viral format, I do same exact thing, right? Like your stuff is like something that is building the format, which is what I I'm looking for. Right? Like I like that. well, and you guys know, you guys know because you guys are working professionals in this. You guys know that the the the place where brands and clients actually pay you and what they pay you for, they're not gonna pay you to do a cheap knockoff of someone else's thing.

Play from 0:07:06 They're gonna pay you to do the Rory thing or the Drew thing or the Jay Boogs thing. And so it really does there's no material benefit to just trying to do what everyone else is doing. Even if it leads to like velocity and views, because ultimately if someone wants to pay for that thing, they're gonna go to the person that's making the highest quality of it, which is usually where it started. Mm-hmm. also c like to your point, like that's such a good point.

Play from 0:07:32 But you're also, if you're not doing something that is leaning into a a crafted or unique style, you're also paired together with all those people that that aren't doing anything like that, right? So then the bar is how good is it versus how different is it? How stylistic is it? How much personality does that have, right? Like so to your point, you know, whether you want to call AI art or not is a whole different discussion, but it's like w you're putting your own spin on it.

Play from 0:07:59 Which is then your identity, right? Which puts you in a different category and consideration when somebody's deciding between 20 people for this role, right? Or you know, they know what they're gonna get with you, right? Versus these other people maybe just have more generalized examples. Maybe they've seen it before. They've done you know, done one off, you know, there's something maybe slightly different about it. So I think that's a really good point to bring up as well.

0:08:24Is AI art? Tyler’s best answer

Play from 0:08:24 I think it's also when I look at your stu going back to the artist conversation, like you it's like again, it's like a it's like a signature style. Like even if it's different video, different, you know, medium, if it's not anime or if it's, you know, the the dog getting its its you know, full full lineup over there. The to me it's like I could tell it's a you video from looking at it. So the same way I look at Drew's images and I'm like, that's a Drew image. Like I know like I can see like the personality in the image, which to me Everyone's like, Yeah, that's not art.

Play from 0:08:54 That's like, uh there's there's something there 'cause it obviously your stuff wouldn't look a certain way if you didn't like a certain thing. And if you you know so here's the take I've I've adopted on the whole like AI is not art stance is because that that's like that's a comment that no matter what you make, how good it is, or whatever, you can't get away from that comment. That comment is always gonna find you on one of the platforms. And my response is like, great, it doesn't have to be.

Play from 0:09:22 All it has to be Is the thing that was in my head that is now on the screen, and that makes me happy. If it resonates with anyone else, that's just a bonus. That well said. Clip it. What's mean, that's his philosophy. I think I like I love that, you know, because that also keeps things fun and fresh and and doesn't quite wear you down.

Play from 0:09:48 You know, when you're working on the things that maybe don't have so many constraints and you've you're just operating from this, you know, creative space in your head, this white space, or, you know, I'm gonna do whatever I wanna do. It doesn't matter, and then I'm gonna post it once I'm Satisfied with where it's at, right? Like at the end of the day, that feels different. Um, and I think like we all need those moments. And, you know, especially when you you get caught up with doing a lot of client work.

Play from 0:10:14 I mean, I crave those moments. I'm like, man, I just need to play. I just want to have some time to just do this or run that or just kind of see where this takes me, right? It doesn't even matter if I get to a point sometimes where I'm like, I don't even know if I need to post this or I want to post it. I just kind of feel like I did some. Some experiments, right? Some R and D, how I think through things, how I improve things from where I started, right? Like just run getting those reps in is always important and refreshing for me.

0:10:35Why clients pay for your thing

Play from 0:10:39 Y you know what I will say, the the benefit of growing the following kind of the way that I did. Cause if if you would have told me four years ago, because prior to AI, I was working as a freelance videographer and editor. And I was I did a bunch of stuff for like Red Bull and especially on like the Red Bull dance side and Sick. Love those videos. it uh you know it's When you work with clients, it's funny, they hire you because you're really good at a thing.

Play from 0:11:13 But then I read the post you made on LinkedIn yesterday, Drew, which completely resonated with me, where it was like, you know, you let go of this contract that ultimately instant in like you is a lot of people would see it as like you walking away from a a 35k or whatever it was deal, but you saw it as reclaiming seven months of your life. And the benefit. To growing the following has now been that as the following has gotten bigger, the clients that want to pay me, which iron ironically enough, now they're paying me even more.

Play from 0:11:48 They want me to come in and just do what I do. Whereas before they might have liked my stuff, but they want me to come in and do what they want me to do. And then when what they wanted me to do, even though it's my version of it, doesn't hit, they're like, Why didn't this? Do the thing. I'm like, because you didn't let me do the thing. Right. Yeah, dude. Come on. You know. it. There's so for anyone who ha who's listening right now and not taking notes on this, if you're trying to create content, there's already been like about six points that Tyler's made and I think that we've probably iterated over seventy episodes here at some point.

0:12:10Stop copying other creators

Play from 0:12:26 Just make stuff for yourself, number one. Just don't follow the trends, just do whatever you like. And like that's definitely gonna separate you than just being, you know, let me make another one of these, you know templated videos that everyone else does and that you can just be repopulate. I think it's probably the best thing to do is like whatever you're interested in, you'll figure out the problems and you'll you'll finish it versus just finishing it if that makes sense. Like so so let's talk about what you just said real quick because this is something that for my Patreon community I reiterate to these guys over and over and I try to make it clear to people up front before they even come in and have the idea of learning with me.

Play from 0:13:01 Like you're not coming into my Patreon or coming to like a live stream to learn how to do what like exactly what I'm doing. I I'm teaching you guys the foundational things you need to be able to to go make your idea or your version. Of what you want to translate, like what I'm doing into. But please, I I tell them like, guys, if you are just trying to copy the thing I'm doing because you see me getting views, you are going to be competing with me, which is not gonna go well for you.

Play from 0:13:33 Like, do do your thing. m it was, you would have already come up with it. And and that's something that Drew and I definitely have hit on a few times here, which is like w it's like you can't even like get tied down to a style most of the time because you can explore whatever you want and it's really har I I have AD like A D D when it comes to that. I'm like, I wanna do this, but I also wanna do this, but then I also wanna try this, but then I never have time to like really continuously like dive into a format or a style to like master it.

Play from 0:14:06 It's like I just I I wanna try anime. I wanna ooh, I wanna try, you Pixar, I wanna try, you know, realism, I wanna try whatever. Like you just want you want to do everything at the same time. sidebar real quick. I can't believe Instagram did not ban me during this era of these videos. Like, I cannot believe it. These are outrageous. protocols to really uh understand, you know, what was being shown, you know?

Play from 0:14:32 you know where the the so like with these things though, where I was like what really made them start popping off was when I realized, like, oh, these crazy outrageous little food image prompts I'm writing, GPT one point five when they still had the Sora platform, right? Like I'm so pissed that that open AI did away with the Sora platform. That was my favorite place to you to actually generate GPT images.

Play from 0:14:58 Um that model generates these outrageous anthropomorphic whatever images on another level compared to everything else. The yeah these these were all GPT 1.5. Okay. It 'cause it's it's interesting 'cause I was I was wondering when you were doing I'm like, where are you not getting where are you not getting flagged on some of this? 'Cause I like any time I do anything that's the remotely close, you know, I I get flagged.

Play from 0:15:26 But does the other was that was this on sorry, d just to clarify. Sora this was Sora two, not Sora one. this no, this this was this so I used the Sora platform to generate the images, but I never really I would only use the Sora video model, the first one at least, for their editing function with their little timeline because that was so for all all these videos you see, the reason why they look so smooth and you don't really see the frame that that jank in the middle was Way back when Sora One came out and they released their timeline editor, I had figured out that if I take um, like I'll I'll take all my clips into Premiere, I'll put them all

Play from 0:16:12 together, I'll export that with the janky frame in between all of them, where you see the jump between the start frame and the end frame. You take that clip into the into what was the Sora editor, and you just chop away those couple frames. where it goes from start frame to end frame and you don't prompt it, nothing. I would just regenerate it and let Sora fill in the blank and it just made everything perfectly consistently smooth and there are no cuts in the video.

Play from 0:16:46 That was the cheat code back then and still no one's really figured it out since then. Like no one's really added that feature. Which I was I remember doing the same exact thing and being like, This is probably the best part of the entire platform. Like, you know, it was funny, but it was also like there was the you know, between that and remix and blend there's some really cool stuff that first go round, even if the video generation itself wasn't like the you know, maybe the top tier at at the time.

0:17:11Viral views vs real conversion

Play from 0:17:12 But about this this whole era for me is that like the velocity on these videos is is crazy. You can look at the numbers on them. Like every one or two videos was doing multi-million views or in the high hundred thousands. But but those six months that I was doing the food stuff, even though it helped me grow my Instagram exponentially during that time, conversion on those things for Patreon, for client work, Absolutely non existent.

Play from 0:17:42 Really? Really? there? That's a here here's a let's you let's do a little bit of I'm curious on the creator side of things 'cause I think there's probably a bunch of people listening that are creators. How are you how are you moving people from Instagram to Patreon? Were you doing like the many chat route? Were you doing like newsletter all so if you if you on literally any one of these videos, if you read the description, they're all uh a an ad for my Patreon.

Play from 0:18:08 Like that that every video gets that same description with that same word, and I just change the first line or two depending on the video. And so the reason why I there there's a little bit of strategy here and maybe it's anecdotal, right? But whatever, we're all trying to figure this this stuff out. Um I don't put the like comment thing at the top, you know, the comment learn. I put that at the bottom because there's I think of the Instagram video as like a lead qualifier for Patreon, right?

Play from 0:18:42 So you have the peop the person that's gonna watch the video and then doom scroll past it in their feed and not even follow or come look at the rest of the page. I don't even want to attempt to to monetize that person in any way, shape, or form because most of the time that's a waste of time. The person that clicks into the video and reads the description, well, that's that's like layer two. They're they're at least interested enough to see what I had to say about what was on the screen.

Play from 0:19:11 Um, and then the person that comments learn is obviously like they're they're completely qualified and they should bounce to the Patreon if they're interested enough that they've clicked through those three layers and fought their own um attention span to get there. It's and was and this was like you were run it were you doing like many chat or something like that where it like just auto DMing the links?

Play from 0:19:36 It's such a such a hack on on Instagram that, you know, you can't do that on like LinkedIn or anything like that. Which, you know, thank God. I mean it it would be just that's all LinkedIn would be at this point. care about people posting on their platform. They can say whatever the hell they want, but they do not care. I mean this the reach share now is just brutal, bro. I mean it's it's insane. Same thing, but like this was see, this is where I I like I love this stuff because I think probably there's some value for anyone here who's, you know, anyone's listening because of marketing purposes, things like that, like there's a there's a whole process to organic social and these are all the micro decisions that go into it from you know, hon honestly understanding like, this worked, let's double down on this.

Play from 0:20:18 How do I vary this? How do I show this in a bunch of different ways to like continue Going downstream with the idea of the funnel in mind where I'm gonna bring everyone in from the top, s sort of funnel it down, but then also you probably found that you went viral a couple of times and you didn't get the right people. So it's like all right I need to shift that to the bottom. Right. Like there's there's so much like I love the intricate like little decision making process that goes into all this stuff because it feels so it feels so like attentive to detail, um, and sort of like mind share.

Play from 0:20:47 if things like this, for example, did better for your conversions, you know, because you've got sort of like a like ultimate outcome that people are aiming for versus your videos are just, I mean, those are entertainment. Those are that's your personality coming through. Whereas this feels more like um educational. the right? So so like I think even that is interesting.

Play from 0:21:14 And and also it's it's a different platform. I have no idea in terms of deals and and revenue from the Instagram side. It's just not a channel like that I've gone hard on. But I wait, pause, pause. but but I think that's like that's an interesting, like those are all the the interesting I like to your point, Rio, like the interesting like decisions of like kind of going This and what you're gonna pull on, you know, because s sometimes if you're trying to do too many accomplish too many things with a single thing, it's not gonna work versus maybe specialized things or you've got different cadence where maybe like one post a week he's doing something like this that's educational versus like, you know, two or three that are

Play from 0:21:59 more just entertainment based, right? Like, I don't know. I think the with with that, what it comes down to is ultimately, if you're going to be doing this stuff full time, you need to be making things that you actually enjoy making. I enjoy educating people on the live stream side and within like my Patreon community and interacting with everybody in my Discord. However, I do not enjoy setting up the camera and making talking head videos, but but when I do want to do that.

0:22:03Building a creator business solo

Play from 0:22:31 I refuse to use an AI avatar for it. Yeah, don't do that. Don't do that. Only if it's a funny one. Only if you like V you know, video to video to your face on like a you know a a uh banana. It's gonna be like, you know what I mean, like one of the one of your characters. But like I think this is also, you know, for anyone who's not familiar with what a full time content creation role looks like.

Play from 0:22:59 Which is where whenever I hear someone bashing content creators, they're talking about the bottom. 95% who don't put in the work like this. You have top a funnel across all the social channels. So you're doing different types of content there. You have the lead magnets and you also have the, you know, you have the many chat funnel, which goes into a Patreon, which probably also is the, you know, it's your community. You're doing live streams. You're doing six hour streams on, you know, every other week.

Play from 0:23:26 If anyone has ever tried to talk for six hours and explain themselves what they're doing, it's not easy. It's not easy to to to explain everything that you every thought that comes through your head. So like not only just the the creation side of it in general, coming up with new ideas, can being consistent, doing that every day, then like understanding the marketing side of the funnel, how to get people there, how to keep people there, and then also how to deliver for them every time. Like there's that's a lot, Pan and that's a that's a single you're you're one person, right?

Play from 0:23:54 That's doing all this. Yeah. dolo for better or worse. Perfect example. This is like the perfect use case for anyone who's looking wants to know what is like how to build a successful creation business, content creation business. This is what it looks like. But it's also like that's the amount of work that goes into it. You don't get there by posting three times a week. Like it's not how it works. Like, yeah. why that's why it's like, I mean, like that was that's also what was interesting about I guess Rory, you and I is I mean, that was the way to get the following, right?

Play from 0:24:27 But it's also not I I mean, at least for me, I I I can speak for myself. I mean, there was a period of time where I did genuinely love to post, you know. But I think at the end of the day too, like w when things get busy, the last thing you want to do is like stress out about man, like I spend an hour thinking about like well okay first of all what am I up to second like how am I gonna communicate it you know what format is it gonna be in you know and then you know early days like I mean even now right like LinkedIn specifically like you need like carousels still bang but how fast can you put that carousel together?

Play from 0:25:01 I mean I know you can do it now faster with AI. Even then though like you're still gonna have to edit and tweak stuff. I'm not comfortable with the design output I get even with design MD files. Where I'm like, yeah, this is ready to go right out of the gate. Like, no, there's no, like it's not. So I still gotta go in and do that. I mean, it just eats up a ton of time. But you know, mad respect to anybody that is like creating full time because again, you have to love it.

Play from 0:25:26 And um, you know, I think like in general, when you build that kind of following, the pr you know, there are a couple of reasons why people do it. But once you have a following, they'll follow you to different channels. They trust you, right? You you're you're being authentic. You can bring some of these things along um with you for this ride. And ultimately, like that's paid off for for all of us here, you know, for sure. Um, but yeah, it's a it's a love hate for me. It's a love hate for me, for sure.

Play from 0:25:53 Wha what no, go on, Roy. No, was just gonna I was gonna ask you how you deal with like fatigue from being from the from the creative side. Like how do you just like 'cause I get it. There's certain times where I come off come off a project where I've been like doing something that's so like system technically heavy. I'm building like, you know, workflows that are it's like multi step, multi checks and balances, huge system prompts. And then it's like, all right, let's go create on social and I'm like, I got nothing.

Play from 0:26:20 Like I got I got no juice left. You know, it's it was different when it was like that was feeding my business, but How do you like how do you consistently stay there in that pocket? So, so let let's talk about consistency for a minute. So before everything really popped off for me, when I was starting on Instagram and I was like, okay, this is this is what I'm gonna do. Um, cause I was at the time looking for where like what path could I go with my career that feels like it's going to have longevity.

0:26:28Burnout, longevity, and going through it

Play from 0:26:49 And uh that still is not a question I've answered. Every day I do this stuff and every time I'm looking for a contract and every time a new tool comes out, I'm Changes. It's fluid. through the lens, as silly as my my content is, I look I view everything through the lens of does this help my longevity or does this distract from my longevity? And those are two very important distinctions when I decide like between what tools I'm gonna use and what makes it into my stack and all that kind of stuff.

Play from 0:27:19 But as far as the burnout goes, there like there's no way to avoid it. And I think that if you're working full time in any capacity, doesn't matter what it is, AI or not, everyone's gonna hit that point. And hopefully the career path in the job you've chosen is one that like when the day gets really long and it's really shitty and you don't wanna do it, it's the thing that like you don't mind being exhausted over.

Play from 0:27:50 And I think back to when I was You know, hustling with the camera and doing motion graphics and doing like VFX and everything. And I was just like, it was fun, but I it I'd much rather be stressed out over this stuff and have creative fatigue over this stuff than what I was doing before. And then on to go even deeper on that, I I look at how I I tried not to be um so insulated and tunnel visioned in what I'm doing.

Play from 0:28:21 That I forget what's going on in the rest of the world and what a lot of other people's situations are. And on the days where I just feel like I'm hitting the wall, I remind myself, like, you cannot take this position you've you've somehow managed to build for yourself for granted. And like this, this is what you gotta do. And the mantra, especially over the past like two years, has been um the only way out is through.

Play from 0:28:49 And so I just Fucking go through it. I love it. think I needed that more than maybe most people. I just needed someone to say that again to bring me back to reality after looking at, you know, a fifteen hundred line code of JSON prompt that's going into a system prompt. And I'm like, is this really where you know I started with making crazy stuff on Midjourney and that that's where I am now. It's like one of those things where you do need some perspective, so thank you for providing that again.

Play from 0:29:15 It's just like you get frustrated the dumbest stuff now. Like I t telling Drew last week I couldn't get C dance to say the word stump. It would it wouldn't pronounce the P. It just ten generations. It wouldn't say it wouldn't say stump. It would say stub. Stum. It would just wouldn't say the pea. I like, I'm gonna I'm gonna rip my hair out. Okay, I'm gonna I'm gonna toss this computer and like and just like light it on fire. that especially with something like C Dance, where it's like, like, I throw so many crazy ideas at it, and I'm like, there's no way it's gonna do this.

Play from 0:29:44 And then those are the ones where it nails it on the first try. And then I'm like, this is gonna like this idea I have tonight is gonna be like low lift, low effort. This is a super simple one, and nothing fucking works. Of course. It's the it's insane sometimes where you're just like, Why? Like, you know, like that used to be like, remember, I mean, probably couple couple iteration cycles ago, where like you couldn't open a door in any of the video platforms.

Play from 0:30:11 It was just like impossible. Or drink. It was like two of the things. Like you couldn't drink from a cup and you couldn't open a door. And it was like yeah, it was like why is this so hard? Like it's there's there's enough training data on it, but now, you know, these are things that we just we look at like, I can Have like a dragon f flying through the sky, blasting a city into smitherenes, like I opening a door is tough. You know, like it does this some of the stuff is like very, very strange. Dude, I can't Omni is killing me with that.

Play from 0:30:38 It's so bad. so we we should give a shout out. So here, you know, give give Tyler a follow here. I mean, we've so he's got 833k on Instagram, which is just insane. I don't even know how the fuck you do that. Um That's that's not does Instagram give you something for a mill? Because they should. That would be dope. so so I so with hitting eight hundred on Instagram, I've now hit like my aggregated million across socials because I have like one seventy eight on Facebook.

Play from 0:31:08 I'm just about at twenty on X. D the the crop the cross post but out of Facebook, bro. I can't even get on there. okay, well one let me just say what the the way that Meta has their whole like like suite set up is just an absolute nightmare. But but but um my they changed something about the way that they they like share from Instagram to Facebook a couple months back when I was in the middle of the food stuff.

Play from 0:31:36 And I would go on Facebook maybe once every two or three weeks just to s see what people I knew from high school are up to, right? Um But then I like logged into Facebook one day and I was like, wait, my my my my personal Facebook page has twenty thousand followers. And then I looked, I was like, shit, I didn't realize it's been posting all the all the naughty food videos here. And Perfect for Facebook. Don't those those videos go crazy on Facebook.

Play from 0:32:05 The boomers love it. my God. Yeah. Uncle Jerry sending him around to all his friends, you know. this out. Here's AI, baby. Um, get I mean, yeah. So I mean, I the point I was trying to make is you got a YouTube 14k there. You've got Instagram, you've got X, you've got Facebook. Follow this guy. He's got great stuff. He's always putting out content. He's got a Patreon too, if you're interested in kind of getting more like deeper into the weeds with this stuff.

Play from 0:32:34 Um Can't recommend that enough. I I am curious, like what, you know, we've we've talked about, you know, the social media aspect of this, but let's talk about sort of like um, you know, what tools are you using now the most? Um, what what is sticking around and what are maybe some of the newcomers that you're like paying attention to and have have really, I guess, become a part of um what you're doing?

Play from 0:33:00 You know, whether that's just creating for fun or for, you know, professional clients. Yeah. So let me let me prelude this answer with how important I think it is. And this is why I think it's great that you guys do this with each other. So you're able to bring two perspectives. Um it's really important if you are trying to be an AI creator, you're doing anything creative with this stuff, to have some sort of community you're a part of. And I'm not even telling you like come be a part of my community or anything, but just like whether it's a small group of friends that are also interested in it, you just need a group of guys or sorry.

0:33:05Tyler’s current AI tool stack

Play from 0:33:34 People who are just as nerdy and obsessed with this stuff as you are because the tool set and the rate at which things are coming out now is so insanely just like overload. And I hear you guys talk about it on here all the time where it's like, you know, update fatigue and tool fatigue. And it's like at a certain point, we just don't even care that the new thing came out this week and we don't even want to look at what anyone has to say on X because it's all the same.

Play from 0:34:04 But I've done a very good job and I try to keep myself very reined in on this part of I look at all the new stuff and I very quickly try to parse out and determine does this f slot into what I'm currently doing? No. What is this good for? Okay, tabling that for when I need that, but other than that, I'm just gonna keep doing what I do.

Play from 0:34:30 Um, my the current stack that has just persisted. For me, is I love making all of my initial imagery in Midjourney, especially since 8.1 came out. I took a break from Midjourney for a while because I got a little tired of like what seven was doing. And then the eight alpha came out and I tried it. I was like, what the fuck is this? And then 8.1 came out and I was like, now we're talking.

Play from 0:34:57 Holy crap. Um, so eight point one, I feel like was a real. energi energizer for me and seeing what it yeah. So mid-journey eight point one for aesthetics and vibe and brainstorming and even when I just need like creative inspiration, just like brain dumping on it until I see something and I'm like that there's a whole story in that image I can run with that.

Play from 0:35:25 Um so I go from mid-journey eight point one, I bounce back and forth between Nano Banano nano ban Nano Banana and gpt2 for the editing aspects um and then gpt two has been my workhorse for my professional and my brand work and i've been able to mostly navigate around the little weird nuanced things that the model does that really give away what it is but mid-journey gpt2 nano and Like I'll always keep Kling in my back pocket for something because Cling is just like, you know, it's like that steady eddy that when you need a good start frame, end frame, it'll it'll do it for you, usually on the within the first four generations.

Play from 0:36:12 But seed dance just has me in a chokehold, man. It j it it just absolutely has me in a chokehold. And then of course, my partner is Claude. And I need to get on Rory's level though, because I have just recently realized. How much I would benefit from just doing everything I do out of clawed code and just having all my context persist. Cause currently my claw is kind of built in a way where I have like individual projects for all the different prompters or types of things that I do.

Play from 0:36:46 But I'm starting to see the bigger picture now. So I need to evolve that system, but there's only so many hours in my day, man. Yeah. hive mind it. You gotta hive mind all those different prompt engines. Be like you need you need to build Tyler, the orchestrator. That sounds like you it's like a perfect, you know, the the offshoot of Tyler, the creator there. and like so they can tap into all that stuff and then just be like, ooh, I want to get a little creative with all these and like sort of mix and match.

Play from 0:37:13 It's d it's it's fun once you get there. You just gotta like like I've always said it's like whenever you have Whenever you have like four or five hours, you know, like it it it's just like I'm gonna I'm gonna build something, I'm gonna like just talk and then like string things together 'cause it's the th the crazy thing is it's like mostly all there. It's just like someone you need to like just work through putting it together, but I think the Which some which sometimes like I actually sat down and had this conversation with Claude last night and I was like, can you walk me through how we can do this thing?

Play from 0:37:47 Because this feels like an insurmountable task with how many different like strings of context for all these different things I do I have spread throughout these projects with you. Can you just walk me through how I can at least start doing this because I know that this will make my life easier? Yeah. Yeah. And it's probably like, let me write like a seventeen thousand line of code plan for you and then it's like dude, all right, whatever, I'm gonna stop now 'cause you you now you're getting crazy.

Play from 0:38:14 I don't need to to go that far. going back to the tools real quick though. So as far as like the the platforms that I I enjoy using this stuff on, because we all have our favorite platforms, right? So um for imagery and all that stuff, I really enjoy working in spaces on magnific aka free pick. Like spaces has just been like I have I I do have like a pro Weavy account, but I go back and forth between Weavy and Spaces.

0:38:15Artcraft, Seedance, and Chinese prompts

Play from 0:38:43 And then as far as C dance goes, I try not to use C dance on either of those platforms because quite frankly, the stuff I make will never get through the sensors. Um I I don't know if you guys have have heard of Artcraft or if you guys have like messed around with it at all. Heard of it, not played with wait, is this this is different than recraft? Yeah, not recraft. So Artcraft is it, I mean, it it's a wrapper, it's an aggregator thing, but um they're they're a startup, they have a desktop application, and they also have like some cool little like 3D workflows they've integrated into it.

Play from 0:39:14 But but I believe they are they are using the Chinese API for C dance. And with that, they're they're my stuff that would literally get. blocked on every other platform I can get through on Artcraft. The only nuance is and I've compared this one to one over and over and over, and it's been true every time. Translating your prompt to Chinese makes the prompt adherence like ninety percent better.

Play from 0:39:44 Ooh, look at that little hack. that. I was I was curious 'cause I know that I had to do that in beginning and it was w working, but I don't know what I'm doing. You know, I can't edit it. Do you know? I don't know what went wrong here. so two notes on that. It depends on what API the company's using. So like I I I use Artcraft and I've also enjoyed using Venice AI over the past couple weeks. Uh Eric Voorhees AI startup. Um, but you know, Venice uses probably the same C dance API that all the other platforms use, which is the American one.

Play from 0:40:16 The English prompts work just fine on there, but Artcraft. Translate it Chinese and you can get anything through. Um, and a little hack that I figure out, and the reason why I I really enjoy using those two guys is they took my direct feedback and I was like, Look, all these platforms are putting a 3,500 character cap on the prompts for C dance.

Play from 0:40:42 You cannot sit there and tell me with a straight face. That this model that can take nine reference images and 15 seconds of video and MP3s and all this, all like all these reference things and put it together and make sense of it can only read 3,500 text characters. Like you gotta be shitting me. And so Artcraft completely removed their character count for me.

Play from 0:41:09 And Venice, on my opinion, um, they made it higher up to 10,000 characters. And I'm telling you guys right now, having those extra characters, the prompt doesn't get tossed out the window after 3,500 characters. C Dance doesn't just not read it. There's no way I could make the complicated stuff I've been making, especially with the anime, without that extra context in the prompt.

Play from 0:41:37 Interesting. That's insight. Yeah. That's so you brought the receipts with you on that, huh? Yeah, that's interesting. my pay my entire page is the receipt. Yeah, that's Well the the so let's let's dive into that a little bit more on the ten thousand side. How are you how are you getting from like I'm using C dance for the first time to I know I need ten thousand characters? What are we reverse engineering? What are you testing to get there?

Play from 0:42:02 Like to to understand that. those ten thousand characters from scratch in an empty box. I'll throw this this up for you guys real quick. Okay, so um this is my Claude. I like using the desktop app, but like I I need to migrate the Claude code, but uh I built this C Dance 2.0 YAML style prompter, and this is my third iteration of it. And uh this prompter has hard coded into it to keep the prompt at 10,000 characters.

0:42:05Claude, YAML, and 10K prompts

Play from 0:42:35 And so the prompts end up looking like this. So let me find like a good one for like the where was the s swamp guardian? So this was I think the the old woman. do how do you have this set up in your clock? Because it's a little bit zoomed out. Do you have this set up as a skill or is this like a project or what are you working with in here? So so I enjoy working with it as a project, but I did also turn it into a skill so that it can like follow me to some other projects that I I work in and like co-work and stuff like that.

Play from 0:43:09 Um and if we actually look at like the the prompts here, so all it is is my um my instructions here and then a master knowledge file. And then prompts or is that sort of like general guidance on what you're looking for? It's reference prompts and then so like I've been building these prompters and trying to like iterate on them and make them as good as possible ever since like Kling two point something and VO3 came out.

Play from 0:43:39 Especi I think I started going hard on it when VO3 came out because I realized how important the detailed prompt was and that I had no interest in sitting down and writing that out myself because quite frankly, that's not what I do. Isn't it great? I it still blows my mind that just something that I I mean, simple is maybe not the right word, but something that like accessibly available to anybody, you know, can be created and save you so much time in this process.

Play from 0:44:10 And it's lasted this long. Like, you know, like when we did this with the with the stills, you know, like I think we knew at the time it was valuable. We also didn't know that that was gonna. still be something that you still use today. You know what I mean? Like that it's it's kind of surprised me that these things are so valuable. And then, you know, I I mean no no shade to anybody like Midjourney. You they've got to describe and you know things like that.

Play from 0:44:35 But it's just like, you know, those those results are wildly different than, you know, results that we might get with our own. And you know, it's I I don't know. I just it's fascinating to me that no one's really like figured that out. And we've still not moved past that. as sort of like this individualized layer, um, in a sense. why AML formatting is particularly transferable. It's a good it's definitely a good format.

Play from 0:45:02 If any do you wanna do you wanna give a shot at explaining that? 'Cause I don't even know how to explain it, but well well I mean I I think about it the same I mean, you know, how if you ask me to explain a JSON prompt, I don't fucking know how to explain a JSON prompt. I know I I know that it takes the thought and compartmentalizes it in a way that these things really enjoy reading it and understand it. And um just so just so like who whoever's watching this has an idea of like when I say prompter.

Play from 0:45:35 I'm not just giving it like the image here, and I'm like, make this woman fight this bull and make it anime style, right? So I I use Whisperflow. I don't type anything because quite frankly, uh the thing I love about Claude and how like this whole LLM process has really um transformed the way that I work. I bro, I brain dump on Claude. I feel bad for my Claude. My Claude has to has to parse through so much weird, crazy shit that comes out of my mouth.

Play from 0:46:04 And make sense of it. And but so like this was the very first prompt for the first scene of the the viral uh old woman with the the bone in the swamp clip that Evolving AI actually just shared yesterday on their page. and like I'm just gonna read this out real quick so you guys have an idea because it's literally this is this is voice transcribed from me talking to Claude. And so all right, we're gonna write a 15-second old school manga animation prompt.

Play from 0:46:33 This has to be in the style of Akira or Ghost in the Shell or Samurai Champlu. Pick one of those that you think makes sense and go with it. We want this to feel very retro. We want to create this world where this old grandmother who guards this ominous swamp bridge and sits on a grave of deceased guardian heads guards the crossing with this massive bone that glows red beneath her. I'm not too sure what we want to say, but let's find a way to include. Wait, I I'm not too sure what we want to say, but let's find a way to include that on the ground.

Play from 0:47:01 And let's really stay inside of this intense color palette that we see in the style of the bottom of the image reference, which this is the um this is the image reference, and I'll explain why I make the image references like this in a moment. boom, where where was I? Uh I'm not too sure, but okay.

Play from 0:47:28 uh I this this reads like if you chopped it down into its pieces, it's like, yeah. That's how you I think we want to open up with her mid fight with a crazy massive Japanese type monster pulled from Japanese mythology. We open on action and she very quickly slays it by clubbing it over the head with her giant bone and then using the bone to absorb the monster into her necklace. She then hobbles and she goes from being hyper dynamic and mobile to hobbling over and crutching her way back to her station on top of the deceased heads.

Play from 0:47:57 All dialogue should be in Japanese. No music, absolutely no music, but we do want great sound effects. We want to start building out a scene here that we can evolve into a longer, more action-packed narrative. So help me fill in some blanks in all of this. Let's start there. Let's remember we want the intensity to open up on a hundred. We want extremely dynamic camera movement as well as interesting compositional shots. You you forgot, make no mistakes. May I I always forget make no mistakes and that's why it makes the mistakes.

Play from 0:48:24 But um see, so the way that I have this built, before it actually gives me the prompt, it reads the prompt and it's like verifying character count against 10k cap. And if it's over, which it usually is, it runs like a bash command and and keeps editing it until it reaches that. And then the prompt comes out in you know, uh code box. So I can just quickly click copy in the middle of the message. And I also have it set a part of ins instructions is it gives me a plain language scene breakdown that tells me what I should expect to visually see from this YAML format prompt because I'm not gonna read the fucking YAML prompt, you know, and make sense of like what it is.

Play from 0:49:09 So it writes this for me. Um, and then I have it give me uh four creative directions that it think thinks I could take things next. A lot of times those kind of suck, but sometimes they they spur some inspiration of where I can take it next. Um and that's kind of like out? Those those next four directions are those like do like a safe route, do like an extreme route, do like a poetic route and do Okay, all right.

Play from 0:49:37 I I like this stuff. And and then I just it it literally is a conversation. And then I was like, okay, the th the creature that it's making in these prompts isn't very good. Let me go make a creature real quick. And then I brought this thing in. I was like, okay, now rewrite that prompt. This is gonna be our image reference number two. this is the thing that she's fighting, and then it you know, it I have it set to refer to everything as at image one, at image two, whatever in the order that I give them.

Play from 0:50:10 Because when I paste this prompt into Artcraft, Artcraft automatically links the at images and I don't have to go back and connect those individually. smart. That's a smart move too. Yeah. And then yeah, this is omni reference for this, right? Just a just a quick interruption. Or you mo yeah, okay. I almost very rarely do I use the C dance um start frame, end frame. I use it to con like to connect really rough transitions.

Play from 0:50:37 Mm-hmm. Good point. So I think that's a good one for I I I'm the same way. Omni's just too good for me not to use it at every step of the way. It's too much context. but but since we're we're here real quick, I'll drop a little bit of sauce on on everybody. So um what I do here for my references, and this was something that I just kind of like stumbled across through my experimentation, and it works very, very, very well. If you give um C Dance a single image from Midjourney and you prompt to do something cool with that image, C Dance really likes trying to lock on to that.

0:50:44Tyler’s reference image hack

Play from 0:51:12 image composition and put that actual frame in the video. And that's really annoying, even when you prompt hard against it sometimes. And so I I like making these character sheets, but I don't want to only use the character sheet because when you make a character sheet from a mid-journey image in GPT or nano, it completely kills the the mid-journey like aesthetic and the vibe that you got out of mid-journey.

Play from 0:51:39 And You can totally tell when you're watching an anime video that was made with GPT anime and I fucking hate it. I absolutely hate it. It doesn't look good. So um I started making these references. This is just a 1080 by 1440 canvas. I'll go into um into Magnific and I will take my, you know, mid-journey image and then super simple.

Play from 0:52:08 make a character sheet with the front back side view of this character on a white background, maintain the retro manga style. And I take that image on the white background. For her, I use a black background just so her um clothing really showed. And then I just come into Photoshop real quick and on the I snap the character sheet to the top.

Play from 0:52:33 But then the secret sauce is I take the image from mid-journey. And I put two copies of it on the bottom, except I offset them. So one is always a little further away, the other one is really cropped in. And then this is my full reference image for C dance. And with giving it kind of like the visual breakdown in this way, it doesn't have any specific frame to lock onto.

Play from 0:53:04 And this gives me maximum prompt flexibility. While maintaining the actual style and aesthetic that I want, with also using fewer reference images, because the more reference images you use, the tighter the guardrails, the less creative it gets, the and kind of like the stiffer the videos feel. Yeah. It's a great point. a very good observation. Yeah. That's this the this is the stuff that I love.

Play from 0:53:30 There's the nerd stuff because you only figure that out if you've done this so many goddamn times. You know what I mean? Like, but yeah, this is Midjourney I haven't I haven't done character sheets in Midjourney in a while, admittedly. Um, but I'm I'm curious if those have uh gotten any better. I'm I'm assuming they're still not quite there as nano and GPT though. I just find that whenever I take a character that I made in mid-journey and then try to expand on that character, it just it kills it.

Play from 0:53:59 It kills the the magic that like I loved about it. And so the a little bit more insight here. Um I have I built a m like a anime specific prompter in it for Claude called the Hardcore Anime Prompter. So here, what what is what's like you guys you guys have like a very loose prompt you use when you test Midjourney.

Play from 0:54:26 What what what what's the phrase you guys use? Which do you there's a formula? thro no, throw just throw one at me. Like when you guys are testing like I don't you guys have a phrase that you use when you're just like testing an image on a mood board or something? Are you looking for something that's more interpretive or like something that's like very matter of fact clear? Or Oh, I mean you could do the work as mysterious.

Play from 0:54:51 You know? that we do. Yeah. here, here. that's what we pulled from uh Severance. Yes. here. Write me four prompts based on the concept that the work is mysterious and make this as absolutely batshit crazy and surreal as possible while staying true to that core principle, the work is mysterious. I like that.

Play from 0:55:16 Yeah. That was a that was a very economic verbal instruction there. I like that. All right, let's see. Yeah. Yeah, yeah. And then more narrative. Yeah. Rory, remember when I was do the last time I ran The Work is Mysterious? It was like an old Asian dude, like doing craftsmanship, you know? So it's funny that it sort of interpreted this.

Play from 0:55:42 I mean, obviously it's anime, but still it's funny that the work is mysterious. so real quick, we're just we're just gonna run these four things. Right now, the mood board I have it on is this mood board I trained on Japanese street photography. Um, which which unironically enough, when combined, so okay, here's the here's the thing I just want to point out before we look at what these images turned into. I've been using my anime prompter that writes things f like hardcore with anime in mind, but using real.

Play from 0:56:15 life mood board things for it and not necessarily anime ones and mid what mid-journey does is just like so okay so these are these are all I was just about to ask you if you're like running this Niji or not, but no. Yeah. Yeah. Sick though. I love the blending of the styles for that. Like it's just smart and I will I do the same thing with that sort of like how much art can you infuse into realism kind of thing.

Play from 0:56:44 You know what I mean? Like that kind of stuff. It's it's great. This one's aw I like that. Sick. I don't know. That's awesome too. Yeah. and This is the best th this is the one thing that just like you can't get away from Midjourney 'cause I don't feel like describing everything. You know, I I don't wanna describe yes, this. Yeah. let's take that prompt because I really like that one.

Play from 0:57:10 And I'm going to put on um this new anime mood board that I just recently made. But then I also want to try it with this combination of these two and let's just see the difference. So this was the prompt with the girl with the mouse with the cat. Alright. Sick. I love it. This makes me happy.

Play from 0:57:36 You you know what this reminds me too of like this you know, this new sort of unlock with you know when you take this into like something like nano or GPT, it may be the same thing, but it's like put this into a different style, just in work, like make this realism. And it's just like turns the exact composition, scene, et cetera, into something else, too. And it's sort of like an interesting testing ground in that way, too. Like you could, these are versatile. You're not locked into this. You're you know, if you you've created this scene that is essentially burned into right the the pixels that you can adapt it to anything.

Play from 0:58:10 Um, so it's kinda like you're telling the story and that story can be retold in different styles as well. This this to me, so you know, I I wish I could say that my creative process was very like a matter of fact. But truth be told, I really like I wanna be stoked about whatever I'm dedicating those couple of hours at the end of my night to making. Cause honestly, by the time my workday is done and I've done like all of the adult things that I have to do, I I have maybe three to four hours at the end of the night to put together something to go on Instagram the next day.

0:58:15Dream sketches and creative prototyping

Play from 0:58:46 And that's just kind of like the chaotic time crunch I operate under because of what everything is. And I refuse to automate my creation process because as much as I love Claude, Claude can't make the stuff that I do and it doesn't have the ideas that I have. Um, and I and I want my ideas to be present in all the stuff that I put out there. But I I go through and I just make like I throw like these batshit crazy ideas at Claude.

Play from 0:59:17 And like I just run images in mid-journey, and then I'll see an image and immediately like I like I'm definitely gonna do something with this one. And I'm just like, there's a whole story behind this image that's waiting to be told, and that's That's the Instagram post. I I just figure out what that story is. You you make me you're making me feel like I forgot who I am, if that makes sense.

Play from 0:59:44 Like that's it's what I used to do. It's just like, I like this I like this image. I'll build the whole video around the image. I haven't done that in so long. So this this is this is this this is like therapeutic for me. Thank you. I'm f feeling a little bit better about myself now. well, it's more tr it's more true now than ever with with seed dance because like literally like even I was just messing around last night. I didn't post anything today because I needed a a night off. But like this is like so fun just to make from a single And that's one that's one single generation?

Play from 1:00:27 Okay. Great. Just awesome. and so like to to also avoid burnout, I have um, you know, I've come up with formats that on a night where I don't quite feel like um doing anything. Uh I I came up with this format. I'm calling I call this series the uh dream sketches, because I essentially take random images that I really like.

Play from 1:00:58 And I run them and just try to get one solid fifteen second generation. And then I just transpose them against each other on top and bottom and with a cool song and with the sound behind it. And it like this one I love because it reminds me of like watching MTV and Celebrity Deathmatch. death match. Yes, there we go. that's a good association right there. But this is how I prototype possible ideas and then see which ones the people respond to well in the comments.

Play from 1:01:31 And then I know which one I can go back to and, you know, turn into like its own little thing and give a story to. Hmm. Also, I could tell you're in post production just by looking at how much stuff you have on the on that on that there's so there's so much oh my god, look at how look at how beautiful that is. The yeah, and it it I don't know if you can tell, but none of the so I I basically exclusively run 15 second generations in C Dance and I think about each 15 second video as like a package of clips rather than trying to get a perfect 15-second cut because I know that I want to go in and edit it and that and that's where the real flare comes in.

Play from 1:02:16 And so like each one of these, these are all you know, sections that I deemed worthy from fifteen second long clips that I just like put together and just like made work. Don't you don't you love when it does something unexpected and then it takes your story in a different direction too? You get like a it's like, that's cool. Let me go that direction also. You know, like it's it you you let it lead sometimes, like what you're doing with some of the image gen, right?

Play from 1:02:45 Like sometimes the story gen from the video side too. I feel like the more you try to fight against against the AI and like you try to like, no, you need to do this, unless it's for like a client project, like you're you're killing the coolest thing about it, which is like I have my my kind of um creative biases. You guys have your own creative biases. We all do. We have directions that when we're presented with a choice, we will naturally kind of gravitate in a certain way.

Play from 1:03:17 Um I want AI to pull me in ways that I would have never thought of for myself. Great point. And that's you know what? I feel like we have to surrender to that sometimes 'cause we've been so hellbent on control. I think probably 'cause we didn't have any control for the longest time, and now we do. And it's like n now that I have it, I don't want it anymore. You know what I mean? It's like now that you can control everything, it's kind of like I want to go back to not having the control.

Play from 1:03:44 But it's uh it's definitely definitely, you know, something that I miss is just like riffing and then going, Do we do we uh do we Lose the stream here? Are we still good? Still good, baby. We're listening to you. We're locked in. What are you talking about? I see freeze frame on on Tyler down here, so sorry about that guys. That's was just sitting still. I was listening. you, brother. Um I'm getting a freeze frame, so maybe that's just me.

Play from 1:04:09 I like, Am I just talking into the void here by myself? Is this my own head? we did mention that we would tap on some of the updates or just like things happening in the space too, on this pod. So as we kind of wind down here, let's talk about some of this stuff really quick. Just give me quick reactions, anything you guys want to add, right? So Reeve. Reeve just dropped 2.0, 4K image. Um, have you guys played with it?

1:04:10Reve 2.0 and layered editing

Play from 1:04:34 Have any initial thoughts? Does this change anything for you? Are you excited about it? Does it yeah? I I'm I'm just curious, like, because I think we again we touched on this last week. I mean, we get in that mindset, like we hear something that's new, we're like, cool. Yeah, fuck yeah, it's gonna change everything. A couple of days go by, nothing changes, right? So, like, d does this have sticking power? I mean, I don't know that any of us here are power reveal users, for example.

Play from 1:05:00 So, does this make you want to go in there and start really generating from it? I guess is the question. I gotta say I was a fan of the first iteration of Reeve. Like I really liked it. I thought it had like a a lot like a very high editorial content and was like uh somewhat had like some of the Midjourney magic where it was like this is I could just type in something and it would go in a lot of different directions. But kinda left it for for the most recent updates 'cause I was like, mm, I don't know how much I like it now.

Play from 1:05:29 But there is something genuinely interesting about this. I'll just pull this up real quick. This is let's see. Yeah, what's cool, it's just like When you generate now, it generates things from code. So it generates in layers. So like you can just select different pieces of it. Like it doesn't do this after the fact. It does the whole thing in layers. So if I didn't want, you know, trees there or whatever, and I just go, you know, prompt of um, you know, rocket ships, right?

Play from 1:05:59 Like you can just change individual layers, right? Like I don't know if that was the best idea, but you're just getting getting the the feel for it. So I think this is Interesting. I don't know how much I'll use it. I just think it's like kind of all right, I see what they're doing here. Like we're going at it image generation a different route. Um there's a a feature that's gonna get stolen. I mean, cause that that that level control is nice, dude.

Play from 1:06:25 I I mean, I so so so I'm gonna answer my own question. I don't know if this chain like I don't know if this makes me want to start using Reeve regularly by any means. I'm not I'm not sure yet. I'm not gonna go that far. Um, but like to your point, Rory, like this this this edit thing is interesting. So you've got this, then you've got the sub layers, right? And then look at this. Like, even if you click face, now you've got sub layers in the face. Mm-hmm. Is it but is is it's it it's separating those layers as it's generating the the image so they're already separated.

Play from 1:06:56 It's not like separating it as you select stuff. No, it's it's happening it's happening via the generation like in real time. So you can see it, like you see it like the way they've built like the UI. Is like you can see like the different things being generated, then it just like populates like this with layers like a Photoshop file. wonder I wonder how it constitutes like what should be on its own layer. It it's interesting, right? Well it's it's 'cause it's writing the whole thing from code. So I think it's I think it's like legitimately coding the entire image pixel by pixel, if my understanding is correct.

Play from 1:07:28 Um, you know, that's that's your non technical opinion right there, of just of face value. But I would say that that that feature is genuinely genuinely useful. Um I uh for all these things though, I I reserve final judgment until I have the chance to put it through like actual working paces because the examples all are always the best case scenario, right? Yep. So true.

Play from 1:07:54 And you never know. Like it's sat like to me it sounds like a great concept, but in like you said, in practice it might not be the same. But I think like this is just something that everyone's kind of been wanting 'cause the the one thing that I find with video or image, the smallest changes are really damn hard. You know, like I I just go into Photoshop half the time, you know, probably more than that. I'm like I'll just do this real quick. But Rory, have you played with this enough to know okay, so if I'm gonna if I'm gonna make an edit to this I here, which I've clicked into, right?

Play from 1:08:25 Am I just typing it in right here? Because that's what that's what So it's making it seem like That's the way I would think about it and then I hit apply, right? Is it not changing anything? I don't know. I mean this one didn't change anything. I mean it did change the eye but it not it didn't listen, right? Um This is just the I So I wonder if I'm doing this right. I probably should have tested this out before we got on here, but I don't know. That I think like that individual layering thing, I think the point we're trying to make here is like that is a very interesting feature, right?

Play from 1:08:57 I mean, you know, to selectively zone in on anything and then not worry about anything creeping outside of that being morphed or distorted or changed, right? Even nano and GPT sometimes fuck up with that shit. So uh, you know. I think Yeah. It's just like a again, you see it. the most annoying part using these tools in the professional way is when you do have to do image edits and you genuinely want to do the edit in nano or GPT and not go the Photoshop route.

Play from 1:09:33 And then just like to change that small thing, the amount of total image degradation degradation degradation degrading that you that you get is is really hard to swallow sometimes. Hundred percent. You're officially now 100 part of the Fast Hours podcast. Whenever we we we trip over at least one to two words an episode. So that's now now you're officially part of it.

Play from 1:09:58 But that's did it I one hundred percent agree. Like it there's nothing more frustrating than like I think video is even worse too when you get like I I made a video the other day with a with a little golf ball that just popped out of nowhere, like it just whole video great, and this little golf ball popped out of nowhere into the screen. And I was like, How am I gonna get rid of this? Like there's I have to go through so much effort. I was about just regenerate it. Hopefully I get it. But you you know, in client projects you can't do that most of the time.

Play from 1:10:24 You have to actually like you have to go the VFX route or the uh you know, I'm gonna have to go segment that and rodo it out. Like it's gonna be a pain in the butt. But uh someone's gonna need to come up with a micro editor that works. I don't know why. We can't we can't nail that. I'm still I'm still I'm I'm still waiting for my Midjourney editor, man. I'm still waiting for it. I'm still waiting for it. I the question, right? Because I'm curious how how how you think that's gonna live up to the high and and will it because if it's good enough, we've we've talked about this before.

1:10:38Waiting for Midjourney’s editor

Play from 1:10:58 Is it if it's good enough, people it's gonna bring people back and it's gonna keep people in this tool. And, you know, like connecting the claw to this and that and the other and all these tools, I mean like that's cool. I mean, you can do it in one place. I still kind of like the old school way of going into the platform personally. Call me old school. But like I I would rather spend my time in mid-journey, the muscle memories there, and get back to that place where I'm doing the majority of what I need straight in the fucking platform.

Play from 1:11:26 You know, I just hope it lives up to the and not just that, I want to edit images with the same mood board they were made on. Like that's that's that's the real thing. Cause like so right now, uh especially for like my split screen series, I will take an image, I will make it black and white, I'll give the people 24 karat gold grills and put gold chains on them and make everything black and white except for the the grills and the chain.

Play from 1:11:55 But I can only do that with nano because GPT, fun fact, GPT is fucking atrocious at making grills. Like they it can't do the grills. It can't do the grills. Nano nails it. It's so good. But um it, you know, just even doing that one little thing takes a little bit of the the the juice away from the mid-journey vibe. And I want to just do it in the platform.

Play from 1:12:23 I feel like the image model, the eight point one alpha, ha is so good. At least it to me it's like leaps and bounds above what seven was. I I I I feel like they could do it. I'd like why why can't they do it? Right? With the editing. We're asking ourselves the same question. We're just like Yeah, when? I I thought the initial thing that we we talked about and what they were talking about on office hours is like after we do eight, add a model, focus on ed model.

Play from 1:12:55 You and I know things changed, you know, that like eight didn't quite land the way they thought it was gonna you know they probably should kept that beta, you know, or close circle for a while because there was so much fluctuation with the results and and the quality. You had some that were like super tack sharp images, you're like, hell yeah. And you had others that were just like Yeah, it doesn't feel like the same model, you know. And then eight point one has been really solid, but now they're, you know, now it's eight point two.

Play from 1:13:21 And, you know, we're talking about nine. But like, what is the edit model? When are we getting it? You know, and what exactly is it? We've talked about it kind of like having some of these features of NAD. We still there's a lot of ambiguity around it. I think partly is that they're playing that close to the vest for a reason, but at the same time, man, I I hope it doesn't let us down. You know what I mean? Yes.

Play from 1:13:47 Resounding yes. r Rora, I mean you you you guys know this, but you know, like there's been plenty of features they've launched that we've waited for and it comes. And then no one uses it. No one uses it and then it's like, damn, I wish y'all would've focused on something else, you know. What was that uh what was that open world Open world. Oh, patchwork. Yeah. That was like a that was like a memorable one.

Play from 1:14:14 That was I remember just remember sitting there being like this is gonna be so cool and then whoever was whoever was going through it and like inst not instructing on it, like demoing it. It was just like what? This goes where to what to which to which world? To back to where to link to what? I'm like, this is you know, it was weavy. W looks much better. And spaces look much better. It's the same same idea in theory.

Play from 1:14:41 They just like couldn't execute it, I guess, from simplicity. in a spaces type thing and like do my generations with my mood boards and edit in there, I think it would make competing with Midjourney, no matter which company you are, very, very difficult. Yeah. I mean, I'd I'd use it all the time. It's like I use it for I use it for environments the most now, like 'cause I hate the environments that ChatGPT and and Nano build.

Play from 1:15:11 Like they're so stale. Like they'll build a somewhat of like a decent character, like if you ask it to, but like the the environments always they like lack on that. And Midjourney can get some really cool like background textures like just overall composition vibe, like it just does it so much better to me. That's just my my thoughts on it. But I tell my community, man, I tell them like, you guys gotta fight with Midjourney to like get what you want and make your mood boards and build up your style and your aesthetic.

Play from 1:15:40 Because I was like, that's that is the sh quickest surefire way right now to be able to differentiate differentiate yourself visually. I was like, if you guys want to make all your imagery in GPT or nano, like that is totally okay. But don't be surprised when your stuff is indistinguishable from everybody else's. That's a really good point. Yeah. It feels like you're maybe a little bit more boxed in stylistically, right? Totally. That's I think that's probably a good way to to close this one out.

1:16:02Closing thoughts

Play from 1:16:09 I kinda like that. Yeah. Yeah. Dude, that was epic stuff. Yes. Like just wanted to thank you guys so much for having me. This was the even the um yeah, like being able to come on here and talk to you guys almost like it it's very validating for me because like I said earlier about having like community and whatnot, you guys, whether you know it or not, are probably the a a very loud um sounding board and voice for people who are trying to figure out.

Play from 1:16:48 Where should I actually allocate my time and resources with these tools and what will actually move the needle for me? You guys are a very good compass for that. So please don't stop doing these because aside from just doing it with each other and like helping each other, I'm sure you guys are helping a lot of people that aren't lucky enough to have found like the friend group or the Discord community that they can work with and they're just trying to figure this stuff out the best they can.

Play from 1:17:18 Dude. Thank you. We with this does this does mean, you know, like we said, it does mean a lot to us that people actually listen. You know, then ac like actually apply some of this stuff and come back and listen to us again or recommend us. Like it it's it's cool. We don't say it enough probably, but it's nice to it's nice to hear that we just don't talk into a void and that there is someone on the other end of the line. Yeah.

Play from 1:17:43 same I feel the same way, man. Someone in my in my Discord yesterday posted some like cool work he he got hired for and did for like a real estate company. He's like I've used a bunch of stuff that like J Boogs has taught us through his workflows and I don't do anything with real estate, right? But it's like hell yeah. Yeah. good. It feels good. It it it's it's nice to help people. And it's nice when people tell you that it helped. I like that's that's also great.

Play from 1:18:08 That feels probably better than doing it yourself, you know. It's it's it so that's the vibe. And if you guys aren't following, like we said before, if you're not following following Tyler, if you're not jumping on a six hour live stream, you know, come on. What are you even doing? So right there. him. Instagram you're saying Instagram is probably your your baby, but you've got fourteen K on YouTube now too.

Play from 1:18:34 Yeah. YouTube I was on until they demonetized me because of because I posted two sexual food videos. Yeah. That'll that'll do it, YouTube. God. And I can't even be mad at him for it, but I I you know, it it kind of it kind of um and I wasn't really making money on YouTube, but it was it demotivated me. However, YouTube is where I um primarily run the live streams. Yeah, cool.

Play from 1:19:00 But well, if everyone has sorry guys, I gotta gotta make a quick little exit here, but if anyone is still listening and you haven't done this yet, you know, just like, comment, do something. I don't care what it is, and then you gotta tell somebody. That's the rule. Tell your tell your Discord buddy, if I'll YouTube employee, it's like Discord buddy, tell somebody. 'Cause you know, we're here. We're gonna keep doing it. So that's the deal. Tell your favorite vegetable on your countertop.

Play from 1:19:28 You know? Like, there we go. Perfect. All right. Well, yeah. Thanks. Thanks everyone for joining us. Thanks, Tyler. Appreciate you, man. Rory, good looking out. And we'll see you guys in the next one. Peace. See ya.