Ep. 63: Midjourney V8 Countdown + Kling 3.0, Seedance 2.0 and Higgsfield Fallout

0:00:00Winter chaos & NYC survival

Play from 0:00:01 8:88:88 0:00:01 What's up, everybody? Midjourney Fast Hours podcast, episode 63. Drew Brucker, Rory Flynn. What's up, man? It's been a minute. Glad to be back. a minute. just, you know, two busy, two busy men trying to just make it through this winter. I don't know if your winter's as bad as our winter, but I'm over it. I'm over it. Yeah, I, you know, it reminded me of, I think last time we talked, you were just like, you know, when they come by and it snows and they do the whole scraping of the streets, they just basically bury your car.

Play from 0:00:37 8:88:88 0:00:37 ah And so it's funny, because I was just listening to uh this comedian yesterday and he had a whole skit just based on that. So I was like instantly thinking about you, but I'm sure that snow's been piling up for a while, man. It really hasn't been that bad here. I mean, we... had maybe two weeks where it was just really, really fucking cold. And then besides that, it's really not that been that bad. I mean, we're back in the high fifties, mid sixties now.

Play from 0:01:08 8:88:88 0:01:08 We the snow still hasn't melted. It's been a month. People still haven't moved their cars. It's disgusting. It's dirty. It's like, I'm over I'm like totally done with it. I'm like it's like enough to make me leave New York for the for for good. I can't. And then we're you know, we finally get it's warming up today. So things are starting to melt. But then it says three days of snow. at end of the week, it's like as soon as it's here, it's gone again, you know, or as soon as it's gone, it's here again.

Play from 0:01:36 8:88:88 0:01:36 And then I'm like, this is if you guys can't tell seasonal depression is a real thing in New York, because we don't have luxurious, comfortable, big spaces. And when you're forced to sit inside for two months, because it's been below zero, and the snow doesn't go anywhere. And your children are getting absolutely psycho. Your dogs are like, I need to go do something other than sit in this apartment. You know, it's like a it's like torture. So And how are you supposed to see the hills on the street when there's snow on the ground?

Play from 0:02:05 8:88:88 0:02:05 Dude, mean, nails, dog poop bombs, where people just apparently during, uh you know, during this time of year, people just let their dogs poop on the street and don't pick it up. Yeah, is this the thing? That was the other part of the skit. The dude was just saying like, I guess there's dog shit. Well, you don't know how many dogs pee in New York until it snows. And then you really see how much pee there is.

Play from 0:02:30 8:88:88 0:02:30 And then you're like, OK, there's the whole all the things are yellow. It's like it looks nice when it's here. It's white. then like, you know, two or three days later, everything is like black and yellow. And it's disgusting. Your car is covered in salt. And it's like everyone's car is disgusting. It's it's not it is not great here this time of year. And it needs to end. I'm going going to London. tonight.

Play from 0:02:56 8:88:88 0:02:56 And it's gonna be about 10 degrees warmer. And I'm like, amped. I'm like, hell yeah. 40 degrees. Yeah, I'm like, Oh, 45. Cool. Like, and it's gonna rain every day. Don't care. Let's go. Get me out of here for a little bit. But it's enough of us complaining about the weather. is like that last point where it's like, this is still totally cool. Like not cold. is fine. We've seen 40 like once in the last 60 days.

Play from 0:03:25 8:88:88 0:03:25 it's, you know, it's not, it's... we're and we should we should we're talking Fahrenheit here, so I have no idea what the conversion is, but yeah. For our friends overseas using Celsius, last weekend it was negative 27 degrees Celsius. So there you go, which is, think, you know, when they start getting close to each other in the negatives, you know, Fahrenheit and Celsius, that's how you know it's bad. So it was pretty rough.

Play from 0:03:52 8:88:88 0:03:52 regardless, uh know, brighter days ahead, I think, probably got, we got a lot of stuff. going on in the space. think everything is... I feel like in the last month, we say this every time, but it's like in the last month, we've taken like a quantum leap to the point of where like, I don't know. It's like the quantum leap like every couple months where it's like this feels so foreign as to where it was prior. Like something about Claude Code, just whenever that released, it feels like everything else just got a boost.

0:03:59AI’s quantum leap moment

Play from 0:04:23 8:88:88 0:04:23 Now we're looking at all these different models, know, Kling 3. uh, C-dance now blowing everyone's mind, um, yeah, that I- Hollywood does not like it. Hollywood should hate it. I mean, you know? you know, they just, got to throw shade on it. Yeah, they, they, and they will. it's it's for one of two reasons, you know, neither of which is a good reason.

Play from 0:04:50 8:88:88 0:04:50 One, they either just don't understand it or what's going on. Right. Like or two, like they're actually they have the sort of nerve, I guess, to blindly just throw it under the bus. Like it's not going to be this competitive thing at all to them. And it's not going to impact them, which I mean, You see that all the time in, I guess, the SaaS world, but that's a dangerous game to play.

Play from 0:05:20 8:88:88 0:05:20 And I think it's just like we've seen, you don't have to take the examples for exactly what they are. You're trying to gain insight from what those examples are showing, right? Just like we've been doing this whole time with images and video, right? At no point up until today has any image or any video been perfect, right? Hasn't. There's always like a little flaw here or there, but we've been... hinting at the fact that it's never gonna be worse than this.

Play from 0:05:46 8:88:88 0:05:46 It's only gonna get better. It's only gonna get better. It's only gonna get better, right? Until we get to this point where now they are direct competitors in every single way, right? And there are different ways to consume content, but I think the fact that now you can have anybody, uh quote unquote, put together visual stories through images or through video or whatever it is, and that's democratized.

Play from 0:06:13 8:88:88 0:06:13 And a lot of us didn't have that capability before, right? You had to be a photographer that maybe worked for Sony or for a cool band or for a cool brand or something like that, right? On the video side, you probably had to be in Hollywood. You probably had to have some real good connections or budget or whatever. And so all those things have lowered and there are great stories that can be told. There will also be some shitty stories that will be tried to be told, right?

Play from 0:06:39 8:88:88 0:06:39 We know that, but I just think like... to not understand that fully is completely naive. I don't know who they're trying to convince, if they're trying to like talk out loud and convince themselves about it or what, but not buying it. I think the closest parallel is like radio to podcasts where like radio was just like, ha ha, look at these people doing podcasts. And then it's like, you know, you realize the podcasts don't have the restrictions that the radio has.

0:06:51Radio vs podcasts analogy

Play from 0:07:06 8:88:88 0:07:06 There's no time slots. There's no, you know, schedules. You don't have to be in person. You can do this remote. You can, you you're not beholden to the advertisers. You don't have to be on for 15 minutes and advertise and then come back and then have sponsors and you know, all this other, the model. sucks. know, podcasts do whatever you want beyond multiple platforms is have to be on one streaming channel like it can be global and not local. you know, there's it feels very familiar as to like that type of transition.

Play from 0:07:34 8:88:88 0:07:34 And again, like podcasters like us, we don't have corporate overlords saying we can what we can do or what we can't do. We can just do it whenever we want to or talk about whatever we want. Right. So, you know, it's like same thing with, you know, corporate corporate news media and independent journalists. Right. Like they They start to see the attacks like, they're not real journalists, like, because they don't have like the access. And it's like, well, they get the access because they have a following now. Well, they're not really that good.

Play from 0:08:00 8:88:88 0:08:00 They don't really uphold the standards of whatever. But it's like, we don't care. You know, it's like we don't care if you just have a name attached to your, you know, your name attached to your name. Right. Like it's do you provide good stuff or do you not provide good stuff? And if you stop providing good stuff, which feels like Holly was just been remaking the same thing over and over and over again with the same formats over and over and over again. And then it's like, Oh, people come around and they're making stuff that's actually compelling and interesting.

Play from 0:08:25 8:88:88 0:08:25 It's like, why, you know, because they don't have the, they don't have the restrictions and we don't have to have a $500 million budget and pay, you know, X amount of dollars to certain actors and whatnot. Not right or wrong. Like you don't have to roar. You don't know how many jobs are getting taken away. I'm like, you know, At end of the day, consumers still make the decisions. We watch what we watch and we pay for it. So, sorry if people want something that's not that.

0:08:55AI series vs Hollywood model

Play from 0:08:55 8:88:88 0:08:55 And whose fault is that? It's not our fault for chilling for AI. making a bingeable series that really becomes, you know, I don't want to say mainstream popular, but gains traction, legitimate traction, right? Where that, that to me is like another inflection point, because then what happens then, right?

Play from 0:09:20 8:88:88 0:09:20 Like to your point, they don't have, maybe we don't have to wait every week for an episode. Maybe they've already got them done and maybe they could release them however they want to release them. Maybe there's, you know, like if someone were to put together that and do storytelling at a high level and then all of sudden do all these things that, you know, these other shows do to monitor, like you could sell merch, you know, on this. You know, there could be communities based on this particular series or whatever it is, these theories of what's going to happen, right?

Play from 0:09:49 8:88:88 0:09:49 Like all the things that we get from a typical Netflix or HBO. series that comes down the pipe. I mean, I was just looking at the, you've probably seen this now, but somebody put together, and maybe you can help me identify who exactly this was. I've just seen it a few times. But basically this John Snow, what happens 25 years later?

Play from 0:10:14 8:88:88 0:10:14 You know what I mean? I mean, this is again, not perfect, but you can see where this is going. And so what if someone were to actually do this and put this together and that gains traction? Also, what happens with sort of, okay, HBO owns Game of Thrones, somebody put together this whole series, takes Jon Snow character, right? Like we've got some interesting times ahead. And I don't know, should we play this in case people haven't seen it?

Play from 0:10:41 8:88:88 0:10:41 Maybe we just play it and see, but. Are they gonna be able to hear the sound? I think so. didn't put my earphones in on purpose hoping that they could hear it. So for anyone that's just listening right now, we're watching a... basically a game of Thrones.

0:10:59What a Game of Thrones (AI) sequel actually means

Play from 0:11:09 8:88:88 0:11:09 I don't even know what the best way to describe it is. Not a spin-off. It's just 25 years later that someone created with C-dance. Yeah. Do it. It's. oh I mean it looks good, that's the thing.

Play from 0:11:34 8:88:88 0:11:34 Is that like? So I turned on, I turned on what's it called the other day? Too Fast Too Furious, right? When was the last time you watched that movie? Long time ago? Yeah. second one. Yeah, it's been a long time.

Play from 0:12:00 8:88:88 0:12:00 I mean, I remember Tyrese. I think that's the one Ludacris is in too, right? If I remember correctly. I remember the muscle cars in that one. Yeah. So there's a couple scenes there that I saw where I've always remembered it because it was so it was such bad. Like, I'm not knocking anyone here. But like, when you look back on it, it's like the first one didn't have like any VFX. And the second one, they had to like VFX some car scenes in and it looks like a video game in comparison to the rest of the movie.

Play from 0:12:27 8:88:88 0:12:27 Like it doesn't look like it fits. And now I'm seeing stuff like this. I'm like, like, this is the solution to that, where it's like, if I was to pull up hold on. Let me see if I can actually, if I search like, Too Fast, Too Furious, know, VFX car scenes. See if I have an image of it. Like there's some really bad. right? Like goes to this whole thing of people, like we've been talking about this being AI end to end, but again, that's where people, again, fall off.

Play from 0:13:01 8:88:88 0:13:01 They forget about the fact that this can plug holes. B-roll, like you're talking about, maybe this is, you know, specific parts of the movie where it just doesn't make sense. You know, maybe you forgot to shoot something and you're in post and you want to add something. I mean, that opens up an entirely new door that probably they didn't have as a possibility before, right? Like it amplifies what you can do with video. You know, it's not necessarily just a replacement for the current, you know, format itself in the way that it's done.

Play from 0:13:33 8:88:88 0:13:33 I like to make that parallel too from the photography standpoint, right? Like you could still be a photographer for a brand and then what happens, like you get the shots back and you're like, I wish I would have shot it from this angle. I wish the lighting would have been a little bit different here. Or we forgot to shoot this color way. we forgot to do, right? Like this is a uh multiplier. This is a leverage multiplier to then fill in those gaps, right?

Play from 0:13:59 8:88:88 0:13:59 You're eliminating uh mishaps or accidents. And then you can also amplify this further on top of that. The same thing can be true here. And people forget that. I think they just always are thinking about one for one. That's like this, this one to me, when I looked at it again, I was just like, Ooh, you know, I go back to, I go back to stuff like I looked at Lord of the Rings.

0:14:14CGI patchwork & filmmaking

Play from 0:14:25 8:88:88 0:14:25 Um, you know, I watch Lord of the Rings like all the time and there's like maybe five scenes in the entire trilogy that you can be like, that's 100 % CGI, you know, and like, looks like it's a little bit CGI, but that was also done in 2000, you know, like there was 25 years ago and the rest of it's just so technically. outstanding because it's the craft is impeccable, right? But it looks better than what they did in the Hobbit, which was like all CGI, right?

Play from 0:14:51 8:88:88 0:14:51 And it's like the CGI now tends to be like a hindrance a lot of times in terms of like the visual continuity of stuff. you know, again, I'm not knocking CGI artists like I understand. I've been trying to do a lot of stuff with 3D lately, and I understand how goddamn hard and intense it is. just saying from an overall aesthetic standpoint matching, right? Like it's we're getting to that point where it's it's pretty good at that. Like if you need to patchwork stuff, like it's not necessarily, you will never have the control that you have with, you know, 3d and doing it that way.

Play from 0:15:23 8:88:88 0:15:23 But can it work for some of these scenes in too fast, too furious? I think it could. No, it's a two second clip side by, you know, side profile, low angle view of a, know, the front end of two cars, like jostling in with each other, like going, you know, who's in front, who's in front. It's not like You need have like a full 3D keyframe VFX pipeline for this. It's like a two second shot. And if you replace that, it looks so much better. Right? so, you know, these are little things, but again, Hollywood's just Hollywood's just looking at the, at the behemoth that they've created and they're, having a hard time grasping the fact that it might not be exactly how they want it to be now.

Play from 0:16:02 8:88:88 0:16:02 And it's not, they don't have the control. Yeah. yeah. about it. They don't want to be the bad guy that started to introduce it into the process. Well, the other thing is that when we get into this Claude stuff later, I basically had it write a full comedy sketch based on a lot of different inference points and lot of different pieces of things.

0:16:14AI replacing exec decisions

Play from 0:16:32 8:88:88 0:16:32 It was basically saying, bringing its own opinion to the Hollywood thing. And I was like, the people at the top are replaceable. It's like, what do they do? They look at numbers and they say, this should work because we've done this eight times and it should work again. And it's like, that's the easiest thing we can do with AI now. It's like the people with the real talent are getting replaced by the people at the top. why are the people at the top not getting replaced? Because they're pointless. know, like they just have money and make the same decisions that they've been making over and over and over again, but are not hard to automate.

Play from 0:16:59 8:88:88 0:16:59 Right. It's like, is, you know, is it hard to say, let's do it, you know, fast and the furious 10? Like is that you know really hard to say is it hard to say a remake Lion King for the fifth time is that you know I understand there's a lot more to that but still it's not really like oh how someone who's technically operating a camera like that's a little bit harder in my opinion an estimation then you know it's sort of making decisions on strategy like that but here we are I'm sure everyone has you know if you got you got difference of opinion drop it in the comment section and also while you're there while I said comment section, you know, do a little like and subscribe.

Play from 0:17:39 8:88:88 0:17:39 We'll also do our little thing at the end. Right. Well, so yeah, man, it's just the sea dance thing is is going to really just it's the new Sora sort of like hype cycle, but better than Sora is, you know, like the Sora thing sort of blew everything up into the mainstream. But this is like, oh, And I really went into, I want to shout out to Tim Simmons over at our buddy over at Theoretically Media.

0:18:03Seedance & model hype

Play from 0:18:10 8:88:88 0:18:10 Did a good video on how C-dance is sort of constructed and it's very similar to Nano Banana. We can get into that later, which was, I thought it was pretty cool to see that because I haven't gotten access to it yet, but it makes sense. And I think that's probably what we can all expect from the future stuff that's going to come out. And we'll talk about that a little bit at length once we start diving into it a little more. But again, we're just... Still just setting the stage here, 18 minutes in. I mean, well, okay, so we had since we last I mean, had clang 3.0 also drop, know, and then you had C dance to, you know, we're just getting and then I would, I would imagine the oh, four will be here.

Play from 0:18:49 8:88:88 0:18:49 think I think I saw something like maybe May on that feels kind of far away. I don't know for four. Yeah. I mean, that's that's still would be two and a half months away. When we remember when we showed sort of like the release patterns, right? Like Kling was dead on to his release pattern in terms of timing. So if he was on, it should be probably in.

Play from 0:19:14 8:88:88 0:19:14 uh Well, it's either that or Nano, another Nano model. It'll be in March, like either early March, like right before Q1 sort of rolls out to a close. So that's my guess. And it's been a while since they've updated VO, so I would assume that they're just either one waiting, like had something pretty dialed up and just waiting for this sort of hype to fall on Seedream, Seedance, and Kling.

Play from 0:19:41 8:88:88 0:19:41 But yeah, think we're getting, you know, we should be sort of prepared for March to be a big drop cycle of things. Yeah, yeah, including including Midjourney, including Midjourney. Still waiting on that V8, by the way. started in December. Here we are going into March. Yeah. are getting close. Let's just open this up while we're here.

0:19:48Midjourney V8 timeline

Play from 0:20:08 8:88:88 0:20:08 Let's talk through this really quick. So there's a rating party, round two happening right now on Midjourney. Ah, round two came out? Alright, I did round one. so here you go right here. You know, just go to updates. It's right here if you want to help rank those images. um But here's the thing, right? We'll be doing two more rating parties this week. So this this just was announced. What was the 16th two days ago? Yesterday, Monday. um Two more rating parties this week.

0:20:15Midjourney rating party (Round 2 + on)

Play from 0:20:35 8:88:88 0:20:35 Then we'll be releasing V8 shortly after that. So I'm thinking next week will be my guess, right? Like I would guess Sunday night would sort of open that door. time next week. Now we're shooting this on a Tuesday, so we're operating still with office hours notes from the previous week, new office hours tomorrow. um we'll see if there's anything that's been updated since then.

Play from 0:21:00 8:88:88 0:21:00 There's a lot of stuff in these notes. I would say it's not all stuff that people necessarily need to be aware of, unless you really, really want to nerd on this stuff. uh But I tried to just sort of highlight a few different parts here. So we can kind of go through that and then uh kind of talk about, you know, what's next here. So in terms of V8, thinking next week, there was a lot of stuff that they had to say with the Style Creator, which I've kind of buried under this tab here a lot.

Play from 0:21:31 8:88:88 0:21:31 Just mostly, you know, experience based around the Style Creator. To me, not that important. I mean, I don't think to you it's that important. I think for some people it is, but I'm just, I'm focused on V8. focused on the edit models, right? So edit models, if you recall, I think were they sort of ballparked could happen either one week up to one month after V8 drops, right?

Play from 0:21:59 8:88:88 0:21:59 So if we're talking next week for V8, we're talking the last week of February, and then we're getting into March. So I think this, you know, this next little week to a month moving forward is, is a big time for Midjourney, we're going to kind of see what V8 is all about. the edit model is all about. um I had some notes. take three to four months? Isn't that just one system prompt with an LLM that they're probably utilizing on the back end?

Play from 0:22:26 8:88:88 0:22:26 It's like we look, we all have this. Doesn't everyone have this at this point? Like personally? Yeah. this point. We could probably do that in a day for you guys, ah if you're interested. um I still think the things that we've sort of reverse engineered from a described standpoint work even better than described in the tool. So... uh if you see like weird little things pop out to you and like that was one of the ones that popped out to me.

Play from 0:22:57 8:88:88 0:22:57 was like, so let's put the yeah, I mean Whatever not that. There were some things from the last office hours though that I noted that were pretty interesting. So I don't know where they landed on this. We'll find out tomorrow, but they were considering a 2K native mode. would love that would be a big, big positive. I would, that would be very into that.

0:23:182K native resolution talk

Play from 0:23:26 8:88:88 0:23:26 And then they also talked a lot, which wasn't as appealing to me, um about this piece here. So they're trying to create a new create flow for V8, but as part of that, they're trying to like give you the ability to just generate a bunch, like mass images, but the resolution will be real low, right? So I guess in the effort that maybe you see something you like.

Play from 0:23:54 8:88:88 0:23:54 quickly and then can act on it, which I get, you know, to some degree, but at some degree, sometimes the resolution does matter and whether or not I actually want to proceed with that image. I mean, there's still a sort of a guessing game that occurs between a low res image and a high res image. I there's a lot of detail to fill in there. So I don't know exactly how that will work, but they basically were saying like batch four is effectively going to be over in the future.

Play from 0:24:22 8:88:88 0:24:22 Right? So this gives you the ability to just like generate at scale. I don't know. What do you think about that? mean, am I missing anything with that? I just thought, eh. I like how Grok does it. I mean, it's fine. I feel like it's a risk for them a little bit if they get rid of batch four and your entire user base who loves you is using batch four and everyone likes batch four.

0:24:26Batch four replacement

Play from 0:24:49 8:88:88 0:24:49 But I get it. I don't like just generating four. I want to generate a million. also, I'd see what you're saying on the resolution side. A lot of things I'm looking at in the images are the small details. That's what... Yeah. Yeah. The small details that it puts in there more than the big overall. Yeah, yeah, yeah. Like the textures, like all the things that like you kind of need resolution for.

Play from 0:25:16 8:88:88 0:25:16 So I don't know. I don't know what it's gonna look like, but I mean, whatever. all these like little thumbnail size low res things. And I'm just like that to me wouldn't feel worth it. Like, I don't I'm more so concerned about the, not concerned, I'm more so excited about native 2K and then also the edit model. that to me is a bigger, ooh, than the bigger patch processing.

0:25:48Edit model improvements

Play from 0:25:48 8:88:88 0:25:48 But. can negatively prompt natively like Banana Pro. So that was pointed out as well. So don't you know, you don't have to worry about saying something in an image and then having it appear, right? Like, and that also kind of gets rid of the dash dash no inconsistency as well, you would think so. That was interesting. in the training, did you rate anything?

0:26:14v8 text rendering progress

Play from 0:26:14 8:88:88 0:26:14 I just did like today. I was just looking at some of the stuff today, but it doesn't feel, nothing in there at this point is that surprising. But if you remember, it's been a while, so I don't know if you do. But remember the last time, I mean, the first two rounds, you weren't really seeing anything that was different. And then it was like, maybe by that final rating round, you started to see these images where you're like, oh, okay, okay. Yeah, that's the other thing is that the one thing I did notice on there, the text was way better.

Play from 0:26:45 8:88:88 0:26:45 Like there was a lot of text and the text was accurate, at least from what I was reading. So, yeah, maybe they've solved that. I don't know if you're going to get full paragraphs out of it, but probably some text will be correct in that sense. But that was interesting to me nonetheless. Yeah. here you go. New architecture supports arbitrary resolutions, down to 64 px.

0:27:03Arbitrary resolution support

Play from 0:27:12 8:88:88 0:27:12 Whoops. Somewhere in between. I don't know. And then the other piece that was interesting was what's gonna be available at launch versus not at launch. So I don't know if this has changed. from the February 4th office hours. this is what they're, know, working at launch, style reference, personalization, mood boards, multiple aspect ratios, things that will likely miss or be limited at launch would be image prompting.

Play from 0:27:45 8:88:88 0:27:45 So same with OREF and then variations, no multi prompts, but we're used to that. So. Yeah. Yeah. Well, again, I'm like, I'm just waiting at this point. Like, I just want to see it before I at this point. I'm just like curious to see what it's going to look like. I've been I've been missing having my my rabbit holes in it as of recent, but it's not much I can do.

Play from 0:28:15 8:88:88 0:28:15 Like, it's it's fun to create like. fun to create cool looking stuff but I can't do anything else with it once it's created so you know without taking it off platform but I just want to I'm like very excited to see how 8 looks overall like what's the vibe because I think that's what we always sort of judge it on is like is the vibe good from the images themselves so you know v6 had a vibe v5 had a vibe v7 definitely had a vibe it was definitely more you know it wasn't as big of a jump as I think we were expecting but it was still had its vibe.

Play from 0:28:51 8:88:88 0:28:51 So that's what I'm looking for. And then you realize that, you know, the one thing you do realize with with BitJourney and how good it is, like from a certain standpoint, aesthetically, is like if you try to type in like simple prompts into any other platform, you just get trash. It's just nothing. Like there's no because they're so coherent. Yeah. we did talk about that. ah But that's where I think I'm really interested in the personalization.

0:29:06Personalization in v8

Play from 0:29:22 8:88:88 0:29:22 And if that's better, you know, they talked about they're gonna do a different way with personalization, but if that's truly better than the way they're doing it now, and we can really dial into personalization, that's another one of those things that is just a total differentiator. compared to other image models. I mean, if it knows what I want and the things that I like, I am very, very interested in that. And to your point, alleviate some of the pressure of maybe, you know, the nice, you know, finessing to get exactly what I want versus it already pretty much knows what Yeah.

Play from 0:29:56 8:88:88 0:29:56 Yeah. That's... Let's hope. We can only hope. Again, I want to that mood board thing I think is important too. If we can get that resolution like that looking good too. That's another area that I think is a big standout differentiator that is very, very interesting. Again, to your point, don't necessarily need complicated prompts at that point. Yeah, it's got the aesthetic down which colors your colors, your textures, know, even maybe even compositions, things like that.

Play from 0:30:24 8:88:88 0:30:24 The characters, the character or attributes of that particular style. And all I'm worried about is the scene, what am I setting, right? Is it subject or are items involved? there background involved that I care about? Done. Yeah, it's good. know, like you realize how good a tool is like mood boards when you go try to build it yourself. You know, like I try to build it myself with Claude so I can use it on other platforms.

0:30:36Mood boards as leverage

Play from 0:30:49 8:88:88 0:30:49 And it's like, there's a lot of steps, a lot of steps that go into it, you know, so it's not just as simple as it seems, but yeah, this is, you know, just like in that apathetic phase, it's just like, we'll see. I mean, let's see what happens next week. I was sort of going to the Alpha site to see if anything had changed here too. Just a little tip.

Play from 0:31:15 8:88:88 0:31:15 Sometimes when you're expecting a release, you can go to the Alpha site and I think, I don't know this for sure, but it seems like they will implement those changes here first. Okay, so sometimes you can get a little peek on some of that stuff. I did notice something change here. It's not a big change, but on the outside, if you click Settings, oh, and now it's not doing it anymore. Interesting. the last time I was on there.

Play from 0:31:41 8:88:88 0:31:41 There was something like that. And then it was gone. Yeah, it was something with settings. putting these settings on this right panel over here. Okay. Yeah, and now it's not interesting. That's a little, that was like when you had access to two mood boards, but no one else on Midjourney did. When you could add multiple mood boards, we were like, what? How does Drew Glitch into a much cooler feature?

Play from 0:32:09 8:88:88 0:32:09 Yeah, I didn't until you showed me and then you're like, yeah, you can do multiple mood boards. I'm no, you can't. I'm sitting here. No, can't do it. can't push the button. But this is yeah, I mean, like, you know, that being said, there's just so much going on. um There's just absolutely ludicrous. I can't, I sometimes have to like stay off X for a little bit or else it'll drive me nuts.

0:32:18AI overwhelm & X fatigue

Play from 0:32:40 8:88:88 0:32:40 It's just too much cool shit that people are doing that I want to do it all, you know. that's the other thing. I mean, that's the other part that kills me, right? um We just don't have time to do it all. I mean, would love to, no, I wouldn't, but I can't abandon my family, okay? I have kids, I have a wife. um I would say this would be a great time to not have kids and not have a wife, but I might not leave my house ever.

Play from 0:33:10 8:88:88 0:33:10 And I might just be doing this stuff all the time, but I think that's a good... call because I've been feeling that a lot. I don't feel good when I get on X or even Instagram sometimes I'm just like, this is cool. I want to save it. I don't have time to even try this, you know, like, and it makes me feel like I'm maybe even falling behind. Right. So it's just like, it's one of those things where they're just so many things happening.

Play from 0:33:35 8:88:88 0:33:35 And the reality is none of us have time to do it all, but it's, it's hard. That's a hard pill to swallow. I'm glad that somebody else's feeling that too besides me. Well, the problem is I didn't really have a lot of LLM stuff on my X algorithm because I wasn't really following it. And then once I started getting back into Claude, now it's like all of that stuff too. then I see all the, it's like all the people that just straight up focused on LLMs and stuff now.

Play from 0:34:03 8:88:88 0:34:03 And like, it's like that, it's like they're at that level of what we're doing with like images and things like that. And I'm like, oh my God, like I see what they're doing. And I'm like, I want to do that. I know. want to set up my agents and I want to set up all that other stuff, which I started doing, which is epic. But like, I wish I'd done it three months ago. You know, now it's can't watch, you know, the guys that are, you know, making Claude bots that are trading on polymarket making like $15,000 a day automated running on their other Mac mini.

0:34:12Claude agents & automation

Play from 0:34:34 8:88:88 0:34:34 Yeah, like, dude, it's crazy. You can but you can see how it's possible because some of the stuff that I've built I'm like, I can see it now. can see how this doesn't just run in the background and it's fine. It'll work. It's not it's like if there's patterns in math and rules, it'll work, right? Like you just have to write the patterns of math and the rules. It's like it's it's really smart. um It totally turned, I think once Claude code got a little bit more.

Play from 0:35:04 8:88:88 0:35:04 Integrated into what everyone was doing and now you see it, you know I think Spotify released the other day that they hadn't they didn't like not one human piece of code was written last year and they shipped 50 features like that's wild, you know, like Yeah, it's like Spotify said that you had that piece that went viral on X which everyone was talking about like something is happening and it was that I don't know if you saw that article it had like like 50 million impressions. It was like some There's like a there's like an article um Let see if I can find it.

Play from 0:35:35 8:88:88 0:35:35 That was like, let me grab it. I know exactly who I sent it to. So let me pull it up here. This is, where did we go? It's just basically, you know, it's kind of how we all feel. Yeah, this is it by Matt Schumer. And it's got. You show us what you're looking at.

0:36:04“Something is happening”

Play from 0:36:05 8:88:88 0:36:05 All right, so has 83 million impressions and it's got 113,000 likes. But basically the whole summary is the guy works in AI, right? He's probably doing what we're doing, building businesses around using AI and doing it for different features and whatever. But his whole thing is like talking to normal people and being like, this is it.

Play from 0:36:33 8:88:88 0:36:33 it's happening right now. Like this is the time like it's not it's no longer like I put in a prompt and didn't get what I wanted. And like, that's it. Like, I was just a joke. Like, it's not a joke. Like, we know it's not a joke. But he goes on to have like this really long description. And then like, at the end, he starts to talk about like all the jobs that are really going to be screwed. And I couldn't have agreed with him more where it's like legal work. Yeah, like totally. Because all legal work is not all legal work is, again, lawyers out there.

Play from 0:37:01 8:88:88 0:37:01 If you're listening to me, I'm sorry, like Contract law. Maybe let's use that as an example. Input data, then there's rules. Like if you run those two against each other, there's your output. Like that's what AI is. Input, rules, output. That's it. So think about stuff like that, like legal work, accounting, like accounting, data in, rules, output. Like anything that's like that is going to be...

Play from 0:37:27 8:88:88 0:37:27 Who's going to hire someone out of college when I can have Claude do it in five seconds? Like I feel bad for this younger generation who's just about to graduate, gonna go look for jobs. the pivot right now? Yeah, exactly. I mean, you might be getting into a career where you're there for three to five years and then all of sudden your life is completely different now. You have nowhere, you have to pivot. this is this one. I genuinely think is like if you're a business today and you're automating your customer service with AI, I understand why you want to do it.

Play from 0:38:00 8:88:88 0:38:00 I think it's possibly the worst thing you could do. If I got on the phone, I was talking to an AI bot. I'm just like, I'm done with you guys. Like I just want to talk to a person. It's called customer service. Like if you just get rid of that. Like, come on, like Claude, for example, Claude. responds with a bot and the bot is stupider than Claude. It's like, I don't want it to like anthropic has like a it's called like fin agent.

Play from 0:38:26 8:88:88 0:38:26 Like I was having some some issues. And I sent it to gave me like this dumb response. And then I sent the same answer the same question to opus 4.6. And it fixed the problem in six seconds while the other one was just like running around in circles. This is the dumbest thing ever. Right. So this, this, like article went long and wide, it's really basically just like a, like, you know, when you have those moments when you just like want to panic and tell people like, stop, like stop and like think about what's happening here.

Play from 0:38:58 8:88:88 0:38:58 And it's like, it's really happening. And then people are like, yeah, I tried to run email with it. It sucked. It's like, okay, cool. This guy over here is running 16 bots that are, you know, trading foreign commodities exchanges, and he's pulling in like 60 grand a day. And because he's been able to engineer this, like, you don't think that's going to become a consumer product and it's going to take your job in six seconds, right? Because you, you said it didn't write an email well one time.

Play from 0:39:24 8:88:88 0:39:24 Um, you know, that's, see what people are doing. And then you have the guy that invented open claw that we talked about last time. He just got acquired by chat GBT for like $10 billion, one person. Like that's, that's what we were all saying. Billion dollar business is built by one person. Here it is. Now it's here. And it's. And it's good. And we're now writing more code than we can ever read.

0:39:45AI skill-building strategy

Play from 0:39:51 8:88:88 0:39:51 I think the whole thing, lot of it, yeah. Like the SAS field too, like you see it. it, right? Though that these these other folks, like, I mean, my wife's one of them. Maybe your wife's the same, right? Like, use it, you know, use AI here and there, but not really, right? Like, I mean, it's it's really hard to convince them because they haven't experienced it really disrupting them yet.

Play from 0:40:17 8:88:88 0:40:17 And they're also not in the same echo chambers that were I mean, we see this stuff. every single day. Like to your point, every time we open up X, every time we open up Instagram, every time we open up LinkedIn or whatever, right? Like our feeds are curated. know, like we've got, we're just in this bubble now where this to us is every day, you know? And again, we probably do feel like in some ways, maybe we wish we would have done this sooner, we're behind on this. The reality is we're still way, way ahead of most people.

Play from 0:40:43 8:88:88 0:40:43 Most people haven't even felt this yet. And uh we still have time to do something about it, right? It's really, really interesting. But like I feel, I feel, if anyone else feels like that, you know, not, you're not alone. Like I feel like the, not like the gates are closing, but it's like, if there's ever a time to really push the gas pedal, it's right now because you just see how smart the autonomous side of it is getting and it's getting better and creating agents is easier.

Play from 0:41:15 8:88:88 0:41:15 And things are, you know, can run autonomously in the background. Like we were talking about with. with Claude Bott and Molt Book and stuff like that. know, the crazy thing, they spun up a website for things that they can't do in the physical world called Rent a Human. So now you have bots paying humans to do things in the real world. Right. Like if I told you that a month ago, you'd been like, that's ridiculous. Here we are. ah So like, I feel definitely like a lot of stuff is going to change and it's going to change very quickly.

Play from 0:41:45 8:88:88 0:41:45 And it's going to happen this year. I think a lot of like I spent And I recommend everyone else do this. If you guys are in the LLM game or like you're just doing a lot of work, I, I purposely last week took an afternoon and said like, I'm just, I'm going to stop doing what I have to do. And I'm going to go and try to build like myself and my thought process, like into Claude for like to agent ties myself, right?

Play from 0:42:16 8:88:88 0:42:16 Just cause I want to see if it was possible. Cause if it is, it's going to say me. a lot of time, and it's gonna save me a lot of headaches. And if it's not possible, then I'll feel a little bit more comfortable, right? That I can't just be sort of replicated that easily. And I did it, and I did it in about six hours, and I was really kind of shocked at how easy it was, and I shocked at kind of the questions that I started with instead of it being, like this is one thing I've been sort of like, in my head lately as to like why I was never asking the LLMs what they like better.

Play from 0:42:53 8:88:88 0:42:53 Like what like how do you guys like to ingest data? Like how do you like how do you like it structured? How do you like it? Look, how do you like it to look and feel verse? Like here's all my stuff. Do something with it. Like I feel like like the best way to sort of make an analogy is like if you give it a if you're a detective and you give someone a case file where everything's like organized, all the files are there. And they can like read it and digest it and be like, okay, I what's going on here. Versus like giving them your detective like bulletin board with like the pictures over here and the thumbnails, the strings.

Play from 0:43:27 8:88:88 0:43:27 And it spends most of its computing power trying to put all that stuff together versus like here's the case file where everything like lives. It you get such a different response. Right. So that was one thing. But I for like this basically built like a whole ingestion machine, but the whole thing was not to build it, like asking it to build agents. I was building skills. So it was like, if I need to do, like, let's think about this.

Play from 0:43:53 8:88:88 0:43:53 Like we need to like an image forensics agent that's going to extract like all these different points, but it's like, no, it's not building agent. Let's build like the skills that are necessary to do that. Right. So like, we need like someone that's really good at this, something that's really good at this, like each individual, like little Claude skills that would go into that one little role. So if you think about yourself, like what is Drew good at? Like Drew's good at photography, Drew's good at branding, Drew's good at, you know, all these different being a dad, like those are all like little individual skills that make you up.

Play from 0:44:24 8:88:88 0:44:24 But like that skill isn't just for you. Like you're not the only photographer in the world, right? But like making a good, like if you can translate that to another person or another agent. So it was like building like a skills library of all these different things that I do. And then it was like, all right, now that we built all this stuff, How do we connect all the dots? Like, how do we bring the connective tissue together? How do we build like an operating structure for like where things go? Then I was like, okay, I'll build like this agent and we'll stack these skills in here.

Play from 0:44:51 8:88:88 0:44:51 And then this agent might be double the skill. Like it might be for two things instead of just like building an agent that does only this thing, right? It's like, then they can like self delegate based on what I'm doing and like, some skills are needed. Some skills aren't like that goes into the first like project management analysis phase of where this stuff needs to go. So I'm sitting there like, what the f- did I just do?

Play from 0:45:17 8:88:88 0:45:17 Now that I can drop, you know. So like the core theory of it was, what I was trying to do was I wanted to take like, if I had in my Cosmos board, I think we talked about it, like 150 images. I want to drop it in there. I wanted like to drop it into, you know, a full file, 100, 150 images. goes into Claude, Claude breaks down the images. into very small points of differentiation.

Play from 0:45:44 8:88:88 0:45:44 And then what it does from there is pattern match. So it takes all that data, then matches into little formulas and recipes. And then from there, we write system prompts from that. And then basically, we can recreate any style of image, or in a mood board style if we want to do that, or if it's a certain thematic style. So it's a very... easy way for me to get from point A to point B now without having to go and run all that process like individually.

0:45:47Pattern matching workflows

Play from 0:46:10 8:88:88 0:46:10 But also I can use the system prompt sort of agent without the rest of that. It's all individuals. I can sort of like pull it together when I need it. But like that was nuts to me that I could do that in six hours. And now my life is different because of it. But I would say like that's a diatribe like really long way to get there. But If you haven't sat down with yourself and thought about some of the processes that you do, a lot of this stuff can be put into more digital format thought process.

Play from 0:46:47 8:88:88 0:46:47 If you have a creative process that you use and you go back to, yeah, you can get there. in your head that at the end of the day takes recall, takes, you know, energy to pull from, takes time to kind of put it back together, rethink, repackage, right? But to your point, if there are objective things or categories or labels or tagging of those things, right, then it's much easier to retrieve that information and get consistency.

Play from 0:47:18 8:88:88 0:47:18 with anything that you're doing in the future. I I think that's such a really good point though. Like that to me though, in general, when you were just talking about, you know, with AI and, you know, like one example is having it ask you questions, right, to get to what you want to get. Like anytime I do like something that's more complex and I've got a lot of thoughts on it, like I don't want to just think that everything I'm spilling out in the moment is everything that goes into it.

Play from 0:47:47 8:88:88 0:47:47 Right? Like, so for example, I was just at the shooting range. I was shooting a gun and I was thinking about getting another gun, you know, and I was like, Hey, you know, I've I've really haven't bought a gun. Like I bought one off my brother a long time ago, but I want to know what goes into it. Right? Like I know the things that I know I want, but what are the things I'm not thinking of that could help me get to the right thing faster? Right? So having it ask me questions, you know, if you're not using AI to essentially help you Or if you're not using AI to help you get the most out of AI, then you're kind of leaving a lot on the table.

Play from 0:48:23 8:88:88 0:48:23 Like to your point, like format, it can help you, right? Filter those thoughts and get to the format or the desired outcome a lot faster with less holes. And I think that's such a good thing, like that's such a good thing to sort of reinforce and tell people because I think a lot of people forget that. Did you ever see Silicon Valley? you ever watch that show? All right.

Play from 0:48:48 8:88:88 0:48:48 but I made it several seasons in. I don't think I watched it till it was done though. It was I used the thought the middle out thought process on a lot of this stuff. And it was it is and that was it actually helped a lot because I was like, look, I was just like talking to talking to Claude. I'm like, everything we're going to is middle out. The inputs are always going to change. The outputs are always going to change. Like we're focusing here first and then we'll focus on the external factors like let's build this and then out.

0:48:54Silicon Valley middle-out

Play from 0:49:18 8:88:88 0:49:18 Right. So that was because then the other thing I did just to just to test, right. Like I wanted to pull some other APIs in which I think this is this is Just as crazy that I spun this up in 30 minutes now that I have like some sort of like idea of architecture Like once I took that agent structure I said I said to quote I'm like alright now build now build a skill so I can like replicate how we just did this entire process But you get all the context from that so I don't have to tell you ever again Like how we did that like and it was like okay cool, right wrote me this thing popped it in and then I'm like I want to take I want to basically take like Andrew Schultz, right?

Play from 0:49:53 8:88:88 0:49:53 If anyone doesn't know him he's a comedian He has a very unique way of delivering information on stuff. um But I was like, yes, I've seen him uh a while back. was uh like right after COVID. there was the thing with this was like comedy is really hard for AI because it doesn't understand a lot of the human connection. So we had to like teach it how to do that. Like how to make how do I make human connection or like this relatable?

Play from 0:50:23 8:88:88 0:50:23 Yeah. Or like at least like from it's like basically, you know, I've built these agents in the same sort of way, but the first step was to go to X and then pull trending topics, right? And like, this is the way that Andrew constructs comedy, right? As it goes with the trending topics and like for like AI, right? It's like you type in, know, what are the trending topics on AI? And it'll pull like probably 50 to 60 of the top trending posts. And it'll go find the ones that have the most discourse in the comment section, like polarizing opinions between like one or the other.

0:50:24Claude comedy experiment

Play from 0:50:53 8:88:88 0:50:53 And then basically build like a hypocrisy map around that of like, you know, the two sides of what they're saying and then like how to sort of like tailor structure to both of those. um To then bring it in to like more of like the comedic writing standpoint, like string together like an entire thought process and chain of ideas. And then I built like this creative analogy engine that was like a way for him to like sort of like tie it back to funnier things in real life with like cultural setting, but.

Play from 0:51:20 8:88:88 0:51:20 Essentially all I do is press a button. go search for this and it just goes brrrr thinking about that a lot from an artist standpoint. Like so comedian would be an example, songwriter, like hip hop artists. you can, I mean, you can literally upload files, audio, videos now, have it assess it. So it can map your tone, you know, like the sort of pacing, the beats, right?

Play from 0:51:52 8:88:88 0:51:52 Like all those things. I wonder how people in that space are using it right now, or I guess who's using it sort of in the most innovative way at the moment, because you don't hear about that. That's like one of those spaces where it's like, people are really using that or not, but they could be in some really crazy ways because you've got so much content to pull from that can essentially map your style or be agentized version of you, so to speak, whatever we're talking about.

Play from 0:52:21 8:88:88 0:52:21 mean, it's got data to train on. It's crazy. The idea that it was like that I was taking like his joke structure from one of his standups and then sort of like applying it to like the connective tissue of how you take like this benign sort of like idea and turn it into something else, which is really hard. Like I think about it because we were trying to do trying to do something with another project that we're working on where it's just standard one input image and then you have to output something else.

Play from 0:52:52 8:88:88 0:52:52 But sometimes the output has to be has to have something behind it. can't just be, you know, for advertising purposes, right? Like you can't just put in water bottle and get out like, you know, water bottle floating on the clouds. Like it makes no, there's no connection to the human being on the other side. Right. So it like wasn't getting that. So I had to teach it like an, like a way to think about advertising, which was like, okay, like if you see a, maybe I mentioned this on the last podcast. I can't remember if I did or not where it's like, if you take one, yeah, if you take like a water bottle, it's not just like a water bottle.

Play from 0:53:24 8:88:88 0:53:24 It's like, hydration. It's like, you know, hydrating champions. like it holds it holds water. It can be used as a weapon. Like, you know what I mean? It's like, like, you have to sort of like apply. Yeah. And then create off the word cloud with idea instead of just instead of it's just a water bottle and you drink from it. Right. It's it's glass. Like, you know, there's a lot of different things you can go to once you create the word cloud and then like branch off of it. But you like that second level of connection, not just make this an advertisement.

0:53:24Word clouds & ad thinking

Play from 0:53:53 8:88:88 0:53:53 Right. the humor thing's hard too, because there's a portion of that, depending on the comedian too, where a lot of it relies on the delivery, the tone, the body language, the facial expressions. So how much of the percentage relies on that for it to be, to hit, to be funny, to resonate. It made some excellent analogies.

Play from 0:54:18 8:88:88 0:54:18 Let's put it that way. Like it made the analogy of AI as the microwave, which I was like, that's interesting concept. And it's like, you know, like, yeah, but we like, we like, you know, a hot pocket at 2pm on a Tuesday or at 8pm on a Tuesday when we've just worked all day and we got, you know, we still got, you know, diapers to change and stuff like that. But when it's Valentine's day dinner, you don't go and get, you know, a microwave meal. You go have someone cook it, right? Like the, it's like the difference is you can have both you can have the Hot Pocket at 8pm on a Tuesday, but Saturday is a reserve for the art, which is the dinner and I was like, that's a good, good tie back and then it was going into all of like, you know, there could there went into this one about it dude had Toby Maguire was talking about like how you know, Hollywood so like

Play from 0:55:04 8:88:88 0:55:04 revolutionary and uh revolutionary and advanced and like, you know, there's such great storytellers yet they've they've done nine Spider-Man movies that they've had three Spider-Mans and Toby Maguire has been replaced. that part drives me nuts. Like, let's just keep changing the character. I mean, this is not James Bond. Like, I'm good with it for 007, but I mean, to have that many Spider-Mans or that many Bat-Mans, and look, like, we just don't need it.

0:55:14Hollywood recycling IP

Play from 0:55:30 8:88:88 0:55:30 You know what I mean? Like, I do like Robert Pattinson, by the way. I think he's a hell of an actor. I think he's super underrated. But it's like, man, we already had Christian Bale, you know? We didn't need Ben Affleck. We didn't need, you know, Michael, I mean, Michael Keaton I like too from back in the day. I think Val Kilmer was a Batman. George Clooney was a Batman. My gosh. yeah, it's like there's a ton. But like this, this came out with me saying, you know, pull public discourse on or, uh, you know, write a bit about public discourse on AI.

Play from 0:56:03 8:88:88 0:56:03 So it just searched all this stuff on AI and it gave me a six page script, but like it went from, you know, talking about the guy in his underwear who made one of the sea dance things, right. To like, you know, Now Hollywood's response to all this was, need y'all to hear this in the most corporate voice you can. At our studio, we believe in the power of human storytelling. It has like the ironic sense of knowing that, where, you know, it gets into the Spider-Man thing.

Play from 0:56:31 8:88:88 0:56:31 It's like, the man went from the biggest movie star on earth to a meme. He's now a meme. That's his legacy. And Hollywood did that, not AI. Like it's, it's got some nuanced takes, right? Like talking about how the, the VFX artist of or the main company that did the VFX or Life of Pi, that excellent movie back that won a bunch of Oscars went bankrupt after that. Right. You know, it's got sort of like some of his some of his mannerisms in the way that he does stuff.

Play from 0:57:01 8:88:88 0:57:01 I talking about the writer's strike is all tied together, like weaving into a narrative. Like it's all one story, which is pretty impressive. Like, you as a microwave, stay with me. And it goes through that whole sort of analogy. You know, as a cover band, right? Like, yeah, cool. Like you still like the song, but you don't really feel much when you get into it. Like it doesn't doesn't do anything for you. Right. Like that's, you know, wrote this in my bedroom at three AM because I couldn't sleep because she left. That's a thing. AI doesn't have a bedroom, doesn't have three AM AI doesn't have a she left like it has AI has data artists have damage and damage is what makes art.

Play from 0:57:34 8:88:88 0:57:34 So the guy in his apartment, one dude laptop 45 minutes made something that looked like a Marvel movie. He's not killing Hollywood. Hollywood's been dead. just the first one to notice the body. was like, oh yeah, let's go Claude. Fucking solid ending there. But to me, it was the interesting thing of taking raw real time data, right? That's out there that people are thinking. And then weaving it into like a full narrative that like flows from one thing to the next with like connective tissue, with pacing, with style, with like the, it's like a really translatable sort of a translatable sort of.

Play from 0:58:12 8:88:88 0:58:12 tool like this can this doesn't have to be for comedy writing do this for marketing like I can go and do this I'd take this whatever brand I'm talking about let's go find the public discourse on Coca-Cola do this whole thing I'm from Pepsi just go write our marketing copy and like marketing angles directions like ways to weave different stories and narratives together analogies this is like a whole data set so like that again the stuff just crazy to me with this and now Alright, I can go take that script and go do something with it.

0:58:41Marketing narrative engine

Play from 0:58:45 8:88:88 0:58:45 I should probably just go try to do one of those like videos he did during uh during COVID, where's the quick hitters. Um, you remember those on on Instagram that like sort of blew him up to begin with? Where's like, it's a minute straight of basically like one liners on political topics. Yeah. Yeah. Which was solid, but listen to his podcast or anything, but, uh, yeah, he's good, man. I've also seen him in person. I mean, was definitely one of the best, best shows I've been to.

Play from 0:59:12 8:88:88 0:59:12 sharp. It's like his connections are sharp. They're witty. You know, it's like not. It's not fart jokes, which I appreciate. But it's a it's cool. If you had a chance to play that's enough with the L. I'm sorry, guys. This is just like, I feel like I have to say some of this stuff because it counteractive to some of the narrative that there's not much going on. Like the fact that I just took that and did that for myself in less than a day.

Play from 0:59:39 8:88:88 0:59:39 is is nuts and then it's replicable that's the problem yeah to that micro tools thing. And again, like creating value in other ways that isn't just dependent on, now I've got 15 second videos versus 10 second video, right? Like there are other ways to sort of, um I guess, optimize your process, but also like package things up so that when that circumstance or scenario comes up again, you already have like a very straightforward, uh thoughtful way to do it in the future.

Play from 1:00:19 8:88:88 1:00:19 Like we said, with system prompts or whatever it is, it's like, okay, well, if I've done this with a brand where we've got a visual identity toolkit and we've got a 30 page document and we've consolidated this down to exactly what the brand wants, and now we're just, uh putting that in and we're primed to create images that are exactly on brand. Now you can take that to the next brand. You've already gone through that process before.

Play from 1:00:47 8:88:88 1:00:47 There's never been a better time to be curious and to your point, just play with pet projects and things that maybe, who knows if they're gonna work or not, but you're learning either way. So it doesn't really matter if it works in the end because you're either learning how to do something correctly or you're learning how not to do it, which ultimately is gonna save you time in the future. think time's like just again, One of those undefeated things that's always going to be the most important thing ever for any of us.

Play from 1:01:12 8:88:88 1:01:12 It's the only thing we can't get back. So I think if you're extracting value from things that you know you're going to do again, repeatable processes, it's a very, very smart way to approach AI right now. you don't have to be a coder to do it, man. They've made this incredibly democratized now. How well can you verbalize what you want to do? the constraints, the do's, the don'ts.

Play from 1:01:38 8:88:88 1:01:38 Again, rely on AI to provide maybe some, the reasoning for you. So you're not just saying, hey, like I want a launch plan. You're saying, well, how should I think about this launch plan? What are the things that could go wrong with this launch plan? What if I had this much money to put toward the launch plan versus this much? What am I not thinking of? You know, like all those things. She's like, this is a great time to do that. It's why, like, I'm just...

Play from 1:02:05 8:88:88 1:02:05 I'm happy that I went down the skills path versus the agent path with this, because now I can build agents with the skills that I've built in Quad, because that's just... I think most people probably look at it from a sense of, I need to build this agent system. How do I build the agents? It's like, well, sometimes you don't need all the agents. You don't need all that firepower, but you don't need that whole system. And if you need to break from that system, then you have to build a whole new system, right? So it's like, it's like trying to be more multiple in what you're doing.

Play from 1:02:31 8:88:88 1:02:31 because not everything's going to be the same. ah But that's sort of like, have to build this broad flexibility. Like, that's a weird way to think about it now, is that we need to build things that are specifically tailored, but also broad for a lot of range of outcomes. And that's sort of like the mindset that you have to be in. Too tight is too tight. Too loose is too loose. But also, like, it has to be repurposable now, also, for other things. So uh it's weird.

Play from 1:02:58 8:88:88 1:02:58 But, you know, this is... I just feel like it's going to be people who are like sort of in the design space now and like just going to get if you haven't been if you've been neglecting the LLMs in the work, like I would say it's a good time to get back in to like use it and just push it. Try weird stuff. That's the other thing is what you said is like pet projects. You'll figure out way more stuff when you try to do something a little bit weird and like you realize that, oh, that has translatable.

Play from 1:03:29 8:88:88 1:03:29 First like, let me build this for a company. And it's like, I don't want to do this. And there's all of these rules. Just do it for yourself. You'll figure it out. And it's repurposable. Yeah. Yeah. right? It's like, if you've got exactly what you're, you can't really be that creative, but if you have no constraints and you're just kind of like, hey, let's just see what happens if I do this or do that, blah, blah, blah, right? That's where the juice is at. Hey, before we bounce, Higgs Field.

1:03:57Higgsfield controversy

Play from 1:03:58 8:88:88 1:03:58 I'm happy. I love it. I'm- with these guys? So Hakesfield, I mean, I've got this nice little article up here. The downfall of... Is it? Well, Forbes wrote an article about them. Yeah. It wasn't as big of a hit piece as I would have hoped, but you know. This is a really good one that I read through in case like, in case you don't know, like I guess like the short of it and I've never really like been fully ingrained in Higgs field.

Play from 1:04:36 8:88:88 1:04:36 So I'm probably not the best one to sort of like tie a bow on this, but essentially a lot of promises that aren't delivered on um and then just a sort of rage baity, we don't care who we piss off kind of ego that this company has. And it's essentially just like a wrapper, right? Like Hicksfield didn't really invent anything here.

Play from 1:05:03 8:88:88 1:05:03 We're just putting things together, right? And putting them in a nice wrapper, use cases. from a standpoint of making things very UX folk, like... here's exactly like use case specific, press the button, like they've done a really good job with that. Beyond that, there's nothing. And so they've also made a lot of promises to like, you know, creators to pay them for this and that and promote.

Play from 1:05:29 8:88:88 1:05:29 I've gotten messages, I'm sure you've gotten messaging, like I've never done that with them, but you know, just shilling for them. And then there's a lot of like proof that a lot of people didn't get paid for this, or they sort of like waited for somebody to post them and they never paid or they just backtracked, they walked. back the offer or whatever it is. But this is a really good article that just sort of goes through it. Here is the sort of username if you want to find this story.

Play from 1:05:58 8:88:88 1:05:58 Maybe I can also link it in the description here, but it's a good read. um Are they back up on X now? I didn't look. I don't think so. Maybe they are. They'll be back. yeah. Well, that's what happens when you break the rules on the platform, right? Like, it's actually, like it's, you had to see it coming. It's like X is a direct competitor of them now, right?

Play from 1:06:25 8:88:88 1:06:25 Like Grok is a competitor and they're, everyone's telling you that they're doing shady stuff, which they are, you know, like, yeah, this is, To me, I don't even want to focus on the content that they put out. It's trash. We know what it's trying to do. It's rage bait. It's just the other ways that they operate, which is so sketchy. It's like the fake promotions, where it's like, Nano Banana free for a year.

Play from 1:06:54 8:88:88 1:06:54 Like that kind of stuff. And then it's like, wait, now you get it. And then we'll throttle your usage of it. It'll take 30 minutes to create an image, but you have to pay more to get that throttle usage back. So like that kind of stuff. really pisses me off because that's just shady, like acquisition focus growth stuff that like I know how to do that you know the different ways the levers to pull. But it's like to be to be that blatant about it. And then, you know, the paying people to repost and the paying people to, you know, just show for them.

Play from 1:07:26 8:88:88 1:07:26 Yeah, not paying anyone like people doing stuff, and then they block them all together and you don't get paid like that stuff. Also, like the the shady, the shady like, like upselling tactics they have this one, just so absurd. they did. To me, the marketing marketing is marketing, right? Like however you got to gain attention, gain attention doesn't mean people have to like it. They don't like it, whatever. It's just like the actual way they treat the customer and the way that like they treat the people that like blow them up.

Play from 1:07:59 8:88:88 1:07:59 That's like cut them out, cut it off the knees. So I mean, at this point? mean, that's got to be such an interesting thing to be a fly on the wall right now. This is like the... there, right? There's enough good companies that treat you right. Weeby, to me, like, I can't say it enough, they treat me like family. Like, you know, if they did stuff like this, I'd be like, see ya.

Play from 1:08:25 8:88:88 1:08:25 I don't know why people are so, like, I mean, we're so stra, I mean, maybe it's just creators are so strapped for dollars that they're just like, yeah, I'll repost for 100 bucks. Which one? Remember this one, they did a apology campaign. So they claimed they were going to replace 100,000 art jobs with employees bowing on camera and they announced that a 65 % price cut framed as we're giving up our November paychecks.

Play from 1:08:53 8:88:88 1:08:53 It's just a lie. texted me when this happened. You're just like, dude, this is ridiculous. But that's the that stuff like the Nano Banana free for a year. It's like you obviously can't do that for everyone. It's just a lie. I guess that's just that's fake advertising. You know what I mean? Like that's I don't care what their content is. It's just like that kind of stuff where just blatant and it's just always promo, promo, promo, promo, promo.

Play from 1:09:21 8:88:88 1:09:21 You know, it's like so diluted. Does Google ever give a promo? Yeah. ruin your reputation, you're toast. I don't even know how they're still, what are they? I wonder what they're doing right now. mean, uh next year when they like who's going to renew that they bought the yearly or the yearly annual subscription costs unless Higgs field is just trying to like pump and dump like get all this and exit and get out like then it makes sense to me but next year if it's like going to forget to cancel their auto renew too and they're going to try to get back and guess what?

Play from 1:09:58 8:88:88 1:09:58 I hate that stuff. That's just, that's what, that's what I really don't like about the whole thing. You know, I feel one way or another about the content. Just do what you want to do. I don't care. I don't have to like it. I don't have to follow it. But this stuff, no go. Don't, don't bite the hand that feeds you. Yeah. The other funny thing is that they don't realize that we can do that stuff too. You know, like their content, like that the creators can just like, oh, you know, like, let's just put their CEO in a bunch of really like compromising positions.

Play from 1:10:35 8:88:88 1:10:35 Like, yeah, we can do that. We've done. There's been some stuff. There's been some stuff. There's a Blacklight's been doing some funny stuff. ah There's there's there's like, you know, I don't think anyone's really caught on to it or like done it a lot. there but there's a Yeah, they what it doesn't realize that you you're doing it to Sydney Sweeney and you're doing it to you know, Tom Cruise and whatever.

Play from 1:11:01 8:88:88 1:11:01 But like everyone can do it to you too. Like your face is out there. You just make as like with your tools, which is hilarious. I mean, I'm probably love it. It goes. That's what I mean. Like it'll go both ways. I think it's just ridiculous that they promise stuff and don't deliver on it and are really blatant about not delivering on it and find ways to like purposely not deliver on it.

Play from 1:11:27 8:88:88 1:11:27 ah But yeah man, that's a there's a lot going on. I'm happy Higgs field is just like at least they have to deal with a road a speed bump on their on their massive rise. Yeah. lesson from these other bumps along the road. I mean, this is not the first hand slap. mean, they just don't care. Anyway, we try to jam as much as possible into this one for you guys. uh So I think let's wrap it up.

1:11:36Final thoughts & sign-off

Play from 1:11:54 8:88:88 1:11:54 Hopefully, this time next week, we're talking about Midjourney V8 and can live sort of do a live demo of what that looks like for you guys. I'm fingers crossed on that one. um Rory, anything you want to say before you sign off for London, man? I know you got a big trip, so hope that treats you well. But we you know, we're as as is tradition. Tell who you got to tell you got to tell your dog flight attendants, TSA tell your local barista your uh who else would we miss here?

Play from 1:12:25 8:88:88 1:12:25 Garbage. We typically get garbage man thinking like you know, bodega owner. um Dog walker. Maybe we've said that one before I'm running out of things we've covered a lot of professions here. Hotel reception. you're about to do so you can kind of plug this as you go. Yep. Yep. Yeah, you know, the conveyor belt operator. ah Like and subscribe, dammit.

Play from 1:12:54 8:88:88 1:12:54 Yeah. the drug sniffer dog. Let him know. Correct. Yeah, let them know. They can come have fun here. This is for them. This is for this is for everyone. But yeah, I'm sure we'll come back here next week and it'll either be mid-journey eight or we'll have you know, some other nuclear drop from some, you know, VO4 or whatever is coming. Just be prepared. But dude, as always.

Play from 1:13:19 8:88:88 1:13:19 Yeah, enjoy the trip. Thanks for joining us, everybody. We'll see you guys in the next one. See ya. Peace.