Ep. 21: Is Midjourney in Trouble? Updates, Discussions, and AMA

0:00:00Weekly update and banter

Play from 0:00:01 8:88:88 0:00:01 [00:00:00] What's up everybody. Midjourney, Fast Hours, podcast, live edition. What's happening, Rory. What's the story, my friend. [00:00:09] No, there we go with those rhymes. Like it, man. Well, it's Friday. We're ripping, uh, [00:00:15] lots of stuff going on. Just finally got back home. It's about, I don't know, 20 degrees outside, which is miserable, but Hey. [00:00:21] You know, here, here we are. So yeah, man, it's been a, been a crazy week back and forth between New York and LA, lots of stuff. I feel like I'm in a time [00:00:30] warp, but you know, what, what else is, what else is new? What's going on with you?

Play from 0:00:34 8:88:88 0:00:34 [00:00:34] Bro, has this week just been nuts or what? I mean, Oh my God, man. You and I were just texting back and forth and talking to each other about like [00:00:45] everything we had going on this week. [00:00:46] I had a lot of client work. Kids are sick. Car. My, my whole tire on my right side just exploded when I hit a curb two days ago and I wasn't dressed for it. And it's cold as hell outside. And I walked a [00:01:00] mile and a half back to my house in house slippers and a t shirt. And then you, my friend, uh, you know, you just did a, a, you know, trip across the coast and back in literally 36 hours.

Play from 0:01:13 8:88:88 0:01:13 [00:01:13] So Yeah, man, we're [00:01:15] just, we're just rocking and rolling. [00:01:16] That was, uh, you know, leaving at 4 a. m. on Wednesday, getting to, getting to LA at 9 a. m. and then leaving yesterday by 4 p. m. and getting back at 2 a. m. last night. Like, I don't even know, I don't even know what time zone I'm in. I don't know [00:01:30] what dimension I'm in right now. [00:01:32] Very weird. Just like, also didn't leave the hotel for like, 28 hours, so, you know. It's like, I was like, kind of like, locked in like, a little jail cell, and then uh, you know, back to reality over here. But it's been, it's been [00:01:45] a, it's been a weird one, but it feels also kind of like, you know, It's just, I hate when you get broken up, when the weeks get broken up, you know, it's like Thanksgiving.

Play from 0:01:51 8:88:88 0:01:51 [00:01:52] It's like, alright, right in the middle of the week. kind of weird, get a little break, then you come back and then it's same thing right in the middle of the week. And then, you know, I feel like I've [00:02:00] done nothing. I feel like I've done it. I know, man. [00:02:02] I'm like, I'm just sitting here. I'm like, I want my week back, you know, like, like give me Monday again. [00:02:08] Like, cause I mean, dude, it was just absolutely crazy. The other part of the equation, right. [00:02:15] Was, um, my kids were sick. You know, my kids were sick and one got sick, then the other got sick and literally today's the first time that they've both been back at school in two weeks. So you know, my wife's traveling for [00:02:30] work.

Play from 0:02:30 8:88:88 0:02:30 [00:02:30] I've got a bunch of stuff on my plate. Meanwhile, trying to like single dad it up and dude, oh man, I mean, brutal brother. Brutal. [00:02:38] Well, do you have a tire yet? That's I think that's, I do have [00:02:42] a tire. In fact, I have four new tires. [00:02:44] [00:02:45] Wow. Brutal. Alright, so you know sometimes it's this forced upgrade, right there. [00:02:50] It's the same thing, I did this like a couple, couple months ago, except when I got my new tires, I think someone screwed, you know, either who was putting them on screwed up my axle, and as I'm driving [00:03:00] over the, you know, one of the bridges in New York, I have like six people honking at me, you know, New York in me is like, Shut the hell up like you know like turn around like I'm gonna kill somebody and then I look out my rear window and There's smoke coming out of my car Just drive it right to the [00:03:15] mechanic.

Play from 0:03:15 8:88:88 0:03:15 [00:03:15] I don't even know what's going on the guys like who did this, you know They did great, New York like Whoa, you know who who changed your tire? I'm like, yeah, I don't know. I mean some guy from a mobile service. He's definitely screwed it up. [00:03:29] [00:03:30] He's screwed it up. Dude, that's a great accent, dude. That is killer, dude. [00:03:34] You might just pull that out, you know? Like, I wish I, [00:03:37] there's kind of sometimes where I wish I had it because it's just so epic sometimes. I've grown to love it over the years, but my god, don't have it in me. [00:03:45] Don't have it in me. But yeah, I'm surrounded by it. I'm in Brooklyn. So, you know, it's true.

Play from 0:03:51 8:88:88 0:03:51 [00:03:51] That's true, brother. [00:03:53] Yeah. Well, uh, yeah, man. Well, hey, I mean, I'm, I'm glad that we get to connect and have a little [00:04:00] time to nerd out here. Talk Midjourney. [00:04:03] We need to talk [00:04:04] beyond Midjourney because there's a lot of shit happening. A lot. And, uh, we'll catch you guys up on that. Obviously. And yeah, we're doing something a little bit different today. [00:04:14] Sorry [00:04:15] for the last minute notice. Obviously our weeks have just been crazy, but Rory and I were just talking, we're like, man, we need, we're overdue for a live session, so let's just, let's just do it and you know, we jumped on last minute and man, for, for some reason, like [00:04:30] LinkedIn makes this streaming thing way too hard.

0:03:55LinkedIn live challenges

Play from 0:04:33 8:88:88 0:04:33 [00:04:33] I don't have a problem with any other channel, but LinkedIn, I got two different links for the event. Um, I'm looking now, dude. I don't even know if the event itself. is pulling up. I think you have to go to my [00:04:45] profile or your profile to see, see the live based on what I'm seeing. I know we're up on X. I know we're up on YouTube. [00:04:52] So we're good there. But, uh, you know, dude, Murphy's law continues into the weekend. So [00:04:57] we can generate an entire virtual world and [00:05:00] reverse engineer an image structure, but we have a really hard time setting up a LinkedIn live. The that's right. LinkedIn, if you're listening, help Make this a little bit. The other thing that's really annoying with LinkedIn is that whenever, like, if you create an event, this is just stupid [00:05:15] conjecture here.

Play from 0:05:15 8:88:88 0:05:15 [00:05:15] You create an event, you invite me to be a co host. I can't like share the event like directly. I have to like technically repost. Your post so yeah, regardless it gets like zero visibility because LinkedIn hates reposts So I'm just like, you know and [00:05:30] then I get invited to every goddamn live event or I have to go through my Notifications and just like delete ignore ignore ignore ignore ignore Then we do it.

Play from 0:05:40 8:88:88 0:05:40 [00:05:40] It's like it's like jumping through hoops. So, you know first seriously [00:05:45] I wish they would just make this way easier than it is. It's one feature still don't like it, [00:05:49] dude How have they not Updated like 90 percent of their product in a decade. Like what is going on guys? Who is anybody working [00:06:00] there? You know what? [00:06:02] They're going to shadow [00:06:02] ban [00:06:02] us. If we're going to stop, sorry, LinkedIn.

Play from 0:06:06 8:88:88 0:06:06 [00:06:05] Love you. I [00:06:06] mean, we, yeah, we, we got too much at stake to, uh, to piss them off too much. So listen, uh, or it could just [00:06:13] be useful feedback. [00:06:15] LinkedIn that's, [00:06:15] that's true. It's, it's constructive criticism, right? So have LinkedIn. Uh, let's, let's get into it, man. [00:06:24] Let's get into, you know, maybe we start with. Midjourney, what's going on [00:06:30] or maybe what's not going on, um, with, you know, the things that they have coming down, uh, in the roadmap. I've got the office hours notes up here. Um, I'm just going to run through them really quick. Feel free to chime in if, if there's anything that pops out, but [00:06:45] you know, V seven, right, we had the confirmation that that's no longer coming by end of year, which not a big deal at this point, cause we're already three weeks away from the end of the year.

0:06:21Midjourney V7 updates

Play from 0:06:53 8:88:88 0:06:53 [00:06:53] So that's coming in January. Um, pinky promise, you know, we'll see. [00:07:00] And, uh, you know, that's going to follow subsequently. With a 7. 1. I think they're trying to get v7 out the door. There's going to be some things that just don't make it in v7 that will then be pushed into v7 one. Right? So [00:07:15] just just something to keep in mind there. [00:07:17] The other piece of this, they're going to introduce a new aesthetic and personalization system with that. Um, they're also planning some like slower but higher, higher quality model variants. Which could be interesting, [00:07:30] right? Like they're always kind of asking people, would you rather have something that's slower and better or faster and cheaper, you know, and, and not quite so good.

Play from 0:07:37 8:88:88 0:07:37 [00:07:37] Right. So it's a, it's an argument that you can get like two of the three, but not all three. [00:07:42] Yeah. [00:07:42] Um, we'll see what happens with that. [00:07:45] Obviously video model as well is going to get pushed, uh, into 25 talking about the same thing with video, right? The affordable versus, um, like a premium model. Would you rather pay a lot more for it and have highest quality possible? [00:07:58] Or would you rather pay a little bit [00:08:00] less and have lower quality, right? I think most of the time What the way i've seen these polls break out is people just want the highest quality possible you know, I mean like I think we're at that point We just want to see the jumps and if it takes a little bit longer So be [00:08:15] it but I would rather roll something two or three times and get better quality than roll it 20 times And do it in 10 minutes like just personally.

Play from 0:08:23 8:88:88 0:08:23 [00:08:23] I mean [00:08:24] I I have a weird I think I might be in the minority You On this one. Are you? I I don't know. [00:08:30] I think I'm just spoiled with Runway. Like, Runway, it's Yes. When you do Gen 3 Turbo, and it just, it pops out in 20 seconds, and it's good. And then I'm like, I go to Kling, and I hit the button, and it's like, Oh, you'll, yours will be done in 12 minutes. [00:08:43] I'm like, ugh. Cause [00:08:45] sometimes, you know, half the, half of what I'm doing on there is testing anyway. I'm not expecting to get the first shot. So it's like, let me just see if this, like, this will work. And if it does, great, I'll just, you know, awesome, I can play with that. Or if it doesn't work at all, I just want to know so I can move on.

Play from 0:08:59 8:88:88 0:08:59 [00:08:59] But like [00:09:00] half the times with Kling, I'll for even forget after 12 minutes, cause my brain's moving so fast that like I was generating anything and totally move on, right? Like it's just like 12 minutes is like a lifetime feels like at a certain point. You know, I'm not, I'm not. knocking anyone for being slow [00:09:15] on the quality. [00:09:15] I think it's just, it feels like everything has to first go in that direction where it's like high quality and slow. And then we move to faster, cheaper. Like it seems like that's a development scale for just about every other one of these video generators. I'm [00:09:27] not, I [00:09:28] mean, I get your, I [00:09:29] [00:09:30] get your point too, right?

Play from 0:09:31 8:88:88 0:09:31 [00:09:31] Like, because you want to stay. I don't want to like do it and then also jump around to tabs and then kind of forget about it for a while. Right? Like you just kind of want to go there, do the job and dip. Right. Yeah. Um, and I, yeah, there is something fun still with the experimentation [00:09:45] part of it too. I just think like with video, maybe that's where it's just a little bit different for me. [00:09:48] I just want like the highest quality possible at this point. Like let's just get there. [00:09:52] We're spoiled with images too. Cause they all come out quick, you know, like nothing takes, six minutes anymore, unless you're in relax mode, which that's like a death [00:10:00] sentence to me. But it's like the, uh, you know, the, the interesting thing is like topaz, right?

Play from 0:10:05 8:88:88 0:10:05 [00:10:05] Like, for example, topaz, when I had a, I just got a new computer, way more power, and now I can run it and things get done in five minutes instead of like, [00:10:15] It seemed like three hours for some videos to upscale. I just wouldn't use it, right? Like I was just like it's not worth me waiting right for it right now now that it can be done in five minutes I'm like alright. [00:10:25] Let's let's go. Let's see how this looks because you know I think it's just the [00:10:30] It's it's tough videos tough in regardless of how it's gonna be But I'm interested to see how they're gonna do it Because it always feels like they do something just slightly different than everything else every every other tool kind of feels like it's following a similar path We all know Midjourney likes their [00:10:45] curveballs and they're you know They're interesting sort of interesting sort of directions that they take on things So I'm amped to see what it looks like.

Play from 0:10:54 8:88:88 0:10:54 [00:10:53] I don't have high expectations I'm gonna be honest like I just I just don't and I'm not that's not a knock on Midjourney That's [00:11:00] just me not trying to be like, oh, there's gonna be the best. Yeah, I get it. I think [00:11:03] that's fair I think that's fair to like, I try to gauge their sense of excitement, enthusiasm in their voice to it sounds like right. [00:11:11] They're trying to do this. This initial version and [00:11:15] then they obviously know they're going to make changes to it and uh, Yeah, I I mean, I I think they're just kind of forced at this point to like push something They've been talking about it so long, right? um, so I think if it were up to david like he would keep [00:11:30] tweaking and tweaking but uh, we'll see man and Where's our, where's our friend Sora at?

0:11:31OpenAI Sora expectations

Play from 0:11:35 8:88:88 0:11:35 [00:11:34] I [00:11:36] have a feeling that's coming. I, you know, we got that, we got that sort of, I don't want to say leak, but it was a leak. And then you saw a couple people posting that [00:11:45] were like, you know, it's probably going to be coming in a few weeks. And we get OpenAI saying we got 12 days of OpenAI. And, uh. Yeah, basically all that means is you just open up your open up your checkbooks because we're gonna make you pay for more things So, you know like yesterday I think was they [00:12:00] released the the new a one pro or whatever that was It's like oh you can get it 200 a month I'm like interesting because you've conditioned everyone to pay 20 a month.

Play from 0:12:10 8:88:88 0:12:10 [00:12:10] Like there's no yeah, that's just an exit really quick Yeah, right. Like I think that's That's a [00:12:15] weird, I understand, like, I would, I'm definitely gonna test it to see if it's worth 200 a month. But, at the same time, like, that's a big jump from, like, a perception standpoint. Like, how much better is it? Then, you know, looking at the [00:12:30] charts they put out, it's like, oh, the regular 01, not the preview, is like 75 percent better, and then that's 85 percent better.

Play from 0:12:38 8:88:88 0:12:38 [00:12:38] So I'm paying 10x for 15 percent improvement, seems like a lot, but, you know, I can complain all I want about this, but I feel like, [00:12:45] you know, I'm gonna pay for some of this stuff at some point. I'm guessing Sora is just gonna, I think it's probably gonna be something like that too, like 100, 200 dollars a month. [00:12:52] There's no way they're rolling that out at 20 a month. I don't, I just don't see it. But I mean, not even at [00:12:57] the beginning though, right? Like, [00:12:59] [00:13:00] I don't know. I just feel like they've been developing it for so long. Like if I'm, if I'm taking this into consideration, right? Like if you can do 60 second generations and like, Well, you know, runway was started with five, like mini max five cling five.

Play from 0:13:12 8:88:88 0:13:12 [00:13:12] They've all kind of like pushed it to maybe 10 [00:13:15] seconds. So it's six times that for the generating power, right? So if that's, that's where we're going, I just have a feeling it has to be more expensive, but you know, it's open AI. So who knows what other updates they have, at least [00:13:30] maybe we'll see them this year. [00:13:31] I just [00:13:31] can't imagine. The video links being that long and being good. And what happens if you've got 45 great seconds and then, you know, right in the middle, it sucks, right? [00:13:45] Like, I don't know. Um, anyway, we'll, we'll see what happens. Well, let's, let's keep going through these Midjourney off servers notes and then we'll, we'll transition.

Play from 0:13:52 8:88:88 0:13:52 [00:13:52] Cause I know there's a lot on the, uh, the AI creative front. So let's go through, you know, we've talked about the video model. There's [00:14:00] also the personalization profiles they've talked about rolling out. So for you guys that are using. Personalization on Midjourney, right? Um, they're, they've talked about sort of giving you several codes, multiple codes, right? [00:14:14] They haven't [00:14:15] really discussed or dissected how that will work exactly, right? Like, will you have a code that's specific for a certain style? Or can you train it very specifically on a certain style? Right? Like I, no idea what that looks like. Uh, it would [00:14:30] be just pure speculation, but I do think that's exciting, right?

0:13:59Midjourney personalization

Play from 0:14:33 8:88:88 0:14:33 [00:14:33] Like I think the ability to fine tune based on the type of image you're putting together is super attractive, especially if you're bouncing around to different types of styles or if you've got client work, right? Like [00:14:45] different. Different bucket areas that you're trying to produce for. Um, let's see, there's that new mood board feature, which we mentioned, I think last time we were together, there's the storytelling stuff, which they did announce that they've got a demo [00:15:00] planned for the next office hours for that.

Play from 0:15:02 8:88:88 0:15:02 [00:15:02] So that's next week. What is that? Let's see, [00:15:05] it seems like there's been a lot of promises lately. There have been a lot of, a lot of delivery. So, so [00:15:11] the 11th, the 11th will be that, uh, [00:15:15] which would be interesting because they're, they basically said, Hey, we're going to let the storytelling team come up and talk and show about show some things, uh, because relatively up to this point, been very hush, hush, right? [00:15:26] Only that we've got the storytelling team and they're [00:15:30] working on something very special. So we'll see what that's all about.

Play from 0:15:33 8:88:88 0:15:33 [00:15:33] We'll see. I mean, this is, I think this is probably, One of the things that I've noticed about office hours, right? Like I write them every week in my, I, I wish I got to tune in, um, you know, more often, but it just comes at a [00:15:45] really bad time during the week. [00:15:46] Like that two, three o'clock timeframe, but it's, I've been writing this in my stupid newsletter every week. It's just like the office hours. And it's just like, yeah, developments made developments, made developments, made developments, man. It's like, what? [00:16:00] What's going on? Like, you just keep saying it and then it's like, oh yeah, seven, seven will be ready next week, seven will be ready in three weeks, seven will be ready by the end of the year, seven will be, it just seems like the same sort of pattern in terms of like, oh, we're working on it.

Play from 0:16:14 8:88:88 0:16:14 [00:16:13] Great. Like, [00:16:15] if you don't have any updates, just don't say anything. Right, right. Just don't be like making progress. Uh, you know, had team members out sick this week. Didn't get anywhere. It's like, okay, like whatever. Yeah. It just, it just [00:16:28] seems like, you know, like we've [00:16:30] talked about it. It just seems like they're getting out chipped left and right right now. [00:16:33] I mean, and obviously they've got constraints. They're not VC backed, you know, they don't have any. Uh, funding coming in. Everything is just revenue based based on, you know, plans and usage from their customers. They [00:16:45] have no outside money coming in that at least as public or formal, uh, in that regard, which obviously there are pros and cons, but I think the tough part for them is they don't have these big bags of money.

Play from 0:16:57 8:88:88 0:16:57 [00:16:57] They can just burn through. Um, I mean, [00:17:00] sure they do have a lot of money with, with the user base that they've got, but it's, it's much more limited in what they can do and how fast they can do it. I did get the sense. And they did talk quite a bit on the last call about how. It is a big priority for them to start [00:17:15] shipping faster. [00:17:16] And they sort of like wanted to test that out in sort of like a micro pull format of, would you rather have fewer, but very, you know, very large updates or more frequent kind of smaller things. But, you know, even the smaller things that we're [00:17:30] saying is like. Style reference, character reference to me.

Play from 0:17:33 8:88:88 0:17:33 [00:17:32] Those were big things. [00:17:33] Those are great. I mean, that was, you know what I mean? Like, so those are the small things. [00:17:37] Yeah, if those are the small things, yeah, give me more of those like, but I think we need, you know, we need faster movement here. They need faster [00:17:45] movement because, um, I mean, it just seems like everybody else is moving so fast and that kind of maybe segues perfectly into. [00:17:53] You know, some of these other tools, right, that are out there and, you know, runway, for example, getting [00:18:00] into frames runway is getting into AI image gen stills and super excited for that. Uh, you and I had a chance to, to sit with one of the co founders of runway and somebody else from their team. [00:18:15] And they gave us sort of a little bit of a preview of that.

0:17:52Runway frames model

Play from 0:18:18 8:88:88 0:18:18 [00:18:18] And I came away. With a few things. One, you know, they mentioned that the, the look was going to be a little bit more cinematically focused. Um, you know, also I think like just from a user [00:18:30] standpoint, it's very interesting now that they've done the video thing first, then reverse engineered to do the still right. [00:18:36] And, and refine that. But, um, you know, maybe what I'll do really quick, I actually have this pulled up is I'll just kind of share this look really quick [00:18:45] that they've got up on their website here. So let's go here. And let's go here. You can see this, Rory.

Play from 0:18:53 8:88:88 0:18:53 [00:18:54] I got you. Yep. Okay, cool. [00:18:56] So this is on their site. [00:18:57] This is on runway site. Um, I'm just [00:19:00] going to kind of scroll through here, but base model. They're specifically focused on, you know, maintaining a stylistic consistency. They're looking for cinematic photography type looks. Um, I mean, [00:19:15] what's on here, right? Didn't feel cherry picked either when we were talking to them.

Play from 0:19:20 8:88:88 0:19:20 [00:19:20] They were just kind of showing us examples of things that they generated right in their feed, [00:19:25] right? In the interface. Yep. Right in the interface. It was not a lot of morphing. It wasn't a lot of [00:19:30] crap, to be honest. It was all somewhat, somewhat polished right off the bat, like maybe one or two reruns per image or prompt. [00:19:38] So that was something cool to notice. About it, that it was, that it was pretty clean. [00:19:43] It was pretty clean. And, you know, [00:19:45] also the other thing that was mentioned too, now that I just thought about this was, was text and them really refining text as well in the imagery and not just like a little bit, this looked very, very good.

Play from 0:19:57 8:88:88 0:19:57 [00:19:57] So going back to that sort of live [00:20:00] example, not cherry picking in their UI, they were just showing us different text examples that they had. And for example, I think there were batches of four, right? It was just like the text was hitting on all four. Right. [00:20:13] And there was, you know, [00:20:15] to, before getting too deep into frames, I want to, you know, I think runway from a marketing standpoint has been rolling this stuff out extremely well because it's, it started with gen three and I thought they did that in a great way where it was just like, you can [00:20:30] just text prompt.

Play from 0:20:30 8:88:88 0:20:30 [00:20:31] That's it. Then it was like, okay, now you can image reference. Okay. Now you can video reference. Okay. Now you can expand. Okay. Now you can do act one, right? Like these little small updates that are seemingly insignificant. One keeps them in the, like the top of mind for people [00:20:45] like trying and testing the new tools. [00:20:46] But two, it also gives you an opportunity to learn them before you have to take a whole like stack of things and try to learn at the same time. So to me, it's very easy. It's like going through like a course. Almost learn this first, get good at it. Then move on to this. Then move on to this. Then we went to [00:21:00] this.

Play from 0:21:00 8:88:88 0:21:00 [00:21:00] I think it's a very smart way to do it. So then this with frames, I would also probably expect the same sort of like rollout in that regard, where it's going to be probably just text to text to image first, then you might be able to add a style reference. Then you might be able to, you know, create [00:21:15] multiple or, you know, Character consistency from what we saw it's just going to be text to image to start it's a foundational model So they built this thing from the ground up which looks awesome. [00:21:23] But that's that's the other key [00:21:24] point, right? Yeah, that's a great point They they built this from the ground up. This is not built on anything [00:21:30] else Yeah, we're interesting you can tell to look at these. I mean these cinematic shots. I mean you can get Stuff like this in Midjourney, from a cinematic standpoint where, um, you know, it doesn't look plasticky or any, anything like [00:21:45] that, but right.

Play from 0:21:46 8:88:88 0:21:46 [00:21:46] Like one of the things that they brought up on the call was just the ability to dial into more of a natural, you know, look with some of the faces, right. That is one of the things that I think Midjourney still has, right. I mean, you'll get a lot of beautiful [00:22:00] faces. Unless you prompt for things otherwise, right? [00:22:03] Whereas these feel more ordinary people ish. [00:22:06] And especially like the, like, even on that, that man with the gray beard, right? Like, yeah, the eye bags, they're inconsistent on both sides. They're not symmetrical. The skin [00:22:15] wrinkles. They're, they're like kind of natural. Like across the face, even the skin texture on his, on his right cheek or left cheek there, like with the light shining off of it, that, that stuff to me is like the, the fine tuning process.

Play from 0:22:27 8:88:88 0:22:27 [00:22:27] And even like the, the definition between the, the [00:22:30] hairs on his beard and the contrast there, like that's impressive. Um, you know, I think that's where some of these other tools fail and it's really in stuff like that, because it starts to look so polished. The one thing I'm going to be cognizant of is like, if we scroll down here, I [00:22:45] think we, you know, I, I want to make sure that this is gonna be a thing because everything is expressionless. [00:22:49] That's the only, that's the only negative I see on this. If that, as soon as you say it, I, we pull up as soon as I say we pull up [00:22:56] one that's got an expression, yo, but everything else is, I'm trying to Do you like that ? [00:23:00] [00:23:00] Everything else is pretty much, you know, that we got blowing a bubble, which is pretty hard.

Play from 0:23:05 8:88:88 0:23:05 [00:23:04] Yeah. Um, you know, regardless in any of the tools. But, as you can see, [name] wants to chime in as well, so everyone knows she's back and, uh, back and, you know, better than ever, [00:23:15] after being away for a few days. But, yeah, I think the one thing that's, that really has, every, every tool needs a little bit more, like, refining on is facial expression, because they're all so model face every once in a while, where it's just that [00:23:30] blank expression.

Play from 0:23:30 8:88:88 0:23:30 [00:23:31] And, you know, that gets to be, unless you're prompting for it, sort of like annoying. Um, so I'm looking forward to this. Did this one specifically, considering the range of styles that we saw and what they showed us? [00:23:42] This might be the most excited I've been [00:23:45] since maybe Midjourney dropped, I think. To be honest. [00:23:48] You know, I'm more of a still guy. I mean, I'm gonna be doing a lot more video. But to me, man, I just love photography so much. So I'm really, really excited that we've got somebody that can go toe to toe. [00:24:00] You know what I mean? Um, and it's, and it's a very original sort of approach. And I like the things that they were focused on within their presentation.

Play from 0:24:08 8:88:88 0:24:08 [00:24:08] I, you know, they were asking us to some questions, right? Like, where do you think there's some opportunity? And I thought you and I [00:24:15] brought up a couple of really good things, right? We talked about expressions, which you just mentioned, right? Yep. Um, That's the default with I think any of these programs, but you know, we use Midjourney a lot unless you prompt for an expression, right?[00:24:30] [00:24:30] Um, you're, you're just getting that straight face, right? Yeah. Um, you're not even getting anything candid. You're just getting somebody looking into the camera and just like, right. And at first you don't really mind it until you realize that that's literally so easy. Everybody can do it [00:24:45] and it's not unique.

0:24:26Creative AI opportunities

Play from 0:24:46 8:88:88 0:24:46 [00:24:46] It doesn't tell the story. It doesn't drive emotion. So I think that's a really interesting thing. And we brought that up. The other thing that we mentioned too, right, Rory was, um, camera angles as well. Right. I think we were just trying to figure out like, [00:25:00] look, camera angles and expressions are two really, really good opportunities to drive. [00:25:06] Emotion and storytelling within the images. [00:25:10] That's that to me was important. Of course, uh, you know, composition was, was [00:25:15] important to me too, in terms of like, not just having like asymmetric, the medium shot, every single shot. Like that sort of look that it seems like, you know, granted, I think flux overdoes it on the medium shot sort of thing.

Play from 0:25:27 8:88:88 0:25:27 [00:25:27] And that's where I get frustrated with it because sometimes it's harder to push it [00:25:30] back or away. But the other thing that they nailed and what I noticed was, and I'm glad that this was in the thought process, cause they brought it up specifically, but I was going to ask the question was like the, the attention to motion. [00:25:42] In a shot. So in Midjourney [00:25:45] and other image generators, it's a little bit harder to create motion based on prompting. Cause sometimes it's just, it's very difficult to, if a car is driving to the left to get everything moving back the other direction, right? So you have like the dust and smoke maybe behind the [00:26:00] tires, or it's kicking up something like snow or water in the background.

Play from 0:26:03 8:88:88 0:26:03 [00:26:04] That stuff's hard, but what they showed us was really good because they're training on a lot of moving images and things like that, which also tend to work better in the generator. So if you can start to add that into the mix, it's more consistent and like, [00:26:15] you know, it doesn't overdo the motion to where we get some subtle baked in. [00:26:19] I'm really looking forward to that because those are the ones I've done a lot of tutorials on this thing. It's like how Much better runway or any of the image generators work for that with like motion baked [00:26:30] into the images because it just you're starting from yeah Better spot. Yeah. Yeah. Yeah. Yeah.

Play from 0:26:34 8:88:88 0:26:34 [00:26:34] This was also impressive this collage stuff. Thank you You want to [00:26:38] dive into that or a little bit really quick? [00:26:39] Yeah, so they were showing us this collage stuff You ever tried to create anything? That's a little bit more abstract or [00:26:45] collage in Like maybe even like a mood board and Midjourney, any of the other programs, like you might get 80 percent of it good, but no, you know, 20 to 10 percent of it will be like super weird and like morphed. [00:26:55] So this, with all these different elements, you know, layered on top of each other, cut [00:27:00] together somewhat coherent. It looks like almost as to me, that's really impressive. That's a, that's very complex. So, and, [00:27:06] and maybe even to have control over where certain things are in the image, right? I feel like that's where.

Play from 0:27:11 8:88:88 0:27:11 [00:27:11] Well, in any of these generators are limited right now is [00:27:15] like you could maybe get so far as to say, Hey, look, you want somebody on the left, somebody in the middle, somebody on the right, right? You've got like a group of people, but you start getting these objects, right? And where you want them placed, it becomes very difficult for that coherence to grasp those things. [00:27:29] So [00:27:30] this type of stuff I think is going to unlock, yeah, some new usages as well, specifically if you're like from like a brand standpoint as well, like marketing. Um, there's some real good storytelling opportunities that exist within here.

Play from 0:27:42 8:88:88 0:27:42 [00:27:43] And they're, they're pushing it. Like you can, you can [00:27:45] just see Runway's vision. [00:27:46] It's very clear. Now it is like we are going, you know, we are going after the enterprise corporate environment, right? Like maybe that's just my opinion, but they're making partnerships with big, with big brands, you know, with big sort of content [00:28:00] providers and. They're really pushing towards, like, you can do an end to end system in here. [00:28:05] You have, you know, this advanced motion capture. You have, you know, video to video. You have image to video. Now you have text to text to image. Then they're gonna do that latent space thing, which I didn't even, [00:28:15] like, comprehend the first time when they laid it out. I was like, wait, so we can create, like, almost, like, multi dimensional timelines?

Play from 0:28:21 8:88:88 0:28:21 [00:28:21] Like, this is, this is a different thought process. I like it. I like it a lot because that to me is like, we're also playing, you know, like [00:28:30] that to me is like, we're, we're playing. Cause no one's really, no one else is really doing that. Um, you know, from a thought process standpoint. So I was like, this is, this is cool. [00:28:38] Like, this is a different way to, I like this sort of nonlinear storytelling idea. So, you know, they have a, [00:28:45] they definitely have, you can tell they have like a creative, but like, this isn't coming from like, NVIDIA, right? Like it's not like super tech focused. Like it's not like a giant tech company doing AI things.

Play from 0:28:56 8:88:88 0:28:56 [00:28:56] It feels like a bunch of creative people doing AI things, right? Cause [00:29:00] that's, that's not something that I would be pretty. That's that I would be thinking of. I think that was a pretty creative idea. I don't even know how to explain it properly. That's how cool it is. [00:29:11] So [00:29:12] that one's exciting [00:29:13] though. I like, I'm real, I'm [00:29:15] real pumped about that one. [00:29:16] What about, um, also, I mean, Magnific is shipping like crazy, bro. Um, you know, I think going back to when Magnific was just this upscaler, you know, I was, [00:29:30] I was a little, I was a little hater on it. I mean, I was, because it was just like, man, I don't need, I mean, you really, It's cool to upscale, but you really don't need to upscale everything, right?

0:29:16Magnific editorial portraits

Play from 0:29:40 8:88:88 0:29:40 [00:29:39] It depends on where you're going to use the image, right? I mean, and a lot of these 99 percent of these [00:29:45] images don't leave your archive. Okay. But, um, as they sort of, sort of like built and built and built and they, now they've got style transfer and, you know, all these other things, but now they released the editorial portraits.[00:30:00] [00:30:00] Pretty impressed with, uh, some of the stuff I got back in there. I think it's a great sort of like V1, I guess, if you will, as they continue to refine it. But I was getting results back that almost felt [00:30:15] too real, where it was like, you know, because I'll do some like juxtaposition or, uh, surrealism or conceptual photography as well sometimes.

Play from 0:30:23 8:88:88 0:30:23 [00:30:23] And with that, it's almost like you don't like you want realism, but you don't want to go like, [00:30:30] All the 100 percent of the way there, you know, like maybe like a 90, 95%, but you know, I was doing my like infamous nun with a gun, you know, with a golden pistol and her golden teeth, you know, and I just get back like this straight laced lady and she's like, [00:30:45] hold the gun. [00:30:45] I mean, it was kind of funny, but I was just like, yo, this is too real. Like it's too literal, you know, [00:30:50] very literal. I've noticed that as well. Um, pretty good though. I mean, granted, I love generating in 4k to start, which is nice. It's [00:31:00] uh, you know, I think the good, what I like about Magnific is like there, they like stay in their lane and it's like, I, I don't use any other tool for upscaling other than them because

Play from 0:31:10 8:88:88 0:31:10 it is just so good and I can get exactly what I want.[00:31:15] [00:31:15] So, sooner or later, the only one feature I would like is either like a masking feature where like something doesn't have to get upscaled or like not in that sense like or like an in painting feature in that because like there's sometimes where [00:31:30] You'll have like a, like a guy wearing like a pattern shirt, right. [00:31:34] Or something like that, or someone wearing like a pattern and then like you'll upscale it and it'll like more, some of the patterns into like another hand or like some like really weird object. And I'm like, I don't need [00:31:45] that. I just need, or like, you know, it'll turn. Like an ear with an earring into like a really fat ear or something like that.

Play from 0:31:52 8:88:88 0:31:52 [00:31:52] You know what I mean? Like it just, sometimes it does that weird stuff, but the editorial portrait side, I was pretty impressed. I mean, I, I [00:32:00] did some, some macro stuff. I did an alligator eye that had such insane definition. I was like, there's, this is crazy. Uh, pull [00:32:07] it up, dude. Let's see it. I was just gonna, I was just gonna scroll through this so you guys can get a sense if you haven't seen it. [00:32:12] Right. So focused a lot on. [00:32:15] You know, closeups, medium shots, macros even, right? So it doesn't even have to be a person. It could be like a bug or a bee or something like that. Right. Um, these are all the same prompt just to kind of give you an idea. I mean, these are pretty incredible. I mean, [00:32:30] outside of this earring, not in the hoop, uh, not in the ear, but, uh, You know, like just the color, the quality, like you said, right?

Play from 0:32:39 8:88:88 0:32:39 [00:32:39] So these, these are already, you can put these in 4k, I think right out of the gate. I think I put these in 2k. [00:32:45] I mean, these are sharp already, you know what I mean? And then you get into, let me see if I got some of those none with the gun idea. These were, these were like, yo, this is just silly, right? That's [00:32:58] like, that's like sister, you [00:33:00] know, sister, sister Jane right there. [00:33:02] Like that looks like someone, [00:33:04] someone who grew up with like, yo, so, you know, these were funny, but at the same time, I was like, yo, it's a little too real. I mean, yeah. Um, you [00:33:15] know, like, dude, it's just wild.

0:32:48Macro photography AI

Play from 0:33:17 8:88:88 0:33:17 [00:33:17] Yeah. I like that though. It's like that, that one's like funny. If you scroll back up real quick, there you go. [00:33:23] Look at the guns, got the grills screwed. Killing it that one right there the one right above it this one [00:33:30] that one which is just like [00:33:34] Yeah, just but also just like her face it's like I it's like every nun I've ever known in my entire life, you know [00:33:41] Right, but you don't you know a lot of nuns, huh?

Play from 0:33:44 8:88:88 0:33:44 [00:33:43] I did have well, you know, I [00:33:45] grew up in a Catholic family went to church a lot How do we not know that about each other? [00:33:51] You know, I went to a Catholic school all the way through high school, bro. [00:33:54] Hey, you know, there it is. It's, it's, we, you know, we had our nuns back in the day. CCD, all that stuff. [00:34:00] The convent. It wasn't called [00:34:01] like a convent, right? Something like [00:34:04] that. Tuesday nights at 7 p. m. Every year, every week for a night. Doing fish fries [00:34:08] and [00:34:08] stuff. You get fish fries? Fish fries! We didn't do any fun stuff like that.

Play from 0:34:15 8:88:88 0:34:15 [00:34:14] Is [00:34:15] that a [00:34:15] midwest thing? I don't know man. That cod was hitting though. [00:34:19] We definitely did not have fun. If we got a random pizza party every like 6 months, that'd be cool. Maybe like a bake sale. But, you know, It's a, it was going [00:34:30] after school to CCD, basically just trying to make it through while we're, while we're learning about the Bible. [00:34:36] [name] obviously doesn't like it right now. Um, let me, let me see this. Let me pull up this alligator eye real quick. I was going to say, yeah, [00:34:43] pull it up.

Play from 0:34:44 8:88:88 0:34:44 [00:34:44] Pretty [00:34:45] damn cool. [00:34:46] [name], our, our podcast mascot. [00:34:48] She is ever present. And very loud now that she's teething. Let's say [00:34:53] you got to drop that video too. Were you, uh, that [00:34:57] one you [00:34:57] sent [00:34:58] me? I, I do have to [00:35:00] put that one up phase duet with Taylor Swift, but I just thought this thing was pretty insane. I mean, I just like the, I, the, I definition here, like a lot of different colors, marbling, like even on the skin, I mean, that macro [00:35:14] [00:35:15] is that's, that's better than, than Midjourney macro Q2.

Play from 0:35:22 8:88:88 0:35:22 [00:35:18] Oh yeah. This is. This is next level, which I was like, I could, I could really do something with this, like all these, this different kind of like macro environment sort of shots, which I love. Yeah, I tested it on a [00:35:30] few things. I tend to think like, here's my, here's my biggest gripe with it is that it, it ends up looking like Midjourney 5. [00:35:36] 2, like where everyone is perfect. You know, like I'm not getting, uh, I was trying to get like a younger face, a, you know, more average face when I was [00:35:45] prompting, but I ended up just getting like, These like model looking crazy sharp eyes thick eyebrows perfect lips like you know it's not That's not the child soldier that I was envisioning when I was doing it I've also found there's a big difference in like six point or doing it [00:36:00] in 16 nine verse doing it in three by four So you get like a little bit more a little bit more lifestyle in 16 nine

Play from 0:36:12 8:88:88 0:36:12 verse actual like portrait style in three point or three four so that was just yeah one thing that I noticed [00:36:15] With it, but you know I was digging it when I used it, I kind of, you know, stopped using it because then I got pulled into every other image generator tool like luma now. [00:36:25] But yeah, let's, let's, I mean, [00:36:26] let's talk, let's talk about that too, right? Like luma's luma's popping too. [00:36:30] Now they've got a lot of things that have changed and dropped. I, I honestly, I was sitting in on just telling you this. They had a, like a creator program call yesterday. I was sitting in on that. I mean, Um, it was just like everybody introducing themselves, but there were, I mean, there were so [00:36:45] many talented people in there.

0:36:29Photon and tool comparison

Play from 0:36:50 8:88:88 0:36:50 [00:36:46] Um, it was, it was just wild to hear everybody's sort of like background and experience and how they're sort of like driving this change in their own respective spaces and industry. So exciting stuff, Luma, definitely something to keep [00:37:00] an eye on with them as well. What was your take on Photon? I actually haven't played with Photon yet. [00:37:05] I like the interface. It feels like they, they took the, the Dali approach where you can kind of like interact and chat with, you know, like your images [00:37:15] as you're generating them in a sense, like, Oh, like, give me this. Okay, let's change her hair color to blonde. Like, you're not gonna get the same exact, like.

Play from 0:37:25 8:88:88 0:37:25 [00:37:22] image, but it's a good way to like, I think like thought process wise, go through it and maybe like, Oh, I want to change this. They have a brainstorm feature, [00:37:30] meaning like, Oh, we can take this in a different direction, but like keep the same theme of what we're doing. So it's like an interactive workspace, which I think is a cool idea. [00:37:37] And yeah, it's like a nice, it's a nice, refreshing way to go about creating within that space. So, you know, image quality to me still [00:37:45] feels like Midjourney 5. 2, right? Like just what it looks like. That's, that's where I think most of these things still keep getting to before they get to six, six point, uh, you know, 6.

Play from 0:37:59 8:88:88 0:37:59 [00:37:54] 1, but I'm, I'm finding it hard on a lot of these other image generators to [00:38:00] get any sort of level of creativity that Midjourney has. That's the one thing that's still just Far and ahead like when I put like when you put a generic prompt in Midjourney, and you get something absolutely ridiculous You know like once in a lifetime photo like we were talking about a couple weeks ago Do you [00:38:15] put that into you put that into photon you put that into flux? [00:38:18] It's just like whatever do you put that into Midjourney you get something crazy stupid because the creativity level on Midjourney to me is Just like the differentiator You know like yeah goes far And you don't realize it till you use other [00:38:30] tools right when you're looking for something super creative like that's the thing that's the You The differentiator plus the level of control, right?

Play from 0:38:40 8:88:88 0:38:40 [00:38:36] It's still on match there. Yeah. [00:38:38] Yeah. And, and that's where it's like, it's really interesting with that storytelling piece. What is that going to be? What's that going to be? Because [00:38:45] here's the other thing. I'm surprised. No one's really, well, there, it might exist. I maybe, um, maybe I got this from somewhere, but how cool would it be to. [00:38:56] Essentially create a string of [00:39:00] images in one go. You're you're, you're, you're kind of, so maybe this is something with storytelling, but you've got a basic, um, or simple idea of a story, let's call it. And instead of getting like a [00:39:15] single image, you're getting groups of three or four images in that story, right?

0:38:46AI storytelling features

Play from 0:39:24 8:88:88 0:39:24 [00:39:20] That connect that to me is something that I haven't seen anybody tackle yet. You know, like, like, remember your Tulum example? Yeah. Right. [00:39:30] What if you were saying something about staying at a hotel and, you know, you just kind of give it an idea or a story or an arc, and then all of a sudden you get, you know, let's just call it the grid of four back, but each grid has three sort of [00:39:45] frames in it.

Play from 0:39:49 8:88:88 0:39:49 [00:39:46] You know what I mean? Like that, that to me is an untapped powerful feature that could be driven further, because right now, when you think about it, everything is very individual with the, with the, with the images, right. Which is [00:40:00] okay. But I don't think that's sustainable over time. Right. Cause then you gotta, okay, I've got this image and I got to create this one. [00:40:06] Right. But it's just like, you know, why can't you just match the style and do everything in one go? Right.

Play from 0:40:16 8:88:88 0:40:16 [00:40:12] Yeah. Yeah. I think it's, you know, [00:40:15] to me, that's like a very interesting way to think about it too, right? I do like that idea and like, maybe it's a similar thing to key frames where it's like, you put this image in, you get this image, then it like generates kind of like the story in between, which, you know, is, is to me. [00:40:28] That's what you have to do [00:40:30] with the video generators anyway, right? Like basically they are creating all these in like 20 frames per second, 24 frames per second in between the two key frames, and then like applying the motion to it, but interpolating it at the same time. So it's like, they're kind of gonna have to go the same, [00:40:45] maybe go the same route with that.

Play from 0:40:51 8:88:88 0:40:51 [00:40:47] I'm, I'm interested to see if the, like the, they're going to add like another piece of the website. Like is the storytelling going to be like its own edit tab or is it just going to live within. The normal UX of, uh, you know, what's going on, right? [00:41:00] That's. To me, that's the interesting part of where this could go. [00:41:02] Cause that's why maybe it's taking so long. What's your, what are your plans in January? God, well, at least, you know, [00:41:09] this is going to drop when you're traveling or I'm traveling, of course.

Play from 0:41:17 8:88:88 0:41:17 [00:41:13] Why wouldn't it? It'll be [00:41:15] Christmas. It'll be Christmas Eve night or something like that. When you can't work on it for three days and then you'll see 75, 000 posts on any sort of plot Midjourney update, Midjourney update with no one actually using it. [00:41:25] It's just like, here's what you need to know. And there's a. [00:41:30] There's, uh, like, I'm just curious as to what's gonna be rolled out first. Like, is it gonna be 6 point, uh, or 7? And then, like, video's gonna come out after? Is it gonna be video first? Is it gonna be like any sort of inkling into [00:41:45] 3D? Because I feel like the 3D stuff is like, it's a pipe dream at this point.

Play from 0:41:52 8:88:88 0:41:52 [00:41:48] I don't know. I just, I felt like they said it was like on track to be before video and now it's like, I remember that. Infin delayed. [00:41:54] Yeah, we don't even, yeah, we're not even getting updates on that. Be very, very curious to see what [00:42:00] happens. Hey guys, while you're with us, um, we're also gonna open this up for questions. [00:42:04] We've got plenty of time here, so. If you guys have specific questions that you want us to address, whether it's Midjourney or Runway or Luma or [00:42:15] Magnific or Kriya or whatever the hell tool you're using. Oh, should we talk about Let's [00:42:19] do Kriya. Oh, okay. Let me just, what were you going to say?

0:42:23Kling and Luma updates

Play from 0:42:28 8:88:88 0:42:28 [00:42:25] What about that hun, hun one? [00:42:27] The hun one. [00:42:28] [00:42:30] Oh, it looks pretty good. The newest, [00:42:31] the new kid on the block. It [00:42:32] looks pretty good. Open source. I don't, I haven't personally tested it. I've seen some uh, some people. I did a little last night. Just a little last night. Oh really? Yeah, how'd it work? How long did it take? I mean, I [00:42:43] thought it, um, it did take [00:42:45] a while. [00:42:45] Okay, like I figured it would be a minute. [00:42:48] If I had to take a guess like 15 20 minutes a gen.

Play from 0:42:55 8:88:88 0:42:55 [00:42:51] Okay. It makes sense. I mean, yeah, you know, it is open source. It is like brand new. So was it, were you trying to do like anything [00:43:00] complex or was it fairly standard? [00:43:01] What was I trying to do? I think I was just trying to do like a first person POV of somebody on a motorcycle, Um, riding through the streets of Matera, you know, Matera, like from 007, I was going to [00:43:14] say that [00:43:15] reminds me of, that was very James Bond of you to do that.

Play from 0:43:22 8:88:88 0:43:22 [00:43:18] Yeah, yeah, yeah. I mean, there was, there was some pretty cool stuff there. Um, but I'll be, I'll be, I'll be honest. I got pretty good stuff on runway too. You know what I mean? Yeah. [00:43:27] Yeah. There's there [00:43:30] like, it's so weird, which one I want, like I choose now, like it's very specific and defined. And like minimax to me still has the best motion. [00:43:37] Like, if you can get the most, the most creative, like if we're talking about creativity scale. Minimax has the most creative, like [00:43:45] creative ability runway to me is the most consistent in terms of like I can do just about really 90 percent of the things I want on there with the tools they have and then cling is I need something extremely specific like that motion brush to me with the multiple like [00:44:00] having that feature.

Play from 0:44:04 8:88:88 0:44:04 [00:44:01] With the arrows drawing, it's so easy now in comparison, that was like a nice update on what Runway did with the motion brush, because it used to be like, you had to toggle the, you had to toggle like the weights of how far you wanted things to go left or right. The arrow is like [00:44:15] ingenious of a way to do it. [00:44:16] The only thing that that doesn't work with, or I've had trouble with, is depth. So like, if I have something that's in the foreground and I want to move it to the background, it's hard to draw an arrow that way, you know what I mean? Because it just looks like it's going up. Right? Like it's very, there's [00:44:30] not enough 3D sort of, uh, you know, plane in there for it to do that, to work on the, I think it's the Z axis?

Play from 0:44:40 8:88:88 0:44:40 [00:44:36] X, Y, Z, yes, I think. So, that's, hey. These are like little things. Oh my god, man, chill. Dude, hey, [00:44:45] I get excited about when you can control this stuff and do what you want. Chill, buddy. Because, hey o. There's, so the, the other thing, I think someone was, I think someone mentioned this up here before. It was like something about the [00:45:00] subscriptions to all of the, it's like, uh, I don't want to pay for all the subscriptions.

Play from 0:45:08 8:88:88 0:45:08 [00:45:04] Whereas this it's tough being subscribed to so many of these video platforms looking for only seven generations or decent. I [00:45:15] like personally, like, I think if anyone's just getting started on this stuff and wants to try a bunch of them, Korea is the best place to do it. Cause you can test all of them at this point. [00:45:24] And. They all have different uses, right? So, I gotta get back into Luma a little bit. I've been, I've [00:45:30] been kind of like fading them. Maybe with this new Photon, it'll be a little bit better. Remember, remember Pika?

Play from 0:45:39 8:88:88 0:45:39 [00:45:36] They've already They've already gone back under a rock. I mean, they had that, they had those 48 hours and then it was just like, geez, okay, [00:45:43] we can, you can squish [00:45:45] something or explode something and that goes viral. [00:45:47] And then it's like, and they're gone [00:45:49] and they're gone. Somebody else ships, ships up the new. So [00:45:54] there were, so the other thing I want to mention here is Korea [00:46:00] has been just. Like pumping. And I, I think it's my, my favorite platform of the moment, just based on their direction of what they're doing. They're making things so easy.

0:46:06Krea Lora training

Play from 0:46:15 8:88:88 0:46:15 [00:46:12] It's like, to me, the beginner's platform, like you can [00:46:15] do, you can test a lot of things on there. You can learn a lot of things. It's not overly complicated in terms of what it is. But like the, the Laura training, even that so easy, so fast. I did one this morning of myself in [00:46:30] 10 minutes. It was done from start to finish between curating my images, uploading it and generating. [00:46:35] But the other thing that's great is the auto is the auto tagging. Um, Cause like, when I look at this also, if we, let me pull this up real [00:46:45] quick. [name] is really just a pull it up on shout out, shout out to [00:46:49] Moody, by the way, this guy, this guy, or girl is just like killing it with the comments. We appreciate that.

Play from 0:46:58 8:88:88 0:46:58 [00:46:55] Um, yeah, it's, it's super interesting. And, and, you know, it's kind of like [00:47:00] when you're just comparing features, right. And like, this goes back to the whole Midjourney community. You need to get the stickiness of a community behind any of these tools, [00:47:15] right? Because just to pique his point, right? You could have something cool and it's cool for two days and then somebody else ships something new.

Play from 0:47:26 8:88:88 0:47:26 [00:47:22] Yep. Right? Whereas it's like, okay, well, Midjourney, I've been sticking with Midjourney for a long time now. You and I both. Why? [00:47:30] Well, one is the quality, but two is also because that was our horse and they've made the updates. It's continued to be good, right? And in a lot of ways, we base, you know, some of our workflows, you know, within it. [00:47:41] And that's where we become comfortable, right? And so these [00:47:45] video tools to that's where there's opportunity, right? You got to build brand affinity. You've got to increase your stickiness. So people aren't just going to choose what's latest and greatest.

Play from 0:48:00 8:88:88 0:48:00 [00:47:57] Yeah, that makes a ton of sense. Cause this is like, [00:48:00] but that's why I'm like really digging what Korea is doing. [00:48:02] I mean, seriously, this took me, I uploaded 16 images of myself. Just put that all [00:48:07] the way. Can you, uh, full screen that full screen? Let's [00:48:11] see. Can you see that now? I mean, I can see it. I [00:48:15] [00:48:17] just, it looks like you could, um, [00:48:20] cause that's, it's really like, Yeah, this one, these are all the images of me, right? Like it was just like me from my wedding day, like nothing crazy.

Play from 0:48:32 8:88:88 0:48:32 [00:48:29] These are just like [00:48:30] my wedding pictures. It was something that I had myself in the same outfit in multiple situations. But you know, when it starts to replicate, right. Like insane that that's not me, like that, that's pretty damn good. Everything from the tuxedo [00:48:45] itself down to wearing no socks on the shoes to when we get into even more stuff over here. [00:48:51] Or my assets like the way that it is able to get the subtleness of my beard Like I know I have a hole in my beard there, right? Like I know [00:49:00] I know that day.

Play from 0:49:05 8:88:88 0:49:05 I didn't I didn't shave the top of my face So I have some like stubble Up on my cheeks and I have a little bit of sunburn because I was sitting in the sun Before I got it. [00:49:08] It's weird that that it picks this stuff up because some of this is so good like I could never know that even we're [00:49:15] just talking about this and [00:49:16] you said I think you said you you based this off of what 12 to 15 images right [00:49:21] 12 to 15 images like it even gets my sunglasses not like nails those The ones that I was wearing that day so like these some of [00:49:30] these images you'd never to me like I would I could pass these off like I could pass that off as an image I could pass that off as an image even gets like this weird pass it all your hair absolutely dude you could absolutely [00:49:41] yeah like no doubt no one would ever [00:49:43] know No one would ever [00:49:45] know.

Play from 0:49:49 8:88:88 0:49:49 [00:49:45] And then even some of the stuff where it's a little bit more generative in nature, where it's not like my, what I'm wearing, where I was dropping the style weight down a little bit. Where's this other one of me in a different, you know, different tuxedo. I just couldn't believe that like this one, it prompted myself for a different [00:50:00] tuxedo. [00:50:00] It still kept my tie with that same little subtle subtle pattern, which I was like, man, this is really good. Some of these are. Pretty, uh, pretty actual into like this one to me is like really actual in terms I have side profile [00:50:15] pictures Even like the the scale of my hands I have large hands everyone not to not to like call that out It's like a thing that I I know about it.

Play from 0:50:26 8:88:88 0:50:26 [00:50:22] So I see it in pictures, too It's like a hold my hands up. It's like damn those things just they look out of proportion. What about what [00:50:28] about shoes? What's shoe size [00:50:30] you rocking [00:50:31] the 12 ish? 12 12 and dress shoes and how tall you like sneakers 511 Yeah. Okay. Yeah. I mean, we're, [00:50:39] we're, we're like the same then. [00:50:40] We're the same. Yeah. We're the same guy. [00:50:42] Maybe. I just, I just do. I just can't believe some of this stuff. [00:50:45] This one, it screwed up the tuxedo, but regardless, I mean, I don't even know what to say. It was like, yeah, it took five minutes and that's probably the best I've seen. That's just trained on flux.

Play from 0:50:59 8:88:88 0:50:59 That's not like one of these specific tools, you know, it's not one of these specific tools. [00:50:59] That's like a [00:51:00] AI portrait generator of yourself to do like work headshots and whatnot, right? Right. But that was. I mean, I'm gonna have to play around with this a lot more now on, on Korea. I just think it's something that's pretty crazy, but yes, you can, Angela. Yes, you can change the entire [00:51:15] outfit. Um, you just have to prompt for it. [00:51:17] I was just trying to get this, get this live and get myself in a little bit of a different like poses and aesthetics to see how it handled that, you know, then the clothes, you can change it. That's just [00:51:30] accuracy was insane to me. Like, like I said, that one little. Patch right here. It even nails that which is yeah juice.

Play from 0:51:41 8:88:88 0:51:41 [00:51:37] Did you see this question from Angela here? [00:51:39] What'd it say [00:51:40] just says can you change entire outfit? Can you see that? [00:51:42] Yeah. Yeah. Okay. Yeah, you can [00:51:45] change the entire outfit It's um, you have to it's like there's weights on how much you can like influence of like the Style that you've input or like the Laura that you trained so you can like bring it all the way up It's gonna look exactly like you and like all your training images or you can bring it all the way down I'm [00:52:00] I would probably suggest using more neutral clothes in the training data.

0:51:45AI tool subscriptions

Play from 0:52:06 8:88:88 0:52:06 [00:52:02] So it's a good mixture because I'm wearing the tuxedo and all of them. It's going to be pretty close to that, like from a model training standpoint. So like, I'm going to have to like work hard to get it out of that. But if you're using like, that's, I think that's just general Laura [00:52:15] model training sort of like best practices. [00:52:17] They're going to have like neutral clothes on like, or your character with neutral clothes, cause something very defined like a tuxedo. It's going to end up looking like that. I'll [00:52:24] tell us, tell us what you mean by neutral clothes. [00:52:27] So, you know, like, uh, t shirts [00:52:30] basics, like base basic stuff, not a lot of heavy graphics, not a lot of heavy patterns.

Play from 0:52:37 8:88:88 0:52:37 [00:52:33] Um, you know, things that are not things that are not going to like really affect the training data. So that's a good [00:52:40] reminder. [00:52:41] Yeah, much easier to work off a gray t shirt than it is to work [00:52:45] off like a Like buffalo check or like a chevron pattern like that's gonna influence a lot of things Yeah, I just couldn't believe it nailed the sunglasses to man like it got it. [00:52:55] Yeah, that's wild frames down the Ombre color of it. I was like [00:53:00] what the hell is going on here So that is are you capable of saving out the training warm up? You know, what?

Play from 0:53:10 8:88:88 0:53:10 That's an awesome question, buddy. I didn't even think of that. Um, let me go check if you can. [00:53:11] Yeah, that would be interesting. Great question. [00:53:14] Shout out [00:53:15] to, uh, I'm going to call a few people out here. See retro punk here. What's up, buddy, Dan. Love you, man. Thanks for showing up. Seems like you, uh, you had some interesting stuff that you just tagged on LinkedIn. Thanks for directing people over to [00:53:30] YouTube as well. LinkedIn Jesus, bro. Um, what else did we have?

Play from 0:53:40 8:88:88 0:53:40 [00:53:36] Anything else? Yeah. I think that's it. Well, let's see. Yeah. Somebody just talking about mini max. Yeah. So, I mean, [00:53:45] crazy stuff. I mean, just crazy movement. So many things shipping and how do you stay on top of it? You know, like, dude, it's just so hard. And I, I even think I'm just like, I do this stuff everyday, you do this stuff everyday, and I see [00:54:00] people like playing with every little thing that comes out and I'm just like, what do y'all do?

Play from 0:54:08 8:88:88 0:54:08 [00:54:03] This? [00:54:03] What do y'all do? I mean, even Runway today, Runway shipped Act 1 with movement. Now, so like, instead of it just [00:54:15] being static, me talking like this, like I can upload a video of myself, like moving, walking in a scene, and then it's going to animate my face for motion capture. Like, come on, it is crazy. [00:54:25] And what was the other, [00:54:26] who, who was the other one that came up with, uh, I'm just, [00:54:30] um, having a blank right now with the. You can animate, um, or you can do video with two K images. Who, who launched that? That was, and it was pretty solid too. Did you see that?

Play from 0:54:45 8:88:88 0:54:45 [00:54:41] No, I think it happened while [00:54:42] you were traveling. So it was like the last [00:54:45] 48 hours. [00:54:45] I'll, I'll have to find that. It's just, or if you know, put it in the comments, it's just drawn a blank right now [00:54:52] and cling, cling, try on Angela. Good call out. That was a good one. Um, I like it. It's not precise. It's not perfect. [00:55:00] I did a workflow on scooter that I got to share. Yeah, I did a workflow on Twitter that I got a share of using, like thinking about it a little bit differently because the try on feature, my, my only gripe with it is that when you render the clothes onto the model, like it [00:55:15] loses fidelity and image quality.

Play from 0:55:20 8:88:88 0:55:20 [00:55:16] So then you have to upscale it and then it upscales like it kind of morphs things out of proportions. Yeah, and then, you know, so I was doing something like, all right, here's the people like, Oh, but you can't do a graphic tee. I'm like, Oh, let's see if it can do it in a different way. Not specifically with try [00:55:30] on where. [00:55:30] I was like, I generated a model and then I generated just like a, or I went and pulled like a blank tee, like a Gildan, something like a Gildan, whatever. Um, just like a blank that, you know, most of these t shirts are printed on. And I just did like the try on there. So then I got the [00:55:45] model and then I brought it into illustrator and took like a vector image and put the live, you know, did, um, the mockup in illustrator, so that you have like the full sort of layout across like the wrinkles and everything like that.

Play from 0:56:00 8:88:88 0:56:00 [00:55:57] And then re uploaded it into create or into [00:56:00] a cling. And it worked that way, and the graphics stayed true. So, like that's, you just might have to take an extra step. I'm sure that, that tool will, that tool will, uh, improve. I'm just, uh, it just didn't really make sense to me as like, Cling, [00:56:15] what, why try on? Of all the, of all the features, like, That was an interesting one for them to, to sort of upload into, uh, into the environment.

Play from 0:56:26 8:88:88 0:56:26 [00:56:23] Do they have a creator program over there, Cling? [00:56:26] Which one? Cling. Yeah. I'm in it, [00:56:30] apparently. [00:56:31] But do they, I mean, do they do any kind of communication? [00:56:34] It was kind of just like, uh, they got, they got to meet, uh, you know, on Twitter and was like, oh, we'd love to have you in the creator program. And they're like, we'll give you some free credits. [00:56:41] And I mean, but [00:56:42] did they do like a meeting or anything like just [00:56:45] sharing roadmap or like office hours? Not that I know. [00:56:47] Not that I know. I mean, Runway is really great at that. I think they have a really awesome creator program where once you're in, you know, you're, you're, they're constantly updating and they're generating, you know, sort of meetups [00:57:00] for us where we can go and talk and like share things and giving us updates on the platform.

Play from 0:57:07 8:88:88 0:57:07 [00:57:04] They're super engaged with their, with their creators like they, [00:57:07] yeah, [00:57:08] they like it. Like you can tell that they enjoy Feeding us and letting us go and do, you know, whatever the hell we want to do. I [00:57:15] think we're going to get them on [00:57:15] the, we're going to get them on the pod here. We're going to get somebody from runway on the pod. [00:57:19] Um, we're going to try to do that sooner rather than later too. So just stay posted on that. I think the other thing too, like we'll, we'll probably do that with a few. We're trying also to get in touch with [00:57:30] Midjourney. If you want to be on the pod, if you want to do anything. You know, community driven outside of the office hours.

Play from 0:57:42 8:88:88 0:57:42 [00:57:38] Let us know we're here. We're advocates. Don't leave us hanging. The people want to hear from you. [00:57:45] Uh, I think Luma would also be a good one. Korea would be also another good one if we can get them on say with Magnific. So all the tools that we talked about today, um, where else you want to go with this, man, is there anything else we want to touch on, uh, guys, if you have any other questions, now's the time pop them [00:58:00] in the chat. [00:58:00] Let us know [00:58:02] Dan. You guys seen 11 labs conversationally. I super cool. Love that. I'm gonna try to get Justin Justin on from 11 labs been in a little bit of a group chat with him. So hopefully we'll get some, you know, some of these people here on the [00:58:15] podcast soon to really go go deep. You know, these might even be good for for lives.

Play from 0:58:23 8:88:88 0:58:23 [00:58:19] Drew, if we get them at some point so we can get some some live questions from, uh, You know, some of these people not saying that would be first run. That's a great [00:58:25] call. Yeah. Yeah. [00:58:26] Hey, here's a question [00:58:27] for you. [00:58:30] Storyboarding, which tool? [00:58:32] Well, I mean, personally I'm using Midjourney. That's just me, but there's, there's a couple of ones now, like LTX has a, has an easy storyboarding type environment. [00:58:40] Um, I think that one's pretty, pretty good in [00:58:45] regards to, um, uh, like how you can actually control it easily. I've seen the interface. I w I worked on it for like a day.

0:58:34Storyboarding and Krea

Play from 0:58:57 8:88:88 0:58:57 And I was like, well, all right, this is like, if you don't, if you just need to get up and running and don't necessarily need control of every fine grain detail, [00:59:00] um, you know, in terms of like the aesthetic, if you want a storyboard, you can do it there and it's easy to like, you can choose your character and place them across different scenes. [00:59:08] You can add props. It's very like click and point verse like type and like be creative with text. So [00:59:15] I would suggest that one. If, if, you know, if you need something easier, like if you want to control every, every piece of it, I would say. You know Midjourney to me still flux for you know, maybe character training and doing it that way But you know i'm more of a they're all they're [00:59:28] all gonna move that way anyway.

Play from 0:59:34 8:88:88 0:59:34 [00:59:29] [00:59:30] Yeah, I feel like [00:59:31] yeah You know, I honestly haven't tried lumas yet. I tried, you know going through and generating images on photon I didn't try the storyboarding, but that's something i'll probably get into over the weekend [00:59:40] Yeah, I think that'll be pretty solid too. I got to try that out myself too. What about, um, current price point for [00:59:45] Korea? [00:59:45] Is it worth it? I think absolutely. You've got all the tools consolidated in one place. It's, it's kind of what we're looking for and we need to be honest, because right to the point of, we've got so many different video models. We've got so many different AI image generations. Like, yeah, some are good at other things than others.

Play from 1:00:03 8:88:88 1:00:03 [00:59:59] [01:00:00] And, you know, like there are strengths of some weaknesses of some. So, I mean, if you can consolidate that, I think that's ultimately. It's going to be more efficient for people's workflows and trying to get stuff done. [01:00:09] Let's just take a look at this real quick because I think some of this stuff gets hidden where CREA [01:00:15] doesn't really update it that much or like really give you the best marketing on it. [01:00:18] So you obviously you have. You can generate images on here with flux. You can do video in any of the models, right? You can, when you generate videos, right? You can choose any of the models, [01:00:30] whichever ones you want to use.

Play from 1:00:35 8:88:88 1:00:35 They have the APIs to them all, which is great. They have the real time generator, which is fun. [01:00:36] Um, you know, they have the enhancer, so you have an upscaler, but they also have like these little mini apps. You can do patterns, you can do logos and illustrations. [01:00:45] You know, if you go into edit, actually, it's not live yet, but But it's going to be like the answer. It's a Midjourney in terms of generative in painting, right? [01:00:54] Like that, it looks really good. I don't know how great it is, but it looked pretty good on [01:01:00] the demo. Uh, the other thing here is. Where is a real time image? Let me see if they have the, cause there's like an infinite canvas in here. Yeah, I'm, I'm [01:01:15] forgetting where it is. See, that's the problem.

Play from 1:01:21 8:88:88 1:01:21 [01:01:17] There's just like, it's so much to keep trying. [01:01:19] Who's doing what? Like, like, dude, you know, please. That's why, that's why I think I don't even know what Korea's current prices, but it's, it's probably worth it. [01:01:28] I think it's like, there's like [01:01:30] 30 and 60. I'm just seeing if there's, if you can bring this into like the infinite canvas here. I haven't played with this in like a couple of days. [01:01:38] So I always forget there, how, where they moved everything. But again, like when you're going through and you're trading, you know, you [01:01:45] can, you can train models. You can also go and add other models that are like within the gallery. Of, you know, community models. So if like someone else is training on something, you can go use their trained styles, as long as they're made public.

Play from 1:02:01 8:88:88 1:02:01 [01:01:57] So, that's just to me. If there's, if [01:02:00] there's a tool where you want to get a good taste of everything, like CREA is exceptional for that. And just like having a fun time with it. They can just do stuff that other tools can't. All in one place. [01:02:12] Yeah, maybe also like a really good opportunity [01:02:15] for those that are just getting their feet wet with a lot of this stuff too Yeah, right because everything's in one place another reason for that. [01:02:21] But um, let's see. We had a question. Will this be your weekly live schedule moving forward Um, maybe a hit or miss I think [01:02:30] rory and I both talked about doing more live sessions That's actually what was you know, really the impetus for us to do this Anyway, as we started doing a lot of um LinkedIn lives, uh, ironically enough.

Play from 1:02:47 8:88:88 1:02:47 [01:02:43] And we, we had great [01:02:45] turnouts. Um, the community has always been super supportive of us both and we're super thankful for that. But, uh, we just got to the point where it was just like, okay, like let's take this off LinkedIn. Let's just do something. We both thought about the podcast idea. And then if you go [01:03:00] back and you watch the first episode, you might cringe for a little bit. [01:03:05] You [01:03:06] know, it, it, we, we have our good apps and our bad apps. Maybe that one wasn't, you know, It was a pretty good [01:03:12] app, actually, you know, but solid [01:03:14] [01:03:15] content [01:03:16] five years younger. So it's like, damn, [01:03:19] AI is aging us a lot that, that and babies. But regardless, dude, I think there's, you know, there's some fun things that we could do with live.

1:02:53Future live show plans

Play from 1:03:31 8:88:88 1:03:31 [01:03:27] We don't know what we're going to do with it at a certain point. We, [01:03:30] we. We have to find the time to prep for it, where if we're going to do some, but we're talking about a live create where we can come on here and sort of like take an idea from, you know, idea to conception, like, or idea to, uh, to completion where, you know, again, that's, that's not an, [01:03:45] that's not like a very easy structure for a podcast, but it might be for YouTube or a little bit of [01:03:49] dead time in there, because [01:03:51] it's going to be impossible for us to just kind of like, you know, talk through all of that and once we're just rambling, but there's a, there's some, you know, We're looking [01:04:00] forward to doing some cooler stuff and hopefully like this doesn't crash when we try to generate video and you know, stuff live, much easier.

Play from 1:04:10 8:88:88 1:04:10 [01:04:06] I mean, your [01:04:06] M four might be able to handle it. Hey, not my, not my M two now all of a sudden, basically obsolete. Let's see questions. Midjourneys, mood board feature. [01:04:15] It's going to let you generate mood boards. Well, I think that's what we were speculating. We don't know a lot about that. That's that was my interpretation of it, but Rory. [01:04:24] Brought up something and I think the last episode where he had a totally different idea as to what that could mean. So [01:04:30] it remains to be seen. Not really sure.

Play from 1:04:36 8:88:88 1:04:36 [01:04:33] Yeah. I wonder if it's going to be like, uh, I wonder if, you know, like this, you see this style, like the style reference code, like you see everyone posting it in the same style of mood board. [01:04:43] I wonder if it's going to be [01:04:45] something like that too. You know, like I wonder if it's going to be like, Oh, You know, here's just one image. You know, I don't know what that looks like. Like I said, I don't know if it's going to be a different set of the different, like UX UI, even for that as well. But it's a, [01:05:00] you know, I wonder if there's so many ways that you can conceptualize it, like even like an auto tune.

Play from 1:05:08 8:88:88 1:05:08 [01:05:05] Like color aesthetic and color grading kind of thing. Right? Like that would be if you had two images and they're not necessarily matched, then it's just like being match them, like in the editor feature. Like that's, [01:05:15] that's something I think for like storyboarding, you know, I don't know how they're going to construct it. [01:05:18] There's so many ways you could do it, but I, that one thing with Luma. I do really like, which is the brainstorm. That's the one thing that's to me was a really nice differentiator was like, Oh, this is [01:05:30] cool. I see where you're going.

Play from 1:05:35 8:88:88 1:05:35 How about you go this direction? And I was like, that's, you know, it gives you a couple options. [01:05:34] I'm like, that's cool. I like that. Um, it's making prompting a little bit more interactive. That's when it starts to feel like the ChatGPT side of things where like has a brain instead of, it's just [01:05:45] an execution tool. Yeah. So that's the, uh, the, yeah, I'm, [01:05:49] you know what, now that we talk about that, I, I am kind of surprised no one's gone that route yet. [01:05:53] Yeah. Isn't that interesting? Dolly had it. I mean, it was just like, Grok, you can do it on Grok on [01:06:00] Twitter, um, but it's never like the same. It's never like, oh, like keep this image the same, but like put that guy's arm up in the air. Like it just generates a whole new image kind of in the same fashion, but it's conversational.

Play from 1:06:14 8:88:88 1:06:14 [01:06:11] to get to that point. Um, I wonder if Sora's gonna be like [01:06:15] that in regards to, like, interactive with, like, the video generations. Someone also made a great point before, which is when we were going back, not to jump all over the place, might have been Moody, but they said, uh, they said, you know, Sora with 60 second [01:06:30] generations, the average shot length is 15 seconds in, like, you know, feature cinema. [01:06:34] Which I was gonna, it was great to highlight that, because I was gonna be like, no one's, no one's doing, like, these, you know, You know, three minute Martin Scorsese shots walking through a crowd anymore. Like you don't see that that often in cinema. So [01:06:45] I'm like, they really only do need to be 15, 20, 30 seconds.

Play from 1:06:52 8:88:88 1:06:52 [01:06:48] I mean, there are some that are going to be longer, but 60 seconds to me feels like overkill that starts to feel like video game more so than actual, like usable shot type, but I just wanted to hit on that because [01:07:00] someone, someone made that point before [01:07:01] Meg balls. Got a nice little zinger. Right. Oh, Hey [01:07:03] Meg. [01:07:06] We have Majority doesn't release so many giant super chill hours. Wow. Burn. I see Meg's getting a little upset too. [01:07:15] Hey, Meg, you missed all, you missed us complaining about this for, you know, a good 10 minutes. It's just like, what are we doing? Like, what are we doing? Lisa [01:07:24] says, I think you can train new personalization codes by using a mood board, meaning a collection of images that have a [01:07:30] certain style.

Play from 1:07:35 8:88:88 1:07:35 [01:07:33] I like that. I like that a lot. If that's the way that it's going to go, that was, yeah, she's, but [01:07:39] I don't think she, I think she's saying that exists. Not, not that she's speculating that's coming, I guess. I mean, let [01:07:45] us know. Um, because if you're saying that exists now, not really, right? Because essentially what you're going to do is anytime you rate new images, right? [01:07:54] And I don't know what the threshold is. You're going to get an updated personalization code. [01:08:00] That's all. That's all that is right now. Okay. Right. Until they come up with whatever changes right now. So, [01:08:05] uh, [01:08:06] yeah, [01:08:08] I would, I mean, I would like to, if it was like train your own personalization code and you just drag and drop like 10 images into a, you know, 10 [01:08:15] images into a folder and bang, it generates a code

Play from 1:08:21 8:88:88 1:08:21 [01:08:17] or let us pick the codes or name the codes. [01:08:20] I think we've mentioned that before. Also, Midgerty did say this super interesting. If you're into style codes, right? Which I think a lot of people are, they did say that they're going to do a [01:08:30] catalog or a library and start to make that much easier for people to go ahead and find things, right? Because as it exists right now, you guys know this, but in order to find codes, right, you got to have somebody that's either posting it.

Play from 1:08:46 8:88:88 1:08:46 [01:08:43] So they're sharing it somewhere or you got to unlock [01:08:45] them yourself. There are a few. You know, third party library sites that have unlocked a bunch, right. And they just kind of put them up and they've got some filters and ways that you can kind of go through there and choose and find something. But yeah, Midjourney said that, uh, that's going to [01:09:00] change here shortly. [01:09:00] They're going to make sure that it's a lot more easy to understand what exists, what doesn't. But, um, yeah, just something to note. [01:09:07] There will be another one in here real quick. The found Photoshop has been Generate in and out paintings [01:09:15] hands down the best essentially like I I still I I agree with you there moody that one I still use Photoshop because it's down to the pixel, right?

Play from 1:09:27 8:88:88 1:09:27 [01:09:23] Like you can get down to the pixel and generate that you can also The only problem I have with it is that it's coherent still sucks. [01:09:30] Yeah, it can't it doesn't like necessarily match aesthetic Well from some of the different generators if you're doing like something that's super clean and photo realistic Yeah, but then you might have to also like repolish it a little bit and Yeah, that's the one thing that I liked with Midjourney's editor is the expand feature like when you [01:09:45] can, you know, really expand and it matches the style extremely well, um, it's good, but then like the finer tuning details is not not easy, like in generative fill within Photoshop still great and speaking of Photoshop, just still waiting on all the [01:10:00] announcements from Adobe Max to come to reality.

Play from 1:10:05 8:88:88 1:10:05 [01:10:01] I don't know how long That's going to take, but we'd love to see it soon. Adobe, you get us all excited. I don't do anything for months, you know? Lisa, [01:10:13] I see her. I see, I see what you're saying. Yeah. [01:10:15] We, so we, we sort of talked about that at the very beginning. We were talking about with the personalization codes, how they're going to change that you're going to have multiple personalization codes, um, for different styles and so forth. [01:10:25] But, um, unless you've heard something different, like they haven't really given us a [01:10:30] real. Clear sense of what that is. Like, I know we're talking about different styles, but there's no indication of how that's chosen or how that's controlled or, you know, the, the details about the thresholds of what you need for each of these style codes that happen on a mat, you know, [01:10:45] like we don't really know much.

Play from 1:10:51 8:88:88 1:10:51 [01:10:47] So yeah, that's, that's interesting, but yeah, it's supposed to be coming out, um, before end of year. So, you know, I just don't get my hopes up any, you know, just like if it happens, it happens. [01:10:58] Just bring back preferred option set. [01:11:00] I mean, it's really what it was. I think [01:11:01] they did. Didn't they? Didn't they bring that back [01:11:04] for the website where you can just install, I can create my own, like, I thought I heard dash dash. [01:11:09] I [01:11:09] could just be dreaming stuff. I don't use that. [01:11:12] Cuz I would like to rename my [01:11:15] personalization code so I could actually remember them instead of like BL2AZR35 sounds like I'm, you know, under centered at a football game. Or just keep it the same [01:11:25] so I can [01:11:25] get [01:11:25] it tattooed on my arm or something, you know what I mean?

Play from 1:11:32 8:88:88 1:11:32 [01:11:28] Yeah, [01:11:30] the, that would be, so I think, I think they said you can create these codes by using mood boards. Like that to me is great. That's a great idea. Cause that's essentially Laura training, you know, like it's very similar to that or like, it makes sense. That's how they do it. Um, cause that, that's [01:11:45] easy. [01:11:45] Then it's a trigger word, right? Like then basically in the same idea. So if they do that, that'd be, that'd be useful. I still wish I had a reverse image search feature in there. I would love to find, you know, some other images. I don't need it guys. I don't need it.

Play from 1:12:02 8:88:88 1:12:02 Midjourney. It's just like [01:12:00] I have so many old images that are, I cannot find in there. [01:12:02] Cause I, I, [01:12:03] Oh my God, dude, it's brutal. I don't know what I prompted. [01:12:07] I don't know where it's like a year and a half old. There's no way to go. Can't even go back. Can't even go back. So [01:12:15] that would be one thing. Do [01:12:16] you remember though, back in the day, how those screens didn't load up so fast? Like when you first launched them? [01:12:22] Oh my God. But [01:12:25] when we had to go to Midjourneys, other website. com when you were still on [01:12:29] [01:12:30] discord, just to see the archive. Oh yeah.

Play from 1:12:35 8:88:88 1:12:35 [01:12:33] That was, that was the, that was the real days. What are the, what did Juan say here? Hot take, but the treasure hunt for SRFs is. Way more interesting than having them around for easy.[01:12:45] [01:12:45] You know what I tend to agree there too, that little gamification aspect of it. It's one thing that's totally different, right? Like it, it is like totally different than every other platform. [01:12:55] I think that's a really great point. One. Yeah. I'm glad you said that. [01:13:00] Um, let's see, Angela, is there a way to look up all escrow, uh, all S ref codes, like you look up P codes that you have generated?

1:13:01Midjourney technical tips

Play from 1:13:12 8:88:88 1:13:12 [01:13:08] Yeah. I mean, if you're asking what I think you're asking, right. You could just take. dash dash s ref and [01:13:15] whatever the number is put that into your search bar, you know, and then you can view everything that you generate using that that s ref. [01:13:23] Angela, there's some there's some tricks I have for this in terms of like trying to hack the system because it's tough to [01:13:30] find all the s refs. [01:13:31] So like, Either one I'll do, I'll create like a smart folder within the, uh, the organization tab and like the, the, I'll use like one token for like mining [01:13:45] S refs, right? Like maybe my, my prompt is like angry, angry samurai dash dash, you know, S referendum, right? So I'll make that, I'll only do angry samurai when I'm generating S ref codes.

Play from 1:14:02 8:88:88 1:14:02 [01:13:58] And then, you know, by [01:14:00] having the smart filter. On the, on the, uh, on the folder, like everything will just filter in there. So it keeps all the S refs, um, organized. That's one way to do it. The other way is to go into, um, [01:14:11] no, don't say rooms. Cause no one's using room. [01:14:14] I know no [01:14:15] one uses them, but I use them as ref mining. [01:14:18] So you can go in there and just do your mining, your S ref mining in there. So that it's always like in an organized place, instead of getting lost in the create. Um, you know, it'll all get funneled into that fold into those folders anyway. And then, you know, if you [01:14:30] find one. That you like, you can utilize those SF codes in the smart folders too.

Play from 1:14:38 8:88:88 1:14:38 [01:14:34] Yeah. So like if you're generating it with a lot of SF codes, you can use a smart folder and just make it the SF code and everything will just filter into there so you can organize it. I [01:14:41] think that's, that's probably, that's probably the best way right now. Like, yeah, I, I do the same thing. I use the [01:14:45] same prompt every time. [01:14:46] I'm mining. Just mining specifically, right? Mm-hmm . Which means unlocking just doing the SF random. And then once I sort of can tell what style it is and if I want to go further, then I have like sort of a short [01:15:00] list. A very simple but varying prompts that I'll run to further explore it. And then, um, I usually like SW two 50 for exploring SRF personal preference, stronger than a [01:15:15] hundred.

Play from 1:15:19 8:88:88 1:15:19 [01:15:15] It's not all the way a thousand. It's not 500. So you get a real nice balance of what you're prompting for. Versus what maybe is infused or baked into the style code, right? Because some style codes could be just focused on like creating teddy bears, [01:15:30] right? Um, if I do that at SW. You know 100 i'm not i'm not gonna really know that but if I do it at like a 250 I might start to see that right? [01:15:38] So, [01:15:38] uh, [01:15:39] that's why I that's how I sort of balance that but [01:15:42] I need to get back into [01:15:45] heavy parameter Variation that's what i've been lacking on lately. Like just I haven't been what do you mean like permutation [01:15:53] stuff? Yeah [01:15:54] Like going higher. I haven't gone much higher recently. My personalization code is, is, is, is fairly [01:16:00] graphic.

Play from 1:16:04 8:88:88 1:16:04 [01:16:01] So like, I'm not going high [01:16:02] either though. I mean, we've talked about this, but like, remember when it dropped, we were going high because we wanted to see how, how the style, like our personalization was influencing the prompts, but [01:16:15] then like over time, it just seemed like, okay, the lower numbers really dialing it in better. [01:16:20] Um, I've been using whatever [01:16:21] reason with personalized a lot. I've been using dash dash S like 75 or stylized 75 through 90. I feel like it [01:16:30] gives me. majority of my personalization style, but still a little bit of the creativity for Midjourneys based model, which I like. It's like a nice little blending of it.

Play from 1:16:42 8:88:88 1:16:42 [01:16:38] So going under the default there, but do you have this where you go all the way? Have you tried going all the way up? Like recently, a lot [01:16:45] on your style or your personalization code. Mine gets super dark, like real quick around your dark 600. I guess so. I don't prompt anything that's, you know, horror or anything like that, but mine crazy. [01:16:57] Real quick. Really? [01:17:00] kind of creepy. Yeah. Like I was running it the other day. Really? I kept doing it. I was like, why is, why is so dark Midjourney? Damn. I don't, I don't do anything along those. I don't prompt for any of that stuff. You're a sketchy guy. I am though. So it probably knows me. [01:17:15] Damn. It took my personality test and was like, you're, you're secretly dark.

Play from 1:17:21 8:88:88 1:17:21 [01:17:17] Yeah. Like I haven't really been doing it consistently enough to really know. Like I've, I've done it here and there, you know, but when you do it, you know, just here and there, you don't really have a sense. I've just been running. Let's [01:17:30] say my default really around like two, 300, um, 300 a lot recently, actually. [01:17:36] Um, you know, I used to be like a 200 guy. Well, I wasn't, I was a one 50 guy for a long time. Then I was two and I was two 50. Now I'm three, [01:17:45] [01:17:45] three on your personalized [01:17:48] three. I'm a run [01:17:49] personalized on everything. I, [01:17:50] it's, it's my default. I don't not run personalized. Yeah.

Play from 1:17:57 8:88:88 1:17:57 [01:17:54] Is there, I'm probably missing this in terms of permutation. [01:17:57] I forget. I feel like I used to do this on [01:18:00] discord where I could run. There was like, if I wanted to do personalize, right? Like if I wanted do personalize and non-personalized in the same prompt with the permutation, there was like, it was like something, it was like, maybe it was like colon, colon dash ashp or something like that within the [01:18:15] parameter set. [01:18:15] I forget. But you can do that. 'cause I remember you could run both. I don't know. I don't think so. You used to be able to do something very similar on discord with that, but I mean, with personalization. Yeah, and it was, but how would you even [01:18:27] set that up with, as a, as a parameter though?

Play from 1:18:33 8:88:88 1:18:33 [01:18:29] [01:18:30] It was something like. [01:18:31] You could run like as a, as a permutation, it would be like one value would be like with the permutation or with the parameter, one value would be like without it. So if that means like, I'm trying to, there's a way to do it. I know I usually, well, because you [01:18:44] [01:18:45] know, if you don't, if you don't put in the personalized, [01:18:47] yeah, [01:18:48] you could just put dash dash personalized. [01:18:49] It's going to automatically put your code in there. So the, what I'm thinking is, is that even possible? Cause I think you would need another value to offset it.

Play from 1:19:00 8:88:88 1:19:00 [01:18:57] Well, you can do P zero and, or, uh, Or, [01:19:00] uh, S0 and S100. There should be, that, that, that's it. So it's S, S0. Yeah, you can style it, yeah. Yeah, that's how to run it in the, in the, That was really hard to get to the answer there, guys, [01:19:15] sorry. [01:19:16] I knew it was, Thinking about it in the way that I used to do it. But then it's, [01:19:19] but then it's not really taking advantage of the Midjourney stylization in that sense.

Play from 1:19:27 8:88:88 1:19:27 [01:19:24] Well, it drops your personalization all the way down to zero. So it's not using it at all. Right. So if you're taking [01:19:30] the style, because you're doing S you know, if you're doing running permutations, dash S as zero would be without personalization and then it'd be as 100 with your personalization, 200, 300, 400, 500, but [01:19:43] how could you run stylized without the [01:19:45] personalization outside of zero, I guess, [01:19:47] but if you Wait, [01:19:50] dude, math is hard.

Play from 1:19:56 8:88:88 1:19:56 [01:19:52] Okay. [01:19:52] No. So yeah, right. But stylized stylization controls like the entirety of your personalization. Right? So right. [01:20:00] If it's if 100 is default, and that's like standard and zero has like no influence of personalization. So then wouldn't it just be without personalization? So what's the [01:20:09] default? So [01:20:11] that's 100. [01:20:12] Right? That's it's already at bro, we [01:20:15] might [01:20:15] have to figure that out. See, that's gonna journey. [01:20:17] Listen, listen to this one.

Play from 1:20:22 8:88:88 1:20:22 [01:20:19] Right? That's, here's a, here's a confusing These if then scenarios, come on. [01:20:23] Why couldn't it just, why do they have to use stylized for this? Why couldn't it just be, you know, PW like P yeah, yeah, yeah, yeah. [01:20:29] I would have been [01:20:30] like everything else. Um, does it make sense to me? But that's a, that's another one for another time, but [01:20:37] another one for another time, I think, are, are we, are we wrapping this, we wrapping this thing up? I think, [01:20:43] do we have any other questions [01:20:45] here? I don't see anything else up. Anything else?

Play from 1:20:52 8:88:88 1:20:52 [01:20:48] No. Scrolling back. It doesn't look like it. Anyone else? Bueller. Bueller. Heh. Going once. [01:21:00] Going twice. All right, [01:21:01] so this will be two if like if you caught part of this or you want to you know Watch it again or anything like that So we're gonna this is going to be the plan Anyway, this is going to be the podcast that is released sunday morning when we [01:21:15] always release podcasts.

Play from 1:21:20 8:88:88 1:21:20 [01:21:16] So check in then we'll have that up in live also also obviously I think if it is up on you know, LinkedIn or you know, you want to go on x or something like you could just watch it back. That's also a possibility Yeah, guys. Thanks for joining us. Thanks for joining [01:21:30] us. Last minute pop up Midjourney Fast Hours podcast. [01:21:35] Yeah on the on the fly MJ FH so it no awesome We appreciate you guys coming to you guys a little bit asking questions [01:21:45] interacting good stuff. Yeah, man Moody I don't I [01:21:48] don't know you man. I don't know where you're man or woman I don't I don't know who you are, but thank you for showing up. You had some great stuff Thanks for adding into the episode guys.

Play from 1:22:03 8:88:88 1:22:03 [01:21:57] We'll see you next week in the [01:22:00] next one [01:22:00] like and subscribe. Thank you. See ya