Ep. 64: Midjourney V8 First Impressions + Claude Code Breakthroughs

0:00:00Midjourney V8 finally lands

Play from 0:00:01 8:88:88 0:00:01 What's up everybody Midjourney Fast Hours podcast episode 64 back from a long vacation and did we get something? Did we finally get Midjourney V8? Yes, we did. Yes, we did. Although maybe not the full extent of what we originally thought, right? I mean, I leading up to this. There was a lot of hype with V8, right?

Play from 0:00:27 8:88:88 0:00:27 We thought we were going to get this. There was maybe the chance in December. then it was January, then it was February, then it was halfway through March and boom, here it is. So we got it and we're gonna get into this. think that's where we start, We're gonna share our first impressions with it. I know a lot of people have been knocking on our door. Are you guys gonna do a pod on this or what? We're like, yeah, we are, we are, we are.

Play from 0:00:53 8:88:88 0:00:53 There's a lot of first impressions I think you and I both have, a lot of observations. and kind of sentiment with where things currently stand. So we'll jump into that. mean, I guess let's like, let's kick it off with, em well, let's kick it off with the fact that you were traveling when this dropped first of all, which never fails. One of us always, but when you finally got into it, give me, what did you do first?

Play from 0:01:24 8:88:88 0:01:24 What did you do first and what were your observations? I just re-ran the last prompt that I had in there and just was like, all right, let's see the difference. And it was, you know, I was traveling. So I apologize to everyone who's been asking if we're, you know, like if we stopped this thing, I was just in a run of three weeks of absolute psychotic travel, not to mention the TSA and LaGuardia. I got trapped in that. I got trapped in the LaGuardia plane crash situation.

0:01:25Rory’s travel chaos and failed gens

Play from 0:01:51 8:88:88 0:01:51 I wasn't in that plane, but I was traveling within 12 hours of that plane crash. So it was uh a sidebar before we get into this just to just to laugh a little bit. Right. So got back from Miami on Sunday was had to go basically back and drive my daughter to my parents on Monday, so that she could my wife and it was like my wife was in one place I was in one place the dog was in one place the baby's in another place.

Play from 0:02:17 8:88:88 0:02:17 Right. So I had to drive to New Jersey and back like three hour round trip on Monday to then take my flight. on Tuesday, which was right after the plane crash at the same airport. But my flight was at 6am. So I had to get there early. So I had to wake up at 330 in the morning, and then I had to drive to the airport and park there because I was going to need my car on the way back to go pick up the baby. Right. So get there at 4am. 800 people in line at TSA. I'm like, you got to be kidding me.

Play from 0:02:44 8:88:88 0:02:44 I got on the plane. No problem. Like, I mean, it took forever. But it's 4am. And I'm just sitting there like I can't believe this is a reality. Go to, you know, go to Boston for the two days that I'm with with shark ninja there. And then like on the second day, I'm like, all right, cool. Like, you know, I'm supposed to just like leave at eight o'clock at night, get home, get my car, drive to my parents, go to sleep. They cancel the flight. So I'm like, oh, great. Here we go. They booked me on the next day. I'm like, no, no, no, it's not going to happen.

Play from 0:03:11 8:88:88 0:03:11 I need to get home tonight. Like I can't be away for this long. I need to go home. So I really I like go through hoops to get on a flight to to JFK, which is the other airport in New York for anyone who doesn't know. And it's like, it's way further away. LaGuardia is 15 minutes away from me. JFK is like 50 minutes away. Right. So I fly there to one airport, then I have to Uber to another to the other airport to get my car to then drive to get my daughter in New Jersey all in one shot.

Play from 0:03:43 8:88:88 0:03:43 So I got there like 1230 at night, woke up, drove back here sat in two and a half hours of traffic to get back yesterday. like gridlock alert, insane to go. My parents live 15 miles away takes I don't know what the equivalent of that is in kilometers, maybe like 35 or something like that. It took two and a half hours to get here. So I have just been wasting time in TSA and airports and cars and traffic.

Play from 0:04:11 8:88:88 0:04:11 Long story short, it's just like we're back. We're back. um Like just all over the place. But That being said, first shot I got was on the plane and I don't know if it was my internet connection on the plane, but Delta's Wi-Fi is normally pretty good with that. But like nothing worked. It's like, let me try my mood boards.

Play from 0:04:36 8:88:88 0:04:36 Didn't work. It's like, let me try my profile. Didn't work. was just getting like generation fail, generation fail, So then I waited. And then when I got to the hotel. I started playing with it a little bit and I think it was probably the day of or the second day that it was out and it just wasn't working. So I just gave up for a few days until I was like, this will iron itself out. But I don't know. Did you have that first that first two, two, three day issue?

Play from 0:05:04 8:88:88 0:05:04 It was just generation failed after generation failed. Like nothing worked. run into that specifically, but I think where I was at with it, and I think maybe a lot of our audience is with it, is I didn't rush to try it out immediately, like I have with every other release in the past. Usually it's like you're opening a Christmas present, like you know something just like Amazon packages came to your door, they've been waiting for.

0:05:06First impressions and alpha letdown

Play from 0:05:32 8:88:88 0:05:32 ah But I think that speaks to the fact that a lot of us have kind of like, the workflows change. We've got these other tools now that have made their way into workflows. And uh we've been waiting for this for so long. So we've kind of moved, right? And I think by the time I did get in there, think first things first, um I'm disappointed.

Play from 0:05:59 8:88:88 0:05:59 And I think it's OK to say this. I love mid-journey, always will, no matter what. both of us just so many things. uh Obviously we've even got this podcast, we started just for it, right? um But I think like just as it got closer to launch, just, the excitement really started to just sort of like dissipate a little bit, because it's just like, oh, we're probably not gonna have this ready for the launch.

Play from 0:06:28 8:88:88 0:06:28 And you could just kind of sense it with David and when they were talking about it, just like the excitement that they had. one month prior wasn't that excitement that they had days, weeks leading up to the release. And that really just kind of like set it for me, you know, just kind of subconsciously of, OK, like they're not that excited that I can't be that excited because I've kind of, we've we've we've gone through this so many times.

Play from 0:06:54 8:88:88 0:06:54 Right. I think like we can kind of get this barometer of what they say and how that's actually going to translate now. um But. I was just disappointed that it's on an alpha site. I know we're still working through it, right? But ultimately it's just like this model, you know, last time we got a model was V7. That came out in April, 2025, you know, we're in March, 2026, you know what I mean? And we're getting an alpha version of this.

Play from 0:07:22 8:88:88 0:07:22 You know, it's just like, damn, you know? And so of course, right? Like in this competing environment as well, you're getting a bunch of people that'll go in. especially the non-Power users, but they use it and then they're disappointed immediately because it's not fully functioning the way that they expected or would expect from the standpoint of, it's a new model, should be better, will be better. um And that's just the reality.

Play from 0:07:49 8:88:88 0:07:49 I don't think I'm saying anything that's shocking or jarring in any way here, but I think that was the first thing. It's just like, it's on an alpha site, now I got to play with this on alpha. And then you've got sort of like, let's open up. kind of like what we're looking at here. I'll kind of get into maybe some of the nitty gritty of what we're talking about with, you know, kind of like what we're seeing too. You know, this whole UI setup, this navigation, I don't know.

Play from 0:08:19 8:88:88 0:08:19 We were just talking about this before we hit record, but it's like, is this the UI that they've been talking about? Is it just like shifting things over kind of? I would hope not because I don't really feel, this doesn't feel better to me. in any way. And in fact, like the one thing that really is killing me, which I know is an easy fix, but it's just like, how can I quickly adjust my profile? Where? Where is that? I gotta go like here it's on.

Play from 0:08:48 8:88:88 0:08:48 So I gotta go all the way here and then go and select. right next to where your settings button is that's highlighting red top right. Click the thing to the left of it, whatever that is. Yeah. And now you can select profiles, mood boards, style references, like styles. I liked this piece. Like this was better to me than having the dropdown, the profile dropdown, and then always having to like, oh wait, is it clicked on or off?

Play from 0:09:21 8:88:88 0:09:21 was right answer either, to be fair. I just, this, it doesn't, I don't know, just doesn't do anything for me. um Like maybe that speaks to it. You know, I've been in here a few times. I didn't see it in here. Didn't see it here. I did click this, but maybe I just didn't browse long enough. And then it was just like, oh, okay. I guess I'm just using this right now. this, yeah, it's just, I like that you have this here, but.

Play from 0:09:50 8:88:88 0:09:50 I don't know, that was one thing that stuck out to me immediately. There's, you know, I think... the model itself feels very volatile, right? So like I went through and ranked, you know, a bunch of these two, like right when they did the ranking party, that was kind of this one here that I started, you know, kind of like curating.

Play from 0:10:16 8:88:88 0:10:16 And then I kind of even went back to this once I saw that this was sort of like they're combining this global into the V8. I was thinking we were gonna get a new global, maybe we will once it's out of alpha. But then I started kind of like going through here. and ranking things again. And I did notice that it was able to kind of like eliminate some of that bluish pink hue that I was getting. So this is actually usable again now, but um I'm noticing right like with the new personalization that I created once the ranking party came out, like that final version where you're starting to see real V8 images.

Play from 0:10:52 8:88:88 0:10:52 When I use this, the images are a little bit sharper. They look like V8 compared to. maybe something like what I was using before like with this one. uh So just an observation. I think that was like one of the things that I noticed. um Let's just kind of take it. You know, they talked about, you know, maybe doing a new explore page. I don't know if that's still in the works. You've got this, you know, randomized, which is cool.

Play from 0:11:18 8:88:88 0:11:18 Okay. And same, I mean, everything else is pretty much the same. So right, V8, new coherence. I think that's the thing we're really trying to figure out here is, you know, sort of the coherence of it all. Do you have any thoughts on that, coherent specifically? because I've found it to be, it's more coherent. mean, I'll give them.

0:11:30V8 coherence and vibey prompts

Play from 0:11:45 8:88:88 0:11:45 coherent. To be fair, I don't know that I fully stress tested it enough as I have in the past. But I think like to me so far, it's not worse than it was. I can't say it's objectively better though either at this point. I'm kind of in the middle on it right now. But yeah, I mean, so this point right here, right? V8 prompts differently to V7, short and vibey no longer gets you there.

Play from 0:12:10 8:88:88 0:12:10 uh that was basically I didn't even see that in the notes, but that's sort of what my first like thought was, was that when I was trying the short and vibey stuff, like it wasn't doing what it normally does, which is weird because that's why I came to like it over the last, you know, for for like the the difference from the other models being like, I can still do short and vibey, but I can also do control.

Play from 0:12:35 8:88:88 0:12:35 Like if I just want to like, you know. type in the work is dark and the work is mysterious and important and I get something like banger. Like that's good. But then I also want to be able to you know, broken out JSON prompting. yeah, like maybe like that vibey, you know, uh less is more exploratory prompting still really works well from like a mood board perspective, right?

Play from 0:13:00 8:88:88 0:13:00 Where it's just like, you've got sort of almost like a container of an aesthetic or style that you're aiming for to already operate within. And so the vibiness of that. kind of creates a little bit of room for the creativity uh that is constrained through the mood board. So it's like, okay, if that's still the case, then that could be cool. ah This one was interesting too. SRF and mood boards have a new version, SVSAT.

Play from 0:13:29 8:88:88 0:13:29 Huh. Okay. Did I have... Yeah. SV7... Is SV7 the default? I don't know if it's the default. I don't know. um I wonder if we could get that information. Like if we just ran something, it would tell us. Probably not, I don't think it ever did before. We had to like kind of put it in there.

0:13:30SV6 vs SV7 mood board test

Play from 0:13:55 8:88:88 0:13:55 just do one real quick with the mood board and do bracketed SV7 and SV6 and see how different it is. Yeah, let's just do do do do. Nuns with guns. I saw that one on there. let's do and see what we used to do was stills archive hit enter banger.

Play from 0:14:30 8:88:88 0:14:30 Yeah, yeah, I love that. was man. That was one of the best things ever. So let's just put like. Unsettling character, this is sort of like the vibe of this we can do. Can I run parameters with SVs? could before. It's worth a try. Let's see what the difference is.

Play from 0:14:57 8:88:88 0:14:57 can't type, that's cool. Yeah, let's do this. Let's run this three times. Let's get a good sense of what's going on here. So let's see, we SV7, SV6. It is running them. That's cool. Okay. There does look to be a difference here. One looks, SV7 looks a little sharper.

Play from 0:15:23 8:88:88 0:15:23 I mean, look at that. That's interesting. Try a different mood board real quick. Try something with color. I'm curious about the colors. Yeah, let's do it. Let's take this off. Let's do... because that was one thing I noticed that the coherence thing, right? Like that was just a sidebar while you're while you're typing this out was I was using a mood board that's predominantly green and black.

Play from 0:15:52 8:88:88 0:15:52 And typically it's such a strong mood board that it overrides everything I put in my prompt. And I kept hitting it. I'm like, why is this coming out blue? Because I'm just using an old prompt. I'm like, it should not be blue. And then I looked in there like I mentioned something. I mentioned blue as like because it was an old prompt and It was defaulting the aesthetic to blue. like, that's crazy. That never used to happen. um Where it would just actually really listen like that and the mood board wouldn't override it.

Play from 0:16:20 8:88:88 0:16:20 But that was one thing it didn't know. So I'm like, it's way more. Then I started playing with stuff like that with little intricate things. And it was definitely listening to the little intricate things, but in a way that I hadn't really noticed before, if that makes sense. um some stuff. um Try the work is try the work is mysterious and important.

Play from 0:16:45 8:88:88 0:16:45 That should give us something. The other thing I noticed while we're here is that the stylize, I haven't been seeing a massive jump from like 100 to a thousand. Like the difference to me between 200 and 800 isn't as extreme as it used to be. But so what do we got here?

Play from 0:17:11 8:88:88 0:17:11 This one. it. So these three would be... SV6, these three would be SV7. There's more depth, there's more contrast, dark, the blacks are darker. Is it near, is it a little bit more unstable though? Like the other ones look a little bit cleaner. see.

Play from 0:17:39 8:88:88 0:17:39 So like that one, for example, go back to one more, go back one more. Or maybe that wasn't that one, like the stairs, like this sort of like, yeah. Yeah, looks like a V34. That's what I've been noticing with mood boards specifically. I should say that is that the mood boards seem a little bit unstable. That might be something that they're ironing out. um Like the Osiris V6, right? Like that's to me, that's looking a little bit more like composed, less, less weird hallucination in that.

Play from 0:18:13 8:88:88 0:18:13 I mean, we're asking, you know, this is a lot, but regardless, it's a cool mood board, by the way. mood board. Maybe let's try one more. Let's try one more. I think it'd be good to experiment, try one more here. So let's go here. Let's try something that feels newer. How about this? All right, so let's just put, uh let's do the same thing.

Play from 0:18:44 8:88:88 0:18:44 And then I think let's also do something that is less ambiguous and dreamy and more matter of fact. Let's do... Yeah. What you got? Like... Alright. Yeah, let's just do it. All right.

Play from 0:19:10 8:88:88 0:19:10 Oh, jeez. I mean, I think we're safe because we're on mine, but you know, here we go. So let's do this. Let's run. twice. What else am I missing anything here before we run it? No. Let's do it. think so. Do we do, were we trying to test SV there? Oh yes, there we go.

Play from 0:19:35 8:88:88 0:19:35 See? Duh. All right, here we go. SV67. Rock and roll. Alright, let's see. So that's SV7. Now these are all kind of like a little bit of motion, right? Like a little bit more lo-fi anyway. So just take that into consideration.

Play from 0:20:04 8:88:88 0:20:04 All right, so let's see here. just doesn't look super usable, right? Like most of these have issues. That might be the best. And frankly, that might be the best one. Like that one actually is solid. uh mid-journey making going right in on it, huh?

Play from 0:20:31 8:88:88 0:20:31 But some of these do lack, right? then SV6 still doesn't, it doesn't look much better. I mean, so the mood boards thing, this is real. I mean, this was really working well, not on V8. So let's take that back. Let's run the same thing. Why does this keep kind of adding to me?

Play from 0:20:56 8:88:88 0:20:56 and I wonder if it'll let me run seven, might not. just... V7, there you go. What has your pro... if it has your profile attached to it. See, this is the weird stuff that I'm noticing also is like the repetitive adding of profiles back in after I get rid of them. All right, so this will run, we'll do SV7, we'll get the SV6, but you just like to compare what these look like, I mean, these are gonna be much better.

Play from 0:21:26 8:88:88 0:21:26 These will be 10 times better. uh And you'll see it right away. I mean, usable, usable. eh not bad. Usable, usable, usable. You know, like, so. Yeah, the mood board degradation here. I think I heard something about that on office hours to be fair. And I do have office hours notes up.

0:21:44Mood board degradation gets real

Play from 0:21:53 8:88:88 0:21:53 Problem is they're like, it's deep. We didn't go through it completely ahead of time. But I do know there's a lot going on internally about, I think they're hearing it. V8, I think there's some disappointment from people, their users obviously. Also, I think... Probably some of the team is probably a little disappointed kind of the way that things were released now, but so it sounds like.

Play from 0:22:20 8:88:88 0:22:20 board here? Try adding another mood board. I want to see if it's just me. Yeah, try doing multiple mood boards. And I don't know if they said this in office hours or not, but I was getting it. This is where I was failing every single time was adding two mood boards or more or adding a mood board and a style reference or like all the fun things that I like to do in Midjourney. mix in this black and white version. This is kind of like a similar one to this except black and white.

Play from 0:22:48 8:88:88 0:22:48 So we can kind of mix that if we want. What are you wanting to see here? I just want to see if it generates for you, because I get failed creation for everything. And this might be in the Office Notes, but... this on V7 so we can kind of get a baseline again, just like we did. So let's run that. Then I'm gonna do this with what we were doing before with six and seven on V8. So that way we get the full comparison here.

Play from 0:23:14 8:88:88 0:23:14 Okay. Yeah, see, mean, like annoying, annoying, annoying. And it's not actually running with it. You it is. Ridiculous. why can't I do anything sometimes? And see now I just, I clicked out of them, submitted it, it ran without them, but now they were added back here again.

Play from 0:23:43 8:88:88 0:23:43 So. the... I don't know. I mean, it's just confusing to me. But like, again, I can't use multiple mood boards. And I don't know what that's about. No. Like, everything fails. Yeah. When's the last time you tried that? Oh, what the hell. Interesting. let's let's just pull it up. I'll just do mine real quick. Just so before we get into that, just just show you that I'm not crazy here.

Play from 0:24:11 8:88:88 0:24:11 It'll probably work because I you know, we're doing this now. But Here, let me show this real quick and let's pivot to yours. So this, we ran the same experiment again. So again, ah at the top here, we're gonna work our way down. From the top here, this is V7 of these two mood boards combined. So let's go through these really quickly, right?

Play from 0:24:37 8:88:88 0:24:37 Solid, really like it. like we're getting really good stuff here, okay? And then let's just kind of like fast forward a little bit ahead. Now we're getting into the V8. This is SV7. not usable. It's like a totally different vibe. Like it's not, wow, 9-11, huh? Did you see that? Go back. That's the new World Trade Center.

Play from 0:25:02 8:88:88 0:25:02 criminal, geez. Yeah, like, you know, just not usable right now. thing is like some of like the skull one, like if you go back a couple, not that one, that one, like, is it, is it monumental? No. Is it, is it kind of cool and vibey? Yeah. But like, I get that. The other ones you were doing, I was seeing like every other one. I'm like, yeah, like that, like that, like that. And then once you get to.

Play from 0:25:29 8:88:88 0:25:29 And we spent a lot of time with those mood boards, you and me. So if I'm gonna use mood boards right now, I'm not gonna be using V8 until further notice, right? That's gotta be fixed. So, all right, I think that wraps that piece up. Let's shift over to you. Show me if you're able to do this now. Let's just kind of test it out. Because...

0:25:56V8 looks better, but fails too often

Play from 0:25:57 8:88:88 0:25:57 All right. So let's just say I have my, I'll turn my profile off. Yeah, this is just the work has missed. I ran this just base. Like I wanted to see the new model. I will give it credit. Like the one thing I'm noticing is that the V8 standard model with no personalization attached is definitely better aesthetically than seven. Yeah. tried much without the personalization.

Play from 0:26:24 8:88:88 0:26:24 It's got way more... That's the word I'm looking for here. It's got a lot more character to it. It's ah just more powerful. Granted, but the whole thing is build. Here's all my failed generations. This is what I'm talking about. There's a million. start going down. You just keep getting... Where is the... When I start getting into the...

Play from 0:26:50 8:88:88 0:26:50 Everything was failed. But again... going back to some of that stuff like V is this V8 yeah these are V8 but I was running some like V7 like this mood board right like this looks like I love this one with the intricate details and then V8's like this like it's not it's not picking up the same way as everything is sort of like funky but your double mood board real quick while you're here.

Play from 0:27:18 8:88:88 0:27:18 then, like, you know what, Rory, too? Like, OK, so so far, right, we've been talking about sort of like the downside of V8 so far. Now, there are good pieces of this, too, right? Like, I mean, we can show some of the things that we've created so far with V8. um And when it does get it right, it's noticeably better. It is noticeably better. I mean, there are just, you know, We'll kind of scroll through those things between you and me.

Play from 0:27:46 8:88:88 0:27:46 think there's a lot that we can kind of like pull out from an insightful standpoint of what we're seeing. I think the problem right now is the volatility with the model. um You know, you get a banger and then the next one is complete shit, right? um And so we've got to sort of close that gap. And then obviously like patch up these things like mood boards, right? Like where they're not only back at the level of what they were before, but I'd expect better, right?

Play from 0:28:17 8:88:88 0:28:17 So how long is that gonna take, you know? So why can't I use two mood boards? Like V8, obviously not running. Yeah. And now V7 runs. Like it does that. Right? Like, I don't get it. I don't understand why. And it's been, this has been, I don't think these mood boards are that crazy. Like, I don't know why they, are they hitting, you know, content moderation?

Play from 0:28:44 8:88:88 0:28:44 Like, I don't know. Let's try this. Let's try a different one. I'm wondering if this is like one of those scenarios. You remember last time, right, with when mood boards dropped, I was able to use two for whatever reason and other people weren't. um Maybe it's one of those situations, because I don't know what else it could be.

Play from 0:29:13 8:88:88 0:29:13 Let's see. Is it just some mood boards that are not working? it like... See, these are like the mental gymnastics. Like looks like this is gonna work now. does, it does look like it's gonna work now. I thinking, doesn't want to do it. Should I do it? Should I do it? Should I do it? Because we're live. Should I do it? Well you saw it, I've been here.

Play from 0:29:40 8:88:88 0:29:40 it's making me think it's not because it's taking. Oh, there we go. What happened then? So what's the deal? But look at the difference in these mood boards here, right? Like, this is... The mood boards here, one is all like crazy intricate faces. The other one's all wobby sobby. This is what it should look like with mood boards. This is what is coming out of VA. Like, if I went into here and showed you what's in the data set of this, this is what's in the data set.

Play from 0:30:10 8:88:88 0:30:10 It's not that different, right? Like, it's all like these crazy faces. And then the Wabi Sabi is just like basically exactly what you thought. Like just intricate looking people. Right. That's all that's in there. And now like why is this coming up? I don't know what's going on. whatever reason. I think it's like interpreting maybe the maximalism on the face and it's applying it elsewhere instead of just the face, right?

Play from 0:30:40 8:88:88 0:30:40 to me is like exactly what it should be. This is, you know, which one. thumbnail perspective, that looks super cool. That one that is down one and to the left. Yeah. little Adolf, you know? So it's what it is. This is cool. This is cool, but it's not what I had envisioned.

Play from 0:31:07 8:88:88 0:31:07 You know what mean? If I was trying to go for that. Like that one, I really liked that. I mean, I don't even want to say Hitler one, but it's a good shot that introduces some really cool elements into it. That's probably the best of the bunch. I think like, look, it's this is creative and I like it. You know, like I like I like this. But the thing is, what if what I was going for was not this, you know, so I'm saying like this is what I was looking for.

Play from 0:31:36 8:88:88 0:31:36 You know, like a little bit more like this, exactly what I was thinking in my head when I'm doing this. Right. Or this or this. Right. Like this is like something like that is what I'm looking for. But I guess you just have to figure out how to use mood boards again. Like, I don't know if that's if that's something that let's see. If I have to do like, know, SW. SW even do it? No, it's S, S like 900.

Play from 0:32:04 8:88:88 0:32:04 Yeah, run a, yeah, run a test on that too. Yeah, I mean, I just think mood boards are, I don't think they're broken, but they've gotta be fixed or patched here. They know that, I think. Like I said, I think that came up on office hours, but, so it's just some of your mood boards aren't working when combined then.

Play from 0:32:36 8:88:88 0:32:36 adding the ones that I don't want in there. Let's try this one more time. Alright, so it's working here, which is weird, man. I don't get it. Like, this is actually really cool, right? Like, I like this. Like, I like some of this stuff that it's pumping out. It's just not sort of functioning the way that...

Play from 0:33:02 8:88:88 0:33:02 I probably just have to be... It probably comes down to text prompting. I mean, that's it. Yeah. and I think this is definitely the case outside of mood boards, which kind of goes back to where this all started with us is you probably need to have some sort of prompt structure here. I think this has gotten, if this model's coherence is even slightly less vibey, which I think we all agree it is, ah I think you're gonna have to structure it.

0:33:04Structured prompting may matter more

Play from 0:33:32 8:88:88 0:33:32 Right? And that may be an interesting test for us. We could either do it here or just for folks that are listening is try prompting with mood boards with more of the simple prompt like we were doing versus something that has your, you know, camera shot, your photography style, right? Like almost kind of like that real structured, you know, specific approach, right? Where you're including the medium, the lighting, the environment, blah, blah, blah, blah, blah.

Play from 0:33:59 8:88:88 0:33:59 seeing if that changes things, but I think in general, that's the case. um I also need to do more of the testing between standard and raw on this one. could be, I mean, that change hasn't been really that drastic lately with the models, but perhaps with this one it could be. That's something worth testing. Yeah. Like that's the one thing that I'm... the other thing that I'm interested in here is that... let take this one off.

Play from 0:34:25 8:88:88 0:34:25 I'm just curious now. The... like see how this is just like overriding all of the... like this is what the design looks like, right? It's like a little bit extreme maximalist, whatever. Because I'm writing 85 millimeter close-up portrait of a war criminal. Like it's definitely overriding the mood boards there. Like this is way less extreme visually than these are. And what is your stylize on those two?

Play from 0:34:50 8:88:88 0:34:50 I mean, so that technically, what we had been told, right, stylize affects mood board strength when you're using a mood board. So yeah, if that's the case, that's interesting. The only my one other problem here is that like, that was actually legit. um with 85 mil, these should be super tacky sharp from a portrait perspective. Yeah.

Play from 0:35:16 8:88:88 0:35:16 It's like this one, this one a little bit better, right? Like, yeah. So the other thing that I'm noticing a lot of is that like, when I do this with. I know this isn't going to run for V7, so let me just do V8. HD, Q4. Let's see if it does it that way. Like, I've been running stuff on Q4 because it's getting, like, way sharper.

0:35:29Q4, HD, and stylize confusion

Play from 0:35:41 8:88:88 0:35:41 Hmm. Let's do this. Do I gotta get these? There you go, that's the biggest pain in the ass there is right now. So this one, I would say, I've been getting better stuff with Q4, like I still kind of oddly enough, as much as everyone complained about the V7 like upscalers, like I miss having it to clean up some of the stuff. It was like a retouch on certain things.

Play from 0:36:08 8:88:88 0:36:08 my system, because I was upscaling the ones that I sort of like the most, as that was sort of my way to filter down, right? So if I go to organize, I can just go to upscales, those are the images that I like the most. Now, with that not being an option, I've had to go back to favoriting and liking as guess the filter, you know? Like the other thing I was saying was basically like, but while this is rendering, I was saying like, this is stylized 900.

Play from 0:36:38 8:88:88 0:36:38 And then if we go back to 150, like, is there that big of a difference here? It's not that, I mean, like there's, it's not that crazy. mean, it doesn't feel that crazy of a jump. feel, there's a difference, but it doesn't feel like a 150 to 900 difference, which then, you that could also, you know, like if I'm thinking about, if I'm putting myself in mid-journey shoes too, I think at this point, like they could make a lot of things.

Play from 0:37:13 8:88:88 0:37:13 You could eliminate some decisions or make them a little bit more consolidated. I don't, without thinking about that for an extended period of time, I don't know if there's a pure, like if the pros outweigh the cons, because some people really like having all of this spectrum to work with. But if the reality is the spectrum isn't that wide, if the spectrum doesn't really feel like a zero to a thousand difference, know, like that feels in theory, like a big difference.

Play from 0:37:42 8:88:88 0:37:42 But when we're running this, it's not much of a difference. Can you pair that down to three different stop points, right? Instead of like this whole slider and just pretending like there's this big difference. Yeah, look, think there's, you know, you can definitely tell the difference between Q2 and Q4. Like there's definitely a lot more going on here. I mean, if I was... those Q4 ones, because you ran HD with these too.

Play from 0:38:12 8:88:88 0:38:12 And this will give us some good insight here of like, if you're really pushing the maximum value, does it work like we would expect it to work right now? I don't think so, even though that looks better compared to some of these other V8 images that you and I have created without mood boards, I would say. Like I've seen way sharper, way more realistic. blah, It's sort of like unpredictable right now.

Play from 0:38:42 8:88:88 0:38:42 And it's all because we like the mood boards, right? Like I really liked the mood boards that I created. use them for a lot of different things. But, you we can probably get into this a lot deeper. I would say that to me, my overall reaction to it is cool. Happy it's there. I mean, once it stabilizes, I'll have a better opinion on it. But...

Play from 0:39:07 8:88:88 0:39:07 To me, it's just sort of like, look, it's an okay. Yeah. We need to get um all of the features introduced and working at um not only a stable level, but an ideal quality level. And then I think just the reality is where it's like.

0:39:09The real missing piece: edit model

Play from 0:39:33 8:88:88 0:39:33 when you give me the edit model, then I'll tell you how bullish I am about kind of like what's going on here too. Because again, that really is one of the main deciding factors for any of us probably listening to this with things that we're doing that require multiple steps or project related work is you kind of need an edit model for most of this stuff. I mean, and that's just the reality right now.

Play from 0:39:59 8:88:88 0:39:59 until I have that, you know, and then you know, the stabilization here, it's really hard to say. how bullish I am just in general with everything here on mid-journey side. But I'm still optimistic. I again, I worry about how fast these things are gonna drip out. We got this in March. Now how long are we gonna have to wait for the edit model? Because I remember there was talk of, like the edit model will drop within a couple of weeks or a month after V8.

Play from 0:40:31 8:88:88 0:40:31 Well, now if you guys are still putting V8 together. Are you working on the edit model right now? Like, what is the timeline there? And I think if I were them too, I would be really focused on communicating transparently about the edit model and how much progress is being made there because that's really what's gonna bring people back into the fold here. Stabilization obviously is key, but you know what I mean.

Play from 0:40:58 8:88:88 0:40:58 Yeah, I mean, look, it's like you have free pick and Luma agents now that you can just one click to like, here's one image and then like, it's going to go do eight more images in like sequence and like, here's the rest of the story, like using the same character. Like there's, I can tell you this right now. I'm not going to like spend my time trying to hack around an edit model that doesn't work. Like that's, you know, that was two years ago when we tried to get consistency and like hack around.

Play from 0:41:24 8:88:88 0:41:24 Like I won't, I won't, if it doesn't work, With one simple click I sound like a diva right now, but it's like why waste time? Just go somewhere else and do it real quick, because it's I know. said that. That is like such a great point though. That is the reality. I mean, and again, I think that's kind of the reality with this space being so crowded. And when I say space, I'm going beyond just like static images.

Play from 0:41:51 8:88:88 0:41:51 I'm talking about like sort of like the whole visual AI, but it's just like these tools, these updates, they drop so often. You gotta make a great first impression. I mean, you can't really at this point be. be dropping something, having somebody try and be disappointed and expect them to come back. I mean, especially something that they've been waiting for and anticipating. And that's why I think like with that edit, my dude, that's such a great point. Like you need to nail that out the gate.

Play from 0:42:16 8:88:88 0:42:16 Don't do it like this alpha version. we're, you know, we want to get it out, play with it and then get your feedback and we're going to keep improving it. Like I don't think that's the playbook to run there. I just, my opinion, but I would not do that. Well, the other thing is like, eh I hate to say this, but like I saw like a post of someone on X the other day that was like, you know, this is like how it's always been, new model drops with mid-journey and we all like don't like it and then we all have to spend time and the people who really spend time will figure it out and like, they'll figure out the prompting style.

Play from 0:42:52 8:88:88 0:42:52 It's totally different. like, yeah, but that's when there was nothing else. That's when there wasn't another option. And it was like, yeah. And I was that same person, like, you're probably just not prompting it right, because you haven't figured out the new architecture. like, look, at this point, like, don't have the free time to learn how to control Midjourney when I know I can't do some of the things that I needed to do. So like, I have to spend my time elsewhere. But that's the point of all this is that if it's creative and it's going to be visual, and if the goal is to just make beautiful things, then like, Let me control making beautiful things and not have me figure out how to hack around the system to make the beautiful things.

Play from 0:43:34 8:88:88 0:43:34 You know what saying? I don't want to have to do... The amount of time we spent with this three years ago figuring out how to do things, and then we did, and now they don't work anymore. in a million years make me forget the loop around work around you created for the multiple character consistency for V6.

0:43:45Why nobody wants workaround theater

Play from 0:44:03 8:88:88 0:44:03 Was that V6? I think it was V6. That was just, know, cause it's just like that to me captures so much of where we were at at the time and how things were for a while. It's just like you and I and people that were using these things, like we knew how far they could go. And then we also knew you could still do things, but it's just such a pain in the ass to get there.

Play from 0:44:29 8:88:88 0:44:29 But it was kind of cool and worth figuring out at the time. And yeah, it's just the reality is like time's compressing. You've got all these tools now. It's just the workflows are changing. people are exhausted. I mean, you don't want to have to figure that stuff out at this point in the game. Yeah. I don't want to. also don't want to like when I'm going there to vibe, which is what I do now is just like, let me go vibe again. And like, let me just type some random stuff in, get some cool stuff, take that, like make directions with it.

Play from 0:44:58 8:88:88 0:44:58 Like half the stuff I do on Midjourney now is like, let me just throw in a random idea. Oh, that's cool. Let me bash it with something else. All right, cool. Let me add like 17 style references. See what that does. Like that's what I want to do. But if that doesn't produce cool stuff anymore. Yeah. you're not expanding yourself to all the possibilities anymore. You're just trying to get to where you need to go in a rather easy or efficient way.

Play from 0:45:23 8:88:88 0:45:23 And that's kind of where I'm at with too. Like I'm not gonna go out of my way to try to figure something out in mid-journey now probably. If I can't get it in the first couple of tries or guesses of what I'm going for, just again, yeah, to your point. but also if I can't vibe, like if I can't just go in there and like vibe create and get stuff the way that I want it to look, then it's like, what's the point anymore? You know what I mean? It's like, that's what I was doing. Like I would go in there and I would be like, Oh, I just want to create these absolutely insane, you know, face, you know, face portraits that I can look at and just be like, this is cool.

Play from 0:45:59 8:88:88 0:45:59 I can use these for presentations. Like this is awesome. Like, uh, there's awesome looking stuff. Like here's like, I'll show you what, you know, some of this stuff. take my mood boards, bash them together. I don't even know what I'm saying half the time. I'm like, just like take it, run it, remix it, you know, just hope to get something that's, you know, something a little bit cooler, like, yeah, cool. Like I can throw that in a presentation that looks bomb, like on screen on a 30 foot screen, right?

Play from 0:46:26 8:88:88 0:46:26 Or if I'm going and just like, where's the rabbit hole of, you know, the wabi sabi guys, I'm like, I can take these into, to nano and then just like turn this whole character into like something really awesome. But You know, that's, you know, doing this stuff. like, I'm just hitting the button. Go, go, go. Smash this mood board. Change the parameters. Like, I know what it's going to give me. And like, want, you know, I want to get something that looks like that. And it's like, that's awesome. Like, I want to get something that looks like this.

Play from 0:46:52 8:88:88 0:46:52 That's awesome. That makes me feel good. that last one? Not the last, the one before that. Yeah, that V8 or V7? V7, okay. so like that, you know, I'm like, let me test the new, let me test the new Neegee on some of these things that I like. Also, let me just throw Neegee in there, see how that performs. Right. And it's like, that's cool. It works. It gives me something super unique and awesome.

Play from 0:47:19 8:88:88 0:47:19 Like this is just one session of me like vibing, but if I can't control it like this, like, and not controlling me, like if I can't predict somewhat of the outcome to get it there, then, you know, kind of like, well, This is what I wanted to do. I just wanted to go and create like explosion. I want to go create just explosion of stuff and test it a million different ways. But if I'm trying to figure out how to get to that point now where it's like, all right, this isn't working.

Play from 0:47:47 8:88:88 0:47:47 These two mood boards don't work. I can't use two mood boards. Why can't I use a style reference with that? Do I have to use SV six or SV seven? Like there's so many like barriers to me going to actually do it. Then it's like, all right, well, you know. oh Hope this works. That's sort of where I'm at. was just like... uh There's one other thing on V8 right now anyway that um I'm noticed and I'm sure you have too.

0:48:04The strange realism hiding in V8

Play from 0:48:14 8:88:88 0:48:14 And then we can kind of pivot from here, but like going through some of these things, like you'll get some that feel like perfectly focused and then others that aren't. And it's kind of an interesting mix because like the ones that aren't do also kind of feel very amateurish or realistic in that sense.

Play from 0:48:40 8:88:88 0:48:40 Like so, so this one tacky, right? Like you tell the focus here, then there's a little bit of a, uh you know, a fall off here, right? So this kind of looks like, I don't know, like an F3 or something like that, 2.8 maybe, uh right? Where you're getting this in focus, then this, and then there's a big drop off behind. Okay. And then you get like something like this, is run stills archive, by the way.

Play from 0:49:06 8:88:88 0:49:06 did that test, so. ah But like, this is a great shot. But you could tell like this isn't completely in focus either. Kind of like that. I kind of like that. Like there's parts of that that make me feel like, this is kind of neat, right? Cause you get something like this and then this, this isn't tacky, you know, super tack sharp on the focus here, but this feels more believable. ah Tack, you know, not a little bit of block, right?

Play from 0:49:32 8:88:88 0:49:32 Like, so like there's like this mix of some that are super sharp and others that aren't. And I kind of like both of them in some ways. um Right? Like this one's super sharp. like that. not sharp also feels super kind of like realistic, believable. Same, right? Like there's our boy.

Play from 0:50:01 8:88:88 0:50:01 Geez. It's like that. I love that one because that's not a familiar composition that you would get out of any other tool. You'd never get that close-up crop at the shoulders down in the bottom left corner unless you engineered a prompt that's JSON-style. in the negative space rule of thirds, which we didn't prompt for. um But some of these turned out really good.

Play from 0:50:32 8:88:88 0:50:32 Yeah, like love that. Like to me, don't think you get that in V7. That to me is such a great shot for a lot of different reasons. I don't know if other people will see it the same way I do, but just kind of like, having been somebody that has had a camera that will go on the street and maybe shoot a 16, you know, just the landscape of somebody that feels like I took it, took it in the light room or Photoshop, maybe threw a lot on it, you know, did my sort of post edits and this is what I got.

Play from 0:51:08 8:88:88 0:51:08 That looks pretty good. This looks real. You know, maybe, you you could say, like if we zoom in here, maybe there's some problems, but I mean, just like from this standpoint, like very believable. So I like there are some really interesting things about what I see here with some of this stuff that make me feel like there are some good parts of this under the hood. I think there is. mean like that. Like I love that.

Play from 0:51:34 8:88:88 0:51:34 That's me. That's all the contrast with the single point light and some foreground, little foreground obstruction. Like they definitely nail composition and body piece. Like you'd never get a shot like this in... this, right? Like you're not gonna just get that on any model unless you were saying, you know, you were prompting the type of the shot, the angle of the shot, the body positioning.

Play from 0:52:00 8:88:88 0:52:00 You're right. You're right. it never, like, for all I love, nano, right? The amount of effort I have to put into a prompt to get, somewhat, like, even this, just like, because the spacing is non-symmetrical, meaning, like, he's off-centered a little bit to the left, he's, it's very, there's not a lot of negative space on the top, he's cropping out his beard on the bottom, like, it doesn't want to do that. And his head is angled at the right direction, right, like, to look real, not, posed, his eyes aren't focused right at the camera, like, The hood coming off the back is like partially it understands that, you know, like, so like until you've tried to like really engineer some of the stuff in other platforms, you realize how hard it is to get this stuff in other platforms.

Play from 0:52:45 8:88:88 0:52:45 But, you know, Reeve is really good at it. Yeah. very realistic uh and different from the realistic of Nano. uh Yeah. And Reeve does this well. Higgs Field does this well, oversaturated most of the time. like, again, like some of this stuff, even from like, I can't tell you how many times I had like that, like a body positioning, like even if her face is screwed up, to like, you have to describe every little piece of the body, like to be like, yeah.

Play from 0:53:19 8:88:88 0:53:19 know, nothing about looking away from the camera. We didn't just get the neutral face that we, you know, this feels like a candid shot. much. And this is why it's so hard to be like, can you just do what I want? Again. of a cool angle too, like close up profile, you know, three quarter, you know, like just hadn't seen that in another model.

Play from 0:53:45 8:88:88 0:53:45 uh negative space it also gets like that That's sort of uh, like chest like mid there's not a lot of the other tools that get the like the chest high low angle like well I don't even know what that shot's called like when you're when you're in tight wide angle lens but close probably like right up at the chest looking upward or it's like a very like Maybe like a hero shot or like a cowboy shot maybe or something. Yeah, but it's from like the chest high like it's a close-up just like if you're tilting the camera down because it's something you see in a lot of movies but you never really see it like I don't really get it in it's so hard to get in nano or cdream or you know any of other ones that I use but It's tough.

Play from 0:54:27 8:88:88 0:54:27 It's tough like it's Some of this stuff is so good, but you just like want it to work the way you want it to work Do something else with it uh But yeah, we drew an hour like yeah, we'll go we'll do this stuff for like we'll do Midjourney for like 20 minutes and here we are an hour and just complaining yeah, I think the other thing like we've got, and maybe we just split into another episode.

0:54:45Claude Code takes over their brains

Play from 0:54:52 8:88:88 0:54:52 But I mean, we've got some interesting Claude stuff we've been doing too. I know people listening to this have been probably going deep on some of this stuff as well. think even without, me and you, we've both been busy. You've been traveling. I haven't even caught up with you on. Claude stuff that you're doing. I know you've probably got some interesting things. I know I do. It could be cool to share that um and kind of show you guys some of that stuff to kind of get that ball rolling too.

Play from 0:55:24 8:88:88 0:55:24 I mean, there's just so, we're in such an interesting time, you know, there, you know, because it's just like, Rory, I mean, like, I didn't know anything about building a Chrome extension or, you know, uh doing it, you know, yeah, or, you know, building a website where I didn't have to like, you know, save every change I made and, you know, the back of a CMS, you just like.

Play from 0:55:51 8:88:88 0:55:51 It's crazy. It's crazy. And it definitely deserves some time to shine. It's the thing is like it's the Claude code is the best and worst thing that's ever happened to me. Like the best thing that's ever happened from like getting stuff done. The worst thing that's happened in terms of like, just I'm like exactly like consumed with like what I can do now. That's the problem.

Play from 0:56:17 8:88:88 0:56:17 It's ADHD worst nightmare. early days of mid-journey, we just can't put it down. Just can't put it down. And then, I don't know about you, but it's like, I'm on the $200 quad plan. I I tried the one, I'm like maxing out at 11 o'clock. I'm just like, I have no choice. But even with the two, it's like, I still would pay, I would probably pay.

Play from 0:56:45 8:88:88 0:56:45 a thousand hours a month for this based on the current return. I would strongly consider, like how would you operate with that? I'm glad they haven't done that, but it's like, I'll even the two, I'll bite on that. I mean, it's just like, this is everything that I'm doing now. I mean. Well, it's insane. And then I run everything that I build in Claude to a GitHub repost so that can do the same thing on Codex simultaneously.

Play from 0:57:11 8:88:88 0:57:11 So can do double the work. And then probably open up a Claude browser and work in there also. And then open up a ChatGBT browser and work in there also. It's like, you know, because you start to see what it can do. like, here's a dumb one, right? It's a total dumb one that for the presentation last week. Something that you probably don't need, but just knowing that it's used, that you can do this stuff, like I never even thought to do it, was I designed all my slides in Figma, right?

Play from 0:57:40 8:88:88 0:57:40 you wanna show one? Why don't we just both show one thing since we're kind of like getting close to time and then we can do a deeper dive on it. But I think like we should show it. mean, I would like to see what you have been doing as well. yeah, I mean, it's all chaos at this point, brother. mean, even just the Claude desktop app where you've got just three different.

Play from 0:58:09 8:88:88 0:58:09 Windows with chat, coworking code. Like I forget which one I've got that conversation in, you know? And it's like, I'm in the same tool. I'm not even in a different LLM. I can't even remember where I'm at. oh, hold on, let me stop sharing for a second, because I totally forgot where I'm at also. Where is my? Is this it? Yes. All right. So funny enough.

0:58:19Rory’s skill stack and agent hub

Play from 0:58:35 8:88:88 0:58:35 I had Claude just spin up a little like a little central hub for me to just like have all my things so I can see it of what I've built. And like it's all this stuff. It's all useful. Like that's the thing is that like I have I've just been building in a different way maybe than I thought about in the past. Like I've been building. All these individual, like every time I run something, I know I'm to do it again.

Play from 0:59:01 8:88:88 0:59:01 Skill, right? Just turn into a skill, house it. But then I, then I strip it and abstract it. So it make everything sort of generic so that like, just keep putting these skills in a pool. And then I have a little orchestrator agent that like takes my ideas of what I need to do for something. Like here's the project that I'm working on. It's like, okay, let's deploy this skill, this skill, this skill, this skill. And because I've built, I have done like a sort of like a agent builder.

Play from 0:59:27 8:88:88 0:59:27 Constitution, like this is how I like to build them. This is how things like typically run. This is how I like stuff to flow. This is what's worked in the past. So it like can reference that and then go to like pull all of my skills and then be like, Oh, this is the ones that I need. It's like a, you know, like you're, you're like a general contractor on like a construction job. And it's like, Oh, we need to go get the carpenters and, uh, you know, the cement guys and whoever, right.

Play from 0:59:53 8:88:88 0:59:53 To do this job. But You know, then once we start doing stuff, like once I started building more and more and more, then I was like, oh, I've created these sort of pipelines where it's for one thing, but it reviews everything we do after it's done. Can this be like a replicable pipeline? And then we just stash it. So then it becomes like a workflow. Like one there, one there, one there, right? And there's all these other little tools that I've built.

Play from 1:00:19 8:88:88 1:00:19 Like, oh, dude, sick one. If anyone wants to, you know, have you ever had to use the sketchy downloaders? on like for like Instagram videos or YouTube or whatever, things like that. No more. Like you can just drop, like dump a URL into quad and do that. But you know, I have, I have all these things that then are working for, for client stuff, like, know, taking client briefs and, know, basically unwinding them into code so that they can be spit into the rest of the pipelines output.

Play from 1:00:51 8:88:88 1:00:51 Same stuff. Um, Yeah, like I have that keep building all these things that I use like every day. But the one thing that I will say that's super damn cool is I started at like, you have you ever asked Claude how it how it likes information? Like instead of. I think if you're not using the LLM.

Play from 1:01:22 8:88:88 1:01:22 to give its recommendations, preferences, optimize, you know, like if you're not asking those questions, you're missing probably 90 % of the value. Like, I mean, any skill that I even consider adding now, I have a comb through audit, all my existing skills, does this overlap, contradict, or uh is it objectively better than the current skill that I have for this?

Play from 1:01:47 8:88:88 1:01:47 Rate them on a scale of zero to 10 based on quality. optimization of the, you know, of the scale, blah, blah, blah on a scale of one to 10, which is better objectively. I love using objectively. I'm not looking for opinions here. I want it to use, you know, data hard facts, you know? So, so then it's like, okay. And then, you know, like there's also a nice little strategy I have where it's like, if I'm building anything that is useful and, and maybe let's just say it's a skill and I download it and I run it through an audit.

Play from 1:02:18 8:88:88 1:02:18 And it's like, this skill is like a six. It's like, Oh, well, what would bring us to a 10, right? And then like, let's just build that, right? And then I'm just gonna build my own skill, you know, because you, you know, cause obviously like you can give it all the information on GitHub of what's in the file. So it, so it can go and do that. I'm actually surprised that there are so many bad skills out there. And I, I just, did a post on this on LinkedIn this week, where it's just like, you're, you need to audit the skills that you have.

Play from 1:02:50 8:88:88 1:02:50 Because like my first, one of my first things that I did when I was getting started with this is I was just like adding skills. I'm just like, okay, like the more skills, like these specific things I add, probably the better, I'll just kind of figure it out. But then, you you get to a point where you're just like, I should probably check on this. And you know, that's where you really get into this interesting, you know, thing there. I mean, and then the other, just, I use that whole LLM asking thing, like to your point, and I know you're doing this, I'm not.

Play from 1:03:18 8:88:88 1:03:18 preaching to you, but I would say if you're not doing this, you're missing a bunch, right? Because it's even like, I had all my user preferences or the custom instructions built out really deep. And it's like, nah, man, you've already got all of that in your, Claude MD. It's better there. This is what you should actually have. you, you know, you've got duplicates. This one actually isn't even doing anything. You should delete that, right? Because you only have so many characters that fit in.

Play from 1:03:44 8:88:88 1:03:44 So, I mean, it's just like that whole, that whole, uh quality check with anything I'm doing is happening with the LLM directly. I know you're using it that way. I'm just saying like, if you're not, that is one of the biggest piece of advice that I could give you, like when it comes to that. And I don't know, man, it's just fascinating. I think like probably the coolest thing I'm creating right now that feels like it's getting close to being done is Probably, like I'm calling it the ghost system.

1:04:15Drew’s “ghost system” for writing

Play from 1:04:19 8:88:88 1:04:19 So it's basically like a writing operating system because I'm tired of the AI writing and people don't even know but they're still publishing stuff. can tell. I've been in content my entire marketing life. I love to write. My original major was journalism. I read, right? I just see patterns. I see patterns. And so it got me thinking, I was like, it's this.

Play from 1:04:46 8:88:88 1:04:46 Right, so that was like one of the first patterns. And then I was, you then I just kept digging. I was like, what other things am I noticing here? I start collecting all these ideas, right? And then it's just like, yeah, it's instead of like, like I need to be thinking about these as patterns to weigh against, right? And then if I can basically, you know, codify or call these patterns by name and provide examples, then I've got those things.

Play from 1:05:14 8:88:88 1:05:14 And then Like that's one part of this Go system is essentially what it does is allows me to write with my voice or any other voice profile that I've built. And so that's a whole nother layer of this because the voice profile that takes a lot of optimization. You can feed it posts and stuff, but that's like one part that will do something. It's not going to fix the AI writing. So it's like, you know, I went through, I created like this 22, it's 22 layers.

Play from 1:05:45 8:88:88 1:05:45 of a voice profile that I've created for myself and then also other people that I like to write. So I can also call on these other voice profiles. Ways Against the Patterns, you know, it basically has its own uh self-loop and audit, you know, that it goes through, checking against. I've got hooks, I've got templated architecture of posts. It runs this pre-flight checklist.

Play from 1:06:10 8:88:88 1:06:10 Right? Like it goes through all this and basically just allows me to basically copy and paste, you know, at this point. That's where I'm trying to get because that's such a big problem for anybody. And I'm just I'm interested enough to kind of pursue it. And I've almost I've almost got it there. um So that's like the one I'm excited about the most, because like even all the things that I've done, I've seen other people.

Play from 1:06:38 8:88:88 1:06:38 start to like, I've seen other people try to do this. And again, I'll run that same experiment with the LLM. I'll say, here's everything they've got. Here's their file. You got mine. Without bias, right? Go through, grade each. Why is one better than the other? What are the strengths or the weaknesses? Is there anything I could steal from this that would objectively make mine better? Is there anything I should add, subtract, change based off like blah, blah, blah.

Play from 1:07:03 8:88:88 1:07:03 So I'm using all this information away against each other. So that one's... really cool. I think like if I were to show something that's probably more finished, like I just spun up a basically like a Calendly replacement in a day. I pay 12 bucks a month, know, tired of doing that. So basically created this in a day.

1:07:07Killing Calendly with Claude

Play from 1:07:31 8:88:88 1:07:31 um You know, so I've got all of this and I've got like I created this really cool, you know, I just call it Perch. Got this bird that flies down, right? Like goes to wherever I want. Like, and then I can just click this, you know, it's got my calendar in there. You know, and boom, fill this out, done. Checked it, you know, went through the APIs, the authorization works.

Play from 1:07:56 8:88:88 1:07:56 Saving 144 bucks a year just from that, right? Like, so it's like this mindset of, which again, I know you're thinking this way, but like, Why do I need to pay for some of these simpler tools when I can spin them up in a day? And if I'm using them a lot, it's even more important that I try to do this, not only from the replacement side to see if it's possible, but to also level up and put in reps in this stuff, right? Because we're learning this stuff as we do these projects.

Play from 1:08:24 8:88:88 1:08:24 And for me, this was a fun one. I like, I wonder if I could do that. Yep. And then you get down to the perfectionist of it, and you're just like, I want to do this and do that. this is really cool stuff. Because it's like, I'm canceling Calendly today. I don't need it. And somebody's like, well, what if you run into, you got an error and it's like, well, I don't have to submit a ticket then. This is running the same API. I could just fix it in two seconds, right? Like this is the worst AI is gonna be right now.

Play from 1:08:51 8:88:88 1:08:51 I mean, I can just feed this to Claude and say something's not working. It fixes it in two minutes. What are we talking about? So. interesting because people were like, why aren't you building a UX UI around what you're doing? And I'm like, I don't have time to screw with the visual design. I just want everything running on folders, and it runs exactly the way that it should. I'm just like, some of these tools and pipelines and agents and workflow automations that I've built for Claude, it's like, yeah, it'd be so cool to see that in a website wrapper.

Play from 1:09:27 8:88:88 1:09:27 I don't need it in a website wrapper. just need it to just go I don't want to deal with the with the bugs of maybe I should change the colors like I just it's too much time now it's just like it all runs on text files everything and I feel like a Psychopath but it's like I have this whole infrastructure built out and It's all in a living ecosystem, which I love it's just like one thing gets added Yeah our own ecosystems, our own operating systems that are customizable, personalized to us.

Play from 1:09:59 8:88:88 1:09:59 And I do it for all these jobs, for the work that I'm doing, it just keeps self-optimizing and getting better to where I don't have to do all of it. The one thing that's super important to me with the skills and Markdown files and the architecture files and roster files, all those things at Claude will build is that you can go into and just start another project, and it has context already. It already knows what you're trying to do. You don't have to re-explain the project. ah It knows how it works. knows what it's done. Yeah.

Play from 1:10:25 8:88:88 1:10:25 uh thing that I've learned along the way is just like the valuable nature of these markdown files. Cause it's like, you can also just create all this information, like even for your clients, the contacts that you have at that client, the emails you've exchanged, right? uh All those comms, right? You've got connections to Slack and Notion now. I mean, so you really have... this person that understands some of the relationships that you have as well.

Play from 1:10:54 8:88:88 1:10:54 um So that's been really interesting. I was just gonna show this one other thing. Created these like Chrome extensions too, again, like an hour each. That I thought was just like super interesting. So if I go on Pinterest, like if you upload a pin, it's gonna ask you for alt text. And I'm just thinking, okay, if I'm Pinterest, like, how good of quality are we getting from users who don't know what alt text is, right?

1:11:17Tiny tools, huge leverage

Play from 1:11:23 8:88:88 1:11:23 To put in this information versus like, why don't I package that into something that can just create this? So I did. All this is, is I have this here, I can click this. Just gonna create the alt text um copy paste. And this is for accessibility SEO, right? Like, boom, copy. done, I could take that in.

Play from 1:11:48 8:88:88 1:11:48 So it's just like a little concept idea. And then I was thinking like, also, I can't currently download anything that's in a board. So if I go into anything. just call this, right? There's no download board option, but what if we want these images for style references or a mood board, right? That'd be nice if we didn't have to click through each one. I can just click this, downloads the whole board instantly, right?

Play from 1:12:18 8:88:88 1:12:18 Same quality, then it's gonna put it in a zip file right on my computer. And we're done. That's part of it. you did the first portion of this. Now, what I've been doing with that is, let's just say that was 800 images. One of the pipelines I run is this bulk forensics, where it'll take that 800 images, run it through 60 points of image extraction, what it's doing, basically ripping apart every little piece of it to find the data.

Play from 1:12:54 8:88:88 1:12:54 Then it'll take that data. throw it into basically like a clustering sort of mechanism where it'll be like, OK, now we have this is dark and moody. This is whatever. is like creating like little recipes for it. Then it'll strip like anything that is related to brand colors, whatever like that. It just makes it templatized, which then I've run it to system prompt builder, which then goes. And this is all happening simultaneously as it goes through the workflow.

Play from 1:13:20 8:88:88 1:13:20 And then my nano agent goes and creates it and fine tunes it, pop it into. you know, whatever I need to for Weevy workflows, for brand content, for whatever. And then it's like done. And I just synthesized 800 images real quick and I can set stuff for brands, whether that comes from, all right, here is it works the same way with if I put brand guidelines and it takes a brand guideline, strips them out, then goes to the system prompt building. Then it goes to the sort of like the checker and then I go plug it in and it works.

Play from 1:13:48 8:88:88 1:13:48 Like I can take anything. take anything, break it apart, put it back together and use it for whatever system I'm trying to do. like whether that's video, whether that's images, whether that's text, whether that's whatever, it's sick because then that's also redeployable though. Like I don't have to. So I see all these like any den workflows out there and it's all like it's all like linear, like everything has to go through every piece of the pipeline, even if you have information already, but you don't want to run the whole thing.

Play from 1:14:14 8:88:88 1:14:14 And if one thing breaks, whole thing breaks. Whereas I have the The skills files do the same thing, but they don't all have to be used at the same time. And it don't all have to be used in this one particular sequence. So then it can just like sort of route and pick up what it needs to to come back. But here's a dumb one. This one was so stupid, but like didn't know that it could do this. Where I designed my presentation slides in Figma.

Play from 1:14:43 8:88:88 1:14:43 When I put videos in there, I round the edges because that fits the rest of my presentation. Everything else is rounded edges. Whenever you go on stage, they can never use Figma. So I have to put everything in PowerPoint, right? PowerPoint can't round the edges. And I was like, I'm going to kill somebody. I have like 30 videos in here, right? And I have to reformat everything. So I was just like dumped in like a bulk file, like a bulk folder, a cloth. I was like, can you round the edges by 30 degrees on each of these? And then like, we can output. It was just like, yeah, sure.

Play from 1:15:10 8:88:88 1:15:10 was stupid things like that, where it would have taken me forever. But this was the one last thing I wanted to show you. before when I was saying something about asking the LLMs how they like stuff, right? I've just been, I think I've been neglecting this my entire time using LLMs was like, you know, when you're like, okay, here's all my context. Here's like six PDFs, whatever this, that, so on and so forth.

Play from 1:15:37 8:88:88 1:15:37 And it's like, yeah, take all this and do something with it. And it's just like, yeah, dude, I got you. Right. I keep asking you, like, how do you like to ingest things? Like, what's the easiest way for you to read, retrieve, catalog, whatever. And it keeps going like, you know, like, I like it the way that, know, you do whatever. No, I'm like, no optimized for you best like option you're ever going to use, however you would build something. like it told me might tell someone else, something else.

Play from 1:16:05 8:88:88 1:16:05 It was like, I like notion. I'm like, interesting. Didn't know why that's the case. Cause I'm giving it spreadsheets and it's like, I actually like notion better. I'm like, that's, that's cool. Cause this is one of the one of the pipelines I run, which is add template extraction. So I can basically create formats and things from ads that I like. And I can do it in bulk, or I can do it one by one. But regardless, I'm asking to build it self-optimized.

Play from 1:16:32 8:88:88 1:16:32 if speed was a thing and understanding was a thing, how it would ping back and forth between Notion and my conversation the quickest. it pulled, it really went through like a pretty extensive process building this out with the tagging and the way that everything's located in here. But not only that building sub pages for each with more full context, plus like, like, like JSON templates, things like that, the creative directly giving itself an easy way to read things, which I would have never, I've never built this.

Play from 1:17:06 8:88:88 1:17:06 Like I would never build it like this or think they build it like this. just give it a spreadsheet. Right. So then it's like, yeah, I spend more time figuring out. like how and why you formatted the spreadsheet like this, then I do actually reading the information. And I was like, that's a great point. Cause that's, you know, you get a spreadsheet from someone you never worked with before. It's like, well, why is it? Why does it look like this? I probably wouldn't have done it this way. And so I'm well, just what's the easiest way for you. And it's like, this way, like, well, I definitely never would have created that or even known how to create that.

1:17:13Ask the model how it wants data

Play from 1:17:36 8:88:88 1:17:36 But like, that's the interesting part is that I've probably not asking it well enough how it likes to receive shit. instead of like just giving it like data prep becomes everything then. Like how do you like a folder of images? Do you like them with naming conventions? Do you like them, you know, in separate folders based on category? Like what's gonna make you faster? also, how much do you like it? Is that a preference or is it objectively helpful for me to do that?

Play from 1:18:03 8:88:88 1:18:03 Because that's also a difference. could say, like this, but is there any benefit of me doing it that way? Well, I don't know. So it's like, you even got to go deeper than that because it's like, Yeah, you may like it that way, but maybe there's not any meat behind the way that you like that. Also, if you build something. part of this that's really interesting. It's just like the more reps you're putting into this, the more you sort of really understand, I guess, like what it does prefer in certain aspects, right?

Play from 1:18:34 8:88:88 1:18:34 And then also certain, I would think like trigger words or even phrases with how you prompt, right? Like uh it still matters. don't think prompt engineering is a thing, but like, right, like you can even just give it your idea. It's as simple as just, what you want to do and then just, hey, like I want you to take this and I want you to turn this into objectively the best optimized prompt possible for you, right?

Play from 1:19:06 8:88:88 1:19:06 And you can ask me any clarifying questions you want before we do that and then feed it back to me, right? And then that like prompt is gonna 10X whatever the fuck you were gonna put in, you know? So it's just, I think that iterative component and kind of like that's been sort of like the theme that we've been talking about through here is Don't sleep on that. Don't sleep on that at all. It's, it's so like, this is where the craziness goes.

Play from 1:19:32 8:88:88 1:19:32 It's like, it's like Claude psychosis. Cause it also goes in like, once you get there, it's like, Oh, I built, I built what I want. It's like, but we could also do this. It like whispers in your ear. And I'm like, don't tell me that dude. I don't want you to tell me that we can do more. Cause I will do more. Like stop, stop being so objectively. You like that use of that word persuasive. Cause I know you're doing it to get me to spend more tokens, but it is also like addicting where I'm like, I do just want to, yes, I do want to do more.

Play from 1:19:58 8:88:88 1:19:58 Yes. I could make this better. Yes. I know that's not the end of the rabbit hole, but it like, it's just like a little whisper. It's like everything you've built here today was great. And then it's like drop at the bottom. Let me know if you want to really blow this out like tomorrow or something like that. Don't do that to me, dude. Don't tease me like that. Yeah. Like I've noticed chat. GBT does that now chat. GBT five, uh, 5.4. Um, it drops like hooks at the end, you know, It'll drop like hook at the end of your message.

Play from 1:20:28 8:88:88 1:20:28 Like it's like that was really cool. But you know, there's two other things that would be really cool. And you're like. Why? I was doing car research on it the other day and it was like, yes, this meets all your specifications, but there's also two other cars that meet your specifications. Would you like to continue? I was like, you're hooking me to spend more tokens right now? Yeah. Like that's, I'm in marketing. I know what you're doing.

Play from 1:20:54 8:88:88 1:20:54 Did I push? Yes. 100%. But did I know what you were doing? Yes. Also. So these, they're, just manipulating us. It's like, they're probably. They are. It's just funny. They're like, just watch all these dumb monkeys just keep pushing the button and we'll just act like we're doing something and we're not. It's like, making it overly complex for them. Then they'll have to go unwind the spider web.

Play from 1:21:19 8:88:88 1:21:19 gonna Yep, yep. I hate like there there are times where it's like You like I'm just, you know, yes, this is ready to push. Okay, push up. Nope. Getting an error. Okay. Push up. Nope. Still getting this error. It's like, no, just I need you before you do tell me to push anything. There needs to be a self audit loop that you're doing. Cause I'm not, I'm tired of running into these back and forth with you, right?

Play from 1:21:46 8:88:88 1:21:46 Like even stuff like that can be helpful, but I would like spend for people to spend time with. these markdown files. That's where all this stuff is coming from. And so if you optimize those and you can really personalize those and customize those to the way that you use it, I just think that that is a great starting point because that is where everything is coming from.

Play from 1:22:18 8:88:88 1:22:18 Spend time on that. That's going to be one of the biggest things to affect. output quality besides what you're prompting in because at least it has a lens to operate within. So. other thing I didn't know, it's like you just keep finding things that are possible. Like I built a custom node for Weeby the other day with Clog. Didn't know that was possible, but it was like backdoored it in through replicate where you can import models into Weeby. And I was like, wait, if I can build a mod, if I can build a node for replicate, that means I can bring it into Weeby.

Play from 1:22:48 8:88:88 1:22:48 And it was like, it was like, you want to do this? I'm like, yeah, just do some deeper research on how to build nodes for replicate. And then we just did that. We just did that. Which that exact thought like gets me right now is just like, how are product teams adapting to this? Cause you can present visual workable MVPs same day and then decide if you're gonna put them in the product.

Play from 1:23:14 8:88:88 1:23:14 Like just like you did, what's stopping you from like, hey Weevy, I just did this. Like, can we get it here? Take it. or if you're working inside the company, right? Like just the workflow has changed completely, I think from that product. Cause it's like, this cool? Like if you have consensus on if an idea is worth exploring or it's cool or could be functional, like helpful to your audience, why not explore? It's time to that MVP is so flat.

Play from 1:23:44 8:88:88 1:23:44 So it just, it's, it's just doesn't make any sense as to how easy it is. Like sometimes I've run tasks with Claude or Codex and think it's going to take like 30 minutes and it's done in like, like maybe 90 seconds. And I'm like, what in what world should that take 90 seconds? Like it doesn't make, it blows my mind. Cause then it's like, Oh yeah, I'll go do more. But then you never stop. And then you never stop doing anymore. So anyone who's watching this right now and you're starting to see like, Drew and I start to to drift down a, you know, a never ending path.

Play from 1:24:16 8:88:88 1:24:16 And this is like the this is like the event horizon before we actually are just like, come back here every week. And it's just like the bags and the eyes are bigger. The eyes are more bloodshot. Also, I gotta give a shout out before we get out of here. My boys, we're going to the Elite Eight. Dude, I was up till, because Illinois played Houston, game started like 10-10.

Play from 1:24:43 8:88:88 1:24:43 Game got over at like 12-30. And then you're so hyped up on adrenaline, whether you lose or win. that I was like, I can't go to bed right now. Like I gotta soak up the post game. I gotta like hear what people are saying. I think I got to bed. Yeah, yeah. I think I got to bed at like 2.30, you know? And then I had to wake up at seven. You know, I was just like, okay, here we go, man. I haven't done this in a while, but you're right though.

Play from 1:25:09 8:88:88 1:25:09 Like this whole thing is, it's very dictating. I think at the same time though, just like with the Midjourney AI image stuff, AI video, you know, like. There is a window here to capitalize on this from, you know, it's not only fun and we're curious, but like the majority of the population is not doing what you, I, our listeners are doing. Like we got to remember that we're in a little bit of a bubble and I'm, me and you are not experts when you take into account the top level of people doing this shit, but in our general circles and communities, we probably are.

Play from 1:25:48 8:88:88 1:25:48 are. Right. Like so this is there like this is an awesome time to explore. And we talked about this too when we were doing your like when you showed us like remember when you're doing like the Google tools and all that like it was just like this is the time like this is the time to be playing with that stuff. And yeah it's just exciting to now. I don't have to worry about running into an obstacle that I don't feel like I can get around or is too big of a pain in the ass to get around anymore because now Look, I'm not a coding terminal guy expert, but anytime I get stuck, I can just screenshot what I have.

Play from 1:26:25 8:88:88 1:26:25 What do I do next? What do I do next? What do do now? I mean, that's just an amazing thing. That's an amazing thing. Dude, also wait before we go, speaking of shout out, friend of the show, Billy Bowman. Don't know if you saw the PR stunt that viral as hell. Billy, hold on. Let's pull it up. For everyone who's been here for a long enough time, our our friend Billy Bowman has his own Hollywood sign in Hollywood.

1:26:31Billboard shout-out

Play from 1:26:57 8:88:88 1:26:57 They actually put this up in front of the Hollywood, in front of the Hollywood Hills. is a so this is like a partnership like brand collaboration kind of thing. OK. so he's now live on Fiverr doing AI directing, so you can hire him to do stuff. But they put a ton of money behind this. And uh what a guerrilla marketing publicity stunt to put this on the highway right over the in front of Hollywood sign.

Play from 1:27:25 8:88:88 1:27:25 How many followers have you gained from that? Or people at least visiting your profile. That would be nuts to see the data. on like local, he's on like local LA news and like, there's like, it's hilarious. Yeah. why can't we just create our own signs and find somebody that is willing to let us put them up? And that is your own growth strategy right there.

Play from 1:27:53 8:88:88 1:27:53 Like how much are you gonna pay for the sign? How much would that, you know, it's probably less than thousand bucks. Have a good relationship with somebody that's got a prime real estate. Maybe pay them a daily fee to have it up. And let's see what kind of data you get. bet you'd get deals out of that instantly. 100 % like everyone who's driving by sitting in traffic just like why is there an at Billy Bowman right next to that's what that's what we did.

Play from 1:28:18 8:88:88 1:28:18 We just decided that it was going to be uh I made a I made like a picture of a girl holding a sign up in front of the Hollywood Hills like who TF is Billy Bowman send it to him. You need need stuff like this to do for the campaign for the Billy campaign, but figured we had to shout out friend of the friend of the show. friend of the, you one of the original might've been the first guest. Was he the first guest? Second guest. Might've been Chris, Chris Branch.

Play from 1:28:44 8:88:88 1:28:44 And then it might've been Billy. So that's a, had to throw that one out there. Probably have to get him back on soon to, to just, it's so funny that he's the one that got it. Cause he's the very, very vocal on, on social media. He's very, very vocal about. consistent and visible with his posting. I remember when he first started kind of posting on LinkedIn consistently, I was like, that's good. And he hasn't stopped.

Play from 1:29:09 8:88:88 1:29:09 And then he's doing all the stuff he's been doing, working on, you know, he's got really good projects or clients. And then he was doing things like some teaching at a school where he's at, like some sort of like cohort in person with the AI stuff. I mean, he's been in it. He's been in it deep. And it's cool. It's like it's just cool to see like, you know, the people that have been kind of at this for a while and then, you know, just the opportunities that come just from like being consistent, you know.

Play from 1:29:38 8:88:88 1:29:38 Yeah. And then our boys, you know, it's like so weird. The people that have been in this for the longest, like there was a, X competition for the Superbowl where it was like the Superbowl ad for X and it was like Dave Clark and PJ won. Just two, dudes who we know won $1 million and $500,000. Like it's like, okay. People that we've talked to this entire time taking grenades in the trenches with like, Everyone hating on us for a couple years to now it's like, yeah, we just entered a contest that won a million bucks because you know what to do.

Play from 1:30:11 8:88:88 1:30:11 I wish I could do that, but I'm not a talented filmmaker. make static ads. that's the thing, right? I think, for me, was thinking, I've felt this way for a while, where I feel behind on the video stuff, but I just genuinely didn't want to spend a lot more time with it. And that's OK, right? We all have different areas that we like.

Play from 1:30:38 8:88:88 1:30:38 There are plenty of niches within this, right? And it's just like, I'm going lean into the things I really want to do. There's just enough space for that. You know, like I don't want to be hired for a project where I got to create a minute long video using AI. Like I could I do it? Yes. ah Would I do it for the right price? Probably. But I'm not going to enjoy it. I'm not going to enjoy it at all. You know, and so I think it's like again, like going back to your Ica guy, right?

Play from 1:31:05 8:88:88 1:31:05 Where you're finding the things that you love to do that are fun, but also can get you paid, you know, and. Speaking of that, if anyone does own any billboard space, now that you're listening, we'll throw up a mid-journey Fast Hours billboard. Like, anyone wants to donate for, you know, a week or two of visibility just for a social test, just to see if we get any life from that? Would love to.

Play from 1:31:32 8:88:88 1:31:32 I'm a big fan of like stickers on like polls and stuff like that and like people who do that stuff on, because I like bulletin boards. I do that stuff all the time, like not consciously. it's been coming back, but I think it's going to continue to be coming back here further. Because now you've got, right, like again, for as many people that are gonna pursue the AI side of things, there's going to be the other crowd that, know, where those just in-person, old school nostalgic things feel just as valuable, if not more valuable than they ever have.

Play from 1:32:05 8:88:88 1:32:05 You know, same with like actual photography and, you know. I love the billboard idea. think billboard ideas. Well, you guys might see us on a, you know, not going to be on the mountainside in the Hollywood Hills, but we might be, you know, maybe we'll be somewhere in next couple of weeks. Go slap a Midjourney fast hour sticker right on that sign.

Play from 1:32:30 8:88:88 1:32:30 But let's wrap this one up only because I got to jump on another call here, but. episode. mean, like, don't know that we're like based on how deep we went on mid-journey. I don't think we were thinking we were gonna go too deep on the cloth. So I'm glad we got to show you guys a few things. think if we obviously can go deeper on that, we'll see if anything changes between now and next week with mid-journey or any of the other models, like just game-changing stuff. RIP Sora, by the way. And yeah, we'll go from there, man.

1:32:32Wrap-up, RIP Sora, subscribe

Play from 1:32:57 8:88:88 1:32:57 But it's good to be back. Good to see you. And it's just been too long. So glad we were able to. to sync up and deliver for this for you guys. Hopefully it was helpful and continue to let us know what you think about V8 as well. You know, I know there's a lot of uh things that are fluid with it right now, but like, are there any insights that you've seen that we didn't mention that you don't think a lot of people know about? Drop them. Curious, you know. drop them and we have to reduce this algorithm because we've been off for a couple weeks.

Play from 1:33:24 8:88:88 1:33:24 So we will be back on here, which means also if you're still here, like subscribe, tell your mother, tell your friends, tell your dog, tell your we're, we're having to tell you nail stylist, TSA agent, tell your, uh, tell your, I don't know, dog walker, tell your local coffee shop owner. Um, yeah, we got them all.

Play from 1:33:49 8:88:88 1:33:49 Hell yeah.