Ep. 61: Live Visual AI AMA: You Asked. And We Went There

0:00:00We’re live, welcome to Episode 61

Play from 0:00:49 8:88:88 0:00:49 Happy Friday, everybody. Welcome to Midjourney Fast Hours podcast, episode 61. Happy Friday. Happy New Year to those that are still trying to get going. Because uh I think for about two, three weeks there, no one really knew what day it was, including myself. And full chaos ensued. Wasn't sure whether I should be working, sleeping...

Play from 0:01:15 8:88:88 0:01:15 I have no idea. changing the world, whatever. new, you just get over ambitious. ah Okay, so we are gonna be taking your questions live today. We think we got this set up the right way. LinkedIn makes this way too complicated. I was just telling Rory, had, nothing makes you more nervous than going into managed streams and seeing two streams that look completely identical and hoping you pick the right one.

Play from 0:01:46 8:88:88 0:01:46 So. one? I think we did. I think we picked the right one. So uh anxiety was at 100 there for a minute. We do got some people in the chat. What's up, everybody? Drop where you're from. And then this episode specifically is going to be revolving around your questions for us. And hopefully we're just going to have a real sort of cool conversation on the state of things. Yes, this is mid-journey focus, but we go beyond that.

Play from 0:02:15 8:88:88 0:02:15 So for those of you that are joining and you don't, uh regularly listen to the podcast. We always sort of tackle mid-journey and then we kind of get into the bigger broader bucket of the visual AI space. So image, video, we get into all the different models, some of the new releases, things that, uh you know, should we be paying attention to this or not? Does this affect our workflow or not? That's kind of thing. So it's good to see everybody here.

Play from 0:02:40 8:88:88 0:02:40 Yeah, let us know where you're where you're dialing in from. Dialing in, that sounds so old. Can you see me? Is my screen working? Yeah, yeah. Is this going to be recorded? I know that's coming, so this is recorded and this is going to go live on the podcast here this weekend. But yeah, go ahead and start thinking of any questions that you have. Drop them in the chat. We've got the LinkedIn comment section open so I can see the questions.

0:02:45What this AMA will really focus on

Play from 0:03:06 8:88:88 0:03:06 Rory, I don't know if you have that, so I'm happy to relay those questions and we can just kind of tackle. I gotta do Riverside window and then LinkedIn window and then mute and hope that we don't get the reverb and whatever else so I can see what's going on in here too. It's like a command center. need two screens for this. Do have a two screen setup? All right. I don't I don't use two screens other than stuff just like this because I would rather have like a small little screen where I can just focus and if I got too big screen it just takes my mind everywhere.

Play from 0:03:36 8:88:88 0:03:36 um exactly my thoughts. Chaos. knew you'd feel the same. And we've got, let's say hi to a couple of folks that familiar names here. Winston Mayo, what's up dude? I think he's in ATL. We've got Spirit here, Lauren here, Chad here, Valerio here, Giovanna here, Alvaro here, Colleen, what's up everybody? Happy to have you guys here. Thanks for joining. We done this in a long time.

Play from 0:04:02 8:88:88 0:04:02 It's been like months, like maybe like six months, I think. Has to be. it been? Because our first one was this, wasn't it? It was just like, let's just do a LinkedIn Live and then we'll just figure it out from there. I think we did like three or four LinkedIn lives and then we're like, and then we're calling each other on phone. were like, we should probably just do a podcast, I guess. Just talk about stuff. There's a lot to talk about. know, luckily, luckily, the industry grew into a podcast format where back then it was just like, let's just ramble about stuff.

0:04:14From LinkedIn Lives to a full podcast

Play from 0:04:30 8:88:88 0:04:30 And now there's, you we get something new and radical every three days. So we have a lot of content now. But, you know, Insane, dude, insane. So why don't we jump right into it? We'll just kind of take this as they come in. Winston, great question right off the bat. David says, new mid-journey will beat Nano Banana Pro. What do you think? I listened to part of the office hours, most of it.

Play from 0:05:00 8:88:88 0:05:00 I don't know that he said that verbatim. My takeaway was uh he talked a lot about the edit. You know, and he just said that like, look, like six months from now that technology may be like archaic, but I do feel like he was saying he feels very bullish on what they're going to put together for Midjourney in terms of the edit model. um When that comes, though, I think that's the biggest question, right? Because we V8 for Midjourney and a new UI that's both coming here in the next month.

Play from 0:05:32 8:88:88 0:05:32 And so that's going to change a lot of things already. We'll talk about sort of like what that means in terms of the jump. But the edit model I don't think is going to be released at that same time. The way that I perceive that is that's going to be probably another couple months after that. So I don't know. mean, we're talking about Google Nano Banana Pro right now. Are they going to have another model before then? Like they released Nano Banana and then Pro two months later.

0:05:34Midjourney V8 expectations vs reality

Play from 0:06:00 8:88:88 0:06:00 So we could be talking about something completely different by then. What's your... What's sort of your take on that? I'm not sure if you got a chance to sit on office hours or not, but just in general, what's sort of going through your mind there? I did not this week and I was on calls, but I read the notes and I always feel like I miss context when I read the notes versus listening to it live because I don't get to hear the tone. Remember when we were hearing David's tone about stuff and he got real excited about video and video was good and V7 just sounded like whatever, we're releasing it and then it was kind of whatever.

Play from 0:06:35 8:88:88 0:06:35 mean, it's good, but it was like whatever. ah So I don't really know. mean, this one, look, I'm starting to get the V8. uh V8 sort of vibes that it was like yeah, it'll be in December. Yeah, it'll be in January Oh, it'll be here February. Oh, then we'll get the you know Then it'll be like the launch off model to the next model which was like the whole thing with v7 So like oh, yeah, it's gonna be great and then it's like our data set sucks But we're gonna do it with the old data set anyway and then so, know, I don't know like I hate to be like down on them, but it's like a it's like a Wait and see approach for me.

Play from 0:07:07 8:88:88 0:07:07 Like I know that I'm not gonna be like overly disappointed. I'm just not gonna get overhyped on timing. Yeah. feel like, yeah, and it feels like a lot of people have started to jump ship for, you know, probably a couple of reasons. Like one, you have more options, like legitimate image gen models now out there. So there's that. But then, you know, now thinking about like just the capabilities that come with the edit model with NanoPanana Pro, like you're already going to spend more time outside of mid-gern just because you don't have that capability right now.

Play from 0:07:38 8:88:88 0:07:38 I know for me, like I've spent far less time in there than I had prior, just because a lot of the things that I'm doing require me to use that type of technology. Once you have that, it's hard to sort of pretend like you don't, right? So I mean, I think from that standpoint, somebody's asking, to Andres asking, what's the best workflow? Start in MJ, edit in Nano Banana Pro, or get it right in Banana Pro and bring it into MJ later?

Play from 0:08:04 8:88:88 0:08:04 ah And then there's still things with like Midjourney, right? Like someone emailed me that took my course the other day, right? And they were like, I'm trying to get this this image of a diver, like diving upside down, like down towards the bottom with like the surface at the top and like Midjourney can't do it. I'm like, look, I mean, regardless, I got it on Midjourney and like the first try because I was just using negative prompting, which I think is like, like I said, I think it's one of my biggest predictions for 2026 is just negative prompting way more important.

0:08:05MJ vs Nano Banana Pro workflows

Play from 0:08:33 8:88:88 0:08:33 But I did it in Nano Banana in one click. know, like in one, it was not even a problem. like, hey, can you do a diver diving upside down, like towards the bottom? Enter, done. Like, you sometimes it's it's like simple things like that, which it sounds like with V8, there's going to be a lot more emphasis on control. And with that, with that side of things, which I mean, look, the aesthetics good enough, just give us control. Right. Like I think that's important.

Play from 0:09:00 8:88:88 0:09:00 The the edit feature, like I'm starting to think more and more that that has to do. with like Y-Nano, Banana and Reeve and C-Dream and whatever are so good, it's because it's 4K output. Like, you can just get more pixels in 4K, and that just works better for consistency. So I don't know if they mentioned anything about that, like bass resolution being upped, because... didn't hear anything about that, actually.

Play from 0:09:25 8:88:88 0:09:25 kind of like if I'm doing math, right? Like it's just more pixels. That means more detail, right? Like that's like that you can see the difference when you run Nano Banana at 1K versus 4K in terms of the detail, right? Like it's significant. I think that's really interesting with the resolution too, because depending on what you're going for, and this can be a trap, like we talked about this before, you don't need to upscale everything. Depending on what you're trying to do, maybe you like that 1K look.

Play from 0:09:52 8:88:88 0:09:52 If you're going for something that feels more UGC or taken with somebody's phone or candid photography, I'd rather go 1 or 2K than go 4K. It's going to be too crisp otherwise. It doesn't match the output that you're trying to achieve. Like there's also that component, but I think what you bring up is especially interesting because I use the 4K specifically when I know there's going to be a lot of text involved, right? I mean, you have to run it up to 4K, especially the small text.

0:10:15Resolution, text, and why pixels matter

Play from 0:10:22 8:88:88 0:10:22 it's just pixels, right? Like if you think about it, it's just you need more pixels to get smaller fine detail like that, which now I'm convinced was probably like the whole thing. Like all along, you know, it's like, why can't we get the perfect text? Like, well, you know, it's just the base resolution is not good enough and the upscalers don't know what you're trying to do. So it just muddies it out. Right. So there's like this weird workflow where you were trying to sort of Frankenstein stuff together and it wasn't working the right way. But, you know, that's That's just my theory on it.

Play from 0:10:49 8:88:88 0:10:49 I'm probably way off. the text thing is interesting because I saw there was something on text on there as well. David, I feel like that was part of the notes. Was that? Yes, we'll be able to do text, but not paragraphs. Yeah. So, I mean, I think that's fair. I just... how much, you know, and to what extent, because I think like that, when I think about the text component, I really think the advantage there lies in sort of like that edit model component with the text, you know, not just trying to get it straight out from the prompt itself, eh if that makes sense.

Play from 0:11:28 8:88:88 0:11:28 But I'm just thinking about it like from that standpoint of... text, like I really didn't realize that I wanted that much text until I had Banana Pro, right? And then all those capabilities started flooding in. And there was that big difference between Nano Banana and Pro too with um sort of the amount of variables that you could also get right with the text, right? Like we've talked about this on past episodes where it's like I had a specific project with a lot of variables I just couldn't do in Banana and then Pro all of sudden unlocks it.

Play from 0:12:01 8:88:88 0:12:01 So I don't know, it would be really interesting to see though what Midjourney comes up with on the edit model side. That is really a big component of potentially bringing people back, keeping people there and not jumping to different tools, right? Because I think a lot of us to your point, Andres, are splitting tools across probably a handful of different things right now. And so that's also another reason why that sort of node-based sort of thing also has been nice because you don't need the subscriptions to tend to 15 different tools.

Play from 0:12:37 8:88:88 0:12:37 So I don't know. The other thing with that, with the text, right? To me, it goes beyond like, oh, I want to have text overlaid or on products. It's like in your scenes, if you're doing something in a city, right? I want the street signs to just say something, or the business signs to not read flabbergash, just something random that I have to go in there and edit. If you do it in a lot of these other tools, it's like, know, it's...

Play from 0:13:04 8:88:88 0:13:04 marshals or whatever on the on the street sign or on the the store sign right where it's just like okay that that like fits I don't have to go in there and like totally like Manually edit and then play with it and the grade the quality and whatever it's like simple things like that sometimes where it's really helpful which I never even think about with nano or Reeve or seed seed dream or whatever um what you think about Seadream 4.5.

0:13:26Seadream 4 vs 4.5 honest reactions

Play from 0:13:30 8:88:88 0:13:30 If you want like a real opinion on it, I'm not as impressed with it as I was with 4. Like I like 4 more for some reasons. Like 4.5 doesn't seem to do what I ask as much as 4 did, even if there is like better... I don't know if it's better aesthetics. Personally, I just like... I like version 4 better. I find I use it way less with 4.5.

Play from 0:13:56 8:88:88 0:13:56 I actually use 4 a lot more. And four to me is one of the ones that can match like Midjourney's artistic style fairly well. It gets the color down pretty accurately. Probably the closest one of the bunch for getting like the mid... Midjourney doesn't get a lot of credit for the color. Like the colors are great.

Play from 0:14:21 8:88:88 0:14:21 Like it's really... And that's not saying it because we're stands. It's like try to go replicate some of your more intricate color work stuff. Yep. Yep. that's, you know, to me, that's sort of where I use it. But 4.5, I've been underwhelmed with. I've been underwhelmed with Flux 2. I know I'm probably not using it at the level that some people are, but wasn't like, you know, what it's like. If I have the choice between Flux or Nano, I'm like, well, why just, why test to test?

Play from 0:14:49 8:88:88 0:14:49 I know what I'm going to get with Nano. You know, it's just like, I know, I know I can do it. So the... And yes, Winston, the new one messes up way more than four. I don't know what it is. It really it annoys the hell out of me because it looks good, but it's you know, messes up product details. It messes up character details and it's not nearly as consistent as four was. But you know, what are you gonna do?

Play from 0:15:14 8:88:88 0:15:14 There's a go ahead. was just gonna say, I don't know if you had any last thoughts there, but Spirit was asking too if you had used Runway 4.5 yet. Yeah. Yes, I have. It's good. mean, can you add an image yet? Like, I haven't tried it since they haven't added, you know, image to image to video because text to video works. It works. You know, it's good. I'm just like, I don't have a use for text to video other than for playing around right now.

0:15:15Runway 4.5 and the Nvidia signal

Play from 0:15:41 8:88:88 0:15:41 um Like, I need some control on that stuff. Like video to me, I'm trying to as it gets more expensive. I'm sure everyone here knows this. Like it's harder and harder to play. than it used to, where it was just like, whatever, I know what I'm getting. Now it's like, I don't just go in there and willy-nilly go on video and just start running whatever I want to, because that's not...

Play from 0:16:08 8:88:88 0:16:08 I'm just wasting money now. So I need to use it when I need to use it. But I do focus a lot on images now. I'm way more back to the image side than I was on the video side. But yeah, Runway 4.5, I like it. that don't have use for it yet. But yeah, there's those I feel like I feel like what they did with Nvidia and I haven't even dug into it more, but I feel like that somehow is an important partnership.

Play from 0:16:36 8:88:88 0:16:36 just saw the headline, but it just feels important. I can't give you guys an expert opinion on it, but I don't know what it feels like. They're both on sort of that same path towards world building. And I think Runway sees it that way. And I think Nvidia sees it that way. It feels like a good sort of partnership in that sense. which, you know, that's really that's always been their focus. So I'm curious to see where that one goes. That's a big that was a that was like a subtle drop, but like maybe a more intensive sort of piece from there.

Play from 0:17:07 8:88:88 0:17:07 Yeah. know they're up to something. ah And there was somebody else. I mean, the question, too, was just, are you mostly mid-journey? It seems like just based on the comments here, a lot of people are still using mid-journey for what they want base-wise and then moving it into Nano Banana Pro. I think that's mostly right. I think that's what a lot of us are doing, just use case.

Play from 0:17:33 8:88:88 0:17:33 use cases are a little bit different, but I think like the default is probably that. um Grok, we talked about last time too, as sort of like this sleeper in 2026. I mean, they've really ramped up a lot of what they're doing on the image and video side. It's gotten good quick. know, there was, they were basic. I can't remember what exactly Grok was at this time last year, but I don't know.

Play from 0:17:59 8:88:88 0:17:59 Did they have image and video at that time? I don't think so. Right. So just kind of like came out of nowhere and I know they're going to continue to double down on that. So I think they're like a sleeper in this Midjourney. I think that's why we're all here. We all want to see it take a big step for V8 and you know, like with with V8 specifically in office hours, it sounded like the data sets a little bit differently than they thought. They're going to have the eight which we'll get in the next month and then they're going to have this really short period of time before they sort of like push into a you know, a second version of V8 or a V9.

0:17:59Grok as a sleeper visual AI platform

Play from 0:18:36 8:88:88 0:18:36 Because I think like just the way things shook out, they're really bullish on sort of like how big of an upgrade that's going to be. I'm just not sure it's all going to make it into V8. Yeah, that again, I got confused with the V8 is like a stepping stone to V9. And I'm like, is that just like a satisfactory release again? Like just normally something to like, to like give us something, you know, like, again, I'm sorry for feeling like down on this, but it's starting to become a pattern, right?

Play from 0:19:08 8:88:88 0:19:08 Like it's starting to become a pattern now seeing it. It was like V6, great, V7. it's starting to have some problems, issues, timing, long waits, you know, it'll be here in a couple days, oh, training issue, it'll be here next month, oh, training issue, it'll be here in two months, you know, then we're just sitting here sort of fumbling. But, you know, it's again, it's the aesthetic engine is number one, it's just missing certain things.

Play from 0:19:37 8:88:88 0:19:37 And it's missing sort of like the, you know, the editing that we talked about, I'm just I just wanted to be so I just wanted to be good so bad so I don't have to like totally ignore it right now. You know what I mean? Like this other than for my own playtime, which is sort of where I find myself is like when I want to play now that's where I go verse what it used to be was like it was everything. ah So feeling a little bit nostalgic.

0:19:42Is Midjourney falling behind?

Play from 0:20:04 8:88:88 0:20:04 would say this too, like, I mean, there are still some things that no other model is doing, right? Like I still think the whole mood board thing that Midjourney has is better than anybody else's doing in terms of like inside of a tool. Same with style references. Yes, you have tools that have style references, but I haven't seen any work as well as Midjourney or as be as variable.

Play from 0:20:30 8:88:88 0:20:30 or have as many options or be as flexible. mean, you can literally do anything in there. And then the personalization aspect. None of these other models have any sort of personalization. They don't know you from the next person that's using it. So I think there are like still some things in there that just overlook, right? And it's still why I think in a lot of ways we can gravitate toward mid-journey, especially on the creative side. Yeah, I think Antidin made a good point over here with reverse prompting and JSONs. Like, that's where I've gone and had a lot of success with Nano, which is like, you do have to do extra steps for doing style ref or like a mood board style, but it's very possible now.

Play from 0:21:07 8:88:88 0:21:07 And then it's like, you can lock in, like what I did with that Polaroid or the Kodak car that we showed, you know, a while back, like that's a style reference and a mood board. know, like that's a whole, that's basically JSON. like style extraction, computer code language to say like do this every time. Right. So you can do it. It's just Midjourney makes it easy. And they do it the best for for that quick, like very easy transition.

Play from 0:21:34 8:88:88 0:21:34 But yeah, just like personalization needs to improve because it's not doesn't feel like it was last time. And I feel like I'm being the I feel like I'm being the negative Nancy on this call. But like I am because I'm just you know, you just. If it wasn't so rapid paced everywhere else in the landscape, I wouldn't feel this way, but I'm feeling, yeah. the odd man because they're doing things completely different than anybody else.

Play from 0:22:01 8:88:88 0:22:01 They're the sort of quote unquote outcasts this entire field, right? Everybody else has investor money coming in, right? Like they are raising uh rounds with investors. Midjourney's not. There are pros to both sides, but one of the cons for Midjourney is they're just not gonna be able to be as fast. as nimble because of the resource bottleneck that they have, right?

0:22:29Edit models as non-negotiable

Play from 0:22:29 8:88:88 0:22:29 um So, but at the same time, right, right. is like, if I still can't do editing in six months, like, I can't use it. Like, I just can't like, there's no there's no point to me. I think we're all gonna need that in six months. That's what I'm saying. I can't not do... At this point, it's also like a... I feel like a father. It's like, let's go.

Play from 0:22:55 8:88:88 0:22:55 Let's step it up. Come on. It's time for some accountability, time for some like, let's move along here. Get it. Scrap the secret projects for now. Just get a model out to the consumer that are feeding the secret projects. It's how you guys are getting money. You're not getting investor money. That's us. I'm paying... yearly subscription fees and then overages from in most cases. So it's like, yeah, just like I said, I'm being negative Nancy on this one.

Play from 0:23:24 8:88:88 0:23:24 So you know, that's my that's my fault. It's just like New Year, New Year feeling like a little bit spicy right now. It's just like I want. It's all we want. It's not like I'm asking for asking for, know, yeah, not asking for a discount, not asking for anything else, just asking for an edit model that we can do consistency. the edit model yesterday. mean, that's just, that is really, I think, standard now.

Play from 0:23:51 8:88:88 0:23:51 I mean, that's table stakes at this point. it. Kling has an edit model. Kling didn't have an image model four months ago. They have an edit model. It's like, what are you guys doing? I will tell you this though. um Okay, so somebody was asking to like, and this is a good topic in general, what do you guys think about FreePik spaces? I think there are more people, right?

0:24:04Node-based tools and FreePik Spaces

Play from 0:24:17 8:88:88 0:24:17 Like playing with, yeah, almost like these canvas node-based interfaces because of scalability, client work, trying different tools that you don't have to subscribe to, all of that. ah I don't know that you or I have really tried pre-picked spaces in terms of like running anything, but we've poked around in there.

Play from 0:24:44 8:88:88 0:24:44 I mean, we did highlight it on a past episode, but I think Rory and even for me, whenever I'm doing anything node based, it's pretty much just Weavy. uh I've poked around in freepix spaces a lot. think it's good. There's a few things that I think would be better in there, but I like the interface. It feels like a more expansive version of Flora, but very similar, simple UX UI in there, to me at least.

Play from 0:25:12 8:88:88 0:25:12 But I do like how you can edit in Node. That's one thing that's nice, to use the image editor or the actual editing in an output, which is just like, let me edit this whole thing. And then that can be the next connection. That's one thing I really like. um There isn't a lot of batch processing room in there. feels very more so like uh less work flow, more like individual project workflow than like scalable tool workflow, which you would do with Weeby and Comfy UI.

Play from 0:25:45 8:88:88 0:25:45 But still, to me, I like to create that way. I know some people don't, but I like seeing everything, like seeing the branches. like seeing like my history essentially as we go through this stuff and then testing, right? Like the testing of different tools to see like, this it's not working with this one. It's not working with seed, you know, seed dream. Let's try nano or let's try Quinn or let's try whatever. Wish they wish there would be a few more things in there just like for my own personal stuff, but.

Play from 0:26:15 8:88:88 0:26:15 got to get it together with their servers. mean, NanoBanana taking, I mean, I could go on vacation and come back and I'm still waiting for images over there. I mean, it's really annoying for client work from that perspective. I don't know if it's faster when you do it in spaces. I wouldn't think so, maybe it is. But I've gotten to the point too, where I have to just, hey, look, if I'm trying to do something quick, I'll go to Weavy and basically just trying to get those images back.

Play from 0:26:42 8:88:88 0:26:42 much quicker than trying to do them in FreePik. I do think FreePik has the potential to be a really great platform now. I mean, they've got spaces in there. I do like the interface. I think it's easy to move around. We did an episode too where we showcased all those sort of like product templates and all these things that they already have kind of baked in there. I haven't even really gone deep into that. ah Maybe we can kind of share too, make this a little bit visual here.

Play from 0:27:08 8:88:88 0:27:08 Let me see. So if I go here, AI Suite, I mean, they've got this whole sort of like set up here, which I think if you're just getting started or you really want to dive into something, I mean, there's a lot of inspiration here. There's a lot of things that are already set up. um think this is.

Play from 0:27:33 8:88:88 0:27:33 interesting. interesting. They change it. I do think they changed some of this. There like I do have a I do have a little bit of confusion with with when I'm in there sometimes. So I stick to spaces. I know what's going on in spaces when I try to like navigate the rest of it. seems very seems very not clunky is not the right word, but sometimes get confused with like the naming conventions of everything.

Play from 0:28:05 8:88:88 0:28:05 You know what I'm saying? I don't know if that that makes sense, but yeah. this point, most of us are using several tools. It gets so confusing trying to master more than a few of these tools at any given time. I mean, it's just hard to navigate around. mean, mid-journey, right? We've all spent so much time and it's like second nature, but all of a sudden you're gonna switch to using FreePik spaces, right?

0:28:07Camera control and multi-angle tools

Play from 0:28:30 8:88:88 0:28:30 Like you just, it's a whole different experience and learning. How many of those can you learn and navigate well? I mean, two, maybe three? So this was, uh I'm skipping around though on this point right here, but our friend over at FAL, Lovis, developed a, developed a QuenLora, that's the multi-camera angle, Lora, but they built a comfy UI node for it, where you just like click and drag the rotation or the tilt to like where exactly you want the camera to be.

Play from 0:29:03 8:88:88 0:29:03 And it was like, you see it happening in real time. I'm like, that's awesome. Like that's kind of what. You know, think we talked about this on the last episode, too. Like, you know, alter camera angle would be something that would be great as a, know, UX tool on mid-journey, where it's just, you know, hit the button. You know, let's go rotate 90 degrees or, you know, tilt 15 degrees, whatever. That that to me looked I just like liked that application of it. There it is. Yeah. So like that was a that was a fun one that I was like, this is see, this is just like people want this, obviously.

Play from 0:29:34 8:88:88 0:29:34 Like you read the tea leaves, read the social. you know, conversation if we listen a little bit. It's like everyone likes this multi-angle thing. That's why we shoot. I guess that's why, you know, everything shoots at multiple cameras. Yeah. Like I was looking at it it was just like, you can actually indicate degrees, right? Like where it's like, hey, change 20 degrees, 90 degrees, 45 degrees. Yeah. And that's, it's, it's definitely, Quinn is like, nah, I'm not going to say it's like the most precise when it comes to it.

Play from 0:30:06 8:88:88 0:30:06 I'm just talking about like the feature in general, where it's like, this is, this is cool. Like, I like this. I want to use something like that. Right. It's like, that's, that's sort of what we're thinking about when we're trying to compose a shot. Right. I mean, like, that's sort of how my mind goes in that 3d space, but now there's like a an option to do it and you'll probably see, know, Higgs field will have a tool next week for it.

Play from 0:30:35 8:88:88 0:30:35 God, Higgs Field's always got a tool. I mean, they just did something too with, what is that? Cinematic, what do they call it? The Cinema Studio. on Higgs Field now. There's too much. When I go in there, and not like I use the tool often, I use it maybe every once in a while to test their new features, but like, what is everything? There's so, so much.

Play from 0:31:01 8:88:88 0:31:01 That's, okay, so that's my point, right? It's like, how many of these tools can you actually like navigate around well? I mean, it's just so, I'm overwhelmed. I'm so overwhelmed. The naming convention, I mean, you throw Google on top of that. I'm getting billed for something from Google. I have no idea what I'm even being billed for at this point. I don't even know the names of the products at this point. Like there's three different names for one thing and then there's 10 different names of the same thing that.

0:31:27Tool overload and UX fatigue

Play from 0:31:28 8:88:88 0:31:28 do slightly different things. I'm so lost and over that. I feel like there's a consolidation effort that is inevitably coming across the board, but I can't wait for that. Yeah, mean, they're just Higgs field just building workflows and then wrapping them like that's all it is. You know, that's if you can build that same thing in spaces, you don't have to play for Higgs field. If you're paying, if you can build that in Weavy, you don't have to pay.

Play from 0:31:53 8:88:88 0:31:53 If you build in comfy, you don't have to pay Higgs field just makes it very easy for people to be like, this is exactly what you do with this exact thing. Right. Which I think is probably good marketing play. most aren't doing that. It's mostly like here you have all the creative freedom you want, but then again no one really knows what to do with it with all that freedom. they've hit they've definitely hit something there, but they are they are you know, it's annoying the hell out of me.

Play from 0:32:20 8:88:88 0:32:20 But it's like that's the one thing with especially now you're seeing it with I think people are liking FreePik spaces. I think they're liking Flora. I think they're liking Weavy. Weavy's a little bit more technical. Comfy's more technical than that. But it's a design space, you know? Like, I think the interface has always been something that's, to me, felt a little weird.

Play from 0:32:46 8:88:88 0:32:46 You know, I never liked chat. Chat's the weirdest interface to me of all time to create images, right? It's just odd. um Then, like, what Midjourney has to me feels like the most appropriate for, like, uh studio, store of, design studio. And then if it's anything else, it's like, just want to use like canvas. It just like what I'm used to with any of the design tools, but that's just me though. There's, I'm wondering if someone's going to develop something new, like nodes feels very, feels very like the hot trend as to where things are going.

Play from 0:33:19 8:88:88 0:33:19 I'm sure someone will develop something new this year. That's, that's, you know, cooler, a little bit more interactive, but that's a. I'm looking at this, what activity going on in here in the comments section? I'm checking this. I'm trying to catch up because I was sharing my screen and I couldn't see it. Let's see what we got here. have professionals using freebie. Spirit AI video looks so fake and glossy at times.

Play from 0:33:48 8:88:88 0:33:48 MJ video makes a gritty photo realistic image into a CGI toy looking video to me. Like I'm, I'm also spirit for this one. I'm starting to think this is resolution based also. Like I think, I don't know if I'm correct on this theory, but I have a feeling that we get this plastic look because we just don't have enough micro detail in the pixels to get like my skin cracks and my beard stubble.

Play from 0:34:13 8:88:88 0:34:13 Right. Like I think that that comes with resolution. So maybe the base outputs for videos start need to be like better. Um, you know, from really 10 eighties, the biggest we've seen, I don't know if anyone does anything, does any video tool output at 4k base yet? I don't know. Maybe, but you have to upscale. I don't think so. Right. I don't think so. Well, I mean, there's probably some open source models that we don't even know about.

Play from 0:34:40 8:88:88 0:34:40 Someone on here is going to drop it in. yeah, can turn it. Comfy UI, duh. Some under-the-radar Chinese model that we've never heard of. Hey, will say this, shout out to Weavy, they don't charge you extra credits for running 4K Nano Banana Pro. Yeah, well, hey, it's. Meanwhile, free picks over here. charging extra for the 4K baby.

Play from 0:35:06 8:88:88 0:35:06 Takes longer, takes an hour. I don't know if you saw this, but we might as well. might as well. ah I watched a nine minute YouTube video on it the other day. It was like this whole Higgs field controversy for their for their like unlimited promotion that they ran. Yeah. I don't know, maybe. just happened to watch, I think we just happened to watch the same video, but I know what you're talking about.

Play from 0:35:32 8:88:88 0:35:32 Like they're banning people or something and like kicking them out of their accounts or something after. It was like they sold Nano Banana Pro like unlimited for the year and then people started using it. They're like, we're to kick you all out. Yeah, FreePik did like this thing where it was just like, yo, you're going to get unlimited generations. And it wasn't even just for like 4K, you know, in relax mode.

Play from 0:35:57 8:88:88 0:35:57 And then all of a sudden, like this pop up appears on free peg. And it's like that ends February blank. And it's like, what, bro? Like, what are you talking about? That's not exact. That's nothing like what you were telling everybody it was. And then they also started downgrading performance and you had to like pay more money to up the speed. So I was like, yeah, you have unlimited, but you know, now we're gonna, we're gonna strip the speed. So it's going to be like, you know, a couple of minutes per generation.

Play from 0:36:24 8:88:88 0:36:24 You can pay to get back up to normal speed though. This is so, so greasy. I just like, can you guys just not, mean, like I get it. People that are just like trying to bleed every dollar, but like, come on. so gross to me sometimes, some of this stuff. Not straightforward, but. got a question. Okay, Meg Balls here.

0:36:43Creative paralysis and decision overload

Play from 0:36:49 8:88:88 0:36:49 What's up, Meg? Ask any specific tools help you guys keep AI exploration from turning into creative paralysis for teams? Biggest headache. We need more context on what you mean by that. Yeah. what I'm saying. Meg, follow up with that. Curious what exactly you mean there. Yeah. What do you mean, Meg?

Play from 0:37:15 8:88:88 0:37:15 I have we have creative paralysis all the time. Like I sit there and think about 70,000 ideas that I have that I can do now. And then I end up doing nothing with it. that's the biggest, oh, that hurts so bad. just like, that's the other thing. I think FOMO is at an all time high when you're trying to keep up with this. You just see other people creating stuff and you're just like, I'd like to try that or I'd like to do something like that or I'd like to, that gives me an idea.

Play from 0:37:40 8:88:88 0:37:40 And then you've already, look down at your list of other things that you already have to do. And you're just like, there's no way I'm gonna have time to do that. Like I'm just not, it sucks. I mean, like that's, that part kills me. because there's so many fun things to try. I do think though, just word of advice, uh I don't know if X is a polarizing platform, maybe it is for some people in here, but I would highly recommend uh if you're not on there to get on there, if you're trying to keep up with things and see great use cases and examples.

Play from 0:38:14 8:88:88 0:38:14 LinkedIn is cool. There are some creators creating great stuff, including a lot of people here, but I just think in terms of like the volume, the variety, uh how novel the ideas are, uh seeing things in real time, like it's not gonna be beaten on X. And I would just highly recommend uh staying up to date and saving use cases and examples and inspiration and things there.

Play from 0:38:39 8:88:88 0:38:39 You just. business focused. It's more exploration and less like comment prompt for the whole workflow. If I see one more of those, I'm just like, my God, I get why everyone does it. if anyone's doing this, good. Good for you collecting emails and collecting different whatever, whatever you're giving away.

Play from 0:39:09 8:88:88 0:39:09 Just don't do it every single post. It's really annoying. Like it looks really, it looks really like it's the word I'm looking for. They desperate, but it looks like it's just baby. It's like salesy. It's a way too, it just doesn't make you look authentic. So it's just a, it's just a, again, I'm full of opinions today, dude. I don't know what's going on. look, people are trying to make a living. I think some people, though, are legitimately gating just to get followers.

Play from 0:39:38 8:88:88 0:39:38 And I'm just like, that is so dumb. People are going to get tired of you very quickly. And I think the other part is with X, you're not getting that. Just people just sharing things and giving, right? Which is... a lot of the mentality that you and I have with most of it. I can count maybe on one hand how many times I've gated something on LinkedIn.

Play from 0:40:07 8:88:88 0:40:07 And uh there are perks. You do get followers. You can put somebody into uh a nurture campaign and try to upsell them on this and that and the other. But guess what? You're going to spend a whole day responding to DMs, sending links, and for what? You know, like, I mean, and you're probably eroding trust along the way. You keep doing that. So I don't know. It's not worth it. Instagram's just so bad right now.

Play from 0:40:33 8:88:88 0:40:33 Like literally, yes, yes, it's more automated. Yeah. with LinkedIn, like, do it. I'm not saying don't do it. I'm just saying don't do it every single post, like every single post when it's like that. It's really annoying. And it's just obviously that, you know, it becomes like you obviously can't have something good every single post to give away. Right. Like or that you're just not giving anything away. You're probably not giving something good away for giving something away every post.

Play from 0:41:00 8:88:88 0:41:00 So like whatever. The other thing about X that I will say with that is that because LinkedIn doesn't seem to update the platform ever in terms of like the creative side of it, like X, you can go deeper on the threads. Like you can just explain more and explore more. Like I can do a video and then I can do an image or I can attach a few things and put a description. It's like a much easier format to follow than sometimes LinkedIn.

Play from 0:41:26 8:88:88 0:41:26 Like I have to like double stack. videos to do an explanation and show an output at the same time like it's a really I have to no you can't do a video and stills, you know, it's just, I don't know. Yeah. but also Jordan, to follow up on your point there, I don't think it's much about gating as it is about increasing engagement and boosting the algo.

0:41:33Gating content, growth tactics, and trust

Play from 0:41:53 8:88:88 0:41:53 It does do that. I'm just saying from a strategy standpoint in building a personal brand or getting business, to me, that stuff is not long term, that's not viable. If you're doing it once a week, good. If you're doing it seven times a week, it's like so cheesy and looks so like to me personally, I'm sorry if I'm offending anyone.

Play from 0:42:18 8:88:88 0:42:18 Yeah. Yeah. That's what I mean. guess what? People are gonna hate you. I mean, I'll tell you, and this is like the marketer's perspective. Should we gate this content? Should we not? I think most of the time it's no. But if you are, that shit better be worth it. You know what I mean? Like we got people over here, you know, gating prompts that are very specific to something. It's like, why do I need that?

Play from 0:42:44 8:88:88 0:42:44 Why would I need that? You know what I mean? Like I could literally take a screenshot of that image right now, drop it into custom GPT or take it over to FreePik. And I'm going to get something that's pretty damn close to what you just got. What are you talking about? You know what I mean? I, I also do think that's pretty underrated right now. If you guys are trying to do things at scale, right? It is not scalable to come up with prompts one by one in this format with clients.

Play from 0:43:11 8:88:88 0:43:11 You need a way to scale that out. So that's either through a node-based tool where you've got the system prompt or you're dropping something in. It's incredible how well that custom GPT of describing images is held up. I still use it all the time. I can take exact examples, dial in how close I want to get to that output. That works well for clients too, because then all of a sudden I'm just dropping something in and I'm getting something back and I don't have to expend.

Play from 0:43:43 8:88:88 0:43:43 any mental energy and trying to come up with that prompt, right? Just iterate, iterate, iterate. There's that's important. I think the other thing just to not to blow past that, but I saw a comment on here from from Vesica says, how do you suggest we deal with our intense dislike of X? There's a there's one thing I've done, which is I don't know if it's helpful or not. If you don't want to use X, don't use X. But it was like I created one like burner account that is only AI.

Play from 0:44:11 8:88:88 0:44:11 I only follow people in AI and my it's the bubble. That's what I call it. Like, there's nothing else. There's no politics. There's no anything. It's just AI content. So it's like, I can go I can just switch to that one and get rid of all the noise. And it's just AI stuff. I did I have that one like that for Instagram also, where I only follow AI stuff. So my whole feed is AI. So that's like, you know, that's just one way, I think, to do it a little bit. If you don't like X, I don't have another better suggestion.

Play from 0:44:39 8:88:88 0:44:39 But that's sort of, you know, one way to get around that. But yeah, there's this. I don't know why why I'm I don't know why I'm on one today. I gotta stop. I'm sorry. I'm sorry. It's new year. There's a like thinking I had too much time to think. That's the problem. come on. That was I was just so I was just so invested in the uh so invested in like the online sort of forensic psychology or the forensic like people out there crazy.

0:44:44X vs LinkedIn for AI discovery

Play from 0:45:14 8:88:88 0:45:14 since Game of Thrones, you know, like where people just have all these theories of what's going on and what's gonna happen and, dude. detail people go into. know, like that was like I could never do that. But like people go into and breaking the showdown like frame by frame for like every episode of every season. Like where you guys are obsessed, but I'm interested in your obsession. You know what I mean? Like that was that was an impressive. And then Netflix sort of like fed into it, which I thought was like, this like, you know, you're like gaslighting like real hard.

Play from 0:45:46 8:88:88 0:45:46 That's the whole this whole community. But yeah, there's there's like There was some really interesting stuff I saw done with AI, like people redoing the final with AI. It was actually not bad. know, Demogorgon Wars and stuff. like, all right, know, fan fiction. I don't know, speaking of that, did you see the Star Wars sort of rendition that was done with AI recently? uh I thought it went pretty viral.

Play from 0:46:12 8:88:88 0:46:12 To me, really good. I don't know that I've seen that I thought it was excellent from a voice matching, face matching, like aesthetic matching. I thought it was very good, but there was a, that went to me, if I can find it here, I'll send it to you. But this one, that one was, that one was pretty like eye opening to me. like, Oh, it doesn't look like crap. wants to know if, uh I think it's Alanka, she wants to know if Quinn isn't Weavy.

Play from 0:46:39 8:88:88 0:46:39 Do you know that, All-Fan? Yeah. Quinn's in there. um Also, this is just a fun, fun trick. um Let me pull this up for you, just in case anyone here is using Weavy and you don't know how to get some models that you're interested in trying in there. Like this is, I'll share my screen. Like uh if you ever want to add a model that you don't see in there, like if you're in Weavy, right, let's just, I need to.

Play from 0:47:09 8:88:88 0:47:09 minimize this. Sorry, this needs to go up here. But let's just create a new file. If you want to import something, you can go to FAL, and there's obviously a ton of models here that you can pull in, depending on what you want to do. But let's just say you want to quen whatever. Say we wanted to do this.

Play from 0:47:35 8:88:88 0:47:35 If this one's not in there yet, if you go into the page for it. This copy link here, you copy this and then you can go, I don't know if this one's going to work or not, but if you right click and you do import model and then just paste that link, typically it comes up. Right, see this one's not ready for Weeby yet, but if you were to go to, you know, let's just say one of these other Quinn ones.

Play from 0:48:01 8:88:88 0:48:01 How do you know if it's ready for Weavy? See how the input just says negative prompt? That's it. Like it's not, like it hasn't been adapted yet. They do a good job over there of getting a lot of this stuff in there. But like if there's, know, Quinn, maybe Jetit, this, like, let's just see if this one works. If I go in here and probably cause we're doing this live, it won't work, but I import model, paste, right? There you go. So now if this model wasn't listed over here, you can bring it in.

Play from 0:48:28 8:88:88 0:48:28 So that's just a little hack to, you can do this with foul. replicate, um forget which other one, which other one can you do here, import model or civet you can do from replicate, foul or civet AI. So that's just a, hey, that's sometimes, you know, they're sometimes they're, not in there immediately. They do pretty good with getting stuff in there relatively quick, but that's one way to, one way to do it.

Play from 0:48:57 8:88:88 0:48:57 uh What other questions we have? feel like I'm missing so much in this chat. There's so much going on. You guys are awesome right now. Let's see. what we got. Are people still using Lora's? Does it hold space and AI at this? I think for really technical uses, yes. um You know, I think if you're trying to get by on a day to day and you don't need like immaculate, like immaculate precision with immaculate sort of use cases, I don't think you necessarily need one.

0:49:11Are LoRAs still relevant?

Play from 0:49:27 8:88:88 0:49:27 But I'd say for, depending on how, like I think when you're looking at things like you utilizing this stuff for, you know, training your own models and utilizing stuff on, you know, local hardware. so that you can get around a lot of the cloud stuff and not having to worry about the legal things. And there's a lot like nuance to it, which is something that Laura's probably uh have a really good space for. like if I just need five photos of my skincare products in different angles, like I'm going to Nano for that.

Play from 0:49:57 8:88:88 0:49:57 There's no reason to waste all that time training Laura. yeah, that's again, these are just my feelings on the matter. I'm sure a lot of people love Laura's and you know, there's a lot of uses for them, but. It just sort of depends on what you're Yeah. like it's, I think we're at a point too where you're trying to do everything that you can in just a select number of tools. You're either doing that or you've got to be operating into a canvas that has all the tools aggregated in it.

Play from 0:50:28 8:88:88 0:50:28 um Otherwise just kind of hopping around to different platforms is just too overwhelming at this point. So I think like, again, from that standpoint, Laura, I'm not gonna go to a different tool or platform just to accomplish that unless I really had to. Yeah, yeah, I get that. There's... We have so much going on in here, like can't even catch up. way behind on the comments. I'm like six minutes behind on the comments. I can't see. I just got Jordan's question about which GPT drew.

Play from 0:50:56 8:88:88 0:50:56 So you may have to sort of like take that over. I would say, Jordan, so I can share it. um Maybe I'll just send it as a comment here in the chat. Let's see if we have, I'm trying to figure out where, God, you guys are awesome. Like I don't mean to say this like three times, there's like a lot of good combos going on in here. um Let's see.

Play from 0:51:27 8:88:88 0:51:27 Let's go, let's just start at bottom here. There's so much going on. Who in here lost their Magnific credits? I got a lovely Christian, lost over 14,000 credits. Hold on. Let's go, let's all go check these Magnific credits and see if they're still there. Mine are still there. I checked mine, but I will tell you, I got Magnific Year Plan half off last year and I completely forgot about it. And then this year I paid full price and I was like, damn it.

Play from 0:51:54 8:88:88 0:51:54 that definitely happened. did that with a lot of tools like Black Friday last year, like Minimax. I was like, what? Why is there such a big charge here? And then it was like, uh oops, didn't see that one coming. Forgot about it. Let's see. Where is my account now? this might not be good. Uh-oh.

Play from 0:52:20 8:88:88 0:52:20 Newbies don't necessarily fall for the comment route. What were you gonna say? Go ahead. this isn't looking my Magnific account. This one got me a little nervous. Let's see. Maurice credits. so many credits to keep track of. I hope not.

Play from 0:52:46 8:88:88 0:52:46 Let's see, next. credits, the new Bitcoin. People just come in and take them. I mean, I could see some of that stuff happening with like... Have you seen like the amount of money people are spending on... While I'm looking at this, this is just like a very random... uh is a very random thought. Have you seen on X how much people are spending on Cloud Code? I've seen someone that was like...

Play from 0:53:12 8:88:88 0:53:12 Their bill was like $200,000 already. It hasn't even been out that long. What? uh You spent $200,000 on Claude code. But I got, like with Claude, I got rate limited pretty hard the other day. I was like, this is like maybe 25 messages. And it was like, you've used all of your credits. I'm like, what? Like, what are we talking about?

Play from 0:53:39 8:88:88 0:53:39 This is the first time I've used it. my month started like four days ago. I haven't even used it. So I don't know what that was about, but. Maybe I don't know if Opus is more expensive. If someone in the comments knows the deal with that, that was like very shocking to me that I had already reached my rate limit after like 25 messages. But some of the quad code stuff I've seen has been pretty, pretty awesome. Wild, Claude code is sick.

Play from 0:54:07 8:88:88 0:54:07 I need to spend more time in there. Another thing I need to spend more time doing. The other thing is, know, Vivian brought up FreePlink just released variations. I just clicked on that link, opened that up. um I don't know anything about this. I hadn't even seen this. this um is...

0:54:40FreePik Variations first impressions

Play from 0:54:41 8:88:88 0:54:41 One visual is all you need to create. Create complete narratives, turn a single image into multiple frames, build a storyboard. Okay, animated. Okay, so I think. this just happened today. That's why. What is it though? So turn a single image into multiple frames, build a storyboard and animate it into a full video. Okay, but I don't know if that's one tool or if this, so here's what I click when it brings me up to it, right?

Play from 0:55:09 8:88:88 0:55:09 So I'm seeing this under tools and FreePik and it says variations, skin enhancer, image upscale, background mover, mockup, change camera and lip sync. Huh, so it's doing stuff like, like Higgs field is? I don't know, but if I click this, what do we, okay. uh Hold on.

Play from 0:55:37 8:88:88 0:55:37 Trying to navigate. Got myself... Alright, let's look now. oh at something here. See if I have like a good example. try to keep up with some of these comments now. All right, let's just bring this up. So use this image. So this is, let's see what happens here.

0:56:08How much creators actually spend monthly

Play from 0:56:08 8:88:88 0:56:08 at Astrid's asking also here's a here's a good one how much do you guys spend a month on credits and licenses uh like a full-time salary unfortunately I would say even just between Google and now I'm going to definitely I think I'm gonna roll back on my chat GBT one which I do have the $200 a month one yeah I mean I use it the thing is I have I have it's so it feels like with chat GBT now this is just my own sort of like analysis of it feels like when you've been using Mac for so long that it knows all of your stuff and everything's connected and it's hard to get off of it like all my project folders and my GPTs and everything it's like harder to migrate than it is to stay on so it's like like between those two at the $400 a month range I think between both of those

Play from 0:56:59 8:88:88 0:56:59 and then Weeby really is like my only other one right now I think the other ones I pay I pay for Kling ah Suno pay for What else Claude? I'm probably in like a seven eight hundred dollar a month range on this stuff if I had to guess Like between everything and then like the overages and on the credits and stuff like that. But again, it is my full-time job So it's like this is like my whole Like my whole life is this stuff.

Play from 0:57:28 8:88:88 0:57:28 So it's I would I like to turn that down? Yes Do I often another thing that I do pretty often is like subscribe? something and then unsubscribe and then resubscribe and then unsubscribe like when something new drops it's like been easier to do that to like test stuff and then just like drop off so i don't forget about it um or just like you know unsubscribe and take the new promotion whatever it is because that's that's definitely a thing like Higgs field i do that a lot sorry Higgs field just like we go.

Play from 0:57:59 8:88:88 0:57:59 mean... What's it doing, by the way? So I'm changing the camera angle so you see this rotate. uh So I can rotate this any way I want and change the verticality of it. let's go down. I wonder if it can't do that on the floor, but we can try. Guessing this is using Qwen This...

Play from 0:58:24 8:88:88 0:58:24 Yeah. Interesting. Not bad. Details off though. Yeah, that's a pretty minimalist product too. right, so this is the thing, right guys? um And this is where it really is tough for us because this stuff is all fun.

Play from 0:58:50 8:88:88 0:58:50 Now, when you have deadlines or projects that you're working with, which I'm sure a lot of you maybe do, it's a totally different standard of what we're trying to accomplish here. I mean, these things gotta be on point. I mean, this one... Okay, glare may be a little bit too strong. It's not bad, but you get some, I don't know, just like where are all the details on the floor here?

Play from 0:59:18 8:88:88 0:59:18 Let's see if that's the same throughout. we shooting this at a, yeah, we are shooting at this at a low aperture, so maybe that's believable. I don't know, this is interesting though. I'm glad you dropped this. This is worth exploring and playing around with 100%. Very interesting too that it just kind of throws it in there. You have no idea what model this is or anything. That's the idea.

Play from 0:59:45 8:88:88 0:59:45 What's the token cost or the credit cost on that? That was unlimited. Really? All right. All right. I mean, like, I never know. Like, you know what's weird is that sometimes the stuff, it's like starts unlimited and then like it just disappears. Like, I feel like that was, I like that was with FreePik Spaces. I was like, yeah, use a Nano Banana for free. And then it was just gone.

Play from 1:00:10 8:88:88 1:00:10 Yeah, then it was gone. But how, mean, obviously it's a promotion. I get it. But that's why I just use. that, and then they're just coupling this with all these other things in here. mean, Image Upscaler they're using Magnific for. Skin Enhancer, maybe that's the same thing. skin enhancer that might be uh that's probably magnetic Sharpen and SmartGrain, that's what they got in Magnific.

Play from 1:00:35 8:88:88 1:00:35 Variations. I don't know, this looks super interesting. oh Have to play around with it. Can you use Magnific in FreePik, like the way you can use Magnific in Magnific? Like with the sliders and all the models and things like that? You can, yeah, you can. I haven't really done it that much.

Play from 1:01:03 8:88:88 1:01:03 So like right here. The last time I used spaces I couldn't use, I could only use like just like select precision upscale or creative upscale and like that was it. um But it's been a while since I've been there. Okay so you can use, all right so you can use it this way. But I don't know if you can use it. Can you can you go to... No. This is just a classic. So this is not precision.

Play from 1:01:29 8:88:88 1:01:29 This is not precision in here. So there's the creative upscale and the precision upscale, but that doesn't have it on there with the model drop down. Huh. See, this is what, not blasting everyone. It feels weird sometimes. It doesn't FreePik own Magnific. Do you want to have all the capabilities there?

Play from 1:01:56 8:88:88 1:01:56 That one's a little odd to me. Yeah. Yeah. To me, like this is a perfect example is that even though they got that, it's like, well, if you don't have everything in there, I'm just gonna go to Magnific's website. huh. Yeah, it's- FreePik, I want to make that feel as seamless as possible to the real experience where no one leaves the. Yeah, this is now again, because I promise we're not we're not on I'm not on one like I'm not I'm not just sitting here complaining about stuff. It's just maybe it's our time because everyone's asking the right questions that it's like these are all bringing up all the things that are like deal with every day like this one.

Play from 1:02:30 8:88:88 1:02:30 I remember specifically like why can't I use like some of the other features from Magnificent here? You know, same thing with Mystic, right? Where it's a because I like Mystic. I've always I feel like Mystic is is underappreciated, but it's a to me. It's a good good image model. Like I like Mystic a lot. It's not bad. It's really not bad. And it can get, again, from the bass generation side of things, it gets really solid resolution.

1:02:493D workflows and what’s coming next

Play from 1:02:59 8:88:88 1:02:59 they got Reaven here too. I didn't even notice that. have Reeve. Reeve has an API, but I don't feel like a lot of people have it. Weevee has it. Obviously, FreePik has it. I don't know if Flora has it or not. But Reeve's solid. Another good one. So much. But let's see. Let see what else here.

Play from 1:03:25 8:88:88 1:03:25 see what other questions we have. spirits got the right got the right mindset over here with Black Friday typically those are the best deals for everything that's basically what all my tools re up is right around that I mean, it seems like every year I just get closer and closer where I'm just like writing a list of things that I'm just going to wait until black. Yeah.

Play from 1:03:50 8:88:88 1:03:50 drop 10k. But a lot of them have been like, you know, like perplexity got me with like, need to, you need to like, you can, you're not grandfathered in anymore. Same thing with minimax did that or Hylua did that last year. They're like, you're not grandfathered in anymore on this awesome deal. Kling honored it. So Kling shout out same price as last year. Didn't have to be like, the models are better.

Play from 1:04:16 8:88:88 1:04:16 You know, it's not like, it's not like when I got that notice from my health insurance company during COVID it's like, due to the rising cost of, you know, hospitalizations here, we have to raise your health insurance. I'm like, cool. You know, thanks. What does that have to do with me? You know, like, that's not, yeah. You know, like the company that I don't feel bad for at all, you know, is my health insurance company who's just raking in the cash off of us.

Play from 1:04:42 8:88:88 1:04:42 But again, not to get political on here, but there's, what was he, have you playing around with, hold on, what did I wanna? There was one. Let me pull up my X saved post. There was something that I saw here that I wanted to bring up while we have like a little lull in the action here. Let me. That's what I'm saying. Like there's so many things I've just got saved where I'm like, I want to try that. I want to see if I can make that work for this.

Play from 1:05:08 8:88:88 1:05:08 And it's just like, man, I'd run out of time. I was playing around with some of the 3D stuff the other day to see if I could get it to work. um Yeah, it was really cool. was building like little... with Hunion, I was actually doing some stuff in Comfy, but was doing some 3D characters. It's amazing how fast you can rig some of this stuff.

Play from 1:05:35 8:88:88 1:05:35 Like, I feel like that was... Again, I know nothing about 3D, but I was looking at... Now I could auto rig some of these things so you can just use like a basically standard motion to it. I feel like that's where things are going this year. Right. Like I feel like there's a lot going to be a lot more 3D. Like we're seeing a lot more of the application start to pop up. It's fun as hell. Like remember when Midjourney was like, we're going to do 3D.

Play from 1:06:00 8:88:88 1:06:00 And then like that just disappeared. Just another add it to the list. You know, it's another one, another idea. But there, it feels like that's the next, that's the next sort of horizon, I guess, to conquer is the 3D space. Cause it's super hard. It's super technical.

Play from 1:06:26 8:88:88 1:06:26 And it's like, I get why it's so hard. Even working on some of the things for projects for clients within the VFX CGI space. It's like, all right, this is super complex. are really damn good at what you do. But no one, as Winston said here, no one talks about Hoonion, Jenmo, Wan Meshy Yeah, I mean, like, in Weavy which I think is one of the better things about it is being able to generate a 3D model and then use it in the compositor to sort of like stage what you're trying to do.

Play from 1:07:01 8:88:88 1:07:01 Rodin in there, Rodin, however you pronounce it, ah Trellis, those sort of like, they're... pretty damn good at getting something usable. And at least like for a reference image for Nano Banana, right? Or anything else that you're going to utilize there. It's a good one. Have you guys worked with any 3D AI projects? If yes, what have your experience been? Like it's, it's tough.

Play from 1:07:26 8:88:88 1:07:26 Like I've been doing some stuff in the CGI VFX space. It's the precision is what's important. It's also like the, the scale and sizing of everything. starts to get like when you're doing stuff with AI like it just especially with products it's like not true so um it's tough like I you know you can see why there's a hole there and why there's a need for more products but certain things like we've been able to do some cool stuff with I wouldn't say was specifically 3d more CGI VFX but like depth maps um you know different different things with like lighting and halation and stuff like that where we're taking or extracting frames from like real videos and then changing them with nano and then doing

Play from 1:08:14 8:88:88 1:08:14 them with, you know, basically like restyling them with Kling to then like lay out. It's a very, it's a very meticulous process to get there, but like you can get there. So it's, you know, I feel like there's someone that just missing an opportunity with some of the stuff to build a product, but it's, it's kind of like, How I feel, feels like how we used to have to do like, uh, consistent characters, you know, like how many, like hoops and like how many steps and hoops we had to do sort of get like one face to transfer.

Play from 1:08:44 8:88:88 1:08:44 Um, that's sort of what it feels like right now, but just. Yeah. Just the 3d stuff for fun though, like to build like little characters and like auto rig them and like, you apply some motion to them. Yeah. Like that stuff's fun, but like I'm feeling like for, for real full scale 3d projects. pretty intense still, but... I mean, I think too, we're getting closer to there being more real world application of bringing AI creations to life.

Play from 1:09:15 8:88:88 1:09:15 You know what I mean? Which is also exciting, right? Like right now, for example, it's like, OK, we could create cool images, maybe we create merch, right? Like with images that we created, maybe we can 3D print something, blah, blah. But I think as the technology continues to grow, we're to be able to turn these into from idea to real things. And I think that's exciting too, just as a creator in general, right? Yeah, yeah. Well, we're going to we're going to move into world building probably this year.

Play from 1:09:44 8:88:88 1:09:44 I mean, the rate of everything's going like we're going to be able to construct stuff. I would guess. mean, you can already. Well, you can already go to a lot of these 3D tools like trellis or Tencent or whatever, and just it's it's text to 3D model. You know, you can just like plant. Boom, there it is. And like you can go and place that in a space, but. I still think you need to use like 3D software to actually like place and structure things and like go in there.

1:10:10Strategy vs experimentation for teams

Play from 1:10:12 8:88:88 1:10:12 So, you know, I think that's where that's where it's going. Because, you know, but oh, Meg, Meg responded to us. Did you see, do you see that one delayed, delayed reaction to the questions question, but we're making way more stuff than ever, but deciding is the hard part. Any tools helping with curation or approvals beyond chat or feedback loops? I mean...

Play from 1:10:37 8:88:88 1:10:37 Great question. I don't know if this is answering the question, but I think it's getting on the same page with maybe it's something as simple as visual guidelines, campaign guidelines, right? It's like then holding yourself accountable to those things. Because I think what you have to do is remove some of the bias when it comes to decision-making because everything's subjective, right? If there's 10 of us sitting in a room, we're trying to decide which image we like best, right?

Play from 1:11:06 8:88:88 1:11:06 we may have 10 totally different images or opinions about which one is best. But I think if we've got some sort of anchor in how we're making that decision, that decision becomes easier, becomes more clear. So I think it's something to do with that, right? And I don't know what the setup or any of the context beyond that, but I think just from like a framework and a model perspective, that's the way that I would look at it. There's never been a stronger need for clear and firm direction, Meg.

Play from 1:11:34 8:88:88 1:11:34 That's what I've been experiencing. Someone needs to be decisive, and someone needs to actually drive the vision of the creatives to get to a certain point that is very locked in and sort of... I think the creative ADD is hitting the corporate side now, too, because everyone knows you can do more, and you can do more fast. So it's like, well, let's just see this. Let's just see this. How about this? What if we did this? It's like someone needs to, it's like, we need like a, like a very strict strategy or like a strict operating sort of process to like get to an end result so that we can like eliminate decision paralysis.

Play from 1:12:12 8:88:88 1:12:12 Cause I think that's a big one. ah The other thing is, is, you know, you might want to build a little, not an LLM, but like a little either custom, GBT or Gemini gem or like maybe a few agents that you run it past to sort of like set some credentials and see like which ones are going to work the best. ah That stuff I did with the other thing too, ah that I did recently on Weeby was create that whole sort of like visual forensics sheet, like that visual forensics workflow with a bunch of different things, Meg, that might be helpful to just like, let it analyze and just pick the winner.

Play from 1:12:52 8:88:88 1:12:52 you because if you're trying to hit certain things, like which one's got the, you know, from a visual heat mapping standpoint, like all the different types of analytics, I can come with that stuff. um You're setting some guidelines and standards there. Now I'm getting really technical, but that's sort of some things that we could do to get around that. we just, think you need some sort of unbiased party or entity or thing that is guiding the decision to help, right? it's like you gotta set like guardrails.

Play from 1:13:19 8:88:88 1:13:19 That there's gotta be like, we do this five times, that's it. You know, like we can't keep going or else you will. Like that's, I've done this on multiple projects. Yeah. will. Yeah, I felt that pain too. think, okay, so one of my clients, one my bigger clients, we have a visual identity toolkit, um which is an entire doc, right? So I hypothetically could take pieces or all that knowledge and then leverage every decision I'm making with that, right?

Play from 1:13:49 8:88:88 1:13:49 And I think that same thing is true with any of the campaigns that we do. So it's all the do's, it's all the don'ts, right? It's all the... subjective nature of personality, blah, blah, blah. But again, that gives us something to fall back on. depending on mood, our decision doesn't change. You could wake up and feel totally different about something one day than you do the next. This gives us something that is static but dynamic in the sense that we can update it, but at the same time is firm.

Play from 1:14:19 8:88:88 1:14:19 So hopefully that answers the question. Yeah, that's it's like you do you need ops you need like a whole operating guide for decision making Like you do like it's just something has to be like a common agreed upon things that you don't get stuck at that loop Because that would just you just keep going and then you find something else that you did that was cool And then it's like well we could do that. That's also cool. And then you know most of the time things are too open-ended, but there's a Hold on one second What do you got?

Play from 1:14:52 8:88:88 1:14:52 OK, so Astrid's got a good one. As a camera woman in real life, I've done lots of picture gen. Starting on video now, do you have any tips? Is there a good workflow on Weeby, which I really enjoy? Probably have one I could sell. I'd have to look for it though for video. The... I mean like... I can't always do this to... Yeah. like step back here too, because like I think I still in some ways Astrid can feel this time to time because I've always just I always go towards stills.

1:15:03Transitioning from image to video

Play from 1:15:20 8:88:88 1:15:20 I do some video, but there are some things that overlap, right? I think like finding a structure for your prompting or some sort of like foundation is important, right? So like if you. figure out how certain tools like to prompt or how you like to prompt and what gives you the ultimate coherence and output that you want, bottle that up, feed everything through that, you're starting to get more consistent results. Now, with you having the expertise in that, I think that's really gonna help you.

Play from 1:15:50 8:88:88 1:15:50 I think you really just need to figure out how the technology is gonna interpret um what you're trying to say. think, you know, just because I've got a photography background, it was really just about me trying to figure out how a tool like Midjourney works, and then all of sudden that unlocks all the expertise, right? So I think it's more or less that, at least in my opinion. Now um Rory has done a lot, a lot of videos, so maybe he's got a different answer there.

Play from 1:16:18 8:88:88 1:16:18 Yeah, would say the stills are equally as important here because they everything's working off. It's a different mindset, right? Like you're going off of your image to then start a generation most of the time, unless you're doing text image. But like a lot of things that can help in that, like I've sort of said this, you know, we've gone over this in the podcast before, but there's like when you're doing any sort of motion, right, sometimes it's easier if the still already has motion baked in.

Play from 1:16:46 8:88:88 1:16:46 Like it's easier for the it's easier for the tools to pick it up, right? If you have someone walking, like having them walking already in stride in the image to start is a much easier sort of transition for it than having someone just standing facing the camera and saying like, have them walk right, so they have to turn right and then walk unless that's what exactly you're trying to do. So if you think about it to like, you know, breaking the fourth wall kind of thing where it's you don't want anyone staring at the camera because that's not how video typically works, right?

Play from 1:17:13 8:88:88 1:17:13 Like starting off with images like that, so you don't have to worry about prompting out of it. know, things in motion, want to have, oftentimes it's good to have motion blur in, in like the images, right? So if you have like a car driving, you know, if you have freeze motion on the car and the rims are static and the entire background is static, like sometimes the car will just like sort of either drive backwards or sort of skid.

Play from 1:17:39 8:88:88 1:17:39 won't, there's a lot of different variables to it, but I would say in prompting just in general, if you're getting started. If you're starting with image, like image is your start point, your start frame, and then you're to write a prompt. Most of the aesthetic is baked in, right? If you think about it that way. So the lighting is kind of baked in already. The colors are baked in. You don't have to prompt for that stuff. So the X factors there are what's the motion like? What's the camera motion?

Play from 1:18:04 8:88:88 1:18:04 What sort of shot are you doing? What direction is that going? What direction or what, you know, is your subject doing? And then what is interacting? I like the interaction pieces is variable, but like, sometimes it can help with a lot like, oh, the, you know, woman enters a room and the light shines on, you know, hits her left side of her face, right? Like, instead of just going insane, like, is it a lot of times you take your video is just making up off the still frame.

Play from 1:18:31 8:88:88 1:18:31 It's like, got constrictions to it. So there's so many different variables with video, it's really hard to systematize, but I typically get away with camera motion, subject action. and then sort of any sort of environmental interaction that's going on. Like this happens here or people are moving. Like if you have a crowded scene, like I love to put simple things like, you know, like ambient motion of the people in the background.

Play from 1:18:57 8:88:88 1:18:57 So like, it doesn't just like keep people static, like they're moving, you know, you can like get a little bit of extra life into stuff like that, or it really just depends on what you're doing. There's so much. And then the starting point. too, it's kind of a good time really I mean like it's gotten so much better than it used to be The physics part of it has gotten so much better. I mean that was a mess for a long time So I don't know that you would have been loving that but I think like The effort to get good outputs now is a little bit easier so I think with your expertise too and playing around Yeah, other thing too to play around with, could be fun for you, Astrid, is keyframes.

Play from 1:19:39 8:88:88 1:19:39 um Veo 3 has it, Kling has it, most of them have it. Just the start frame, end frame. Like sometimes it's easier just to figure out how this stuff works. It's like, I want the image to start here and end here. And then you prompt for what happens in the middle, right? So it's sort of a way for more control, but also a way to kind of like not get as many random outcomes as you could on the other side. I tend to like it. works. fairly, you get a lot more consistency that way from like, if it's a picture of me here and then like a picture of me back here, it's going to keep me consistent because it has both context of start and end.

Play from 1:20:13 8:88:88 1:20:13 So that's just one, it's another like little piece to it. But to me, that's a good one to start working on. What else do we have here? Let's see if we have any. looking through the promotion capture. Who you think is best runway clings, smaller specialty companies. Feels like feels like Kling is the winner at the moment. Sorry, what was the question? I missed that one. Well, I think depending on use case. For motion capture, who do you think is best?

1:20:21Motion capture, Kling, Veo 3

Play from 1:20:39 8:88:88 1:20:39 Runway, Kling, or the smaller special? I think you got to throw VO in there as well. I mean, I've been pretty impressed with Kling. Yeah, motion for the motion capture, meaning like the like input video of me talking motion capture or just like, yeah, wasn't that's the thing we tie it because it claims ones. It's pretty good. mean, I like it. Yeah, like I like claims new claims new.

Play from 1:21:04 8:88:88 1:21:04 I feel like one also is pretty good with that stuff. A little bit more, I think, complicated to use than Kling makes it. But, you know, Kling makes it pretty simple to do it. So I would say. probably leaning towards that one. I don't know if you've ever tried one, I don't know that I have tried one. I've tried, I mean, I, no, I haven't.

Play from 1:21:29 8:88:88 1:21:29 I've played around with... um I've played around with most of them. I'm kind of surprised, like, has Minimax been doing nothing? I mean, was a period of time where they were right near the top with a lot of the stuff that they were doing. I don't know what's going on over there. um fans like Midjourney. think there's like Midjourney or like Minimax lifers. Like, yeah.

Play from 1:21:55 8:88:88 1:21:55 Yeah. I think Kling and Veo 3 right now are pretty unmatched. mean, again, just depends on the use case, right? total. here's a good question. Or from from Vivian D. Are you still feeling that old school creative muse that started back with MJ when you know, ah back with MJ then or are FOMO non-stop updates, monetization and efficiency kind of murdering the vibe?

1:22:21Has AI killed the creative muse?

Play from 1:22:26 8:88:88 1:22:26 Awesome question. Good. You want to take that one? Because I I mean, I kind of feel like a little bit of the latter. mean, how much time do you have on your hands? Because um when you're busy and then you see the other updates and everything that everybody else is pushing out and you see the, know, it definitely consumes, right? Like I was just thinking about this yesterday.

Play from 1:22:51 8:88:88 1:22:51 And I think what you have to remember is there are a lot of people in this space, right? If you're watching one particular person, It's a lot different than if you're watching a echo chamber of people dropping all these cool use case applications. Those are individual people. Not everybody has infinite, no one has infinite time on their hands. So maybe those people throw out a couple updates a week. Maybe they throw out an update a day. I try to like level set myself that way too, because I haven't been creating as much.

Play from 1:23:21 8:88:88 1:23:21 I've been doing a lot more just work. in this space and so I haven't been creating as much content that does give me FOMO. I haven't been doing as many creative projects lately. Also gives me FOMO. I also know that I have to have time to create to stay passionate about this space, right? So times when there are no expectations, I don't have a brief, right? So I think it's almost like finding a way to separate exploration time from production time.

Play from 1:23:49 8:88:88 1:23:49 And that's the way I've been trying to think about it. uh trying to hold yourself accountable to both those things. I totally feel like the latter is the case though. FOMO nonstop updates, monetization, efficiency, kind of murdering the vibe. I mean, there's so many things I wanna do and it doesn't help when you have sort of like the ADD brain like both of us do. um Like you said before we hopped on is like when you get a little bit extra time, then you're like.

Play from 1:24:16 8:88:88 1:24:16 start a different like side hustle. You like start going down these, dude, mean like staying in these like lanes. I mean, it's hard, man. I think with this one, I've felt like the non-Spark three months ago, probably, when I was so busy and it was all like, just get this stuff done, versus have any time to play, which I don't know how anyone survives without playing.

Play from 1:24:46 8:88:88 1:24:46 You gotta be able to do, I'm not gonna go my full 100%. on things that I'm just like trying to execute on for efficiency versus like something I really want to do to figure out like a creative vision. So, you know, I think that's I needed needed this break over the holiday to like just be like, okay, I can just do what I want. Like, I'm happy I created this little this little thing on lovable is I never have time to just like vibe code.

Play from 1:25:15 8:88:88 1:25:15 And I was like, let me vibe code on lovable and just make like this database for all my system prompts and client projects. I built an operations hub for myself, for my clients. And that was, as much as that didn't seem like was big vibes creatively, it's so helpful now. It's like I had the time to do it, but I was like, wanted to do it. And I needed something to do, to also push.

Play from 1:25:41 8:88:88 1:25:41 But you gotta find creative time to do it. There was, you have to find play time. was gonna say is it's tough when you try to master it all. That part can drive you nuts. I think knowing that certain things exist in some ways is good enough. I think it's almost like the matter of what is your specific lane? What is your niche? Where are your efforts best spent?

Play from 1:26:10 8:88:88 1:26:10 I mean, I'm not gonna be a master on every single tool. I'm just not. And I would be dumb to think that I could do it at the rate that these things move and the velocity that takes place. But there's power in knowing what the current capabilities are. And if that does become a use case or an application, then I can explore further, right? So I think it's like centering yourself on what are the things that you really are passionate about and want to do or you have to do. And just kind of like that's where you spend the most of your time, right?

Play from 1:26:38 8:88:88 1:26:38 Like it's the unfortunate truth. We all want to do it all. It all looks cool, right? We do. And that's the problem. I've, I think there's like a one, I think there's one like really good practice you can get into if you do want to keep up with it you do want to try it all. Like having it, sounds so, I don't know, boring, but having like a quick like testing protocol.

Play from 1:27:03 8:88:88 1:27:03 Like I have a couple prompts. I wouldn't call it a protocol, but I have like a couple prompts and visuals that I go to for just about every new tool to just see if they can do it. Right. Like, I use mood boards for images. When I'm trying to prompt for a specific mood board, I want to see if it can handle all the complexity of a mood board, like the color palettes, the textures, this, that, that, like all in one image, right? If you can do that, I'm like, okay, cool. Like this is useful.

Play from 1:27:29 8:88:88 1:27:29 Same thing with videos. I'll do, you know, test physics. I always do a prompt with like a girl dancing in a room of mirrors, because that'll just show me right away, like where the flaws are. So things like that, mean, like just a... m do it on new models and just see like what the deal is. I have a couple of them that I run. This is the way to feel like is this this good or bad? You know, is this going to work or is this not going to work? Or take something complex that has worked before on another tool, like something I've done with nano and then apply it to to flux too.

Play from 1:27:58 8:88:88 1:27:58 And I'm like, didn't work like, all right, so I know it can't do that. It's the way to sort of like trim what's out there quick. I think you waste too much time trying to test everything and see if everything can work by the time you're done testing or something. So that's a different one, different sort of way to think about it. Ralph just asked, sometimes I think, my God, how easy this all is now. How much time did I waste learning to prompt, relate anyone?

1:28:13Was learning to prompt a waste of time?

Play from 1:28:25 8:88:88 1:28:25 Yes. I think in some ways it feels easy because you kind of already did the hard part though. You know what I mean? Like there are still plenty of times where I'm sitting in a room with people that just have no idea about anything. I mean, they're... There's so much value in having been early because you've kind of really put in so many reps to understand how the models interpret things and how just using one word and putting a different synonym in its place can completely change things.

Play from 1:28:58 8:88:88 1:28:58 It's like almost the micro dynamic. I think it just gives you a huge advantage. think it also obviously gives you the ability to describe visuals much more precisely than other people as well. There are a lot of variables that go into that. So yeah, I wouldn't either. I wouldn't either man, enjoy that if you were early, like most of us were here. It's a huge advantage.

Play from 1:29:25 8:88:88 1:29:25 I still think there's plenty of leverage even right now. There's a lot of companies that have not even gotten started with any of this stuff. They don't have anybody in-house that even knows the first thing about doing any of this. uh I come from the B2B SaaS world. which is typically a little bit slower. mean, people will straight up use the ChatGBT image generation and it looks like shit. They don't know anything about Nano Banana Pro.

Play from 1:29:51 8:88:88 1:29:51 And if they do, they just know by name, they don't know how to use it. They don't know anything about prompting. You know what I mean? So we're still early, enjoy it would be my thing. Yeah. has been useful. Like all the grinding. You had to go through and get something that looked good back then. And I say grinding, like, you know, not like sitting here like a sketch artist, like trying to make something good and like learning over and over again.

Play from 1:30:19 8:88:88 1:30:19 Just like it was hard to figure something out that looked good, that wasn't screwed up, that didn't have all the artifacts, that didn't have 30 fingers, right? Like that was a good... like set for now, like when we couldn't do consistent characters on anything, right? And like figuring out how to like get around that. Some of those tactics still work now. Like I use multiple image inputs as a, you know, or like sort of contact sheets as inputs now that work as little mini Lora's because that's the way that we did it for characters, right? So it's, there's always like little things that somehow keep coming back around.

Play from 1:30:53 8:88:88 1:30:53 And I think it's good that we all had to go and sort of just like, and then also get exposed online a lot too is how like crappy stuff was. I had to get beaten down a few times on stuff that didn't look good. uh To then get to a point where it gets better. So yeah. But I do see what you're saying though, where it's it's gonna be way more. I feel like Midjourney has mentioned that a few times where it's like it's not going to be prompting anymore.

Play from 1:31:22 8:88:88 1:31:22 Yeah. From that standpoint, it feels like, oh shit, you know? But yeah, I don't know. I think there are just some things that are naturally gonna carry over as leverage that we know that most people won't because we were early. Trying to see if there's any other questions here before we wrap up. Trying to make short dance films. Weight of the movement is a bit of an issue. The movement is beautiful, but the momentum is wrong. Looks like they're moving in water. I still not at that level, or do I just not know how to do it properly?

Play from 1:31:48 8:88:88 1:31:48 I've tried Veo 3, runway. I think, Rory, we'd love for you to chime in here, but I think, you know, dancing specifically is still gonna be. Video to video is really the only way to make it look truly, truly authentic to what you're trying to make it look like. um That's just been my experience. You can obviously tell the movements are way too smooth and polished.

1:31:56Dance realism and motion problems

Play from 1:32:13 8:88:88 1:32:13 Now, that whole motion control thing, Like that unlocks some things. Like that's really, really interesting. So I would think there's some familiarity with that here, but let me just kind of show this really quickly. If not, so like this, perfect example. Right now you can, you think about this too, like how future shows and films are like could be built.

Play from 1:32:44 8:88:88 1:32:44 I mean, like this guy's just in a room doing all the movements. Now you can transfer the motion seamlessly. um So, you know, what you're talking about, you almost need somebody that could do something like this, or it's gotta be the right kind of dance. Like I still think that's sort of like where we're at right now with the tech, but I would think by this time next year, that's probably gone. Yeah, this is this is gonna like video to video is gonna increase, you know, that's motion capture.

Play from 1:33:13 8:88:88 1:33:13 Really? I think that's the best way. Honestly, if you need like the dance to look super authentic, like that's really the only way I would trust because the models just aren't that they're not there. Look, if that makes sense, unless you're gonna rig it with, you know, 3D, but it's still then it's still like motion dancing, you're taking the motion that's already there. mean, like... That's a. trying to get things on cue with music if not, that, yeah, that's just not, not there.

Play from 1:33:43 8:88:88 1:33:43 But like there, like you could, I mean, look, you you got the motion, you're just plopping it. It does synchronize dance well most of the time, which is weird. Multiple people doing the same thing. It's a great prompted mid-journey, by the way. Just synchronized dance in general. dance. Look at you pulling out the tokens at the end of the show here. Just use synchronized dance. Then you can use just run the generation and hit auto prompt on the video and it'll look good.

Play from 1:34:12 8:88:88 1:34:12 I promise. Did it a million times for a project. How do you see the creative AI scene in five years? How about that one, Drew? Want that one? I can't even, I don't even know. I don't even know. I don't know. I just, yeah, there's gotta be something where typing is no longer the thing, right? um And yes, we have like speech, like just, yeah, there's gotta be some sort of quicker control to what we want.

1:34:21Where creative AI goes next

Play from 1:34:42 8:88:88 1:34:42 um You know, like I said, with that, GPT thing right now, like the ability to not have to think about every single prompt I'm creating is like a superpower in itself. I mean, before you had to do that stuff, right? And it's amazing how that is still held up. Two years late, I didn't really think, and any of us that have created, anybody can create those by the way, I'm not doing anything special, but for how much effort that took versus the payoff that you got, I mean, is insane, right?

Play from 1:35:13 8:88:88 1:35:13 So I don't know, hopefully there are some other things like that, but I have no idea what this looks like in five years. I have zero idea. I mean, we can barely imagine what happens in one year. um think the ability to go from, hey, can I have a water bottle? And then the water bottle, just auto 3D renders, and then we have a 3D printer, and then it's just out in five sites. I think it can get there.

Play from 1:35:39 8:88:88 1:35:39 I think we can have fabricators like that. It doesn't seem that crazy now, seeing all the tools that are available in disparate areas versus bringing it all into one. Five more years of Google innovation? They have so much. Rory, what was your biggest update of year from 25? I think we covered this maybe last episode, but give it to me anyway. I mean, there's like a personal and then like a business one.

1:36:00Biggest breakthroughs of 2025

Play from 1:36:06 8:88:88 1:36:06 I would say like personal, personal is weavy because that was as soon as we became available. I could do everything that I've wanted to do scale wise for a really long time. And it was possible for me easily to then translate it to the other people that I'm working with verse having to use comfy and get like too complicated to get other people to do that with you. So That one and then, you know, really Nano.

Play from 1:36:32 8:88:88 1:36:32 mean, that's I don't know, Nano Banana 2, like really the next evolution. Again, that took what I would say, Weavy 1, Nano Banana 2, because Weavy was the like the engine and Nano was the gasoline. So like that's that's definitely it. But you could say a million things, could say VL3 just being like taking six tools and putting it into one. That was also pretty insane. But I don't use it as much as Nano or Weavy.

Play from 1:36:58 8:88:88 1:36:58 So. That's that would be my answer. Yeah. Nano for me too. I would say one and two, Nano and Nano Banana Pro. Nano kind of gave you, I mean, you could say maybe ChatGBT sort of rolled out that capability from an edit standpoint. Oh, did it? It just, yeah, just, mean, but. you know, like just chat.gbt ruined me with that warm hue and like it was very inconsistent with the direction.

Play from 1:37:27 8:88:88 1:37:27 You only got one image at a time. You had to wait two plus minutes for the one image. You know, it was just a mess. The nano moment was great, right? Because it was just like, oh shit, now I can do this and I can do it fast and I can use natural language, right? It really was like that merging of LLMs and AI image models into one, you know, like that idea, which goes back to the whole thing that Rory brought up about negative prompting, which again, I think is one of the great things that we're talking about here, it's not so much even just everything that you want.

Play from 1:37:57 8:88:88 1:37:57 Now, what don't you want? What are the constraints on it? And you can put that into these tools now. That's the part we could really dial it in. And it's almost, in that sense, just as important to optimize for the negative part of the prompt than it is the actual prompt. I think about it like when you talk to your significant other and you're like, what do you want for dinner? And when the answer is so vast, it's like, just tell me what you don't want first.

Play from 1:38:25 8:88:88 1:38:25 It's like, okay, I don't want pizza, I don't want sushi, I don't want, okay, great. Much easier now to visualize what you don't want and what you do want. Because when there is infinite now, infinite capabilities, it's like, we have to start to limit it down and there is... Nano has it Midjourney has it flux has it. ah Seadream definitely has it. There's a very, there's a very standard look when you don't put the details in right like they all want to they all want to do a certain way flux has flux face and a very oversaturated tone.

Play from 1:39:02 8:88:88 1:39:02 Nano Banana has a very generic person face that they use all the time. Same thing, like same shoulder length hair for women brown hair, brown eyes. Like it's a very constructive Midjourney, same thing. Like there's a Midjourney fade. So kind of got a, and everyone's, no one's overweight unless you ask for it. No one's like, everyone's sort of like the total, there's so much data now that everyone is the mean of everything.

1:39:11Negative prompting and visual defaults

Play from 1:39:28 8:88:88 1:39:28 So it's like everything's, right? It's just whatever the most data is out there, it's gonna take that and just make it the median. That's also been the different part of using Nano Banana Pro versus another model. Yeah. It just, you know, like that real candid look has been dialed in now without having to like, you know, I don't have to put like shot on an iPhone photography, shot on blah, blah, blah. You know, like I don't have to put five qualifiers in there to get it candid anymore.

Play from 1:39:59 8:88:88 1:39:59 I can just put candid and most of the time it's basically there. That's that's important. It definitely is important to have that because it's like we used to have to jump through so many hoops, but I just don't look at the damn camera. Don't look at it. Facing side facing left, it still be like, like trying to peek over the shoulder at you. I want to look. or you zoom out and all of sudden you get side eye in there, because they always throw stuff in when you zoom.

Play from 1:40:25 8:88:88 1:40:25 Yeah, spirit also does nano-respond-well negative prompting. um I'm using it more so in system prompts and things like that. It does work very well for me in that sense where it's giving the constraints as much as the do's. A lot of times I'm creating multiple prompts from one prompt. It's like, here's my idea, but I want to go like 50 different ways with it. And it's like, we don't want this to be the output.

Play from 1:40:52 8:88:88 1:40:52 We don't want this to be the output. Don't want this to be the output. I don't want it to be... I don't want it to be centered. If you look at the prompts that I shared in either the Polaroid workflow or the Kodak car workflow, the one on the nine image grid, a lot of it is like never, because I wanted to get this real disposable camera look from it with total amateur, not amateur, but total non-traditional angles and non-traditional lighting and non-traditional sort of like...

Play from 1:41:20 8:88:88 1:41:20 So was like all the things in there are never centered. never taken from a forward perfect angle, always taken from a dynamic tilt using unique cropping and framing techniques. Most of the time, these tools don't like dynamic cropping. If I say I want a man in it, they want the whole... they want your head to be perfect.

Play from 1:41:45 8:88:88 1:41:45 If we're looking at the frame, they don't want it to be like this. I would say like with that, Nano Banana, you could do it two ways, right? Like if you've got a lot of constraints, you set it up in a way of like do's, do nots. So you could do that. If it's more of like you just wanna kind of have a prompt and put no's in there, you can also do that. So I'll give you like a perfect example. Like I'm doing a lot of candid shots for a client.

Play from 1:42:13 8:88:88 1:42:13 And I know based on For example, like visual identity, they don't like background blur. They don't want anything shallow depth of field, right? And they don't like this one. I don't want people. I just put that in the end of the prompt and does great with it. Like the coherence to that is excellent. So just a heads up. think there's like a couple of different ways that you can do that. And that's why this is a little bit different too, because it's not so intuitive in the sense that you feel like you have to get everything.

Play from 1:42:43 8:88:88 1:42:43 in a prompt format, you could have a prompt up here, right? So here's the prompt. And then you could also go down here and you say, do not ever or do not, you know, and then you've got a list of A, B, C, D and, you know, E, right? And then you've also got maybe some other subjective things, right? Like, so it's not just the prompt either from that standpoint. So this is why it's a little bit of a different thing. But I like that, like the coherence to the nose is great.

Play from 1:43:11 8:88:88 1:43:11 Like you can build it right in there. let me see if I actually have, let me see if I have this example with the nose. Because this is actually pretty, like I went pretty hard on this one. Let's see, like is it here?

Play from 1:43:38 8:88:88 1:43:38 I know if I have it in this one, I'll try to share it in that continuity. Let's see. cracking me up. what she got. is just. We got so We have our absolute continuity requirements characters scenes. What do I get to the? Only variables allowed to change Here's like the critical instructions the asymmetry subjects must never be perfectly centered must only use like certain words in terms of framing You know reject standard portrait framing cut off four heads put subjects in the corner Because you know like this is like this might not be negative prompting in traditional sense but this is me saying like I know what you're gonna do which is make everything perfectly centered and perfectly lit and perfectly this like I don't want that right so

Play from 1:44:29 8:88:88 1:44:29 I'm like rejects the standard framing like I want cropped stuff I want weird tilted compositions you know I want you to put subjects in a corner I don't want them to be like perfectly framed you know texture shots they can't be lifeless they must interact with people's bodies uh let's see what else we have here anchor. No, this one like reject cinematic perfection, embrace disposable camera, we want direct flash, red eye motion blur over exposed, like pitch black backgrounds, cut off heads, center nothing, like I'm saying it a lot of different ways.

Play from 1:45:02 8:88:88 1:45:02 But it's really like to me, that's, that's what makes these actually show up and be, you know, get the car in the top right corner. You know what I mean? Like that, that allows it to get there or you know, get Donald Trump with red eyes. But that's really, again, what I'm trying to get with here is this non-perfect type aesthetic, which is what gives it a little bit more life.

Play from 1:45:29 8:88:88 1:45:29 Because a lot of times it just comes out like this. That's it. That's all that comes out. Perfectly centered. If you drew an X through the middle of it or a cross through the middle of it, it's right at the perfect vanishing point there. So I think you're just going to see a lot of people that just do the mean. It's the same thing where we started with Midjourney, it was like you did it. where that, we kind of, somebody touched on a other question, but I think like it's that idea of having a visual profile for the brand or the project or whatever it is.

Play from 1:45:58 8:88:88 1:45:58 What are the do's, what are the do nots, what is the personality, right? And then work everything backwards from that because yeah, the default is what everybody else is gonna get. yeah, I mean, let me see if we have any other questions here. We're running real long. Probably going to wrap this baby up. I think that's probably it.

Play from 1:46:24 8:88:88 1:46:24 You guys have been awesome. Thanks for sticking with us, Rory. cool. This is great. Because rocked the comment section. Yeah. That was great. There was so much interaction. Like I saw there was stuff going on like not even like asking us questions just going on like full conversation mode. Sick. So much fun. Guys are awesome. Thank you for listening to this stuff, by the way for tuning in to our podcast if you're a regular listener.

Play from 1:46:50 8:88:88 1:46:50 Yeah. podcast every week. It's not always on LinkedIn Live, but you can follow us on YouTube, Spotify, Mid-Journey Fast Hours, the name of the podcast. We try to drop every week. We go into deep stuff. We're talking about current tools, our process, our workflows, things that we like, things that we don't like, all that good stuff. um Let me drop that link in here. So that will, even if it's not there, let's drop it in. Look at us. Or you guys can just comment podcast and we'll reply to you later.

Play from 1:47:18 8:88:88 1:47:18 I'm kidding. So please don't. We're not doing that. Yeah, exactly. Let's see. Here's our MJFH podcast. YouTube. in there. Let's see if there's anything else here.

Play from 1:47:43 8:88:88 1:47:43 think that's it, guys. I think that's it. uh Yeah, we'll be here next week. We'll go into detail on the rest of it. uh Regarding mid-journey, V8 upcoming, new UI upcoming, so just be prepared for that. Mid-journey also in their off-siders just mentioned that whichever comes first is whichever one they're gonna drop. So there's a world where we could get the new UI and then V8. There's a world where we get V8 and then there's a whole new UI and we gotta figure out how to use it.

Play from 1:48:11 8:88:88 1:48:11 So. Just be prepared for that. That's coming up and we'll keep you guys in the loop on how to work through all of that. We'll probably do a live walkthrough once that drops so no one's lost and we'll see you guys in the next one. Awesome. Thanks for coming. See ya. Thanks guys.