Fix It in prompt - Ep 003.mp4

Hello everyone and welcome back to Fix It In Prompt. My name is Alex Villabón and I'm here with... Doug Hogan. Nice to see you everybody. It's been a little bit. Yeah, turns out summer takes time sometimes. Yeah, we've been on vacation for, it was two, three weeks now and it's great to be back. Alex and I were just spending probably like the whole length of the podcast already just catching up because it's been, I mean, we were kind of talking about this too, like how weird it feels to unplug from the news for a little bit and then come back into it and there's so much to talk about. Yeah, yeah. It's like I was telling you Doug, I still do the newsletter every week. Yeah. But not, it's one thing reading about it but when I do the newsletter I actually try to test as much as I can. That wasn't the case for at least a couple of weeks and it felt great. I have this fantasy of if we could just freeze releases for a month, how good would life be? Imagine when we were just waiting for the 0.0 drops of our software every year and now we're like drowning on unreleases. You know... So it's good. It's very good to unplug. I recommend it to everyone for a bit. Yeah. Yeah. It's really hard and not impossible to stay up to date but yeah, it's good to connect back. We have tons to get through today because we have this document where we just dump interesting links so that we can talk about them. Yep. And it's going to take more time than we have. So let's see. So we can jump right into it. Yeah. I love your idea of pausing all this stuff though too. Maybe we can, maybe some of the AI companies out there, we can get a, maybe an international holiday where we just pause for a week. Yeah. Yeah. So everybody agrees. Everyone. Let's reach out to Jensen. Yeah. See if he has the muscle to make it happen. Let's make it happen guys. We need a break. Yeah. Yeah. There's, there's been so much going on. Oh gosh. Where do you, I think personally the thing I was most excited about right out of the gate was obviously some of these new model releases. Which ones were jumping out to you? Like the ones that you're just like super pumped about. I mean, to be honest, to me, what is really, I just can't sleep waiting is the release of flux three and now mini max H three. Yep. Because the era of the first video. So one 2.1, right. That was the real, you know, big boy model that came out and the community embraced and we did so many things with it. It's been going on. Yeah. It's been improving 2.2. Okay. It was better. LTEX, you know, everybody loves it. Tons of great LORUS to it, but we've hit sort of a quality ceiling that when I look at everything that flux three and mini max H three are able to do, I'm like, whoa, okay. This is, this is not a 0.5 upgrade. It looks like a completely different thing. Yeah. No. And it, it, I was really excited about that as well too, because like we've been talking about this, the, the kind of the cadence of the, the, the closed source model releases and then the open source model releases. And the, the way I'm kind of thinking about it now is it's almost like the closed source model in this AI space that we're working on for VFX, the closed source are kind of like a sneak preview. You get like, like, this is where all the, the, the, the open source models are going to be in like three, four months. And, and that gets me really excited because the, you know, like I, one of our, one of my favorite model and closed source models is obviously a C dance and C dance has a new model that's going to be coming out very, very soon with 2.5. And one of the dangers of that is like now when I'm using C dance too, I've noticed I'd have no data behind this whatsoever, but I'm getting more API errors. I'm getting more kickbacks than normal. And it starts worrying me that, you know, if you build a production around a closed source model, you are always going to be subject to the gremlins. No, no, that's, that's it. I don't think responsibly you can do anything past advertising with online APIs. Yep. Because it has to be open source. Imagine if you built the pipeline for a film around C dance too, and then C dance 2.5 comes out and they either nerf or retired 2.0. What happens then? Like you can't, you can't have that. So that, that's where, I mean, that, that's why we love open source, right? That's why we do it. And it's our bread and butter. Open source, open waste. Cloud stuff, you know, cloud models are terribly exciting and they'll continue to be. So I think especially in the advertising space or, but we need, you know, open source, open weights. We can train, we can, and then that's, you know, that's a success of one. That's a success of LTX and flux and, you know, the list is long, but. Well, and now Minimax, Minimax is promising right now it's select access, obviously, but the, they've promised open weights. Yeah. And the, the test that I've been seeing from, from Minimax H3, I guess, how, how do you pronounce it? Haloo? Haloo? Haloo? I don't know. Haloo? Haloo? I don't, I don't know. Just. They're, who knows. It's a, it's, it's a cool sounding name. Very, very cool. But the, you know, you can, the quality and the tests that I've been seeing, they're really, really close to those closed source models that were three months, four months behind. So, you know, as the space continues to develop, I think we're going to keep seeing this trend of closed source models, you know, giving us a sneak preview of where it's going. And then these open source models that are continuously getting developed and getting better and better. And, you know, my favorite phrase is like, this is the worst it's ever going to look. I mean, it looks pretty darn good. It looks incredible. It looks incredible. And I just want to share real quick. You just, we have the technology now. Yeah. So here's the, I saw this post and this is where I found out yesterday. So it says native open weight support coming soon. That blew my mind. They said it. Because I had already seen some examples of, of the H3 model. I was like, wow, okay, that's cool. You know, it's not C dance, but you don't need C dance for everything. So that's cool to have an alternative. I was not expecting this to be open source. And that really, really surprised me. So I was very, very happy to see that. I love that trend. I think it's, I think it is the way to build responsibly in this space. Um, you know, uh, Jensen and NVIDIA, like he's very much open weights is the way to go. Uh, he very publicly came out and, and, uh, had a whole, whole, whole talk about that, which was really cool to see that. I think this might be the trend going forward. Um, and you're seeing these really incredible models from, from now mini max LTX, uh, like just it's more and more players are coming into the space and they're all open source. And I think that's a place that artists really win, especially when you're operating on a high production level, you know? Yeah. Yeah. No, it's, it's super cool. It's, it's great to see. I just want to see some results of what the local model can do. Yes. Because I, I honestly, I, I hope I'm wrong, but I have a hard time believing that all the demos that we're seeing from flux, all the demos that we're seeing from, um, from H3 are what the open weights model is going to produce. I hope so. Do you have the same feeling? I just, the jump is just so vast. It's, it's, it's, I mean, it's a quantum leap. That's kind of what it feels like. The physics are a lot better. All the things that are really, really hard are getting really good. Um, I saw a, uh, uh, one of my favorite tests is called the deck of cards test. And a lot of people will run, um, run a simple prompt where somebody fans out a deck of cards, like you're at a poker table and you start dishing out the cards and most models start failing right then. Uh, you'll see multiple cards spawn when that is, you know, it's just not going to happen. Um, I'm wondering how these models are going to produce on, on those types of benchmarks. I've been seeing things that make it look like they might've solved those, you know, those types of, uh, physics issues, but I do worry that they might be cherry picking. You know, we've, we've, we always have to test these things because there's a lot of hype in this, in this case. I don't, I mean, there, obviously there's always cherry picking when you see what people are posting online. But to me, it's more about everyone that's posting now is posting from the API results. Yes. No one has the models locally yet. So that's the bit, because I want to show you another example here that absolutely blew my mind and this is flux three. Yes. Four CCTV cameras at once. And the action syncs pretty well. I'm not going to say it's perfect. You see the bottle that was dropped in the water appears here as well. The guy that was walking through it's this, you can't prompt right now. There is no way. That's absolutely no way. That's some world model understanding. Yeah, exactly. This is a different thing, different thing. Wow. I hadn't seen that test before. That's, that is amazing. I've, I have been seeing a lot of tests. We were talking about this before too, of flux three is also really good at doing 1980s, like hair metal videos and 1980s, like sitcoms that are wild and hilarious. That's amazing. I think who, who knows what the training data was for this? You know, since it's open weights, you know, we might get a peek under the hood, but yeah, it's, it's very curious. Is it going to be kind of like CRIA when CRIA was released and it's an extremely good model, but you have to add some stuff to get to the quality level of like the API access? I, I wonder, I don't know. This is, we're going to find out soon. So I keep hearing that we should expect H3 to have the level of control of Cling Omni. Okay. When run locally. Yeah. Okay. Yeah. Okay. Let's see it. I'm down for it. Let's, I hope so. I love reference videos. Yeah. That is my favorite way to work, you know, with, with a lot of these. I hope they, I hope you're right. Yeah. I mean, I, I hope they're right. I don't know if they're right. I just hope they're, they're, you know, uh, they'll deliver. We can just hope it's just, I think we're going to be drowning in the next couple. I imagine, you know, flux has been talked about for, what is it? We're just over a week now. Yeah. So I would imagine next week is release week, hopefully. Yeah. And then, um, H3, it says coming very soon. So maybe they're going to battle it out next week. Yeah. We'll, we'll profit from that bottle. We'll see. Well, and it's, it's funny too. Cause then C dance at the same exact time, you know, they were calling this like the C dance killer. You saw them in marketing a lot. Yeah. Yeah. Um, and then C dance then like announced like 2.5 is going to be coming out. Yeah. It, it, it, well, it started to feel kind of like when, uh, like when we were growing up as kids, you know, like you would have like those movie releases, uh, deep impact would come out and then immediately again would come out. And then like, or, or like the B movie schlock, uh, you know, the, the jaws movie will come out and then you have, you know, uh, then you have shark NATO and stuff like that. It would always be these little, you know, back and forth. I wonder how much we're going to see it like that. I don't know. It's a, I mean, we'll see. I, I, I mean, I don't know. Obviously C dance is incredible, right? No news there. Yeah. 2.5. I was looking at some pricing. I saw someone tweet out. It's going to be around $11 for 30 seconds. Yeah. Okay. You're going to have to read that prompt at least 10 times and make sure it's good. And it's. Well, and I think that's what we're going to see. Um, this is why I'm excited. There's so many players in the space because, uh, then they also, I saw pricing for, for H3 that was coming out compared to C dance for a similar generation. And it was like a fourth of the price. Like, I mean, if, if C dance would cost you four bucks, this would cost you like one buck. Uh, it was, and I think, um, that is really what a healthy market is going to be doing with this stuff. The tech is going to be amazing. Everybody's going to first go to be the early adopters, but then you're going to have these open source kind of counterbalances coming out that are going to then get that price down. So you're going to pay for the preview, but overall it should start getting cheaper because I, I mean, that is just a crazy price per generation. And what if, what if it doesn't give you what you want? And what if it airs out? Do you get the credits back? Are you just like, yeah, no, it's, I mean, we'll say it's sort of, uh, early days, right? Early days technology is always very expensive, but. This is a different league of expensive. It's like, if it's like 10 plus dollars a pop, oh boy. Yeah. You better have a lot of control so that you can more or less know what to expect. My hand is going to be shaking before I press that run button. No joke. No joke. And my supervisor yells at me. It was like, you just spent how much, what do you have for him? Like, yeah. I don't know. Some weird generations that didn't quite get there yet. Exactly. Exactly. Um, yeah, we'll see. We'll see. It'll be interesting. It'll be interesting. Um, so the Odyssey came out. Yes. Have you seen it? I'm going today. Uh, oh, lucky you. Yeah. We found a 70 millimeter. Uh, we got to drive like an hour, uh, being in Louisiana. Uh, we don't have, uh, it takes you a bit to get to a good theater, but we're going to go see it in 70 millimeter. Nice and IMAX. I am so excited. It's, um, it's, it's kind of a fun palette cleanser, right? From all of the AI stuff. And, you know, tech, he does a lot of stuff practically. And, um, I was, I was reading, did you, I posted a thing in the doc about this really cool rig that, that Nolan talked about mirror, the mirror set up. How brilliant is this? I see the IMAX camera was starting to affect the performance, not because like, um, technically it was getting in the way, but it's such an imposing piece. It's a refrigerator in front of you. Right. Yeah. And, uh, Tom Holland talked about how he was having this one-on-one with John Barenthal and the emotion just wasn't there. Cause he's just staring at the soulless camera. And so they rigged up this mirror system. Which is a problem for all cameras. If you think about it, right. It is just this one is a lot more intrusive than normal. Yeah. Yeah. The size. And for those, you know, that don't know the Odyssey was shot is actually, I believe the first movie that was completely shot in IMAX. Totally IMAX. And the thing with IMAX is you can only record a few minutes at a time. It's a massive rig. It is, it's extremely loud. Yeah. So they had to work with IMAX to, to come up with a rig that has an enclosure so that it doesn't make that sound so that you can actually hear the actors talk. It's that loud. Yeah. Of course the camera is big and add an enclosure on top of it. The thing is that when they were recording, it's like they're blocking the face. So they, they put these mirrors like dog is saying, just so that even though you're looking at the camera, you have a mirror right at the, at the height of the lens where you can actually see the, the other actor. So you can act off of the other guy after his face. It's super cool. Super cool. It's really neat. There's a whole, and they even match like the eyeline. So like the exact eyeline was, was perfect. It's the attention to detail is what I'm really excited about. It's it's, it's, it's really cool technical solutions that are kind of old school that I don't know. I'm very pumped. I saw, I read that article and I thought that was just the coolest thing. I want to see it. And it's, I mean, just movies are so fun. Right. And this is not about AI. We're just geeking out because movies are fun. Movies are fun. The thing is you get to, to explore more if you're doing commercials, like in the era of AI, you get to use more things because film has to be very stable, but it's just, it's just cool to see. Like this is a once it's the first time it's happening and there's a lot of people hating on it. I don't have, you know, that, that's fine. You don't have to like it that if you want the full experience, you have to go to IMAX. I have a hard time finding IMAX here in Spain. Yeah. Let alone the, the big one. Yeah. So. Yeah. Yeah. The hate doesn't bother me. Like, no, it's okay. You know, you can either watch those extra pixels or not. If you can, if you can't, I'm sure it'll be fine. Yeah. You know, we'll survive. It's almost like we're really comfortable being in a space that is very polarizing. Yeah. And like VFX AI, those two mixes have been challenging. And I, I mean, honestly, it's, it's a great opportunity to thank everybody for their comments as well. We, we do believe it or not. We go through and we read it all. Alex made a great dashboard. Did you, like, did you publish it through Vercell? It's really smart. Uh, no. So I have, I have a Hermes agent. It's so great. And I just said, I have these episodes. Please go through the comments, put them all together so that Doug and I can talk about them. We can check it out. No. So I'm not the smart one. The machine. Yeah. Hey man. So what we, I guess the, the idea is we'll, we'll have an episode where we go through, because there's a lot that's going to be that, that we want to be cover, uh, the comments, you know, uh, so challenges. It just, it doesn't have to be all fun and games because there is a lot of careers on the line and decisions on the line. And there's, it's not, it's not easy, but we're happy more than happy to chat about it because that's sort of why we started this, right? It's like, uh, started with two. So let's add more people to the mix. Yeah. No, let's get out of the echo chamber. Like, uh, we really appreciate all the positive comments. It's, it's, it's nice to see people, uh, are finding this podcast useful. We're having a good, uh, constructive conversation. I think part of the constructive conversation is tell us what you're struggling with. Tell us what you're worried about. Um, it, odds are we're probably struggling with or worrying about very similar things, you know, being, we're all in the same industry together. So, um, let us know cause we're, we're going through it. Uh, and Alex's Hermes made a really great dashboard that it, nothing falls through the cracks and we're really enjoying the questions that we've been getting. No, it's fun. We, we've already talked about a few, so we're going to save it so that we do it properly structured. We show you exactly what we're talking about. Um, so I, while we're talking about that, I guess we should say that, uh, yeah, it's only been two episodes. This is episode three, which feels weird. It feels like longer cause we talk so often. Um, but, uh, we've, you know, some companies have been reaching out and so we have friends and non-friends from other companies reaching out wanting to be on the podcast. So expect guests in the very near future. Actually, the first is going to be very exciting to probably hold that news for a bit, but, uh, don't, don't tell anybody I'm really pumped for it. I, it's super cool. Yeah. We had a great conversation the other day, um, in prep for all of that, that stuff. And I just left it just really excited about where everything is headed. Not, not for, you know, I'm excited for the podcast and the response reading and getting to talk to people like this, but just the tools and the way things are going. I think, uh, it's going to be a really fun, um, a really fun and informing, uh, podcast for people. I think they're going to get a lot of value out of it. Yeah. And if you're out there watching and you feel like you have something interesting to say, please reach out, reach out, right. Send us an email, send us a message on LinkedIn, post a comment. Yeah. Um, there, there's a lot, there's a lot to talk about and we definitely can't cover everything. So you might have experience or exposure to something that we're not, not talking about. Let's talk about it. Yeah. And I like Alex, you said, we're going to have friends on, uh, new friends. Uh, you know, it's a, you know, I think it's going to be a really, I want this to be a very diverse, uh, uh, place where everybody can kind of come and talk about stuff. Um, you know, one person I want to give a shout out to, we talked about this in a second. It's like, I want to, I want to give a shout out to Andy. Uh, it's a little, uh, where he's, he is, uh, without any of you knowing, uh, the, the reason we're named, what we're named, uh, Andy is, uh, is, is a good friend of ours. So I promised him a shout out. So I, I want to have Andy as a guest. It's like, I want to, we got to work it out. He's Andy is great. Like I, I, I told you, I've told him it's like, I think he's one of the best humans I know. Like Andy's amazing. So I, I, and then he's like super friendly, super fun, super funny. Yeah. So we'll keep it, we'll keep it, uh, mysterious about who is this mysterious Andy? Yeah. Until we have him on because we're going to hook him because if he wants to, yeah, come on. Let's do that. Let's do that. Yeah, no, that's great. Well, one, one thing I wanted to talk about too is, uh, so we've been talking a lot about these models, uh, that were released during this time. Um, all, a lot of them are kind of familiar with what we've been seeing before, but one really exciting, uh, model that I want to talk about is the, the Lucy 2.5 model. Did you, so it is a real time diffusion model, uh, real as in it has still about like 30 milliseconds of latency, but have you checked out any of those, uh, any of those demos? They're really fun. I mean, yeah, I, I, I don't think, hold on me, let me share my screen here, but I tested it that night. So I, I, some, some of you know that, uh, I do the newsletter on Wednesday nights and I try to test as much as I can so I can properly talk about things. And this, this was cool. This is a, a real time diffusion model where you can just use your webcam or your phone and then you can edit, uh, what you're looking at. And it, it does what it describes. Is it C dance quality level? No, it's not, it's not meant to be, but I think it's, it's quite interesting what it can do. And I, I think you can just go on the website and have a minute or two for free every day. I think you can, I think it resets daily and it's, it's works. It does what it says on the tin. It does, it does, uh, it's quite interesting. I think Doug has something to show us with it. Oh yeah. Oh, did I didn't, did you love how I teed that up? This is, well, we're all about testing these things, right? So, um, you know, it's, it's actually very easy using the tools we have now to kind of spin up a little application to, to try this out. Um, I've played around with Lucy 2.0, uh, quite often for, uh, some experiential projects, virtual try on things for clients. Uh, but 2.5 is really exciting. And let me, let me, uh, share, I'm going to have to share my entire screen. So we'll, we'll edit this part out so people don't get freaked out. With the infinite mirror, but there we go. All right. So very quick little electron app that we got and you can do things like, I have like slime on head. Let's see. Ooh, there you go. You've got, you got slime and it interacts with you, which is. Yeah. Yeah. Like if you shake around, I'm sure it'll have the proper physics and that, that's what's, that's what's super cool about it. It's like, let's see. Head on fire. Ooh. Yeah. We're starting to get. Hair on fire. This is a, and I mean, this is, this is sort of like what I was thinking about. It's like a Snapchat has always been a toy in terms of the filters, right? Like that was sort of early days machine learning. Yeah. Super cool. We were on vacation, right? Holiday. Yes. There you go. This is, this is where, you know, you can start thinking about these tools as actual tools. Like this is just very early days still, but it's, it's fun, right? Like not everything has to be a production tool. So go, go on there and, you know, make yourself blonde or bald or anything. Right. Yeah. It's just fun. It can be just fun. Some things can be just fun and that's okay. Yeah. No, I have a very big, I, I honestly, you know, everything can get very serious in production. Like you were saying the, the, I think artists should always be playing with these tools and having fun with them. Cause you never know what's going to come out of that. Like I could see this being, uh, uh, you know, have virtual production applications. Um, very similar to like what we saw on, um, like on Zemeckis's and Tom Hanks is here where they had that real time, uh, de-aging that was happening. Yeah. Um, you know, you could see these types of real time diffusion models once they get that audio sync together, right? Um, having these, these actual like, uh, real production use cases. So it's really interesting. The, the different models that are entering the space and like what they're unlocking for artists and artists should play with them. Yeah. Yeah. Actually, I have another one that we'd actually, we don't have on the notes, but I just think it's quite cool, which is, uh, and you, you posted about this on, on, uh, LinkedIn as well, which is the IDV2V from Netflix. And I can't remember who was, who the other studio was. And then there's Paul Dubovic as well. Yeah. So super cool. It's, uh, this is big if it does, you know, Netflix has, they did, what was the name of that other in painting void? Void. Void's excellent. Void. So void's great. And now they're open sourcing this. Actually the code is already out there. So we're just, uh, and, and Kijai is already, uh, working on, uh, on the comfy integration. So I imagine, you know, give it a week. The guy obviously doesn't sleep, uh, super cool, right? Like it keeps, uh, the foreground intact and you can change not only the environment, but also the lighting. So. Yeah. Well, and what's really interesting about this too. And, uh, Paul Dubovic for, for anybody who we talked a lot about VFX history. Uh, I mean, uh, he is responsible for the, the HDR, uh, uh, file, uh, format. Uh, he's responsible for a lot of photogrammetry work. Um, from, uh, uh, you, you should, you should, you should Google all this Paul Dubovic photogrammetry and you'll see his early film that he put out, uh, while at, uh, while in California. And he's now working with Netflix and iLine to do these, do this release right here and, and, and a variety of others. Um, and he's, he's, he's really prolific in the VFX space. He, if I have a dream guest for this podcast, it was so cool. So inspiring. And his work that he's doing there is, uh, is really cool. And it's, um, there is no comfy wrapper currently. You have to run this kind of in a CLI, but, uh, but I give it a week. Yeah. I mean, I very easily, uh, wrote a little comfy wrapper, um, that, uh, using Claude just to kind of like spitball this stuff together. And I can tell you that they're not cherry picking results. Like I was very excited working with it. Um, it's, it's, it's a really great model. Yeah. Yeah. So Paul, if you're listening, come on. Come on. We'd love to chat. Yeah, no, it's, it's cool. It's, there's just so much. There's just so much. Um, there's that we're, we're being showered with icy Loras, which is great. Um, it's, it's, uh, yeah. Never, never ending fire hose. Um, so on, and other news, uh, did you see the Nightborne, uh, short film from Neil Blomkamp? Yeah. Uh, yes. Another prolific, uh, person in the VFX space. Uh, he's playing around with AI tools and, um, you know, the, the shorting put out not perfect, but neither was the, uh, you know, uh, the, the, the test he did for the district movies. Like when he, when he did all that, he's a, he was experimenting with new technology. This is just kind of, it's in his blood. He just does it. And, uh, he got so much flack for this. And I thought, uh, I don't know, personally, I thought it was very unfair. Uh, I don't know what, uh, what you thought. So I, no, I obviously nobody deserves hate. He, this is just a guy like you and me found something interesting. He wants to play with it. Yeah. Why do we have to judge him? He's not charging you to watch it. He's just putting it out there. He's just enjoying it. I have to eye roll a bit at people complaining about this. It's like, oh, we don't, this is not what we want. We want a new district movie. It's like, well, buddy, this probably cost a couple of thousand bucks. Yeah. That's not how it works for a proper movie. Right. So I, I think this is fine. I think people, so there's two camps, right? There's people that do short films with C-dance and they said, this sets a new bar. This blew my mind. I am not in that camp. Yeah. I think, I think it's okay. I look at a lot of issues that it has and I, and I wonder why or who, like, I wonder why he didn't have VFX team. Just saying that doesn't look good. We can fix it very quickly and nuke or something. Yeah. And if he did have people, then I don't think they were the right people to have because there's a lot of over sharpening. There's a lot of bad chromatic aberration that's present in one shot and not the next. Yeah. There's a lot of continuity issues. But again, maybe that's, he just wanted to put something out and that's okay. Right. Well, maybe those are the constraints that he was challenging himself with as well. Exactly. Can I do it completely? Right. You know, what's the, what's the state of everything? Yeah. That's a valid, that's a valid thing. Yeah. And this, this is what's annoying, right? Like it's, you see the up votes versus down votes on YouTube. That's what we were talking about. I don't think that's fair. I don't think that's fair. Like, I don't, I think artists should be free to experiment with this stuff. Have no preconceived notions over whether or not it's good or bad. Does it work? Does it fail? Well, I, I applaud him for putting it out there because I mean, you know, he knew that he was going to get flack. You'd be naive to think that, you know, you wouldn't. But it's just, it's interesting because, you know, he put that out and got a lot of hate. And then just last month, Scorsese came out talking about, you know, how he's using generative AI for storyboards. And he was celebrated. So it doesn't, I, I don't know the rhyme and reason for it. I don't know if there is rhyme and reason for it. It seems like it's just kind of a, the, the fear is still very prevalent out there. And it's also an easy target to jump. And I mean, you know what? I, if you, I, I don't know why we do these things with social media, but I did spend a bit of time reading the comments. Just like annoyed a bit mostly. Yeah. I mean, there's a few people just like, you know, clapping and being happy about it. It's like, yeah, cool. It's like, let the man play. It's like, you can only learn if you play with it first. So there was a comment that I said, this guy, this guy gets it. And he's like, I see all these people complaining, but maybe what we're looking at is just a preview at a new universe that he's going to come up with. It's like, cool. Okay. There's something there. Right. So what if you put this together? Obviously he doesn't need to persuade studios as much as a no name to get, to get a movie sign, but he could just start having excitement around the IP and maybe he, he does the proper movie. But every three months he releases a 10, 15 minute short inside the universe. Right. That could be a thing. Like that's cool. Yeah. Like that's feasible. And isn't that kind of like, uh, I mean, won't that pave the way to more greenlit traditionally filmed projects, more, you know, traditional VFX pipeline work, more, more work just in general, just because the, the, you know, the creativity doesn't require so much capital to get from the director's head out to kind of say, this is what I'm thinking. What are you guys? Give me some feedback. Let's yeah. No. And I think it's a good point that you bring up jobs, right? Because you say, it's like, Oh, but we wanted a new, you know, district movie. And it's like, okay, so let's, let's think about it this way. Let's say he gets the next district movie greenlit. Yeah. That's obviously going to employ a lot of people. So this is not taking away from that. Right. Even if he did a 15 minute short, it's not going to take away from that. Yep. But if this becomes a trend of them releasing something every three months, six months as a buildup to the movie, you're going to need VFX. Because as I was watching this, I watched it the first time and it was like pausing and it's like, okay, no, no, no. I'm just going to watch it. And then I'm going to watch it a second time. And I, and I'm going to do like as if we're in Bailey's. I'm going to start taking notes. Like, I don't like this. This is wrong. That's great. And then I started looking at it and it's like, okay, I didn't think about it in the moment, but now that I'm talking to you, it's like, if this is a thing, that's going to employ people because people need to fix all these things that I was looking at. Right. So it's like, if there's continuity or like I was saying before the, uh, the sharpening or upscaling being too aggressive, uh, yeah, all sorts of technical things that weren't really taken care of. We need to do it. When I say we is the, we, the VFX industry. Right. So that could be another vertical of money for them, but employment for us. So I think, I don't think we need to just immediately go for the pitchforks. It's just a bit annoying, to be honest. Are you saying Alex, that they're not going to fix it in prompt? Yeah, exactly. Is that what you're saying? They're going to fix it in post? I think you are correct. I think, um, I think a lot of people, um, are, uh, are not aware of the, the already existing protections that are, that are in place that require, you know, uh, that if you have real actors, how much of their performance can be, uh, changed or manipulated. Um, all of that stuff goes in when you're operating, when you're working on an actual high-end Hollywood production, there's a lot of safeguards already in place. Because we went through all those strikes for a reason. And they came out with some pretty meaningful, not perfect, but meaningful protections for, uh, for real artists that I think, um, that's the boring, uh, that's the boring stuff. The, the, the things that don't make a lot of big headlines. It's easier to talk about how AI is taking away jobs and, um, you know, it's gonna, uh, now he could just do a whole movie like this. Well, not really. Uh, there's, there's actually a lot of red tape, a lot of friction to get, uh, to a fully released AI film from a large director that, you know, um, that I don't think a lot of people are aware of. So I think we should not just jump automatically on the narrative of just, this is the death of Hollywood. This is, it's just, it's not accurate. I, I agree. And, uh, yeah, it's just, we'll see. Let's, we have to test it. We have to play with it. Let's just give it a bit of time to mature and then we can complain, but we can't. Yeah. Let's get a taste first. Just getting started. Um, and then I just want to say that they're also going back to that video. It looks like he started a new AI studio. So it says an AI film studio by Neil Blomkamp. So that's interesting. So you're going to see more of this pop up, which is why I wanted to, to just mention it. It's like, don't be angry. It has a studio name to it and there's some people prompting, but there's going to be people fixing things and doing things creatively on top as well. So it's not all doom and gloom. And I think they will be surprised if, uh, if anybody does, I won't rename names, but, uh, but I think people would be surprised if you start going down the list of, uh, some major Hollywood directors and the relationship with plus AI company. A lot of them are spinning up their own studios. I mean, how you can blame them for it or not, but I mean, um, I mean, one that is very public that, uh, that you can't talk about Tyler Perry. Like he had a whole studio that he, uh, was planning to build right before the strikes, shut it down because AI was going to replace everything. Now, uh, it was, it was announced in variety, I believe of now he's building a whole new, uh, studio again for virtual production. That's entirely running AI tools to do it. So, you know, I, you kind of see these things in cycles in this space when you, when you spin it up and you really spend some time in here, um, on the surface, it looks like everybody hates AI and nobody's, nobody wants to play with it. But when you're at the studios and you're talking to people, I mean, everybody kind of knows this is the way it's, you know, this tool is not going back in the bottle. It's far too powerful. Um, and the risk is that if you ignore it, well, now you don't have a say in how it's actually being used. And if artists are the ones that are controlling the tools, I guarantee you, it's going to be more artist friendly. It's going to be more made by, you know, by artists for artists. And that's a space. That's how I want to see this new AIB effects industry evolve into for artists. The thing is, if I think the key, and this is obviously just an opinion, but I think the key here is understand how to do the AI bit, understand how to add on top of it or fix it with VFX, right? With a proper VFX tools. And you'll be in a much better position than just doing one, right? Yes. Because does it mean that the end, that's the end of the traditional VFX? Like, no, that's, you know, it's obviously going to be still alive for a long while still. But if you want options, I think that's the position to be in, right? You want to understand where the limitations of one ends and the other one can take over and vice versa. And you want to be able to do the whole thing, right? So you don't want to just be the, okay, this is what I prompted for you. Here's my cake. It's like, well, but it tastes bad. It's like, can you fix it? It's like, no, I can't. So we need a maker now. But I just spent $20,000 on Seed Hands 2.5. Yeah, exactly. So give me this cake. Please fix it. Right. But if you can tackle that problem holistically using everything, that's so hard to find. It's so hard to find. And I think that's what we just say. It's like, play with it. You don't have to bring a reel doing a new avatar on your own using Seed Hands. It's like, just have the understanding of the technology and understand where one helps the other. Yeah. That is key. And that's incredibly hard to find right now. Yeah. It is. And a lot of people don't know where to start. I know that's been kind of our last few years. You know, I've done some courses, Action VFX for FX PhD. You put out a great YouTube series of I watch your like while I'm doing my courses, I'm watching your courses. You know, we're all watching everybody's stuff just to kind of like learn about these tools because they're so rapidly developing. I think it's really cool that we're all sharing all that stuff, whether it's, you know, buying a paid course or doing free stuff. And I wanted to highlight one thing real quick that we were talking about because, you know, it's no secret. I also work for doing some work over at Comfy. And one of the things that we're doing is just last night they put out my first module for Comfy 101, General Creative, where it's an introduction to Comfy. It's over an hour long for this one module. So we go down exactly like full no-gaff basics for this first module. And it's completely free. Go and check it out. It's literally just there to help people learn like where, what's available and start getting an idea of how to use these tools. And the coolest thing for this too is that it doesn't require a high-end machine because we're using all Comfy Cloud. So you can go and just, it's really turnkey. You can jump into it. And if anybody's interested, go check it out. Let us know what you think in the comments. We're listening. It's informing what more training we do down the road as well because we want artists. We want to get this in the hands of artists. That's a huge push. I saw, like I told you, I saw that last night and I added it to my watch later list. It's so, it's such a good feeling to see that because for a while, and I'm saying for a while, I don't know, a year and a half, two years ago, when I was really like struggling to learn because it was a hard road. Now it's a lot easier because there's a lot of people covering it. Yeah. So this is like great. It's like people are always asking me, oh, where should I start? It's like, well, just go on YouTube and just do one thing. Yeah. Learn how to prompt an image. But I also had a lot of frustrations with Comfy because it felt for a while that they didn't know exactly what they were doing. Or like for a while, I was like, oh man, just mature into something. Realize what you have. And all of these small initiatives that have been happening over the past few months, it's like, yes, this is the stuff. This is like when I told you probably on episode one or just months ago when we were chatting, I was like, I was so happy when I saw someone mention Asus and EXR in the Comfy stream. Yeah, in the Comfy live stream. Yeah. So that's where I thought like, oh man, it's like, okay, they're getting it. And this is like, okay, it's like, it's starting to feel like a company and that's great. It's really good, really good. Because I, well, you know, I'm big fans and it's my daily driver. So I'm super happy and I can't wait to watch it because I still watch introductions to Comfy all over the place because there's so many nuggets that you can't learn everything. It's too much. No, I do too. Like, I think that the trick to all this stuff is just continuously trying to be a sponge with all this stuff. Like, you know, I had been using Comfy for a few years before you put out your tutorials. And so like, I knew everything, you know, how to do it. But what I was interested in was like, how do you approach it? How, like from your lens, how do you operate in that? And it felt very much like, I mean, like how we learned VFX, how we learned Nuke. Best way I learned Nuke was by senior artists sharing their Nuke scripts with us and getting to poke around and kind of like see like, how did they approach this problem? And it's exciting being over at Comfy because, you know, that is a big push of we're trying to listen to artists, you know, the community. What are you struggling with? What is not making sense? I still don't think people understand how powerful the new desktop instancing system is for Comfy Desktop to spin up like an R&D section, to set your input and output directories, to like choose where your models are being stored. Do a video on that. I'd love to see that. I'd love to see that. Yeah, and, you know, that's the type of feedback we want to hear. And it'll go into more videos down the road because the goal here really is, you know, they're an open source company, just like we're seeing with a lot of other companies spinning up with model companies. And they want to hear from the artists of how it's operating, how they use it, what are they struggling with. And then we're here to support it and make it better because, you know, one of the big reasons why I wanted to work with Comfy was from an early meeting I had with them when they were asking questions about like, hey, we know we didn't do this great. Do you have any ideas? Like, how can we make this better? That's great. Yeah. And that's really exciting when a company is open to that type of stuff. And it's a cool space. So, yeah, go check it out. It's pretty cool. We got a lot of stuff. Definitely. Definitely. We'll definitely put a link to it in the comments if you want to check it out. But it's on the official Comfy.org. Yeah, this is official Comfy training. It's kind of cool. It's got the logo on it and everything. Nice. Nice. So, on the note of Comfy, I think we've talked about this, but I think just because we're talking about training and there might be people that obviously haven't heard the previous episodes, I just want to say again. That people are afraid of learning because it is difficult. And what you just said is so true. It's like I watch introductions to Comfy all the time or just, you know, even watching a video about something I'll recognize how to do because I want to see how other people do it. Nice refresh. That reminded me of something. When I, I don't know, 10, about 10 years, actually, yes, 10 years ago or 11 years ago when I first joined Framestore, I was, after a few months, I was put on a project to do mini comps for stereo conversion. Right? Right? So, you needed to go in. We had our shots, but then whenever you had time, you need to do a few mini comps to help along the queue because every shot needed to go through it. So, you would, for those that don't know, creating mini comps is when you have a shot that needs to be converted to stereo and at a different shot. And you need to go into your script, make sure that everything is properly tidy, no internal tools. So, you need to pre-comp where you see internal tools doing things so that they can replicate the results on their end. To me, like, I remember when I was first assigned to it, I had no idea what a mini comp was. And I started asking around and people were like, oh, yeah, no, like mini comps are the worst. And I was like, oh, shit. It's like, okay, this might not be fun. So, I got into it and I understand why people potentially don't like it because you're just getting to open other people's scripts and sort of tidying their mess up, which isn't fun. But I remember I learned so much by looking at what other people were doing to get their results. It's like, I was like, oh, man, this is a cool tool. I'm going to save it to my tool set. It's like, how do you get inside? And then I look at a terrible green screen. It's like, how the hell did they key this? Yeah. And I go through it because I have to for the mini comp. It's like, oh, oh, okay. I'm going to take note of this, note of that. Oh, yeah. To me, after doing mini comps for, I don't know, a month or two on that project, I was like, can I get some mini comps? I was like offering myself. Oh, you're craving them. I was just like looking. It's like I'm learning from all these great artists around me that I felt like, you know, there was so much to learn. And I mean, that's definitely a profile thing. Like I crave that sort of stuff. Yeah. But that is sort of the reason why I also, when I see a tutorial online, even though I have already a workflow for it, I want to see how you do it. And I actually messaged you. You put a video on the ComfyX account, actually on YouTube as well, about face replacements, right? Yeah, yeah, yeah. Or head replacements. That's like, oh, let me take a look. Let me see how Doug does it. It's like I have my workflows, but I want to see how other people do it. And you gave me some really good notes. It was like, oh, you know what? This is actually, if you add this, it makes this even better. And I was like. I mean, but that's the beauty of it, isn't it? And I save everything that I test just because there's always something interesting. Even if it's not the best possible route. Yeah. It gives you an outlook into how other people are thinking when they're building their workflows, right? Yep. And I find that incredibly, like, it blows my mind that I can do that for free. I can just go on YouTube. It's like, oh, yeah, you can just grab my workflow. It's like, yes, please. Let me take a look. It's like you have too many custom notes, friend. But okay, let me install them and let me just see why you have that. That's cool. I really like that. I encourage everyone to just don't think about building your own if you're starting. Grab other people's. Break your comfy. Well, hopefully not anymore with the desktop app. Yeah. I'll show you in a future one how to set up a nice little section where you can break that with all the custom notes that you want and try it out. There is a future module that we talk about. Like, a lot of feedback we got was people getting intimidated by that big blank workflow screen. Like, where do I start? And you really can. There's so many workflows out there, so many people sharing their own techniques that you can treat it more like Lego blocks. Like, you can grab this section over here. Ooh, that's interesting what they did over there. I'm going to copy that over here. That's it. Ooh, I'm going to take this. And that changes that. You can wire something up really quickly. And then when you approach it that way, it's a lot less intimidating, a lot less like I have to memorize every node. You don't. There's only like maybe – I probably use maybe a dozen of the hundreds of nodes in Comfy every day. You kind of go back the same ones, and you start seeing those patterns. So, yeah, we talk all about that in some of these future modules. Yeah, it's sometimes. I actually not – I think about this often, and I don't think I'm going to do it, but I think about it often. It's like there are so many nodes, and there are so many custom nodes that I just think life would be so sweet if I could just have my Alex nodes. And in there I have merge, and I have transform, and I have blur, and I have all the names that I know. Because sometimes I'm like trying to build something. It's like, what's that one thing called? It's like, oh, okay, I know I have it on that other workflow, so let me go ahead and open it so I can remember. There's just too many, but I sort of hold myself on not doing that because then it's sort of – it's like I'm not – this isn't that other package. Yeah. I should learn this one, right? Yeah. But it's just a lot. It is a lot, even if you've been doing it for years. So actually, you know what? A few weeks ago – I said weeks, I don't know, maybe a month, maybe a month. How long has it been? I don't know, man. So I put together – so I've always liked the templates launcher from Comfy. Yeah, yeah. And I said, like, I like that you have really nice previews and you have all these things. And I never really liked the way the workflows are saved in Comfy because it's just a string, and it's very hard to remember what it did, what it was. Yep. Let alone version control doesn't exist. So I've had this idea for a while, and one day I said, you know what? I'm going to do this thing. And I basically built my own workflow vault where I can save things, and I can look at the before and after like you do on the templates. And I can, in there, have – it'll do automatic version control. It'll save as many inputs and outputs so that I can replicate results again. That's cool. It'll save prompts. It'll save everything. And I highly recommend that if you're getting started, like if you're seasoned, you might not want this. I want it because I sort of save everything, and I want to be able to go back to it. I'm going to do rabbit. If you're just getting started, I highly recommend it because I always recommend go get that workflow from whatever YouTube video. And then I always say the first thing you should do is add a note, a note node inside of your Comfy, and put the link of where you got it from, whether it's from Reddit or YouTube, so that you can always go back to it. Because otherwise you open it in a month, and you're like, the hell is this? Yeah, where did I get that? Yeah. So this solves that, where you have actual examples inside. You can put notes. You can have version control, and you can just go back in time if you want to. You can save inputs so that you can get the same result out. So to me, this is like, I just can't live without it, and I'm doing my best to keep it up to date whenever I find issues or bugs. And it's like, you know, it's just, you're going to be opening a lot of stuff, so you might as well know what that stuff is. Yeah. Dude, that's so valuable. And that's, I mean, that really showcases too, like, again, the power of open source, being able to build your own tools. You know, and Comfy encourages people to do this. So, like, that's why there are so many tools and nodes out there to fill in those gaps that are super useful. And I think it's great people are sharing their stuff for free. I think it's great the people are putting together paid courses and paid tools as well. I think there's room for both. I know I purchased my fair share of courses. And just, again, like you were saying, just to go through and just kind of like, how are you guys approaching this? How is this going through your lens? How is it, you know, Victor Perez has a great course as well that goes through. Very good. And, I mean, he covers stuff that I don't cover any of my stuff with super deep, which is hardware as well, which I don't know if you saw his rig that he posted a video while we were gone. Yeah. Holy cow. That's crazy. It's like six, like, hyper-threaded, like, Blackwell 6, that, like. I don't even know where you plugged that in. Like, to be honest, like, I have a 5090 and I have the Pro 6000 on two separate boxes. Yeah. I need to plug them on different plugs if I want to run them both at the same time. Yeah. I don't know how or where he plugs his in. Like, you probably need, like, a special outlet. Like, that's just too much electricity. And can you imagine what it sounds like, too? I don't think, like, I kind of filter it out when I'm doing, when I've done some of my courses. But I'll mention it just like, oop, it's starting to heat up. Yeah. Because, like, you can start hearing it going, and you're like, you kind of take a step back. Kind of like when Egon, like, turns on the proton packs in the elevator. And they kind of back away from it. But it's, I can only imagine the sound of the raw power going through those things. Honestly, we should, you know, we should really have a whole episode that we just talk about, like, hardware. And, you know, what it takes to do a lot of these workflows. I know we got a lot of comments about that. We'll talk about it. But there's, of course, hardware. But then there's also cloud options that are really accessible to a lot of people. There's so much to talk about. That's why I love this podcast. There is. We have to dive into it. And I just want to mention that Victor's course is, I mean, Victor's a friend. So I just want to say that up front. But what I, so when he came out with the news about his course, I reached out and it's like, oh, man. It's like, this sounds incredible. And I'm all for it. And he's like, but why? Hey, you already know how to use it. It's like, well, I want to see how you do it. And it's like, you know, that's what's interesting. What I love about his course is the fact that he came out immediately and he said, you're not going to be opening comfy for a while. Yeah. We're going to be talking about the tech. We're going to be, so we're going to understand it holistically rather than going, oh, this is how you generate a thing. It's like, this is not the course for you. You got to be patient. It's still rolling out. It's still rolling out with episodes. It's great. It's great. And I think Victor is a great teacher. He's been doing it forever. You've been doing FX PhD courses yourself. Victor has been doing it for ages as well. So I encourage you. That's a good one to look at for sure. There's so many good ones. There's so many good ones. There's both free and paid. Yeah. We're all standing on the shoulders of people that come before us. Like we were talking about Paul DeBevick. I mean, a lot of the technology that we're working with right now is based on standards that he set. And now it's exciting to see he's playing in this AI space as well. Victor Perez is obviously one. All of the FX PhD people. That's why I was excited about going over there was because I learned so much from them when I was a student trying to figure this stuff out. But we're all learning from each other. And I think that is just kind of circling back to that Neil Blomkamp controversy. Nobody should be shamed for trying out new tools and experimenting and sharing what they're doing. I think one way is kind of dark where people don't want to talk about this stuff and it gets super secretive. Nobody's sharing information. The other way is super exciting where even if they don't work all the time, sharing those results, you learn a lot that way. And I think everybody should be playing in this space and experimenting. Whether or not they want to use in production, that's entirely on their choice. But they should be informed because that's the best way to kind of fight that fear of it. Second, you learn about this stuff. You start seeing where the gaps. You start seeing where it fails, what it doesn't. You start testing it like we do. You're like, oh, well, this doesn't really work that great. Or that does. So play around with it. Yeah. I guess it's just under an hour. But actually, I do want to talk about this. The Foundry releases Smart Roto. Yes. Yeah. No, that's another OG company. We've worked in Nuke for a couple decades now. They are a foundational institution to the VFX industry. And they're starting to play with these AI tools. What are your thoughts on Smart Roto? So I haven't tested it, to be honest. I have seen a lot of feedback on it. I've seen a few videos on it. Actually, I recommend everyone go see a video by Compositing Academy. It's titled How Smart is State of Art AI Rotoscoping. Yeah. And they go in and they test it on a different series of shots. And I think it does, you know, what it says. It's meant to be doing. It's meant to speed you up. So I think that's fine. What I really like is the fact that you build your shapes. And then you can go to another frame and it'll, through machine learning, snap your shapes to where they should be now so that you can go ahead and adjust. And then it's sort of, it reminds me a lot of Keen Tools. If you've used Keen Tools, where you put a few keyframes, then it sort of goes through and it tries to adjust them for every frame in between. And then you do the same in the middle. Yeah. Sort of like traditional Roto, but it just, I think they claim 4x speedups or something like that. So yeah, it looks like it does that. What I don't understand, and maybe it's something that I'm failing to understand, it's why it's being treated as a separate software in the sense that, well, a separate plugin, right? I believe it's $500 or $600 a year. Ooh. I don't understand. Why isn't this just a NukeX feature? Because if they came out and they said, this is going to be hitting an API, okay, you have to pay for those calls. So it makes sense that we pay for this extra module. But if it's just a model that you're running locally. Yeah. And they didn't do that with Copycat and Cattery. Yep. Why are we doing that now? I don't understand it. If someone has the, again, I'm giving them the benefit of the doubt. I'm sure there's a reason for it. So if someone understands it, please let me know because that's the only thing that, where I looked at it, it's like, oh, cool. And then it's like, wait, I started seeing people complaining about it. It's like, what's wrong? It's like a new thing. It's like, oh, we have to pay for it. Not even NukeX includes it. Yeah. Well, that's, we should, you know, we have friends over at the Foundry. We should, you know, one of the guests I'd love to have on is somebody from the Foundry to ask these questions because I think they're very valid questions. Especially given how expensive Nuke already is. What's the, what is the reasoning behind it? They might have a very good reason behind it. But I think some stuff is getting lost. That message is not coming through, it seems like. And it does, yeah, that's crazy expensive per year. Per year. It's like, I mean, I don't know. Is there an option for per month? Can you try it first and see if you want to come in? I don't understand why. I mean, even if they wanted to do kind of the DaVinci Resolve and then Resolve Studio aspect. Like, I pay for a studio, but it's a one-time payment. Like, then you get this great, you know, it's not another subscription you have to go, you have to manage as an artist. Yeah, we should have them on and talk about that because I think it's very valid. Yeah, and I know you and I have friends there, so, you know, I'm sure we can probably talk about it. I think some of them are watching, so reach out because we want to know what the thought process is with this stuff. And, you know, when we are seeing a world in the VFX space, you know, they were dominating the compositing world for a very long, very, I mean, they still do. They are the software that is used on feature film productions. We've used it for a long time. I owe my entire career to Newt. Same, same. So it's, but the industry is evolving. Now we have open source tools that are coming in. We have competitors, you know, that are hopefully coming to the market that, like, we have a healthy market of competition, of closed source models, open source models competing to get the price down. So it gets, it's not out of the hands of artists because I think that would be the worst thing that can happen. Oh, yeah. Artists get priced out of the tools that they need to use. That would suck. And I think, I mean, I think the product is great. It's like, does it make Roto faster? The overwhelming answer is yes. Yeah. Yes, please. A hundred percent. Yes. Like, so that's a thing that is successful in doing what it does. So I'm happy that it released because they were talking about it for a while. Super happy that it's out there. Like I said, I haven't tested it. I believe you said you hadn't either. So this is, you know, this is a step forward. So I haven't paid the $600 a year. Yeah. I haven't done it yet. No, not yet. Not yet. But anyways, it is a new thing that you can go out and you can actually try it for free. So, okay. There's that. There you go. There's that. First taste is free, everybody. Then you gotta. Well, this is great. That we're, yeah, we're coming to the end, but this is, we're going to be releasing again, these, what's our cadence now? It's every. Every two weeks. I mean, for now it's every two weeks. We'll see that. Like, there's a lot that we want to talk about. There's a lot of people that we want to have on that have, that we either have ideas for or have already reached out and are, we're already planning for those to happen. So we'll see for now. It'll be every two weeks. Summer is sort of still here, but for now we're around. Like if we need to take another week break, we'll let you know. Yeah. We'll still be around hosting. We're now on Spotify. We're on Apple Podcasts because a few people requested that and it was easy enough to set up. So we have that. You can find me on YouTube. You can find me on LinkedIn. So send me messages. Always happy to have a chat. I have the weekly newsletter, Alex Villivan, VFX and AI. I'm coming up on issue number 50 very soon. So excited about that. That's awesome. So yeah, we'll be around and wait around for that guest that we were teasing because it's going to be a cool one. It's a really cool one. Yeah, you can find me all in the same places like Alex was saying, LinkedIn, YouTube, go to DougHogan.com. I even have a thing. I might regret adding it, but I added a book call with me. So if you ever want to like just chat about, you know, for a little 30 minute free call, I like to, again, the goal of this podcast is to help people kind of get guided through the weeds. So I offer some time to do that as well. Yeah, but it's great doing this podcast. Keep the comments coming. I mean, again, don't feel like they have to be all positive stuff. We're leaving it all on there. We want to know what you're frustrated with, what you're struggling with, what you love, what you, you know, all of this stuff. This is a free, open, safe space to explore these tools and also explore it honestly because we, I mean, like we just said, we have, you know, we will always put things to the test and give you our honest feedback on it. And hopefully that's the value of this podcast is we're going to, that's the, that's the mission statement, right? That's it. We're not peddling. We're actually doing the thing. So we're actually doing the stuff guys. All right. Great. Well, uh, all, uh, Alex and I will, we'll, we'll definitely see each other sooner than two weeks, but we'll see you all just a little bit. All right. Bye everyone. Bye everybody. Bye.