0:00 I know AI is a touchy subject for 0:02 creators. That's why I asked you on X 0:04 about an episode on UFN's new MCP tools. 0:08 The response was an overwhelming yes. 0:10 So, I brought in Mark Price from Play 0:12 UGC. You may know him as Warford to 0:15 explain what's new and what these tools 0:17 can actually do. What you use is your 0:20 call. Now, with MCP, I have like a skill 0:22 that's like migrate features over to a 0:24 new map or whatever, right? He also 0:25 walks me through the tools he built for 0:27 getting assets into UEFN. 0:29 >> It reprocesses and reformats them, puts 0:31 them into UFN the way that they're 0:33 supposed to be in the proper formats, 0:34 indexes them properly, and that process 0:37 is just solid now and just takes 0:38 minutes. Follow me on X at immature 0:40 gamer to help shape future episodes. We 0:42 get into what the MCP can actually do in 0:45 UFN, how Roblox compares, and what went 0:48 into Mark's 1hour prototype. Let's get 0:50 into it. 0:51 >> This is Creative Juice. I think we're 0:55 live. I have a very very special guest, 0:58 somebody who is goated in the game 1:00 development space and also in the AI 1:03 space. We have the new MCP tool that has 1:05 came out in UFN. But first, I want to 1:07 welcome my guest Mark Price and then uh 1:09 War Forge and uh or Play UGC is also his 1:12 other company he runs. We'll get into 1:13 all of that today, but first Mark, 1:16 thanks for coming on and I would love if 1:17 you could just intro us kind of like 1:19 what's your background and everything. 1:21 >> Yeah, thank you. I appreciate appreciate 1:23 being here. Well, I'll just I'll give 1:24 you I'll give you the the full history 1:25 here, but I I uh I've been in games for 1:27 a long time. I grew up playing games and 1:29 when HalfLife came out in 1998, they 1:31 released I I believe it was called like 1:32 the Hammer World Editor and I started 1:35 playing with that and dabbling in some 1:37 C++ stuff. It got me into games and then 1:40 when the app source came out with the 1:41 iPhone was invented, you know, I got on 1:43 that, started making games and apps 1:45 there and, you know, built a consulting 1:46 company and then I really really wanted 1:49 to get into the game industry deeply 1:50 though. So, I ended up taking a job on 1:52 Call of Duty Ghosts. 1:54 >> Let's go. 1:55 >> Yeah. And I I built there, they had this 1:58 app at the time where you could change 1:59 your loadouts in between matches on your 2:02 phone. Like, so all the 3D weapons, 2:03 everything, you just do it from your 2:04 phone. And so I built that I built that 2:06 app for, you know, iPhone, Android. It 2:08 was with Unity. And I and we deployed 2:10 like cloud services and stuff. 2:11 Eventually got rid of that feature. But 2:14 anyway, so that was that was fun. And 2:16 then I ended up starting a company 2:18 called DevSlopes where we trained people 2:20 on programming and app development and 2:21 game development and grew that company 2:23 for about 10 years and then sold that a 2:25 few years ago for about $5 million. Wow. 2:28 And then after I sold that, I was gonna 2:31 take a break. And then I saw Tim Sweeney 2:35 announced this thing called UEFN at the 2:36 time. And and I had just been coding I 2:39 had just been coding this game that I 2:41 wanted to build because I there was this 2:43 idea I had I wanted to turn into a game, 2:44 but I couldn't do it by myself at the 2:46 time because it was so big. So I ended 2:47 up writing a book. It published it as 2:49 The Last Hacker, this like apocalyptic 2:51 cyberpunk thing in the future. But 2:53 anyway, um but I was programming and I 2:56 was like, I have to build the 2:57 multiplayer system. and I have to do all 2:58 these things and I was building it all 2:59 from the ground up and I hated it and 3:00 then UFN came out. I was like, "Oh my 3:02 gosh, yes, we need a platform where the 3:04 multiplayer is handled, the features are 3:05 handled, everything's handled. You just 3:06 focus on the gameplay." Um, and that's 3:08 what got me into that. And I started 3:10 teaching it. I made a a pack to myself 3:12 to be the very first person to teach uh 3:14 the verse programming language and and 3:16 UFN, which I did. I got a tutorial up 3:18 right away. 3:18 >> That's how I found you. 3:20 >> Okay. Yeah. And then did that for a few 3:22 months. eventually met up with a typical 3:25 gamer YouTuber in Fortnite as well as 3:27 Chad Mustard whose family helped you 3:29 know kind of create Fortnite and make it 3:31 what it is. 3:32 >> Yeah. working with Tim Sweeney and then 3:34 we created you know started Joo got 3:36 millions of players lots of success 3:37 there and then about man just a year 3:40 year and a half ago just about a year 3:42 and a half ago I exited the company and 3:44 I have been heavy into AI and AI 3:48 pipelines and specifically in relation 3:50 to games and so doing that with Roblox 3:53 as well as Fortnite as well as generic 3:55 game development. 3:56 >> Yeah. And uh now I have Play UGC, which 3:59 is my company that helps AAA studios 4:01 port games over to Roblox or build 4:04 internal prototypes for shareholders so 4:06 they can move into Roblox because Roblox 4:08 is so huge and growing right now and 4:09 everyone's talking about it in the 4:11 industry. So I'm doing that and then I 4:13 have my my Play UC school where I train 4:15 developers on AI game development. So 4:17 the next wave of generation of game 4:19 developers is going to be with AI and 4:21 I'm I'm helping people get there. So 4:23 that was the five minute speech of 4:24 [laughter] 4:25 20 years or so. So yeah, you did it in 4:27 like three minutes. Perfect. No, look, 4:29 there's this is the perfect guest to 4:31 have today to talk about all of this. 4:33 And dude, funny enough, so Call of Duty 4:36 Ghost and people I get roasted for this 4:38 because but like Call of you know, 4:40 obviously Modern Warfare, all everyone's 4:42 favorite Call of Duty, but Call of Duty 4:44 Ghost is actually what got me into Call 4:45 of Duty. Like I was ranked number three 4:48 in the world in gun game. And like I 4:50 used to like it's a weird flex I know 4:53 but like I that got me started that got 4:56 me actually into start making content on 4:57 my channel and that's what actually kind 4:59 of like started everything for me. So 5:01 that's like a crazy call back dude. I a 5:04 lot of people didn't there's some like 5:05 hate on Call of Duty Ghost a little bit. 5:07 Man, I actually love that game. I 5:08 thought it was well made and and I 5:10 remember that app actually. Dude, that's 5:12 crazy. That's that's a full circle 5:14 moment. That's crazy. But anyway, man, 5:17 like I thank you again for coming on. 5:20 You have so much experience in all of 5:21 this. And I think there's, you know, 5:23 there's both sides of this. Everyone has 5:25 their opinions about AI here and there, 5:27 but I think overall today, what's really 5:29 important is just we got some new tools 5:31 in UFN and I really wanted to talk about 5:33 it. I know a lot of people in the 5:34 community actually was requesting this 5:36 and so it's this is perfect. And also, I 5:38 want to, you know, this is not just 5:40 Fortnite, this is UGC, you know, content 5:42 as well on this podcast. And so who 5:45 better to have also to talk about the 5:46 other side like Roblox is there I know 5:48 they're even like next level you know 5:50 with their AI integrations and stuff. So 5:52 we'll we'll get into all of that but 5:54 let's just start off here you know 5:55 because I think I really want people 5:58 kind of to understand who are listening 5:59 you know what does it actually means 6:01 when you know MCP when they open up 6:03 their editor. So Epic just put in a 6:06 brand new MCP in the UFN is actually I 6:08 think less than two weeks. It's very 6:09 new. I think most creators just think 6:11 it's like Claude or Codeex, you know, 6:13 can write verse for you now. Like what 6:16 can an agent actually do inside UFN that 6:18 makes, you know, this a different thing 6:20 entirely? 6:21 >> Yeah. I So I want to be really clear. It 6:24 a lot of people use that term all over 6:26 the place. MCP as if it's like some, you 6:28 know, magical thing. It's just model 6:30 context protocol. Rather, it's a it's a 6:32 it's a protocol, a way a handshake, you 6:34 know, between AI. like a generic system 6:36 of protocols that you know allow any 6:38 agent to kind of talk to your system. Uh 6:41 and so just to be clear, it's like no 6:42 different in some way. Sometimes it's no 6:44 different than having a CLI or an API 6:46 where an app or a game just talks to a 6:49 API talks to a server. And so it can 6:51 seem magical when it's not, but but 6:54 really the the power of MCP is with and 6:57 and with UEFN, but is it is this? It's 7:00 that you know, let's say I own a 7:01 product. Let's say I would I owned, you 7:04 know, Epic Games UFN, right? And 7:06 normally when you release an API, you 7:08 have to release the new API, potentially 7:10 deprecate old APIs. You got to send out 7:12 documentation. You got to tell people 7:13 about it. You got to get the word out 7:15 that it even exists, right? Like for 7:16 instance, I'm in Roblox right now, and 7:19 there are some amazing features that 7:20 most developers don't even know about 7:22 because they haven't received word that 7:24 it even exists, right? That's how it's 7:26 been in the past. Now, a company like 7:28 Epic Games can have an MCP where you 7:30 just call it, any AI calls it, and then 7:32 at runtime, that minute that call 7:34 happens, the the MCP will actually show 7:36 you ever all the tools that are 7:37 available, everything you can do, things 7:38 that you as a human could never even 7:40 think about. And then they can keep 7:41 adding more and adding more and adding 7:42 more and updating. But you as the 7:45 developer, you don't even know, you 7:46 don't care 7:47 >> what they're doing. All you care about 7:48 is that the robot can find it and then 7:50 use those tools. And that's the magic of 7:52 MCP, you know, is is your robot being 7:55 able to tap into this thing that the 7:57 other company created. It doesn't have 7:58 to think. It just follows this and then 8:00 it uses the tools that are provided and 8:01 then you never have to as a developer 8:03 wonder what's been updated, what's not 8:04 and things like that. It just it just 8:06 always knows. So that's kind of like the 8:08 foundation layer if that makes sense. 8:10 No, that makes sense. And again, we have 8:12 probably all skill levels, people who 8:14 don't use AI to people who are probably 8:15 very, you know, use it all the time. And 8:18 so that that was a great explanation on 8:20 how that works. You've had a chance to 8:22 come from like both sides, you know, 8:24 Roblox, UFN, all whatever, all game 8:27 development, right? I want to get your 8:28 take on like the tools themsel. You 8:30 built your own verse MCP actually server 8:34 before even epic ship, you know, their 8:36 what is their what is their verse coder 8:38 called? I forgot. 8:39 >> Yeah. Their Yeah. Their assist and all 8:41 that stuff. Yeah. 8:42 >> Yeah. Yeah. So, you know, talking about 8:44 that, like what did Epic, you know, get 8:47 right and what is the one tool you wish 8:48 they had exposed that they didn't, you 8:51 know, at at those times? 8:52 >> Yeah. Yeah. You know, honestly, I'm I'm 8:55 very glad they released the MCP because 8:57 it makes things much easier. Uh, you 8:59 know, things that you typically have to 9:01 click on, right? You know, for instance, 9:03 right now, you know, well, here's a 9:05 perfect example. One of the biggest 9:06 problems that all Fortnite developers 9:08 have is that you want to make a new 9:10 project, you want to duplicate features 9:11 you already built, right, into the new 9:13 project. But guess what? You can't 9:15 because there's no migration feature and 9:17 there's so many problems. And so what 9:19 happens is you write all these scripts 9:20 and then you manually bring them over 9:22 and then your developers spend all day, 9:24 two days reconnecting all the devices, 9:27 ret basically you broke your game and 9:29 now you have to reput it back together 9:31 and it loses its trust, loses its 9:32 reliability, all kinds of things, right? 9:34 And you know that's just cuz bad tooling 9:37 or whatever. But you know, so now what's 9:39 really cool is with the MCP, right? You 9:41 know, in the past maybe I'd have to copy 9:42 those objects, parse out all this data 9:44 from their properties, their paths, 9:46 things like that. But now with MCP, I 9:48 have like a skill that's like migrate 9:51 features over to a new map or whatever, 9:52 right? So let's say I wanted our 9:54 achievement system and I wanted, you 9:56 know, my shop system and I want the 9:59 combat mechanic system. Well, now I can 10:02 just say, "Hey, port these to my MCP, my 10:04 AI. I can say, "Hey, port these over to 10:06 my new map. Here's my new map. Here's my 10:07 old map." It will it'll use the MCP. 10:09 It'll read the Outliner files. And if 10:11 you know anything about UFM, when you 10:13 copy those, there's like a they have 10:15 what's called T3D under the hood. It's 10:17 like it's it's Epic Games like language 10:19 scripting format, template format for 10:21 like copying data, like position sizing. 10:23 And so, it copies that and it's like, 10:25 "Oh, these paths are wrong. What's the 10:27 path of the new project?" to grabs those 10:29 paths, replace them, paste it into the 10:31 new project, connects all the devices. 10:33 So then all your features are instantly 10:35 working in the new game perfectly, you 10:37 know, and that wasn't really possible 10:39 before. I mean, you could do that with 10:40 the uh 10:41 >> when they released the Python plug-in 10:42 tools, which wasn't that long ago 10:44 either, but my point is now those things 10:46 are possible, whereas before they were 10:48 not. They it was it was it was chaotic. 10:50 But a robot moves much faster than a 10:52 human. So 10:52 >> yeah. No, that that's that's perfectly 10:54 said. And I I I definitely tried out 10:57 like I've always anytime something new 10:59 comes out, you know, that's what I do. I 11:00 try to educate the audience with what's 11:02 new on our platform and I tried out the 11:04 Python and I had it making like Red 11:07 Versus Blue. Like again, it was still 11:08 very slow and janky and it was way 11:11 faster for someone to just go do it 11:13 themselves. I feel like that might not 11:15 be the case anymore. And that's what a 11:17 few months like that's 11:19 >> Yeah, 11:20 >> that's crazy how fast things are flying. 11:22 I mean, but no per perfectly said and I 11:24 think you publicly talked about this 11:26 like you made a Roblox game in like 1 11:28 hour and I want to spend a little time 11:29 on this cuz I think this is fascinating. 11:31 So, you said you built like this 11:33 playable Roblox game, one hour with, you 11:35 know, AI generated the 3D assets, 11:38 building the level, running the first 11:39 play test, you know, and then you had 11:42 your daughter, I guess, come in to be 11:43 the first human to play it. Actually, I 11:45 think that's very really cool. Like, can 11:46 you walk us through that hour? Like, 11:48 what surprised you and, you know, what 11:50 was still janky maybe at the end of 11:52 that? 11:53 >> Yeah, I mean, and this this this will 11:55 apply to, you know, all AI and and tools 11:57 when you're building with with AI. You 11:59 know, there's a one-hour game, right? 12:00 And then there's and then I built a game 12:02 actually in a week which I did publish 12:03 which which which was much better. But 12:06 so we use those kind of as buzzwords 12:08 like in an hour. But in reality the more 12:10 you can plan and the more tools you have 12:12 right up front the better it's going to 12:14 be. And so when I say I built a game in 12:16 an hour you know I did and I can I can 12:19 get a core loop in there but at the end 12:22 of the day you have to have the right 12:23 tools in place right? If I just tell my 12:25 robot out of the gate, you know, to do 12:27 that, it'll get it'll get some good 12:30 results. Even Astra, which is pretty 12:31 amazing right now, we'll get some good 12:33 results, but it won't be fun. It it 12:35 really won't. And 12:37 >> and uh so you know what you have to do 12:40 when you're leave, by the way, I when I 12:42 this again, this is just for fun, right? 12:43 I don't go around building one-hour 12:45 games, right? Like 12:46 >> Yeah. Yeah. 12:46 >> You know, but but what you can do is if 12:50 you have the right tools and systems in 12:51 place, right? You can give it 12:53 specifications of your game, the core 12:54 game loop, and then I use an 12:57 orchestration loop. Some people call it 12:59 a gauntlet loop because that's the kind 13:00 of the buzzword that's been going 13:01 around. But what happens is I I'll plan 13:04 one agent to be like the manager, the 13:06 orchestrator. And then I will tell it to 13:08 have other agents, you know, verify the 13:10 work, but not just verify, like 13:12 visually, right? So I have a skill where 13:14 it trains on like screenshots, takes a 13:15 screenshot, where it can like move the 13:17 player, it can it can do those different 13:19 visualizations. And then I have let's 13:21 say a a UI and u UX expert agent or 13:26 skill that's a third-party agent and it 13:28 will verify. So then the orchestrator 13:30 sends the image to the third party and 13:32 their job is to say no no no and this 13:34 needs to pass review. This needs to be 13:35 looking good and then it will and it'll 13:37 it'll say this is what's wrong with it 13:38 whatever and then it'll even try it on 13:40 different screen sizes. And it's that 13:41 loop is what allows you to make a game 13:43 very fast that turns out good with like 13:46 the core foundational stuff on there is 13:48 having third-party agents verify the 13:50 work of the others. Otherwise, the 13:51 agents tend to lie and be 13:53 overenthusiastic and optimistic. And so 13:55 that's how you build like systems very 13:58 quickly, but it doesn't happen 14:00 magically. It happens when you have all 14:01 your tools in place ready to go. If that 14:04 makes sense. That makes perfect sense. I 14:06 mean, I think even, you know, AI has 14:07 been out for years. And I think a lot of 14:09 people it's like how you prompt, right? 14:11 Or whatever. That was like the big thing 14:12 of getting the best results. And I think 14:15 that's probably like the next level 14:17 where you're come in, where your 14:18 business comes in, what your school 14:20 comes in is just like I think there's so 14:22 so many things, so many tools, there's 14:24 so much noise about do this, do that, 14:27 set it up. And I and and and when you 14:29 see these people make make oh, I made 14:31 this game in and out. You see that all 14:33 the time or whatever. We've seen this 14:34 all week with Astro, right? People 14:35 making Rocket League, whatever. And and 14:38 it's just like but they didn't they 14:40 probably set it up right up up front 14:42 first, which I think probably 90% 14:45 probably 99% of people don't know how to 14:46 do or haven't thought of of just that 14:48 when they think about they just think 14:50 someone's typing a prompt and then it's 14:51 done. There's a whole system of it, 14:54 right, that you have to set up. And I 14:56 think you said it very well, like it's 14:57 just like thinking about a game 14:59 development studio. You have all these 15:00 people. You have a a programmer. you 15:03 have a model or you have UI, UX, all 15:05 those things. And so you now you just 15:07 make many dev developers, right? Or that 15:10 can do those specific things and then 15:12 have a manager on top of them 15:14 overlooking. So it's like building out 15:15 that whole system first will then give 15:18 you the better product, which still 15:19 needs probably still some love. We're 15:21 not at that point yet where you can one 15:23 prompt a game and make it, but we're 15:26 getting close. 15:27 >> Yeah. And I don't think I don't think it 15:29 really needs to go there. Like I I I 15:31 talk about AI all the time. And 15:33 >> here's the thing. It to me, this is not 15:35 some like religion for me or whatever. I 15:37 I like producing things. Like I'm a 15:38 creator. That's what I like to do. And 15:40 and everyone has their different things 15:42 that they like. But for me, right, when 15:44 I wanted to make my video game, like 15:45 this idea that I had that's in my brain, 15:47 right, I wanted to make the game, but 15:49 that game required multiple humans, 15:51 probably like a $5 million budget and 15:52 all these other things. And so 15:54 therefore, I was unable to produce a 15:57 thing that I wanted. and and and so 15:59 that's where AI really really changes it 16:01 here. So we talk about those like those 16:03 oneshot games and things like that, but 16:04 at the end of the day, whether it's the 16:06 player or the creator, you know, people 16:08 want to have experiences that are fun, 16:10 that are unique, that are interesting, 16:11 and and for me, AI just allows us to do 16:14 that that quicker. And so what I 16:15 actually tell my students and anyone 16:16 who's new getting into AI with games is 16:19 is not even to, you know, think about 16:20 those systems, those tools, or building 16:22 the perfect setup. Do that over time. 16:24 And honestly, the best thing you could 16:26 really do is just right in the beginning 16:27 go back and forth with your agent on one 16:29 feature. This that this that you test it 16:31 and that may be slower, but guess what? 16:33 You you've already probably just with 16:34 one agent replace three different roles 16:36 that were in the past necessary, right? 16:38 Because now the agents are getting 16:39 really good with uh user interface and 16:40 UI for games and even chopping it up, 16:43 you know, in the right formats. And so, 16:45 >> you know, now you can work with Astra, 16:47 for instance, and its image gen feature 16:49 and you get unlimited image generations 16:51 if you're on the pro plan. And so I have 16:53 a template in in codeex, a GPT, the 16:56 desktop GPT app where basically I have 16:59 one for let's say for Roblox that it 17:01 knows everything there is to know about 17:03 Roblox UI, how to chop up the nine slice 17:05 images, what formats, things like that. 17:07 And so I'll feed it mockups and then it 17:09 will extract every item from the mockup, 17:10 recreate it, cut out the background, 17:12 create the nine slice images, import it 17:14 directly into the game, assign it to the 17:16 UI, and all that happens with one 17:18 prompt. So even new people getting into 17:20 games with AI can do that just back and 17:23 forth next and next. So while I prep 17:24 panel a lot of things, you don't have 17:25 to, right? It's it's a lot easier now. 17:27 It's because the tools have become so 17:29 much better. 17:29 >> Better. Yeah, for sure. I think I mean 17:32 that's where it's going. I mean, it's 17:33 getting a little easier for I I actually 17:35 think this MCP is probably that for a 17:38 lot of developers. I think people have 17:40 tried it probably using AI and game 17:42 development as some at least Fortnite 17:44 side and probably been like ah getting a 17:46 average results or or whatever and I 17:48 think the new Astra the new MCP tool is 17:51 probably opening a lot of people's eyes. 17:53 It it and I want to point out with that 17:55 too like one of the most powerful things 17:57 with that I would say because there's 17:58 some problems but I will say the most 18:00 powerful thing with that is is the verse 18:02 programming language. It just does very 18:03 well with it. I mean, some people 18:05 probably disagree to some to some degree 18:07 on this, but honestly, you can not have 18:09 to learn to you can not have to learn to 18:11 code. You cannot learn verse and and it 18:13 does so well with that. Now, the 18:14 problems with the UFN MCP are other 18:17 things we can talk about later, but that 18:18 aspect alone, if you don't know how to 18:20 code and you want to make like a 18:21 Fortnite game, for instance, 18:23 >> the MCP is very great with verse. In 18:25 fact, that's the number one thing that 18:26 it's done the best at with AI is the 18:29 programming language itself. And so, 18:31 that's really that's huge. You don't 18:33 have to learn a programming language now 18:34 to make a game. And I I think that's 18:36 worth pointing out for sure. 18:38 >> No, that is crazy. And it's it's such a 18:41 weird time because I remember I think 18:42 you've talked about this openly. I think 18:44 it's pretty interesting. It's a it's a 18:46 good conversation. It's just I remember 18:48 1.0 Fortnite 1.0, right? I was a top 18:50 programmer, visual scripting language. 18:52 There was only a few of us that actually 18:53 really like understood the whole thing, 18:56 right? And I maybe I tied my identity to 18:58 to to that. And I remember when Verse 19:01 came out and I I really just I had this 19:03 like like I just spent five years 19:05 learning this thing that's gone 19:06 [laughter] gone. You know what I mean? 19:09 Um and I I think we're there. We might 19:11 be coming there with I mean in 19:12 programming in general, right? Um and I 19:15 I know you have some fun takes on that 19:17 on that concept and just like around 19:20 thinking that because that's you you 19:21 have the same thing, right? I mean, 19:23 you've learned all of these different 19:24 programming languages and I'm not saying 19:27 you're out of a job now with not knowing 19:28 that cuz it probably helps you still. 19:30 There's probably some still reasons to 19:31 to know that those things, but like you 19:33 know what what's your take on that? It's 19:35 like it's crazy. 19:36 >> Yeah, it's it it is crazy and it's it's 19:38 exciting. It's also scary and and and 19:40 and to actually be fair, I you know, I 19:43 so I've programmed in many languages and 19:45 keep in mind my the two companies ago, 19:47 my whole company was training people to 19:48 become engineers. 19:50 >> Yeah. 19:51 And my goal is to help them get jobs. 19:52 You know, what I what I guess what I 19:54 would call bluecollar programming, which 19:56 is, you know, oh, front-end developer, 19:57 backend developer, you know, iOS 19:58 developer, like, you know, you're not 20:00 you're not reinventing the internet or 20:02 anything, but you're doing complicated 20:04 engineering tasks all day long. And 20:05 that's fine. And in a way, those roles 20:07 are getting are getting replaced. Like, 20:09 even though I've been building, let's 20:10 say, iPhone apps for x amount of years 20:12 or games x amount of years, my ability 20:14 to get a job doing only that is probably 20:16 impossible at this point. if even if I 20:18 with my resume background, everything. 20:20 Why? Because if somebody's now going to 20:22 hire a human, let's say to be a gameplay 20:24 engineer, a C++ gameplay engineer. Okay. 20:27 Well, there's now 40,000 developers 20:29 available because the games industry is 20:31 imploded. Okay. Well, guess what? Now I 20:34 for the same price as before, you know, 20:37 I can get a top engineer, one of the 20:39 best in the world, and they they better 20:41 be the best person ever, right? and and 20:43 and those companies have that ability 20:45 now because the pool of of talent is so 20:47 huge and there's not enough jobs for it 20:50 right now. So it it is dangerous in that 20:52 sense if you're a programmer and all you 20:54 kind of do is program in one thing 20:57 because if you lose that job like how do 20:59 you get another job and I already know 21:00 so many game developers and and other 21:02 engineers that are struggling. So I mean 21:05 it is a real thing. It's a scary thing. 21:07 But also truly the solution is truly to 21:10 become an architect and master AI. I 21:12 believe I believe the architects are the 21:16 future of of creators. And when I say 21:18 architect, you know, I'm talking about 21:21 creative vision, you know, in your in 21:23 your lane. So, let's say I was a 3D 21:27 animator or let's say 3D character 21:29 artist. I was amazing at like the kind 21:30 that Epic Games hires, right? Yeah. 21:32 >> That was I'm amazing at it. and let 21:34 let's say I lose my job, you know, hey, 21:36 it's hard to get in and you're just this 21:37 amazing AAA character designer. Well, 21:40 what I would do is I would actually 21:41 master AI and all those tools and take 21:44 my years of experience and skills and 21:46 distill those into my robots and and 21:49 now, you know, instead of hands-on 21:51 clicking in in Maya or Blender for days 21:53 to make your character, imagine you just 21:55 talking with GBT voice and your tools 21:57 are in place and you're like, "Okay, now 21:59 make the hair look like this. Now do 22:00 this." And you're just talking like 22:01 freaking Iron Man, which by the way, 22:02 that's available as of right now today. 22:04 >> I've got systems that do that myself. 22:07 >> Yeah. 22:08 >> So you but you're you're architecting 22:10 the skill, the experience, everything, 22:12 and it goes forward. And so I do think 22:13 that the middle layer of of us 22:15 programmers and engineers who've been 22:16 building those products, that's going to 22:18 be done by robots. I it's already 22:19 happening. 22:20 >> Yeah. 22:21 >> But there's a huge opportunity to take 22:25 what makes you special as a creator and 22:27 then expand it with with with AI. Yeah, 22:30 and that makes sense. Look, I think it's 22:32 a fun conversation to and if you think 22:35 about it, like let's say we get to the 22:37 point where we are one prompting a game, 22:39 right? We all can do it, everyone has 22:41 it, whatever. But like, is everyone 22:43 going to be playing your game? Like, no. 22:45 No. Like, it really does take someone 22:47 that understands gaming. It also, do you 22:49 have an audience or do you have an IP? 22:51 Like, do like people want to that's why 22:54 Fortnite Roblox people go to play with 22:56 their friends. Like, if there's no one 22:57 there to play, they wouldn't be on those 22:58 platforms. we're going to still have 23:00 those same problems. So, I think if 23:02 anyone's looking at this with a little 23:04 bit of doom and gloom, I still think 23:06 there is 23:07 >> It's still not easy. 23:08 >> Yeah. No, exactly. Exactly. And just 23:11 because it makes a game, there's so I've 23:14 help I help out lots of creators I 23:16 consult and help grow their successful 23:18 their games that's been successful or 23:20 revive them, all of those things. and 23:22 they have the game, they have the 23:24 audience, and there's still so much that 23:26 goes in after you publish a game and 23:28 understand your audience. This is an 23:29 update and the platform. There's this 23:32 whole, you know, algorithm on how to get 23:34 your game discovered. Like, we know some 23:36 of it. We all try to figure it out. And 23:39 even though you have all those things, 23:40 there's still so many other factors that 23:43 it's not just going to be like, oh, it's 23:45 over. Like, there's still all those 23:46 things in place. So I think people that 23:49 are staying ahead of the curve, know the 23:51 platform are still going to have, you 23:53 know, for at least for a little bit for 23:55 a while, you know, I guess, you know, a 23:58 leg up. 23:59 >> Yeah, I would I would agree with that. 24:00 Yeah. 24:00 >> Yeah. 24:01 >> And and evolving evolving those skills 24:03 is is always incredibly useful anyway or 24:05 any any skill. I mean, any market that 24:07 evolves, you got to evolve with the even 24:10 everything. So I I agree with that. 24:11 Yeah. 24:12 >> Yeah. No, it it's interesting. I just I 24:14 like to point that out because I still 24:15 think everyone has that that like doom 24:17 and gloom outlook and it it is a scary 24:20 place but in some some regards but if if 24:23 you're in it learning the tools and a 24:25 lot of people have said that not not 24:26 just yourself is just like what you need 24:28 to do is learn like the people that are 24:30 going to get the jobs are the people 24:32 that are going to be the the people that 24:34 know know how to do it and also know how 24:35 to you use AI with it to you know 10x 24:39 themselves right so let's get into the 24:40 UGC pipelines we talked about a little 24:42 bit you Let's talk about, you know, 24:44 doing these things over and over because 24:46 that's where it gets interesting for a 24:48 creator who's trying to get work out 24:49 every day. So, your UGC pipeline took a 24:52 workflow from about 3 hours per item 24:54 down to under five minutes with one run 24:57 finishing in 48 seconds. You know, what 24:59 part of that pipeline was the hardest to 25:00 pull the human out of? 25:02 >> Yeah. 25:02 >> And then would you trust it to publish 25:04 on? 25:05 >> Yeah. So, okay. So, let's talk about 25:08 this for a second. And I'm going to 25:09 specifically speak in terms of technical 25:12 art. And by the way, it I don't mean to 25:14 shed too much doom and gloom here, but 25:16 that is the one slice of game 25:18 development where there's not going to 25:19 be any jobs, you know, at at the low 25:21 levels is tech art because it's being 25:23 replaced so quickly and and and let me 25:26 give you an example or become a person 25:28 who can master AI with technical art. 25:30 And you got you got yourself something 25:31 real special here. But let's talk about 25:33 this. Uh I'll talk about UEFN and and 25:35 Fortnite. The first game that I 25:37 published was with Typical Gamer and 25:39 Chad Mustard and we we published uh I 25:41 think it was only up time travel 25:44 >> during the only up. 25:46 >> Yeah. 25:46 >> Yeah. It did really well. It it was it 25:48 was the second most successful game in 25:50 Fortnite for that. Well, I mean actually 25:52 I think those two games were most 25:54 successful on Fortnite in the world in 25:55 terms of total place. But so that game 25:58 was interesting because you know the 26:01 well the other team members you know 26:02 manually place let's say the objects to 26:04 make it fun and all these things but 26:05 where where I ended up spending most of 26:07 my time was in oh my gosh the technical 26:09 art managing Unreal Engine the memory 26:12 management of it all you know lods on 26:14 assets you know memory management all of 26:17 these things and keep in mind I would I 26:19 was going across the internet finding 26:21 assets you know 26:22 >> I had to find you know a typical gamer 26:25 wanted a very specific Egyptian asset 26:27 that they only had on like Turbo Squid 26:29 and it was like a million polygons and 26:32 >> I was like well we don't have an artist 26:33 to do all this stuff right and so I 26:35 literally had to take that I had to 26:37 learn some Blender stuff decimate the 26:38 object get it into the game and so I 26:41 spent again 90% of my time managing 26:44 assets for that I mean none of that 26:45 would have been possible if I had not 26:47 done that work that game would not have 26:48 come out but my point is like importing 26:51 assets formatting assets all those types 26:52 of things and those are repeatable those 26:55 are some of the easiest est things to 26:57 have AI do. And so now, here's a perfect 27:00 example. So, I released my own little 27:02 play, it's called PlayKit. It's it's 27:04 create.playc.com, but it's a series of 27:06 skills for UFN, and I have a a Roblox 27:08 one as well, too. And so, for instance, 27:10 now I have for asset management for 27:13 UEFn. And so, I can now take any assets, 27:16 whether it's FAB or wherever, and 27:18 whether it's FBX or GB and whether the 27:21 formats are all differently, and it goes 27:22 through all those things. It 27:23 reprocesses, reformats them, puts them 27:26 into Unreal and or UEFn the way that 27:28 they're supposed to be in the proper 27:29 formats, indexes them properly, and that 27:32 process is just solid now and just takes 27:34 takes minutes. And and of course, we try 27:36 to use deterministic or repeatable 27:39 things whenever we can in relation to 27:42 like, you know, editing polygons or 27:44 importing things like that. So the AI 27:46 will write scripts. So the AI is not 27:47 thinking every time and costing you a 27:49 fortune. We're building these tools once 27:50 and we're just repeating them. And and 27:52 so now I have I also have that for 27:54 Roblox and Roblox is way more 27:56 complicated in management of assets 27:57 because not only they have to be 27:58 formatted a certain way, they have 27:59 different polygon limits and different 28:01 management of of things and then 28:04 >> you have to upload them to your 28:05 particular creator account or your own 28:07 account and make those assets available 28:08 and publish them and all these things. 28:10 But that's all automated too. And so 28:11 now, man, I wish I had this on a demo 28:13 for you here, but I I could literally 28:15 open my phone on my Grockbot and I could 28:17 I could tell it right now. I could say, 28:18 "Hey, I need to make a new uh a new 28:20 handbag 3D object for my my closet 28:23 cleaning game on Roblox." It'll and 28:25 it'll know where it is. So, can you just 28:27 go ahead and have Codeex make me the 28:30 image using our game design image tools? 28:32 And what it'll do automatically is make 28:34 it in three force image view with a a 28:36 solid background so a 3D asset generator 28:38 can make it. And then I'll say and then 28:40 just send it up to Meshi API and get it 28:42 into the Roblox game and then publish a 28:45 private version of the Roblox game and 28:47 so I can test it. And so it'll do all 28:49 that maybe five minutes and then it'll 28:51 tell me, hey, it's done. And I'll be on 28:52 my walk and I'll just pull up my phone, 28:54 go to Roblox and I will play the game 28:57 with my little new little test bag in it 28:59 and literally on a walk 29:01 >> and that's a real thing. Yeah, that's a 29:03 real thing. and and so we are we are 29:05 there and I do that for oops sorry I do 29:06 that for studios as well as for my own 29:08 tools but 29:09 >> and it takes some setup but yeah that's 29:11 that's like where we are with with tech 29:13 art in that in that regard. 29:14 >> Wow. No, that that is wild. That is wild 29:18 and it's I don't know. I'm I'm kind of 29:19 speechless for once, but like so 29:23 you can just talk to Grock and then it 29:26 will connect to other AIs, connect to 29:28 other I don't know, not repositories or 29:30 gits, but like the MCP is for me or 29:33 whatever. 29:34 >> Yeah. 29:34 >> And then create it, put it in your game, 29:37 publish unlisted so you can play it on 29:39 your your phone. That's 29:41 >> and specifically Grockbot. 29:43 Grockbot. Uh which some people used Open 29:46 Claw in the past. I did as well too, but 29:47 nothing is as good as Grockbot. You can 29:49 literally manage your your life. It even 29:51 hosts a a virtual machine on your 29:53 computer. It can log into websites for 29:54 you. So, so long as your computer's on 29:56 or you have it on a server somewhere. 29:57 You can do all of your normal work just 30:00 through speech on your phone. And this 30:01 I'm really nerding out here, but I'm 30:02 actually about to order some Raspberry 30:04 Pi kit stuff. So, in my house, 30:06 >> I can have these little robots in 30:08 different places of the house where I 30:09 can just talk to my Grockbot all the 30:10 time, anywhere in my house, and it'll 30:12 just do the things. Yeah, [laughter] 30:14 you could definitely. I mean, there's so 30:15 many angles here for automation you can 30:18 do with AI. No, that's pretty crazy to 30:20 think about. And I've always kind of 30:22 counted out Grock, to be honest. Like, 30:24 so that's it's interesting to hear 30:26 though. 30:27 >> Yeah, Gro Grock bot. Okay, there's Grock 30:29 4.6, which is great for coding and 30:31 stuff. Grock bot is like that harness 30:33 that that exe Elon Musk Space X released 30:37 specifically for kind of managing your 30:38 business in life. And it's so freaking 30:41 good. Like I just can't overemphasize 30:43 that enough. But that's that's how I'm 30:45 able to automate things so well is cuz 30:47 they do most of the work in the coding 30:49 on that, not me. 30:50 >> Yeah. No, as if I need more things for 30:53 on my phone to take away [laughter] from 30:55 my attention and now I can just 30:56 constantly be at my 30:58 >> Yeah. 30:58 >> on my phone but working my PC. That that 31:01 blew my mind. That's crazy. Well, I want 31:03 to stay a little bit on the same topic, 31:05 but it's kind of like building teams 31:07 today now. You know what I mean? And you 31:08 kind of just gave us like the perfect 31:10 like a perfect example of this. So you 31:12 know you were CTO at Drogo and you 31:14 launched with a team around 20 people or 31:16 got there more more right. 31:18 >> Started with us three grew grew to that. 31:20 Yeah. 31:21 >> Grew to 20 which that's crazy. So like 31:23 if you were starting that exact studio 31:25 today with the agents you have now like 31:28 how many of those 20 seats would you 31:29 actually hire? 31:31 >> Oh boy. 31:32 >> It's a tough question. 31:34 >> Yeah. So, I mean, I really can't get 31:36 into this too much, but I mean, that is 31:38 that is what I'm about to do and and 31:40 going to be working to launch here in 31:42 the near future in relation to a studio 31:44 and creating and and and that is going 31:46 to be my my plan and path. And so, you 31:50 know, my you so I'll just tell you my 31:52 what I think is is for my skills and 31:55 what I think, you know, we're going to 31:56 need, but any any studio uh who wants to 31:58 start off, but you know, having a 32:00 technical person like myself who 32:02 understands AI and can manage them, but 32:04 also understands programming, I I think 32:06 I think that's really important because 32:08 it's one thing to say, "Hey, make me a 32:10 game." It's another thing to say, "Hey, 32:11 we need to optimize this and we need to 32:13 make this faster." Like what is our draw 32:15 like draw call count? like we need to 32:17 you you need to know certain things. So 32:19 basically CTO technical person but all 32:22 all of you know probably for the first 6 32:25 months to a year you know all the 32:27 engineering work will be by my agents 32:29 and even just by myself and then the 32:31 other person that I think is really 32:32 important is an art director 32:35 >> but an art director who is is pro AAI 32:37 and AI forward and uses AI tools and 32:39 pipelines because they're going to have 32:41 success. They're they're going to know 32:43 what needs humans, what needs this, what 32:45 needs that. They're going to organize 32:46 it. They're going to make sure it all 32:47 stays in there. And I I mean I literally 32:49 believe that between an AI forward art 32:52 director um and a like um an AI forward 32:55 like CTO or engineer, you know, you can 32:58 actually build a successful version one 33:00 of a game and actually make money with 33:01 it and and move it forward and you know, 33:03 you can still outsource of course as 33:05 well too if you have assets and things 33:06 because now at the end of the day still 33:08 at the the hardest thing to do which 33:10 these tools are getting better with is 33:11 you know maybe making very custom 3D 33:14 very custom animations that are really 33:16 amazing and and AI is doing okay with 33:18 that. But you know those are things that 33:19 you know I might outsource in the 33:21 process but they would have all the 33:22 specifications and images and things 33:24 like that. And but honestly though now 33:26 even then like I might be more tempted 33:28 to get like a RCOO like bodysuit 33:32 animation thing and do my own animations 33:35 because then that solves all the 33:36 problems that AI couldn't do in the 33:37 first place making things look lifelike 33:39 and that's just data you know. 33:42 >> Yeah. But so anyway, and and the UI and 33:44 UX at this point, like man, I I don't 33:48 want to like hurt anyone's feelings, but 33:50 like I'm I'm okay at a double AAA level 33:54 shipping the UI that 33:56 >> you know, some of the new modern AI 33:58 makes if it's done correctly. And so 33:59 yeah, I think that three-person teams or 34:02 or five maximum is more than enough to 34:05 create a 10 or 20 or $50 million studio 34:08 in games these days if they know the 34:10 tools if they're using them correctly, 34:12 you know, all those things. But uh I I 34:14 do not believe any company ever again 34:16 should go get a massive amount of 34:18 funding and then hire a bunch of people 34:20 and launch. I I think it should be lean 34:22 teams, AI forward people, because you'll 34:24 be surprised at how fast you go and the 34:26 quality doesn't change. You still want 34:28 the same goals, the same fun. You're 34:29 still in it, living it, and breathing 34:31 it, and playing it, and testing it. 34:32 You're just doing it faster, 34:35 >> saving a lot of money, and and and 34:37 scaling yourself. So, that's that is 34:38 what I'm going to be doing. Yeah. 34:40 >> For the very few. 34:41 >> So, you're going to prove it out even. I 34:42 know we're like speculating here a 34:44 little bit, but like I mean, you're 34:45 going to try you're proving it. You're 34:46 going to prove it. So, like that. 34:49 >> Wow. 34:49 >> Yeah. I mean, that Yeah, that's exactly 34:51 that's exactly what's happening. But 34:52 I've already proven it in the different 34:54 areas. Yeah. 34:55 >> Now it's just like time to like put it 34:56 together and actually build something, 34:58 you know, uh really just amazing 35:01 >> with these tools. So 35:02 >> yeah. No, that that is wild. That that's 35:06 that's crazy to even think about right 35:08 now. And I'm sure a lot of people are 35:10 like whoa listening and look whatever 35:12 side you are on AI and there'll always 35:14 be I think there will be pullback on AI 35:17 in the sense of some people want to make 35:18 their own games and they everyone has 35:20 that right. You know what I mean? But I 35:22 think it is important to understand what 35:24 is going on in the world. What these new 35:26 tools cuz there's always new things 35:28 getting released new. It's hard to 35:30 understand like what does it actually 35:31 mean and they everyone says oh this 35:35 industry is over this is industry 35:36 whatever all these things but just like 35:38 what is is it true is this actually 35:40 happening right I am speechless again on 35:43 this podcast and and to think about u 35:45 you know you could probably make a AAA 35:47 game with three people. 35:48 >> I it's crazy. I mean it at the end of 35:51 the day it's it's all these tools right 35:53 like Tim Sweeny's right about a lot of 35:54 these things when he speaks about like 35:56 computing power and and what broke the 35:58 game industry all these things you know 35:59 talking about you know all the focus on 36:02 better technology right like oh we want 36:04 better rendering better rendering we 36:06 want more realistic graphics and then so 36:08 even if you think about it like the 36:10 original teams like on Doom and stuff 36:11 were just a handful of people 36:13 >> game companies used to be small and and 36:15 what I was uh telling the other M on the 36:17 other podcast is like I met a guy who 36:20 worked on like one of the Marvel games 36:22 on Super Nintendo like in the '90s and 36:24 they had like some rush like job where 36:26 they had to build the whole game like in 36:28 seven days and they shipped it you know 36:30 you know worldwide and and I'm and he 36:32 was talking about how he's doing the 36:33 pixels and he's doing the sound and the 36:35 art and he was a programmer and you know 36:37 and and you know you see but those games 36:39 succeeded and and then you know come you 36:42 know let's go to 2015 where it's like oh 36:45 300 person team you know or like for 36:47 instance uh Sledgehammer Studios, you 36:49 know, they they made one of those Call 36:50 of Duty games and and I I heard those 36:51 hor horrific stories of where they would 36:54 like scale up 300 environmental artists 36:56 like at the very end of the game and 36:58 crunch them and work them dry and then 37:00 fire them all off the project after and 37:02 they repeat that process all the time. 37:03 And I was just like like this is doomed 37:06 for failure like like you know and and 37:08 so and then guess what? Roblox comes 37:11 along and the new generation doesn't 37:13 even freaking care about, you know, uh, 37:16 you know, RTX ultimate rendering, you 37:19 know, on this and that and and ray 37:21 tracing and, you know, like they're just 37:23 like, 37:23 >> I just want to have fun and play with my 37:25 friends, right? And so those things 37:27 combined have just almost created this 37:29 dysmorphia in the game industry in 37:32 Western games. It's it's different in 37:33 Japan and other places. We we kind of 37:35 went crazy over here in the US for some 37:37 reason, but, you know, it's now it's all 37:39 changed, right? There's 40,000 people 37:40 who have lost their jobs and it's hard 37:42 and the industry is not putting money 37:44 back into new studios. Even Microsoft 37:46 themselves, you know, they're going into 37:48 streaming. They Xbox, we don't need a 37:51 console. Let's just put it on your TV. 37:53 Let's get it on Apple, you know, where 37:54 you can just pull open your phone or a 37:56 controller and play the game from 37:57 anywhere. And, you know, that's kind of 37:58 what people wanted anyway. 38:00 >> Yeah. 38:00 >> And so, the console wars, you know, are 38:02 changing. Everything's changing. It's a 38:04 little chaotic. Um, but you can do so 38:07 much still. And but keep in mind also 38:09 like this is like 1% of the industry 38:11 like that knows about all these tools 38:12 that we're even talking about. Yeah. If 38:14 you go talk to your neighbor about AI, 38:16 they be like, "Oh yeah, yeah, I used 38:18 Chad GBT once. It was great." Like I uh 38:20 like I was trying to find some flowering 38:22 tutoing 38:24 tutorials, you know, and and I was just 38:25 like, you know, they don't they don't 38:27 know. So we are in a bubble for sure, 38:29 but it's a tech bubble and we live it 38:31 and people are evolving, losing jobs and 38:33 so you have to take it serious in the 38:35 same breath, too. So 38:36 >> yeah. No, that there's so so many 38:38 nuggets there. And I think that the 38:41 biggest one was just that like the 38:43 highfidelity like I think UF when it 38:45 came out that was kind of like the AAA 38:47 UGC kind of that's what everyone was 38:49 kind of going for and we all have fallen 38:51 into that too like 38:53 >> let's make red verse blue HD or whatever 38:57 you know what I mean just like high qual 38:58 and and at the end of the day and still 39:00 pro like everyone always likes to say 39:03 you know like slot whatever you want to 39:04 say slot games this or that about pay 39:07 bread what whatever there's a lots that 39:09 go into that to make those games 39:10 popular. I don't I don't believe that. 39:11 But just because it might look very 39:13 basic and at the end of the day, I think 39:16 we came up like when we grew up it was 39:19 always about the best graphics, the best 39:21 game. I wanted the newest graphics card 39:23 cuz I wanted to play this and that like 39:25 you know that was how we grew up, right? 39:26 And so naturally 39:28 >> we didn't have anything better, right? 39:30 Like I remember Jurassic Jurassic Park 2 39:32 came out on Super Nintendo and they had 39:35 an a cut scene in cinematic with actual 39:37 audio voice and it was like let's go go 39:40 and I was like I was like there's voices 39:42 on Super Nintendo how do they do that 39:44 cuz there typically isn't it's like you 39:45 know 16 bit music and so you know it's 39:48 almost like there was a point where we 39:50 stopped caring like we got enough like 39:52 we we just don't care anymore like we're 39:54 great like we're good. I think too at 39:56 the end of the day it just a lot of it 39:58 comes down to platform devices. 39:59 Everything's so expensive now and like 40:01 what runs well on your phone or my 40:03 Switch and what are my play friends 40:05 playing, right? They don't necessarily 40:07 it's a it's it all comes back when 40:09 you're creating games like where's the 40:10 fun? Is this fun? What's your game loop? 40:12 Is this fun? Is it going to get people 40:14 to come in and and want to play with 40:15 their friends and invite them and and 40:17 that's still that's what it is all 40:19 about. None of that other stuff matters 40:21 at the end of the day. 40:22 >> And that's my outlook too. that and 40:23 that's why I don't like I don't really 40:24 even I care about AI definitely but 40:26 that's also why I don't care. I'm just 40:27 like I think the thing I came here to 40:29 make games and have fun like I end up 40:31 becoming like a martyr and a profit of 40:33 AI at the same time like but I don't 40:36 even care. I just I just want to work 40:38 faster and you know and make the games 40:41 that I actually want to make you know 40:42 like that's what it's always been. So 40:43 you know the the motive doesn't have to 40:45 change. Nothing has to change. You know 40:46 there's a bunch of tools. 40:48 >> No for sure. Look, there's and there's 40:49 always been tools and we might not have 40:52 had them in UFN, but there's been a lot 40:54 of uh plugins that help on here as 40:56 Python script. There there's been other 40:58 things that have done similar things. 41:00 Now, AI just makes a lot of that easier 41:03 today. But again, everyone has a choice. 41:05 If you don't want to use it, you don't 41:06 have to at the end of the day. But 41:08 again, if you do, you shouldn't also 41:09 feel bad about it, I don't think. And 41:11 and you were very early on. I mean, you 41:13 were very early on like back in the day 41:16 where like death threats were going all 41:18 over online if you said anything about 41:20 AI. 41:20 >> I got voodoo threats like people cursing 41:22 me with [laughter] 41:23 >> curse. 41:24 >> Yeah, maybe deserves that. 41:27 >> Yeah, no one deserves a voodoo curse. I 41:30 don't [laughter] 41:30 >> No, it didn't work, I guess. 41:33 >> Or maybe it did. I don't know. 41:34 >> Maybe. Maybe time will tell. That's cra 41:37 That's crazy. Well, let's get back kind 41:39 of back into the questions, but again, 41:41 me and you could talk about this all day 41:43 and I I'm sure again this is just I 41:45 don't know a lot of this is new to me 41:46 too, right? I'm I haven't been that in 41:49 that side of of development very and and 41:51 when I am it's just for like oh there's 41:52 an update, let me just check it out 41:53 really quick, you know? Yeah. 41:55 >> So, it's it's fascinating. But let's 41:57 talk about Roblox and Epic. I want to 41:59 kind of bring it back to like tools 42:01 creators have in front of them. So like 42:03 if you're building in Fortnite, it's 42:04 probably good to also understand what's 42:06 possible on Roblox side or any other 42:08 UGC. And I I will always say this like 42:10 you shouldn't just probably corner 42:11 yourself to one UGC, you know, but like 42:14 Roblox has had MCP now for a while 42:16 longer. They have mess generation, 42:18 procedural, you know, models, AI, 42:21 >> they have every they have every they 42:22 have everything. 42:24 >> That's crazy. So, I guess I mean that's 42:26 the question. Is Roblox, you know, ahead 42:28 of letting AI build games right now? And 42:30 like what should Epic I guess co copy 42:32 first or like what what do they need to 42:34 do to you know get get there? 42:36 >> I I mean I I have so much to say on this 42:38 subject, right? I I I just uh you know I 42:42 want to I want to respond in a way that 42:43 doesn't like soapbox preach anything 42:45 here, but 42:46 >> yeah, 42:47 >> you know, I here's some things that I 42:49 can I can say everyone pretty much 42:50 agrees with, right? like, you know, the 42:52 audiences aren't really changing on on 42:54 Fortnite, right? Like, if you look, you 42:56 can just do an analysis of the top games 42:58 and and I mean, I kind of been out at 43:00 Fortnite for a year and I came back and 43:02 it's all the same stuff. 43:04 >> Yeah. 43:04 >> And I would I would actually say it's 43:05 it's a little bit worse. Like there was 43:08 games that I used to play that I can't 43:09 even find now in terms of like the the 43:11 types of games. 43:12 >> Yeah. 43:13 >> You know, and and so I I do a lot of 43:15 research on on UFN and Epic Games 43:17 because I spent three years in this 43:19 space. very passionate about it. 43:21 >> So very frustrated when the leadership 43:24 says things and they don't come true and 43:26 then they don't acknowledge the things 43:27 that they didn't do and the things they 43:30 did do, you know, and then they, you 43:32 know, fire thousands of employees and 43:34 then, you know, Tim Sweeney goes on 43:36 stage talking about how Roblox is taking 43:38 over the world and I'm thinking like 43:40 this is not what you want to be telling 43:41 your your shareholders, you know, like 43:43 your people are invested here. He's 43:44 like, you know, they're just taking up 43:45 too much market share, you know, more or 43:47 less, and we need to worry about 43:49 Monopoly. I'm just like, did you just 43:51 like do anti-Roblox monopoly speech from 43:53 your own like conference? But my point 43:55 is, as a technologist, I cannot 43:57 comprehend what Epic Games is doing with 44:00 UFN. I've done a lot of research on what 44:02 they're doing, their time, like all 44:03 their board members. I researched 44:05 everything because I was again, I was 44:06 invested like how they operate, you 44:08 know, and you know, for instance, Tim 44:09 Sweeney, he gets all the he gets he can 44:11 call the shots. No one can fire him. He 44:13 can't be fired. He calls all the shots. 44:15 At the end of the day, whatever he wants 44:16 will happen. He he has he has a majority 44:18 vote. So that's a fact. And so with all 44:21 that in mind, you know, there's no like 44:23 other companies have boards of directors 44:24 like put pressure on people and things 44:26 that and they do have pressure because 44:27 they have investments from other 44:28 companies and stuff. But my point is 44:29 what are they doing all of a sudden? You 44:31 know, they announced, right, that 44:32 Fortnite's struggling, UF's struggling. 44:34 It didn't pan out the way they wanted. 44:35 And now we're gonna do this big huge 44:37 thing where we merge Unreal Engine with, 44:39 you know, UFN and create this like kind 44:41 of mix, you know, these singular tools. 44:43 And I'm sure 44:44 >> Yeah. And I'm sure they will be like 44:47 agent more agent focused, right? 44:49 >> Yeah. 44:50 >> But, you know, it's this thing that's 44:51 launching so far from now. And anytime 44:53 they've ever launched anything, it's so 44:55 far behind schedule. And so anyway, I'm 44:57 supposed to believe you, this is my me 45:00 just as a human living on their 45:02 platform. I'm supposed to believe that 45:04 even though they've never delivered on 45:06 anything, they've promised like on 45:08 schedule ever and and when those things, 45:10 you know, come, they they break most of 45:12 the time. And then the mass firings 45:14 twice and then rehirings after, right? 45:16 These are all just data points. And then 45:18 I'm supposed to believe amid all of that 45:20 that this is all going to continue and 45:22 magically work two years from now. amid 45:24 the launch of GTA 6 with all their UGC 45:27 tools with the rapid expansion of UGC 45:30 with some top former employees of Epic 45:33 Games going now building an agent 45:34 focused game engine from scratch you 45:36 know like people see things here but 45:38 what Tim's not talking about that he's 45:40 not he's not talking about any of this 45:41 stuff he's not like we're not hearing 45:44 hopes for the future of like hey 45:45 Fortnite developers here's where you're 45:47 going to you're going to be able to do 45:48 this and this and you're going to be 45:48 able to make money and build your 45:50 business guess what they've stopped 45:51 talking about that they're they're 45:52 throwing out updates and stuff. But to 45:55 me, all signs point to to UEFN in terms 45:58 of the the metaverse where you're making 46:00 money launching games. To me, that seems 46:02 like it's going to get shut down because 46:04 here's where we are. Here's where their 46:05 competitors are. Here's what they're 46:06 saying they're going to do. Okay. But 46:08 that still doesn't solve. Okay. What 46:10 about the millions of players of all the 46:12 different genres? Right now, you got 46:13 Fortnite players, you know, whatever 46:15 ages 12 to 16, same audience, right? You 46:19 could never launch a dress to impress on 46:21 Fortnite right now. it would get zero 46:22 players even if they put a million 46:24 dollars behind it. You know, there's 46:26 there's not five and or six to 10 year 46:28 old girls just ready to play games, 46:29 right? That like where's the audience? 46:30 Where's the people? Right now, I can't 46:32 open my phone and play a game instantly. 46:34 Like, I got to do a 2 GB uplight. Those 46:36 things haven't changed. And so, I'm 46:38 supposed to believe that two years from 46:39 now it's all going to magically work 46:41 when there's no progress even that I see 46:43 to get there. And so, me as a business 46:44 owner, like I'm, you know, I'm very, 46:47 very cautious. Whereas like right now in 46:48 Roblox, I can make a game update it 46:50 instantly. Push it out. I get players 46:52 every day. I'm watching my revenue 46:53 increase every day and I'm treating it 46:55 like a business. Oh, let's add this 46:56 feature. Let's do this. Let's add this 46:57 feature. Let's do this. And every genre 46:59 you could ever think of is there for all 47:00 different types of ages. And they send 47:02 it out to you for free. And if your 47:03 game's good, it starts growing the 47:04 ranks, right? And so I have a real cause 47:06 and effect on both platforms. Whereas 47:08 right now, if I wanted to go on UEFN, 47:10 I'd be like, "Okay, well, let me find a 47:11 combat map to make and try and, you 47:14 know, kind of hack the algorithm because 47:15 you can't really make anything special 47:18 succeed." And so, like, it's like it's 47:20 really hard. Like, I I see all those 47:22 things and I'm just like, no one's 47:23 there's nothing that's convincing me why 47:25 I should be spending more time here 47:27 unless I was a fanatic about Fortnite 47:29 itself and its game systems and stuff. 47:31 >> Yeah. 47:32 >> But anyway, that's kind of where I am. 47:34 >> Yeah. No, no. I mean, and and again, 47:37 look, thank you for being so candid. I'm 47:38 sure there's some people that feel the 47:40 same way that you do. And it's hard to 47:42 Monday quarter quarterback. I think EP 47:44 like Fortnite is a blessing and a curse 47:47 cuz people come to Fortnite to play 47:49 battle royale. 47:50 >> Yeah. And it's fun. I played it. I've 47:51 been playing all week and I love it. 47:53 Like it's great. It's great 47:54 >> for sure. Only 40% of people right now, 47:57 if you look at the stats, are playing 47:58 are trying UGC games in in in there. So, 48:01 like they built like it is hard to 48:04 compare Roblox to Fortnite because it 48:06 was a game first and then they're trying 48:07 to add this in. And so it it's a 48:10 different playing field for sure. And 48:12 yes, there there are definitely are 48:13 features. And that was kind of the 48:14 question was just like, you know, this 48:16 is is this too late? Where's other 48:17 features and AI do they really need to 48:19 bring in to kind of compete with Roblox 48:21 if they can? But it is a different 48:23 audience. We know we mobile. We got a 48:26 they they did a whole lawsuit with 48:27 mobile and I lost 50% of my audience 48:30 back then when they got in that lawsuit 48:31 with mobile back you know five six years 48:34 ago. So it was a big deal. 48:36 >> I think 48:36 >> if Fortnite had mobile back still then 48:40 maybe this is a different conversation 48:41 too like with the with the players and 48:43 how that works. But they have a huge 48:45 like there's a I don't say it's a maybe 48:47 it's a huge problem but they have to 48:48 figure out that that issue. And is UE6 48:50 going to do that? Is it going to allow 48:52 us to maybe publish our games 48:53 independently? Do they show up Fortnite? 48:55 Is Fortnite back? Like how how this 48:57 whole new, you know, things going to 48:59 look? Who knows, right? But yeah, it's 49:02 definitely the last few the last few 49:04 this last season has increased plays. So 49:06 that's been good. But it has been kind 49:07 of on the decline and it's been a rough 49:09 rough go for sure, like speaking 49:11 honestly. And you know, it's just it's 49:13 fun to kind of talk about the future and 49:15 where where we're going to end up. But 49:17 yeah, it's very um I don't know. I still 49:20 I have faith. I do I still have faith in 49:22 it. And there it's interesting because 49:24 they do have a brand new I have heard 49:26 some of the other things some players 49:28 have been getting surfaced some better 49:30 games but overall you're I think you're 49:32 you're what you're saying is like the 49:34 games that you like to play you just 49:36 they will not work in Fortnite right 49:38 now. They work on Roblox. It's just like 49:40 no one cares or like the audience just 49:41 isn't interested. And so there one is no 49:43 no one's going to either if they make 49:45 them they're not going to update them or 49:46 you know it's not worth you your time 49:48 right now to get your investments back. 49:50 And so, and that I think is something 49:52 that a lot of creators have said for a 49:54 long time and that's kind of 49:56 >> so there's that and then and there's the 49:58 business aspect of it like 50:00 >> you know I'm all like I I get it like 50:02 for I was new to Fortnite community in 50:05 that sense like and there's there's some 50:07 major loyalty there and and that's all 50:08 fine but for me it's it's about you know 50:11 I have enough I have x amount of time I 50:13 can spend in a day I have to pay the 50:14 bills for my family and my kids like how 50:16 do I spend my time right and and so 50:18 sometimes when I'm talking about 50:19 Fortnite I get DMs on on X and and 50:21 people will be like, I think you're 50:22 wrong about UFN and that's fine, 50:24 whatever. 50:24 >> Um, you know, but I I just I believe in 50:26 it and this and this and this and I I 50:28 this is a real thing. This has happened 50:29 about 10 times in the last like three 50:31 months. And then I will say, okay, I 50:33 just want you to be real honest with me. 50:34 How much money have you made in Fortnite 50:36 in the last, you know, year? Every 50:39 single time it was zero dollars 50:41 >> and like, well, I just create because I 50:42 like it. Well, that's fine, right? But 50:44 like, right, maybe that's your goal. 50:46 That's cool. I I don't know. But my my 50:48 goal is to make enough money doing what 50:51 I love so I can pay my bills. And so 50:54 like when I put that scaffold in UFN, 50:56 like starting a business on, you know, 50:58 with Fortnite or UGC, it's like, okay, 51:00 well, the odds are not in my favor by 51:02 any means. And if I actually want a 51:04 chance to make a dollar is going to have 51:06 combat and it's going to be mostly 51:07 Fortnite mechanics. 51:08 >> Yeah. 51:09 >> You know, and so already out of the 51:11 gate, it's like, oof, you know, and and 51:13 I had it lucky in the beginning. I'm 51:14 honestly just really lucky early. A lot 51:16 of those guys are, you know, early, 51:18 right? And not no trash on them. They're 51:20 great. Like Hive, all those guys like 51:22 the live ops is they're doing amazing, 51:24 right? Yeah. 51:25 >> Still the same people though, right? 51:26 Still the same people, right? And and 51:29 and that has nothing to do with their 51:31 scale or things like that because I've 51:32 there's some games like I I did a search 51:34 on Fortnite the other day for Clean the 51:36 Library. Uh it's a it's a sorting game 51:39 that made, you know, millions of dollars 51:40 on Roblox among some other types of 51:42 cleaning games. And it's amazing on on 51:44 UFN. Everything is so good. It's so fun. 51:48 Beautiful. I think it's called like 51:50 clean the anime library. You put all the 51:52 anime books away. Amazing. Like top 51:54 quality. You know, there's no plays. 51:55 Never really got any plays. And then, 51:57 you know, I've posted about people like, 51:58 "Well, you got to do marketing and 51:59 things like this and that." I was like, 52:00 "Actually, this to me is is the essence 52:03 of a broken ecosystem. If I make an 52:05 amazing fun game that humans want to 52:08 play, it absolutely should end up in the 52:10 hands of humans. Yet, this has zero 52:12 players on the store. Uh, and you know, 52:14 so there's like a signals. I call those 52:16 just the signals. There's so many 52:17 signals to me that anyone should look at 52:19 who's starting a business, you know, 52:20 it's like, can I can I provide can I do 52:22 this? And of course, the the top shops 52:24 in UFN are now also doing Roblox. You 52:27 know, there's there's a a reason for 52:28 that, you know. 52:29 >> Well, yeah. I I again like I said I do 52:32 think it is important for you to try all 52:34 UGGC's and and not corner yourself to 52:36 any of them. You know what I mean? I 52:38 think that's a very common thing you see 52:39 online is just like you know I can't 52:42 these games I want to make I can't make 52:44 because they won't surface. And again I 52:47 think that's just part of Fortnite part 52:49 of VR. It's a lot of practice, a lot of 52:52 little quick mini games, but it's still 52:53 all attached to Fortnite. And they have 52:55 to solve that problem, I think, for 52:58 their UGC ecosystem to grow. 53:01 >> And dude, I don't want to pretend like I 53:03 know how to solve that problem because I 53:04 I I don't know. I that's it's going to 53:07 be hard and all, you know, take away BR 53:09 out of Fortnite. Think about that. What 53:11 happens to UGC? 53:13 >> You know what I mean? 53:14 >> Yeah. Exactly. I mean, you're right for 53:15 sure. Yeah. 53:16 >> Yeah. or or vice versa, you know, just 53:18 like, all right, UGC is now on epics or, 53:22 you know, your games are there. It's 53:23 like, what what no one's playing box 53:25 fights on there like, so it's a it's 53:28 definitely a crazy problem. And yeah, 53:31 I'm getting back more into the tools a 53:32 little bit, you know, I want I want to 53:34 make this kind of useful for somebody 53:35 who's, you know, maybe trying it after 53:37 listening like I want to get into how 53:40 you actually ask these tools to do work 53:42 maybe. So what is the worst way you have 53:46 seen someone use these tools so far? 53:48 Maybe is that like a is that easy? 53:50 >> Yeah, that's a good that's a good 53:51 question. 53:52 >> Yeah. 53:52 >> Yeah. So too many people want AI and and 53:55 this doesn't have to do with Unreal MCP 53:57 per se, just in general, but too many 53:58 people want AI to do too much. 54:02 >> And and while it it can perform a lot of 54:04 things, you know, at the end of the day, 54:06 you're making a game. You're making 54:08 something that's fun, you know, and 54:09 special. And so someone will be like, 54:11 "Hey, make me a gun game." or or hey, I 54:13 want a gun game that does this, this, 54:14 and this. And again, again, those new 54:15 models are really cool. But here's the 54:17 reality, right? You know, there's Roblox 54:20 UFN, they have their algorithms, they 54:21 have the types of games that work, the 54:23 types of games that don't, the types of 54:24 things that are important for a new 54:25 player versus not, and a thousand other 54:27 things to consider, right? And not to 54:29 mention the fun and and fun is very very 54:32 hard to codify. And like we can study 54:35 it. We can have the AIS research it like 54:37 and I do like I I'll have it build, you 54:39 know, like gray box maps that have been 54:42 proven in the past, you know, in terms 54:44 of distance to run and corners and edges 54:46 and it does a good job with those. But 54:47 still, at the end of the day, you have 54:48 to it has to be fun. And so really, you 54:52 know, I would say to combat the bad way 54:54 of of working with AI is simply to put 54:57 like pure intent into whatever request 54:59 you're making. You know, like that extra 55:02 that love that you need in there, right? 55:03 So like in in my case here the other day 55:05 with UEFN I I I was building a skill so 55:08 other people's could work with the 55:10 materials a little bit better. But 55:11 basically I you break a brick and then 55:13 like a resource gem appears, right? And 55:15 I wanted it to have a white stroke 55:17 around it with some sparkles. And here's 55:18 the thing. If if I just told the AI, 55:20 hey, whenever I break a block, you know, 55:22 drop a resource and make it look cool, 55:24 right? It it would not look cool. 55:27 >> Yeah. Yeah. 55:29 But if I say, "Hey, drop the resource 55:31 when they break a brick." And then I 55:33 want you to use materials to actually 55:36 create, you know, a lowcost, you know, 55:38 outline, white outline on it, and then I 55:39 want you to use Niagara to make some 55:41 sparkle effects on it. But I don't make 55:44 them too powerful. And I don't want, you 55:46 know, fake five point stars. I want more 55:48 of like a glistening star. And, you 55:50 know, make sure that you play a sound 55:52 effect when it's picked up, you know, or 55:54 maybe give it some options of sound 55:55 effects to play. All of a sudden now, 55:57 you know, I'm gonna get a much richer 55:59 feature. And so, you could do one 56:01 feature or you can do 10 together, but 56:03 that level of effort you need to put in 56:05 for the request is what you need to do 56:07 for everything that goes into the game. 56:10 >> And that's a great way for beginners to 56:12 work with AI. Just make sure that you 56:14 have really thought through. Don't leave 56:16 things for the AI to figure out for you 56:18 in terms of the design, the output, the 56:21 mechanics. Who cares? let the robots do 56:23 all the coding and stuff, but the output 56:26 and intended design like that's where 56:27 you spend your time and you will always 56:29 get great results if you do that. So 56:31 that's what I would say is is the the 56:33 best way for new people to start working 56:34 with AI. 56:34 >> That's that's a great answer, man. Great 56:36 answer. Like piggybacking on that, you 56:38 know, you have taught, you know, 56:40 hundreds of thousands of people to code 56:42 and most of them, you know, got good by 56:45 struggling through bugs, right? Like if 56:48 AI removes that struggle now, like where 56:50 does the next generation of great 56:52 developers come from? 56:53 >> Well, it's at you know what's crazy 56:55 about this whole thing is I get I get 56:57 the discussion point, but also it's it's 57:00 it's not quite a discussion point at all 57:01 because at the end of the day, why did 57:03 we ever program anything in the first 57:05 place? Not to be an engineer. We program 57:07 things to do things, to create things, 57:10 you know, like why when I drive down the 57:12 street do I see 50 people working on a 57:14 house with hammers and nails, not 57:16 because they want to go have a party, a 57:18 build party. It's because somebody wants 57:20 to buy a home, right? And it requires 57:23 those 50 people to do that. 57:25 >> And so really, it's about it's about the 57:28 output. So, and what I what I mean is if 57:30 I talk to my robot and I don't know any 57:32 programming and I say, "Hey, make me a 57:34 first-person shooter multiplayer 57:35 mechanic where I can shoot the other 57:37 player and it makes it right. Okay, 57:39 well, let's say it's got lag and all 57:40 these other problems and things, you 57:42 know. Well, now I say, okay, hey, so we 57:44 really need to fix the latency of this 57:46 game. It seems like it's it's going 57:47 slow. So, make sure you're using best 57:49 multiplayer practices." And and me being 57:51 an engineer, I might say, you know, go 57:52 look up like predictive, you know, 57:53 multiplayer, you know, systems or, you 57:57 know, maybe we should consider using, 57:59 you know, sending packets over TCP or 58:01 something like I might add some things 58:02 like that. But at the end of the day, as 58:04 an architect, you know what you want. 58:07 You want the output. You want a 58:08 fast-paced multiplayer game. And so, 58:10 you're testing, you're building, and 58:11 just focus on the output. And and then 58:13 and guess what? Now you're an engineer. 58:15 You're you're an architect engineer. And 58:17 there's still going to be those top 58:18 level engineers who you need who are 58:20 amazing with math and computer science 58:21 who people over at you know Open AI and 58:25 Claude are hiring and and you know or 58:27 working on you the next game engine at 58:29 at the lowest level sure but even a lot 58:31 of those roles right the robots can do a 58:34 lot of that hard math and so you still 58:35 need those architects. So I'm just 58:37 saying the next generation of 58:39 programmers is actually going to be the 58:41 programmers that we needed always. 58:43 >> Yeah. it because we're just producing 58:45 like it was never about the means. is it 58:48 was always about you know the output and 58:50 unfortunately yes some jobs will be 58:52 displaced but also I me personally I 58:55 don't tie my human value to work I guess 58:58 I I believe humans are inherently have 59:00 value and you know what you spend your 59:02 time doing doesn't necessarily define 59:04 you and it's okay for it to change too 59:07 >> so focus on the output if your game's 59:09 slow make it tell it to go faster like 59:11 focus on the specific things and and our 59:14 voice is becoming more powerful every 59:15 day with AI and the back and the forth 59:17 and and like we're not far away, you 59:20 know, probably six months away where 59:21 like there's even a tool released on top 59:23 of open AI or whatever where you can 59:25 just or glasses you're just talking back 59:27 and forth and all your systems are just 59:29 building right in front of you. We saw 59:30 that in the demo with open AI so output 59:32 the end result, right? Just focus on 59:34 that and and guess what we've kind of 59:36 cir the the robots are the engineers now 59:38 you know. 59:38 >> Yeah. So the next generation is going to 59:40 be coming for from developers that are 59:42 being uh you know putting resources in 59:44 AI, learning AI. 59:46 >> I mean yeah at the end of the day. Yeah. 59:49 [laughter] 59:49 >> Yeah. Yeah. To answer that I mean last 59:52 question here you know looking ahead 59:54 2029 u this is this is like this is a 59:57 fun prediction. I know no one knows this 59:59 but this world is going to get crazy 1:00:01 right? Like what is something we still 1:00:02 do by hand maybe in UFN or Roblox today 1:00:05 that looks that's going to maybe look 1:00:07 ridiculous? um you know in 2029. Oh 1:00:10 >> geez. I mean I mean one of the the 1:00:12 things I'm even I would say worried 1:00:14 about but to a degree I am worried 1:00:16 slightly and this goes kind of Tim 1:00:18 Sweeny's fear is I'm worried that these 1:00:20 the bigger companies is so good at 1:00:22 creating right that it's like hard to 1:00:25 find your place in in the world of of of 1:00:27 where you are and what you're doing you 1:00:28 know. So, for instance, if you cuz like 1:00:30 for instance, Roblox already it's it's 1:00:31 in beta right now in uh New Zealand and 1:00:35 Canada and Australia. I think they where 1:00:37 you can actually build Roblox games from 1:00:39 the Roblox iPad app, like the app 1:00:41 itself. And they've got a design thing 1:00:43 where it actually spins up a server and 1:00:45 you just type or talk into it and it 1:00:47 builds a game. You play it instantly. 1:00:48 You keep revising it with your your 1:00:49 talking or back and forth. And they're 1:00:51 basically bringing the power of creation 1:00:53 to their to their own people, right? 1:00:55 And so you know these big companies 1:00:57 whether it's you know maybe a rockstar 1:00:59 go you goes full on UGC and becomes 1:01:02 massive this comp you know Roblox but 1:01:04 you know there's this you know we are 1:01:06 middlemen at the end of the day in terms 1:01:08 of people who build and create to some 1:01:10 degree you know between a user and what 1:01:12 they want. So if a player or a user of 1:01:14 software literally is just talking, you 1:01:16 know, in 2029 talking and things are 1:01:18 being created, right? Well, it's like, 1:01:19 okay, well, we didn't need to hire that 1:01:21 person to build my website. We didn't 1:01:22 need this. We didn't need that. They've 1:01:24 it's become so easy for the average 1:01:26 human to create that it it becomes 1:01:29 instant, you know, then it's I mean, and 1:01:31 Elon Musk talks about this quite a bit, 1:01:33 but it's like what do you what does 1:01:34 humanity do with it with its time? What 1:01:37 like how are we innovating? How are we 1:01:38 pushing boundaries? And so I do think in 1:01:40 2029, I mean, you will literally be able 1:01:43 to just speak in your house wherever and 1:01:46 things will happen. You know, the 1:01:47 hardware comes later, but in the 1:01:49 software world, like 1:01:50 >> I mean, Iron Man level Jarvis is is 2029 1:01:54 or sooner. Uh, you know, 1:01:56 >> I mean, I'm I'm on my I'm I'm on a walk 1:01:58 with my Grockbot and I'm almost there, 1:02:00 you know, so 1:02:02 >> be native and everything. 1:02:03 >> Yeah. 1:02:04 >> Yeah. No, it's always fun to kind of try 1:02:07 to like predict the future or look look 1:02:09 forward. I didn't know that about 1:02:10 Roblox. That's wild. They're they're 1:02:12 definitely like going all in on AI. 1:02:14 Sounds like 1:02:14 >> Yeah. You make a game and you instantly 1:02:16 just share it with your friends. They 1:02:17 come in and play it and then it's 1:02:18 published and it's just like, oh my 1:02:20 gosh, they just they bypassed a thousand 1:02:22 steps of game development. Something I 1:02:24 see is just like, you know, our path 1:02:26 into game development, you had to work 1:02:29 very hard. You had to learn a bunch of 1:02:30 different skills, maybe go to school, 1:02:32 all of those things. And then and that's 1:02:34 just to make a piece of a game maybe or 1:02:36 or maybe you were solo dev, you made 1:02:38 your own game, you grind it, whatever. 1:02:39 But like my kids, they're super 1:02:41 interested in me and my game building. 1:02:43 And they want to now naturally obviously 1:02:46 now make their own games, right? And 1:02:48 they can just talk to a phone and make a 1:02:50 game. 1:02:50 >> I mean, I I can see I could see how that 1:02:53 takes over in in some ways. Like at 1:02:56 least, you know, whatever. Those games 1:02:57 might not take over. They're kid games. 1:02:59 But look, think about if you were 1:03:01 freaking 10 and you could just talk and 1:03:03 have a game you've been dreaming of 1:03:04 right there to play. 1:03:05 >> I mean, it's it's fantastic. I mean, 1:03:07 that's and that's what we do as game 1:03:09 creators anyways. We want to make these 1:03:10 experiences. Yeah. So, 1:03:13 >> it's really it's kind of mindboggling. 1:03:16 >> No, Mark, look, interesting times, man. 1:03:17 And again, thanks for coming on, being 1:03:19 so candid. I think I that's what I 1:03:20 appreciate you most. Like, you're not 1:03:22 you're you're not going to sugarcoat 1:03:24 anything. You just kind of give it to it 1:03:25 to it straight. And with this brand new 1:03:27 tool coming out in UFN, um, and you 1:03:30 being just on the forefront of AI and 1:03:32 and and game development, I think, you 1:03:34 know, I really appreciate not a lot of 1:03:36 people want to come on and talk about AI 1:03:37 right now. 1:03:38 >> I mean, I I' I've been punched so many 1:03:40 times in the face over the last 20 years 1:03:42 being a a public creator. That's not 1:03:45 quiet, I guess. So, it's it's all good. 1:03:47 It's just it's part of the part of the 1:03:49 lifestyle, I guess. I don't know. 1:03:51 >> I guess I guess that you That's crazy. 1:03:53 No, look, brother. Look, again, awesome 1:03:55 conversation. We might have to have you 1:03:57 come back on again, man, if more crazy 1:03:58 stuff comes out and talk about it. This 1:04:01 things go so fast. I would love it. 1:04:03 >> And I can definitely talk at more 1:04:04 technical levels of UEFN MCP and the 1:04:06 things I've done with it later if you 1:04:07 ever want to talk about that because I I 1:04:08 have been spending the last two weeks in 1:04:10 it like from morning till night running 1:04:12 on this computer here while I'm working 1:04:13 on my other one. 1:04:14 >> 100%. We'll definitely have you come on 1:04:16 again and talk talk about more 1:04:18 especially I know Epic's even announced 1:04:19 they're up updating it again. So there's 1:04:22 going to be more features added to it 1:04:23 soon what it's going to be able to do. 1:04:25 But yeah, man. Uh if you would shout 1:04:27 shout out your business, what what 1:04:28 you're up to and where people can follow 1:04:30 you. 1:04:30 >> Yeah. Uh you can find me on on X as 1:04:33 playc also playc.com and there you can 1:04:36 find my school that's spelled sk and 1:04:39 basically playc on school. and I I train 1:04:41 people to build Fortnite games, Roblox 1:04:43 games, and just games and AI in general 1:04:45 and all the things that I'm learning and 1:04:46 doing. Uh you can learn it there. And 1:04:48 then I also post like my scripts and my 1:04:50 my skills and things like that. So you 1:04:51 can get be building games like like way 1:04:54 faster. Like just one quick example, 1:04:55 like for instance with UEFN right now, 1:04:58 like the AI struggles a ton with MCP 1:05:00 finding assets. Like what's in the 1:05:01 gallery? If I say build me a a cool dark 1:05:04 dungeon, it doesn't know how to find the 1:05:06 pieces. It doesn't know any of that 1:05:07 stuff. So I actually build a tool that 1:05:09 like ripped all those assets, found 1:05:11 their paths, analyzed the images and 1:05:13 actually made keywords and notes and 1:05:14 stuff. So now when I give it a mockup, I 1:05:16 say build this, it can actually the MCP 1:05:18 looks at it and finds all the assets 1:05:19 visually that it needs and then it 1:05:21 actually builds it perfectly in the game 1:05:22 like in minutes and right. So those are 1:05:24 the types of things you can build with 1:05:26 skills, but like that's kind of the 1:05:27 things that I offer and teach people as 1:05:29 well too. So 1:05:30 >> very cool. All those links will be in 1:05:31 the description if you guys want go 1:05:33 follow him and dude that's an episode, 1:05:35 man. Thank you so much for coming on 1:05:36 Mark. Yeah, I appreciate it. Mark, 1:05:38 thanks for coming on and walking us 1:05:39 through your process. If there's a part 1:05:41 you'd like us to explore further, let me 1:05:43 know on X. You can follow Mark at play 1:05:45 UGC on X as well. All the links are in 1:05:48 the description. And if you enjoyed this 1:05:50 episode, do me a favor. Follow the 1:05:52 podcast, subscribe on YouTube, hit the 1:05:54 like button, and if you are listening on 1:05:56 audio, please leave a positive review on 1:05:58 your favorite podcast app. It generally 1:06:00 helps out more than you think. I'll see 1:06:02 you in a Fortnite, which is in 2 weeks. 1:06:04 And remember, don't hate creation.