Heroza Zhang

Hello, today we met Heroza Zhang. This is water. I know. I mean, today we met your name. Wait a minute, okay, I'm here. So recently, one of your tweets has gone viral. So what's that? Can you tell us about it? Yeah, so we are building like our product is called GoRes. It targets for anyone to create AAA games, so it's like a multi-agent. So the user will just like a game designer and like from your, show your like what you want in your internet to work and like the agent can directly synchronously work everything out for you. And we really open source like 2D like spreadsheet products for like user can just using codecs and connect with our like pre-will agent and they can directly get like the 2D spreadsheet they want, which is like playable and put it inside of the background. So and it becomes a little bit like gather many attention online. So we just like publish this result. It's pretty like kind of tricky. Yeah. How many views? I think it would be like 500,000. Yeah, because we did not like spend any like cost inside of it, but there are many like influencers from like TikTok or Instagram or like Reddit, like almost everywhere we find those like reports about us, but yeah. So what's the funny of full viral? I think not very enough, so it need to be more like attention, please. Yeah. Okay, so I think that's quite a lot. So what you were building, can we explain it simpler? Like sample? Simpler. Oh, what you were building. So it's just for, I mean, I'm not sure like everyone here like has ever played video games or enjoy playing video games. So I can use another example. It's like, so we are making something which is like, like relatively similar. This is like everything you can experience in your real life, like maybe the architecture you see and like normal yourself or like everything or other people are talking around you and maybe those kind of, we call it like interactive world. So in our product, you can just easily create a world like that and just tell it what you want and you can get it. So that is what we are working on about. So natural language, dummies like me, talking to your product and it can create a 3D world. Yes. Like any AAA games before. Yeah, yeah, yeah. And because there are like many models recently, like 3D model generator or like world model, like those kind of model. And actually, they like a platform, the backup agent which combine them together for user to easily like really use it and apply in their real life, like simulation, nothing. So that is what we are creating about. It's about like inside of it, like combine of them together. So for user to easily create a 3D world. Okay, so and the product is for what kind of users? So at the very beginning, because we wanted to make the result to be high quality enough and because our team are all from the game industry like background. So we started from inviting many professional game designers to help us to like, let's say what they want and show the feedback to us. So that is the very beginning. But the target is just make sure the final result, the final output will be professional enough. But actually, because following the trends of the AI products, so like every product need to be simpler for people to understand and use. So there should not be any barriers. So in the later future, it's actually everyone can use it. So that is how we define it about it. So right now it's for the prosumer. And if I want to develop my own game and use your product, it basically do everything from generating the 3D world to have the pilot, the game and even launch go to the store, everything. Okay, so it's a one stop from idea to launch the game to Steam to the other platform. Everything you can do. Okay, that's cool. And so I think in the future, everybody will be a creator for games. And I think that's huge because... Since the world was last time with TikTok, everybody can be an influencer. That's because the technology develops so we can easily use our camera to film everything. But today, with the new development of the technology, everybody can create the virtual world they want and put games in that without any effort. Yeah, because like previously, people like in their spare time might just like scan in their TikTok and use it as entertainment, but because like according to like AI grows so fast and maybe like AI will take your daily work and you have more spare time and people have more spare time, so they can spend on like entertainment. But I think the next generation of entertainment will be not be limited by like just watching those like short films on TikTok or something. There should be a new one and also because the technology is improved, there should be a like better, more immersive and like interactive environment to exist. So that will be the like future trends, like people can have something more in. So game was called the nice art, right, the nice, or the horse, the nice. Nice art. Nice art, yeah, I can't understand. Before, the popularity of Roblox and Minecraft already proved that, that is the user want to develop their own world. Yeah, but the technology was not that so good, so we can only see that kind of, and you think in the future, everybody can develop their own player number one. That's a vision, right? Yes. Actually, yeah, so actually our like initial like consideration about that is like my personally just wanna like understand like everything, like every people everywhere, like this theory about like whole universe and I'm thinking about there is a potential for achieving that by like the ability of AI to predict something and also the potential of like work generations. I'm thinking about if we can allow that everyone to create a like relatively simulated world as fast, so then like that can be like achieved like maybe in the future and because I really enjoy in like creativity. And actually, game will change the way like people interact with each other because maybe someone just feel like, okay, give me something I like spend my spare time on this and just having fun, but actually like when people are designing a game, so you can like put your the real world perspective inside of it, like everything you see, you will put it in your design and also like the things you feel and like the way you interact with the work, like you put everything inside of it and for other people to play it, then they will feel like fully understand about what you are experiencing in your daily life. So actually the way one sec, everyone create a world and then like everyone can understand each other, like the way will be totally different. Like so, and also about the like connection will be changed. I mean, like people's like connection used to be changed by maybe like WeChat or Facebook, and now like there's gonna be a new way and like because of those kind of infrastructure is like emergence. So it's not invite you to my home, but invite you to my, to my game. To my mind. Oh, I like that, because the physical world is limited, it's very small, but with a virtual world, I can create so many things, everything there. And it's a celebration from the physical world to the virtual world. Okay, so what's the cost? What's the cost? I guess before, like for the GTA 6, it's still maybe 10 years, still no launch. So what's the cost of change of this? You mean like how much should be spent inside or like you mean the time or the cost? For example, I'm an independent game developer. Before I cannot be, but today use your product, I can be. So how much money do I need to invest to have my own game launch, user product? Okay, actually that is depends on like how detailed the user was. So for our business model, we just doing the like simple subscription and also if we're developing like extra products, they can just like spend inside, but they can also like directly insert their models inside so they will not need to be like generated, so for example. So actually, if they wanna like, for example, create a simple demo and like show their symbol, that would be like within like $100, so it's not, not a. I think it's just depends on like how big the user you want about the things they're printing. This question, I want to understand the structure change in the industry. Like I think before a AAA game would cost about 200 million. And today, because the technology has changed, maybe it cost like, yeah, hugely. So it's 1%, 5%, 10% of before or more? Yeah, maybe about 1%. 1%? Yeah. So it's 1 million. Yeah, because like actually people did not, like used to be like impossible for, because do you know why like AAA is called AAA? Why? Yeah, because like there are three A's, like a lot of like content, a lot of quality, and a lot of like time. So we can define it because it takes a lot of cost, a lot of time, a lot of people to work on that. So actually, maybe there are small team has that ability. Everyone knows how to make something like that, but they might not spend too much time or like cost on that. So they did not build a game, but actually they have a great idea inside of their mind. And now it becomes possible. So once we want to like make this to be possible, so it will be like this cost because like AI is really like easier and really fast. Like it's already like 100 times faster than the previous workflow. So it will be actually like this one minute, maybe just like a few thousand and you can make a pretty good game just as like, you can like maybe similarity to the like video generation. It will be pretty simple for people to build like a short movie with like video generation model, but just like with a thousand dollars. So I think that's how the industry is changing. So it's kind of similar. Oh, okay. So these days I watch a lot of AI generated short videos. It's getting fascinating. And your prediction is AI generated games would also go as popular as AI generated videos next year. Maybe more. Yeah, more. Okay, I sort of understand because Kodak Cloud really make our apps development so cheap and fast. So let's talk more about yourself. Why you want to do this? What's your background? What's the relation? And what's your dream? Why your dream? What triggered you go into the gaming industry? I think I mentioned before, just now, I say like I want to understand like the theory about the whole universe and like, so it's not only the physics thing and also the mental thing. So like how each of the people are thinking and how the world is rotating. So I was thinking about I want to really create a like universe model, so which is simulated to the reality so I can understand everything, like predict the future and understand the history. And I'm really enjoyable in like creativity things. And I find that, so game is just like create an interactive world and which is like not only matches my like final vision and also my like hobbies. So I mean like the things I really like enjoying is about creating. So I'm thinking about it's the best thing I can do. And there's a platform for anyone to create something like that. So it's gonna be, yeah, that's exactly the future. So. But you have been an indie gamer developer since you were very young. Yeah. So what makes you start to do that? So because honestly, I, I'm good at writing novels, so I, yeah, I'm good at like writing stories. And so, and I was founded a game developer community like six years before. So after I graduated from the high school, I have a gap here and I just go to join, like make this community. And I know many people who are in the industry because, you know, like in China, like people who are making indie games, they all know each other and they are all a group of people. So after we make the community, we, I have many like connection with them. And like, so then I go to do like product manager intern in Beijing in Perfect World. And there's some people who is working nearby like those different like company who is, which is making games and they invite me to like, so, so then just invite me to join a game jam. And then, so I say, at that moment, I never make game before. And they just say, I say, I'm good at writing stories. So they say you can like do becoming the narrative designer for the game. And then, so I, I just do that. And that game, we achieved like 350,000 will sum up. That's my 350,000. 350,000. Yeah, well spoken. That's 3 million. That's a 0.3 million. Okay. Yeah, so that's actually, that's my first game. So, yeah, and then I just like, because I think my, the people who is building that with us, they are already experienced. So I just like learned a lot from them and I keep building things I'm interested in. Yeah, and like turned out. So the community we're talking about, how big is the community? So for our team who is doing that, like in total, because we have many like volunteers who is helping us. And like for people who are covered by that, I think we'll be more than, I think there'll be thousands of people. Yeah. And like, yeah, and we used to help some like game jams or like seminars and yeah, like both online and offline. Yeah, and also used to like collaborate with like Roblox and Tencent. So we held a like game jam for them. Yeah. So it's something just like that. So the community, thousands of people. Yes. Online community. Yes. Okay. So including the audience, everything. Like everyone is like almost like most of them are like game designers, like developers or artists. Yeah. And we also need to like, because we need to collaborate, we also need to get like connect with those people like from studio or company or like publisher. So yeah. And yeah, so for, I think it's more like for comment people are joining our events. So it's just something like. Okay, so it's each time thousands of people are totally thousands of people. Totally. Totally. Okay, great. So you grow up in the gaming industry and started in Perfect World. That's a tier one game studio in China. And I played a lot of their games when I was young. And you have the early community users as well. You have connection to have your early users. So what's that after you do your intern, what's that? Okay, that's actually I was intern in Perfect World. I was year one. So in our game studio, it's not intern. So it's actually a like long term collaboration. So it's like part time things. Yeah, so my intern is full time. It's just for summer like in Perfect World. And that was the time I know other people like in our studios. And so after I finished the intern, I actually, because I used to want to like create startup as the very beginning. So after that, I go to join like internship program called, if you guys hear like Y Combinator China or Miracle Plus. Yeah, so I was like an investment researcher there. Okay, and in gaming industry. As for the VC, it's for Miracle Plus. Yeah, I just can pick some like industry I'm interested in. Yeah, so I just pick at that moment. Yeah, I pick some. And some of the research is about games, but I can. Okay, games. And then? Yeah, and then, so I think I was the only people who graduated like within three years from my undergraduate. So in our like department, the computer science department. So I think the second year, I joined the, because I need to finish my final year project. So I joined our like laboratory and they are making the like 3D point cloud segmentation. Yeah, and also like in my, during my last year, I also went to another research. Yeah, it's called InnoXchange. So yeah, so I think I am my target. My focus is about like game and hardware. Yeah, I have like, I do like four years research about that to like understand more about the industry and like find a good way and also. Four years, that's four years. Four months. Four months. Yeah, yeah. Okay, so how come you can graduate from computer science in three years? Because I like every year I, because like we can maximumly like take seven courses each semester. So I just like plan them earlier. So every, like every semester I have like full schedule. And like I put my capstone earlier for one year. So I'm finishing within three years. Computer science is hard. Yeah, I think it takes time. It's not some easy. Are you making? No, I mean, similar engineering, but engineering. Is it similar? Engineering. Okay, and I heard you were the top tier of Chinese chess. Yeah, so last year, I joined the World Chinese Chess Competition, and I won the fifth place. Yeah, and I was also the first place for like, I mean, nationally, in Chinese chess amateur championship. Yeah, and I actually, I learned it when I was really young, like when I was 10 years old, I was the first class national athlete. So yeah, I just kept learning it, but in my spare time, because I did not join the professional team of learning Chinese chess, I just used my spare time doing that. Okay, I got it, you're quite smart. Oh, thank you. So that's the program. Okay, so then, when did you start to study Cornell? Are you graduated yet? Yeah, I graduated in last December, and I came to here as like January, yeah, last year. I have a one-year program. It's a master? Yes. Okay, in Cornell, and study product development. Like engineering management. Engineering management. Management, yes, M-E-M. So it's a program designed for people to learn how to manage a product or something. Yeah. And before go to university, because actually, I graduated from my undergraduate in July, but I came here as like January. So there's a gap here. That's because I was bombed by a car, and I just take a rest for half a year, and I came here. Okay. So the future is everybody will have their own game with your products, easily build new games. So what's the bottleneck right now in this industry? Bottleneck? Okay. So I think actually, I think comparing to the previous engine people are using, because currently, they are using Unity or Unreal, the most famous kind of game engine, but they are all currently provides the AI feature. But because their structure are developed like 10 or 20 years ago, so it's a really old structure, old software. And if you want to add AI inside of it, it's really hard to manipulate those kind of software, because they are all like, they are not designed for AI, they are designed for human. And also, for those kind of things, there's still a barrier for people to learn, to get started with, for no one to learn usage. So, and there's also for about the back-end, about the AI native, because currently, Unreal are just advertising, say they're gonna cancel their blueprint. It's because blueprint will not be readable for AI, so that's why they are considering of cancel it. And also, that is exactly true for why old engine cannot just directly add AI become AI native enough. So there has to be a new platform for that. So, and that is where we started. And that is also a bottleneck that previous engine cannot be get over. Okay, great. I used a lot of products recently, and I don't think they can be AI native. It's totally different. However, one lesson to here, people would have a question. That is, sounds like you're building a cursor for the games. You can understand that in that way, but just don't understand that we are building some IDE for writing codes for game codes. But yeah, if you understand that, just an agent for making games and that is. And when people are talking about that, people would think about, okay, cursor was very popular, and later CodeX and CloudCode come to take the market. So how do you think of that? Because in our industry, we don't have a CodeX or Cloud. Or if you mention about, for example, Fable, like they can generate games really fast, but actually for games, it's different because the most hardest part for gaming is about rendering, and it will be hard for... Rendering, what's that? Rendering. Rendering is just like, like Vero part, the light, or like everything. I mean, like, for example, the Vero effects of how the... If you can know the AAA differences between a simple game, it's just like the light, the reflection, every detail of the rendering. Because, for example, you have a cyberpunk, but you cannot play it in a MacBook. Or maybe MacBook does not support cyberpunk, or maybe just like a tiny computer, Windows cannot support it. It's just because your computer's abilities, capacity is limited because the rendering is really need a higher capacity or something like that. And also, there are some computers that cannot run the Unreal product. So if you can get it, so actually, if you're just simply testing and maybe Fable generates a game for you, it will not generate a simple game for you, but there has to be a structure to combine them and to well plan it to make sure the result can be high quality enough. And it will be an easier way to understand. It's just like, so actually, people can always write games from using Python or just C-sharp or just C++. But why do they still need to use an engine? Because engines combine those different systems together, so to make sure they can develop something based on it. Or it will be an easier way, it's just like, you can actually use zero or one to coding, but why use languages? Coding languages to code is like, why use Python? It's because there has to be something structured based on the most initial ones. So that is also why there is kind of infrastructure should exist. Okay, so rendering is, I have different computers, there are different computers, so you need to be compatible to everyone. That takes a lot of work. Is that right? Yeah, so I mean, it's just like, if you just like Fable generate a package for you, it will be not high quality enough, like for like AAA levels, so you have to find a way to combine them. Okay, so people would ask, where will the model company start to, so it's more about the model itself or about something behind the model. It depends on whether cloud code will replace cursor or what else exists in there, or there's a lot of things to be in there, but the language model is just a token input. I think cursor might not be the best example because cursor is a new company and even like Anthropics, they are not for like more than 10 years. So like AI is like still like a new industry and new trends. So if, for example, if cursor already, because there are still huge amounts of users for Unicursor, even though we have cloud or GPT, but if cursor already been replaced one day and nobody uses it, it's not because this strategy is not applicable. It might be because like cursor might like miss some good point about like continuously growing. Yeah. So, but I, I hope like they are still like keep growing, so I don't mean anything. And yeah. Okay, I think that question in fact is boring. That question is, what if OpenAI and Stroke do it? It's quite boring question. Let's skip that part. Okay, so tell a little bit about your team. What's your big strength? And I already buy it. Everybody will have their games. It's just like everybody will design their website because it's so easy today. And next year, everybody will have their photo video generated by AI. Maybe it's a year after next year, everybody will have their own games. So, yeah. What about your team? So, okay. So my co-founder used to work in fine, like Meta for years and then she become her own indie producer. She created games and launched it on Steam and then she got, yeah, she just switched to do this program right now. Yeah, so she's really like experiencing in like backend structure and like coding. And she's our CEO. And my another co-founder, he used to, used to learn like data science in Berkeley, but he created his own startup. Like in one month, he got like $3,000 like MRR for his product and then he just switched to other direction. Yeah, and because he have experience about how to make the like markdown for like AI and how to make them to build structure for user use, so I think we all have like advantage other members does not have. So that's why I like make it together. Yeah. Okay, so let's back to the startups. How many users do you have right now? We have, we actually, so it's like two weeks ago, we have 2,200 users for our studio platform. How many? Two weeks ago, we have 2,200 users. 2,200. Yeah. Yeah, just 200. Yeah, not 2,000, 200. Oh, sorry. 200, yeah. We have 200 users two months ago, and also three studios collaborated with us. So it's more like a kind of like could be at my beginning, and there are also like a few studios, like three more studios are just collaborating. But now, because we open source like a 2D platform recently, and it becomes really famous because we gathered like more than 1,000 stars within like 10, like 14 days. And yeah, so there are many people are watching it every minute I just open my GitHub. So I think, and the users who are gonna also wanna use our story, it's also double. So I did not count the new like kind of users at this moment, but yeah, I think just like. So now it might be over a thousand. Yes. And you say the 200 originally was not independent. You talk with some studio, other studio using it. So that's what we are like currently used to have our user in different way, because actually we can gather more than 200 users, but we just invite indie, like professional designer to use it, and also invite like someone who have studios or studios to use it. So they are actually two different groups of people. And for 200, they are like indie designer, and for like studios, they are different, different people, but we just make sure they can give us like professional feedback and they can spend time on that and finding what I'm saying like what they want for this product. So are they paid? Yeah, but just to prove they are, they think it's worth to pay. And also like signing ROI for studio. How much do they pay? How much? Oh, that's a secret. Okay. So you have early users, early sample users, and they give you feedback. Do they like the product? Actually, we were kind of like shocked more than especially my co-founders feel, like when user, like they always, because I just asked my like few friends who are doing game design, say like, can you test it for me? They say yes. And I just like posting, I mean the conversation, but like maybe one week later, I still a really long like letter and they write a lot. They say they spend more than six hours to test the product and what they want and how this like change their workflow, like everything. And my co-founder just say, do they really think so? Why do they have like so many like positive feedback that we expected? Yeah, because it's not just happened once, it's happened multiple times. And yeah. But people are nice. People are nicer than you think. Yeah, and yeah, you see the real demand. But yeah, like they spend time like every week for using that. So we can see their active members. So it actually higher than our expected. And also for the like the 2D one, it's also like higher than I expected because we just like published it for like gathering some attention and we find it's so maybe more than we thought they need that. So it's just kind of interesting. So there's a demand, there's a sample user, and there's some content go viral and you have a good team, good future. And you're smart. Thank you. What else? Let me think about it. There's like something interesting about that or? Yeah, what else? Okay, so I think we can start a like a conversation or like about what do you think gonna be the future interaction between like, I mean, not limited by mouse or keyboard, right? Yeah, so and also I think you're used to have a concept about like no UI kind of concept, but you changed, I know. I have to change my belief. Yeah, so can like, I mean like you can share about your belief right now and what do you think people gonna interact in the world, like which kind of way, what's the trend for that? I'm still a VR guy. I bought not a Oculus 2, Vive Pro 1, and I still use a Vive Pro. It's pretty good. It's just there's not too much content in it. And when you are embedded by with Vive Pro, everything is virtual and much more beautiful. I think. It will still come, the player one, player number one. I can't just get into that world, not the real world, but it was like in the movie, so I still like that concept. Yeah, got it. So, like, how do you think about, because people use, like, computers to, for example, doing modeling, like in, like, CAD or the software, it's because, like, if they really, like, maybe put some, like, block on a block, it's really, like, tedious and hard, like, manually working on this. So they're gonna, like, plan it on a computer, but now if people are switching to, like, VR interaction, actually people are back to, you know, manually putting a block on a block. Do you think that's interesting? So, like, yeah, you may, like, have a vision pro and then you just, like, put a block on that. It's actually back to the initial way people interact with the work. Yes, it's more direct, and you don't need to learn anything more because I can say it, like, put that 100 animals to that mountain, or give me five different kind of mountains, so, and then it was generated. The keyboard and the mouse can make things very accurate, but with today, I think the AI is getting more and more good with the taste. It almost can predict what kind of stuff should be what kind of place because they have the very top taste of the designers. Maybe I just say it, so create a world with create a world like Fate. That's a Saber and Archer feature. Yeah, maybe I just do it. And I say, just create another version without the copyright. Yeah, true. Okay, so that's all.