Building giant and ambitious games | Brendan Greene interview

Elvis Presley once said, “Ambition is a dream with a V8 engine.” Brendan Greene, the creator of PlayerUnknown’s Battlegrounds (PUBG), has a lot of ambition. His battle royale game, inspired by the Japanese film Battle Royale (2000), has sold more than 80 million copies.

And one of Greene’s ambitions is doing something important like that again in video games. And so he just announced that his PlayerUnknown Productions is resurfacing after years of development with a three-game plan to bring on the next generation of survival games. And it’s ambitious.

I talked to Greene, who is known as PlayerUnknown, about it in an exclusive interview. It’s down at the bottom of this introduction and I hope you like it. At the end, I asked him about ambition.

Greene got the idea from the movie that he could stage a battle where 100 people would compete with each other. With each player eliminated, the battle space would get smaller until the last two were battling it out in a very small circle. The last one standing was the winner.

Greene first created a “mod” called DayZ in the Arma universe. Then he teamed up with South Korea’s Krafton to make PUBG. The game debuted in 2017, disrupted shooter games like Call of Duty. On the strength of PUBG’s 80 million in sales, Krafton went public and Greene became wealthy from that. That gave him the money to work on something even more ambitious.

I had a front row seat to this plan. Greene went off on his own to create a new startup, PlayerUnknown Productions, in 2021 to make a gaming survival world that was a lot like a metaverse. Then he gave me a scoop on his ambitions.

Without anything to show me except a screenshot at the time, Greene said was creating a world called Prologue that had a huge amount of terrain — about 100 square kilometers. That world, bigger than just about any existing game world, would be a test where players would drop into the world and try to survive until they exited the world in a given spot. It would be different every time they dropped into it.

Now Greene has released a video that describes his intentions more concretely. Prologue now has a real preview in the video and the world looks very realistic, with trees and grasses swaying in the wind. And it’s still a huge world, fashioned with machine learning and AI tools. The aim is to release it sometime in the middle of next year as a single-player game for people to try to survive.

The challenge is that the open-world of Prologue will be an emergent place, where anything can happen and the weather will get progressively worse. It may seem simple to get to the exit point on the map, but it’s likely going to be hell getting there.

Then there will be something else. The company will do a shadow drop of the company’s free tech demo, called Preface: Undiscovered World, showcasing its in-house game engine called Melba. Preface will be able to generate terrain for an Earth-size virtual world, using very little in the way of computing resources.

This demo aims to provide users with an early look at the innovative technology that will power the subsequent titles in the series, and eventually a third game called Project Artemis.

Project Artemis is the large-scale end goal project of the series. As described in the past, Greene sees this as an Earth-size world where players can drop in and create their own gaming experiences in different sections of the world. We don’t use the word metaverse so much anymore, but that’s what it seems like to me. The journey to get there could take another five or ten years.

In the video, Greene said he embarked on Prologue three years ago and “then life happened” and it has taken three years to get it into a solid and breakthrough shape. Now the company can start sharing it and getting feedback “to make it into really something different.”

In our interview, Greene said that the team started pulling together when Laurent Gorga joined as CTO. About a year ago, Gorga started putting in motion a process that enabled the team to make a lot more process. While they were making the tech, the team would now create frequent builds to test the tech on a granular level. They started making enough progress so that they started scheduling the timelines for Prologue and Preface. And they talked about it in a video stream on December 6, during the PC Gaming Show. It made a lot of jaws drop. Prologue is expected to drop into early access on the second quarter of 2025.

“When I started this I was trying to make a larger open world experience than most people made, and we tried to provide a couple of years and we found a way to do that,” Greene said. “We essentially reinvented how you create these worlds using machine learning technology, using natural earth data to generate” the terrain.

Now the company is ready to test this terrain, which will form the basis for the larger worlds. He said the team broke the journey into three stages. The first job was to fill out the terrain of the world. The second was to fill that terrain with lots of interaction when scaling up. And then third, the goal was to pull a bunch of those players onto the world, Greene said.

The company will keep enhancing Prologue with its current game engine and then it will move it over to the next version of its game engine.

Prologue started off as an experiment in Unity and then it moved to Unreal a couple of years ago and the tools have proven to be a solid foundation. The proprietary tech will eventually be able to generate a world with millions if not billions of objects in it, with the help of machine learning.

“It’s more about the large scale and again machine learning is very good at it because it will capture the patterns that we teach it,” Greene said.

The physics will be realistic. If the ground gets wet, the terrain becomes a slippery mud and rivers can form, and these will have repercussions for players as they try to survive in a wilderness. This will make the game challenging, but it can’t be unbeatable, Greene said.

“We’re discovering what is fun, what is not fun but at its core it is about survival. I think the more we can test, the more we can get the feedback from the users or the players, and that’s one of the reasons why we are going to early access,” Greene said. “The more we can actually engage with the community and get their feedback” the more it can reshape the models in the right way.

Meanwhile, the company is working on Melba, the in-house game engine. Using machine learning, it should be able to generate worlds and then regenerate them for the next game.

“The way that we build the engine is allowing us to scale up to large agent interaction,” Greene said. “We have an Earth-scale planner with some various biomes and some simple systems to allow you to explore it.”

The company is working on two projects at once — one with Unreal and another with Melba — so that it doesn’t develop tech in a vacuum, said CTO Laurent Gorga, in the video. Unreal and Prologue will generate a piece of the world. Preface will help achieve the scale, and then Artemis will be the full expression.

“I want to get our tech into the hands of the people out there to help us perform what this tech will become,” Greene said. “Like this terrain tech is interesting, but I really need, I want to leave it open. I want to leave it moddable.”

Greene said this may be a five or 10-year journey, but Prologue could be available on Steam in the second quarter of next year. There were a lot of details about what he’s doing that we talked about. Here’s an edited transcript of our interview.

GamesBeat: I was very impressed by your demo. I saw the Discord event, as well as the announcement.

Brendan Greene: It’s been a busy six months. We finally got it out the door.

GamesBeat: I remember the original vision and how you went about doing it. It sounded like there was a big technology pivot or approach pivot you made. What did that involve, from the time you were first talking about it? How has it turned out?

Greene: We found Laurent Gorga, who we appointed as our CTO. He’s in the video we released. He wanted to make more of a product, rather than a research experiment. Try to focus our efforts on releasing something. He said he doesn’t believe in developing tech in a vacuum. Laurent, Kim, Scott, Petter, they sat down and figured out how we could leverage the great team and tech we had, and the ideas we had, and make it into something we could release.

He posted only last week on our Slack. He said, “A year ago I joined the company, and said that in a year’s time we would release something.” Not to the day, but in a year’s time we released something. It’s a credit to him and the team for making it work.

GamesBeat: Is there an easy way to explain what the approach is, and how it differs from what you had tried before?

Greene: It was the approach that Petter brought to the production of Prologue, but also that Laurent brought–we brought both projects into production rather than keeping them as research experiments. That was the previous tech lead’s view, that we should prove it all out before we move into a more production stage. Laurent really believed–I remember Petter joining and asking the game team, “Let’s play the build.” They said, “Play what then?” And within a week we had a playable build together.

Since then we’ve shifted mentality, from experimenting and playing with ideas to–now that we have really strong leadership in tech and production. That’s put us on the right path. It brought in more traditional techniques. We have a seven-week sprint. We work fully remote, more or less. We’re experimenting with how to make the teams work together well. We have a good synergy between all the different departments now. We have a core engine team. We have our art team. They all work together in conjunction on all the projects.

It’s a credit to Kim, Laurent, Scott, and Petter. I have the vision. I have the dreams. But they’re the guys that really make it work.

GamesBeat: How many people did the team grow to now?

Greene: We’re 60 people now. That’s fully staffed for Prologue.

GamesBeat: That’s higher than the original plan called for.

Greene: Yes, I think we were around 50 or so. But now we have publishing. We have finance. We have a game team of about 30 people. The core engine team is about 10 or 15 people at the moment. It’s a really tight team now. The team itself–we have a presentation and Christmas party in a few days. We’re doing five-year anniversary presentations. That’s quite something. A lot of the team have been with us for years. I’m very happy now that we have leadership in place that can do what I want to do, rather than telling me we can do what I want to do and then not really having a plan.

GamesBeat: The vision sounded the same. You’re going to build this world, and then the players will figure out what the game is.

Greene: The vision really hasn’t changed. Even when I looked at some old pitches I did from four years ago, when I was first pitching it internally to Krafton–again, it was a three-game plan. They came back with slightly longer time frames and slightly more realistic goals, but it was still this idea that we’d prove each stage of the tech with each game we’re building. The vision is still the same.

I don’t think anyone is serious about building a metaverse. I think everyone’s building IP bubbles that will sometimes have to talk to each other, I guess. I don’t really see the metaverse as described by the people building it. What we’re doing, it’s open. We have it in Discord. People are already modding and hacking it. I see Artemis or Melba, that engine being hopefully an open-source world creation engine that will power some form of 3D internet. It’s not just one world. It’s hundreds of worlds, thousands of worlds. I see every world as like a web page.

Since we did the release–they have those things, deep links. You probably saw them in Discord, where you can hop around the planet. I had this flash in my mind. Maybe that’s what a hyperlink will be. There’s this idea that you don’t have to travel there on the planet. Someone will just send you a link to something cool on their planet or your planet or Tom’s planet. Then you can click and it will open up the app and bring you there, much like a browser will in today’s internet. It’s just a 3D location that has something interesting, or not. It might just be beautiful. The vision is still going for that.

It’s not meant to be like a game world. It’s a world with game-like experiences, I’m sure, but ultimately it’s just a huge world for players to come and build or view or share. I’m not really sure what they’ll do yet. I know I’ll give them lots of tools to do stuff. I always thought that the world we’ll provide, or the example we’ll provide, will be like Minecraft survival. That will be our slice in all the worlds. That’s more just a big Earth-shaped thing that looks like Earth and has basic survival mechanics. Let’s say civilization mechanics. You can do lots of stuff to eventually build communities. But again, that’s 10 years away, I think.

GamesBeat: I didn’t quite grasp what the three games meant. Prologue is a geographically limited game. Preface is more like a demo. But I didn’t know whether you counted that as one of the games. And then you have Artemis.

Greene: Preface will be the final game, probably. Prologue was just us testing the small-scale systems, player interaction, and the terrain tech. The reason we have three games is that each is solving one step in the process, or one problem. The first is terrain. Prologue, we have our ML tech that powers the terrain, generates the terrain. We can leverage Unreal to test that in this box called Prologue. We can test out lots of player interaction systems. How do we store that? How do we have persistence? All this using this ML agent.

Game two will be testing the ML agent on a bigger scale, making bigger terrain. Hopefully the terrain tech will be relatively mature at that stage. And then thinking about multiplayer. Not on a crazy scale. Just what’s usual at the time. But then lots of agent interaction. It’s going bigger and testing the terrain, the systems, stuff like marketplaces on a slightly bigger world, before we finally go to massive multiplayer, where I hope hundreds of thousands if not millions of people, in 10 years, on this massive terrain, which should be generated locally–that should be well mature with all these other systems that we’ve tested through Prologue and game two. It’s all just iterating on the vision.

GamesBeat: Will each game then be a separate product that gets to market? Or do you see them more as demos?

Greene: Prologue will be a product, for sure. There’s a story that we have, that I would like to leverage during early access, or after we launch into a full product. But it serves a purpose. I don’t want to put every bell and whistle on it, but it will still be a product. Then, once its life cycle is over, we’ll evolve it into the next stage. Prologue will move into the next game. Maybe you can play Prologue in the next game. I don’t know. But it’s kind of like Rust. As we go bigger, the products will be separate products, but they’ll bleed into each other and iterate on top of each other. They’ll stand on each other’s shoulders, so to speak.

GamesBeat: If you have a story, it sounds like you’re going to make your game within that game world. But you’ll also make it moddable so that other people can play with it and figure out what kind of game they want to make. Prologue can be that directed game where–it seems like it’s important for you to design a game, as opposed to leaving it all up to consumers.

Greene: When I thought about this many years ago, when we were thinking about whether we could generate a terrain every time you press play–that’s an interesting idea. What’s the easiest thing to do here? I thought about a simple survival game where you get from A to B across a map. It’s you every time. The weather gets worse, wave-based weather. It just keeps hitting you. Prologue is essentially that. It’s not that I’m making a game. I said in the Discord chat that I want to build games with the community, not for the community.

This is an interesting way of generating game worlds. We have some simple systems in it, but already, during the playtest, people are suggesting, “How about this? How about that? I want to stay in a cabin for four hours and play guitar and watch the weather outside and not do anything else.” I’m not trying to make people play a game. There are things you can do within Prologue to get to the other side of the map, get to the finish, and learn a bit of what the game may be about. But otherwise you can just sit in the cabin for five or six hours if you want.

I’m not trying to force people down a particular path. That’s why I want to get the community involved early. This way of creating game worlds is interesting and exciting to me. People who love survival games more than me will give some really good ideas when they get a chance to play it. That’s why we have playtests already. People are already finding weird and wonderful things about the game. That excites me. Sharing this tech early with the community and getting their input now is how we make this a great game. It’s not just me directing everything. It’s pulling feedback from people who really care about these games in ways that I haven’t thought about.

GamesBeat: One thing that I wonder is what kind of variations you can have if the game is–I don’t know if you call it procedural. You regenerate the world every time you log in, is that what you’re actually doing?

Greene: It’s machine learning procedural, but it’s machine learning. The ML agent generates a low-res map at the start of the game. Technically, mathematically, we can do 4.2 billion-odd maps, or generations. If a million of those are interesting, I’ll be happy. But you can see in the background, this is the ML map, but with us generating mountains. These are going to be impossible to create. You won’t be able to traverse them. But the idea was, we want to get the weather station up here. How can we make it more interesting and get it up in the clouds? They got very excited when we generated this, but no, it’s not going to be traversable.

The idea that it gives us a base to work on in Unreal–the maps we have, I’ve seen a good deal of variation. Even now, it’s very early days with this tech. The guys are discovering new ways to manipulate the PGC system, the procedural generation system in Unreal, to create more interesting biomes, to leverage our tech to create different rivers, masks for rivers and mountains. It gives a pretty good variation of worlds. We’ve seen some interesting worlds from the generations already, and that can only get better over the next six months.

Before we did our very first playtest with the Dutch Game Association, we had gotten cabins spawning in the week previous. This is all very new for us. But it’s still exciting. This looks cool. It’s not going to make it into the game because it’s far too high, but still, this kind of landscape, to me–yes, I want to go explore that. I want to get up to the top of that. That’s why we’re doing it.

GamesBeat: There’s the thrill of exploration that you can have in a world that generates over and over. What about the feeling of familiarity that some people may want? I can see myself thinking that I just want Earth, so I know where everything is. Or something that remains persistent that I can go back to and explore different parts of it. Is that going to be possible? Or will it be different every time you log in?

Greene: Melba and Preface is meant to be persistent and deterministic. If you go back to the same place, you’ll see the same things, always. That’s the aim. With Prologue, it’s seed-generated. We can hopefully eventually share the seed of the map you just played with friends, and you can play that same map. There will hopefully be a meta-game. Maybe you can even race people. But that’s probably DLC content down the road, because for the first launch it’s too much to expect from the dev team. This is not a fully-featured product. I don’t want to split dev resources. I want to focus Prologue on what it’s there to do, which is test the terrain tech and make an interesting systemic survival mechanic or game loop that we can carry over.

It’ll never duplicate the Earth. Nvidia’s Earth 2, that kind of thing, our terrain tech isn’t designed like that. It’s not designed for replication. It’s designed for Earth 5, Earth 10. It looks like the Earth. It might have the same feeling, the same biomes. But if you go to Barcelona it’ll look a lot different. It’s not Barcelona. It’s just that part of the world generated in a new way. Also, I just think Earth’s been done. So many other people are generating duplications of these things. Go on Google Maps and you can see the world. I want to create unique spaces. This is going to be Earth-like, of course, but it’ll be not-Earth-like as well, depending on who’s putting in the design input. This will all be open.

GamesBeat: Some of the variety is going to come from how many biomes you can create, then? If you come up with 1,000 biomes, you can have wide variation in the terrain.

Greene: Exactly. But again, you look at NASA data, and there are 20 defined biomes on the Earth. That fills the whole Earth. They’re very high-level definitions of what a biome is, though. Tundra, this kind of stuff. Within these you’ll have sub-biomes and so on. Earth data already provides us with a huge amount of data to try to train these agents to give us the right combination and depth. We still style and theme the worlds. We decide on how many biomes, how frequently they should mix. That kind of thing is still decided by us rather than agents. We’re still guiding their hands, so to speak.

GamesBeat: If somebody wanted to re-create your battle royale inside Prologue, do you think that would work?

Greene: Prologue, you won’t be able to do that. It’s Unreal. It’s a single-player game. This is a survival game. We’d like to open it up for modding, but I don’t know if that’s on the table right now. Whereas Preface, the tech demo we released, that’s being released with an open mind. We’re leaving the files unencrypted. The models are there for you to play with if you can. We’re not trying to hide that. I like to say it’s HTTP version 0.01.

It’s funny. If you think about biomes, there are already people in our Discord who say, “I’ve been going for hours and it’s still just the same rocky desert.” Yes, because the Earth is big. The true scale of the Earth is massive. It’s going to take time. The internet was pretty empty at the very start. I see the same thing with Preface. Right now it’s empty. There’s not much happening. But people in the Discord really see the possibility. You can see them getting what it is, or what it could be.

GamesBeat: By Artemis, then, you have that world where anybody could create anything. You could do your battle royale there. But maybe you want to rope off territory and say, “You can only play in this area.”

Greene: No, not necessarily. One of my earlier ideas–say I discover this forested area here, and I want to do a motocross race. I should be able to just pull up something on my wrist, paint where I want the track, and the game provides the rest. The game enacts a motocross race for me, adds everything there. That’s what I would like. We’re probably 10 years away from getting there, if not longer. But ultimately I would like that ease of creation. You can just wander around this big planet, fly around doing whatever, see something cool, and say, “Yes, I want a battle royale there.” Or a motocross race or whatever. The game should make that easy for you.

That requires whole layers of thinking, different networking layers specific for those types of game modes. They’ll probably lift and shard off that part of the world from the main world. As I said, five or 10 years. Probably longer.

GamesBeat: If you look at what everyone else is trying in these different ways, there’s the Nvidia Earth 2. There’s Hello Games trying something with a planet-sized world. There’s Flight Simulator doing it by adapting photos of the Earth that planes or satellites can take, getting their hands on all that available data to generate an Earth. Are there any approaches you’ve seen that you’ve thought about or found interesting? It seems like everyone is doing something different.

Greene: As I said, I like our approach. I think we have a pretty good one. We use three agents to generate the world locally. Most of the stuff I’ve seen from even Epic’s big world stuff is server-client. I don’t think that’s how you create massive worlds. You’re always dependent on a performant internet connection and all kinds of things that a kid in Africa doesn’t have. How do you generate a world for everyone that half the world can’t access?

Our view on it, which is, you do the simulation as much as possible locally on the device, rather than worrying about server farms handling that for you–I just think the future is local anyway. Ultimately I would like to have all my data stored locally and give it out to the network when I need to. Otherwise it’s here, rather than worrying about what server it’s on. Again, five or 10 years–for what we’re trying to create with Melba and the platform, these kinds of things are important to think about. They will come into play in a very big way. Trying to solve them with Band-Aids is not the way to do this.

GamesBeat: The good thing is we’ll have much more storage by the time this is ready. The interesting thing I talked to the Flight Simulator people about, if you added up everything they created for Flight Simulator 2020, it was about 500 gigabytes. Then they decided to shift almost completely to the Azure cloud. Now they have just 50 gigs on the local machine, and everything else streams in. That led to some hiccups at the beginning, trying to deal with so many players coming in, but that seems to be under control. But I wonder, why would that way of building a world be harder to do than the approach you’re taking, where it sounds like most of it will be on the local machine?

Greene: I’m not familiar with how they do things. I guess the core difference between their tech and our tech is that it’s still generating game worlds in an old way, where you need to understand what they look like. Our tech understands that inherently. It understands what terrain is, what mountainous regions are, what biome placement is, what trees to place in various areas. That’s all done generatively and in real time around the player, rather than having everything baked. That’s why you have so much data, whether 50 gigabytes or 500. Our world, which is 500 million square kilometers, is 3.6 gigs. That’s all generated locally on the player’s side. It’s just the way they’re thinking about doing it.

We have three patents on what we’re doing because we’re making these breakthroughs. How we’re doing this is a new way. We’ve seen other attempts at using inpainting and all kinds of stuff, using ML in other ways to create these worlds. But I’ve been happy with what we’ve been able to do. We’re generating millions of worlds in Unreal now, eight by eight, and they look pretty good, pretty high detail, not super fake. They look natural. It really excites me. I think this can open up games to a lot more varied experiences, rather than replaying the same map over again.

I saw that The Long Dark is coming out. But also Don’t Starve. That was a great game, super procedural, a different map every time. It was exciting to play. But we’ve never really had that in a single-player game. Maybe we have and the internet will shoot me down. But I really want to create this kind of replayable single-player game that focuses on exploration. We were even putting maybe a tent into the game, because people had said, “Maybe I want to sit on a hill until the weather changes and see the vista.” So let’s put a tent in so people can survive there instead of being cold. There’s this kind of lovely back-and-forth with the community already.

The dev team is excited. The community Discord is excited. I can’t wait to see what we can do in the next six months as we ramp up to Q2.

GamesBeat: I remember when we were talking about the metaverse before and what happens when you try to go between worlds, different worlds. There’s one question there. Did you consider breaking up something like Artemis into a bunch of worlds? You have so much territory here, something planet-sized–

Greene: But I think it will be eventually. It will be millions of worlds. It’s like the internet. It won’t be one single page.

GamesBeat: You mentioned that when you cross a border, AI is going to translate your stuff from one world into the next world.

Greene: I would hope so.

GamesBeat: I thought that was crazy at the time. But the last year or two years of generative AI–it seems like it’s made that possible. Has that become important for your plans?

Greene: I wouldn’t say very important, but there’s definitely been some advances that we can leverage. For example, texture generation. For a whole planet, to ensure we have a variety of textures, ML generation is great. It gives you infinite variety, basically. It also speeds it up and lowers the cost. You don’t need to store hundreds of texture files. It’s all generated on the fly as you go through the world. Stuff like this, we can find specific ways for it to make the world run better, with a smaller footprint.

Doing the photo to a 3D object, that kind of stuff is exciting to watch, but I’m not all in on AI yet. Even though I’m working on it quite a bit. There are some great possibilities. It’s an exciting future. But we want to be careful about committing too hard in one way or another. We’re pretty happy with what we have right now. But some advances in the last few years have filled me with a bit of excitement as well.

GamesBeat: I was trying to think of game spaces within these different projects you have. With Artemis, it seems like you’d have those millions of different kinds of spaces. People can choose to have very small game spaces, like a town where you could have a gunfight, or very large ones too. How many people do you envision in one game space? Is there a maximum you’re thinking about?

Greene: I don’t know. In the shared experience I want millions of people. Having a massive Earth-scale world, you need millions if not billions of people. But I don’t think that’s–again, solving the network problem. We’ve solved the terrain issue, generating massive planets. That’s not that hard. It’s not that costly anymore. We can do it locally. It doesn’t ask for a lot of disc space. It generates pretty nicely. It’s the same for multiplayer. We want to make sure the protocol, the layer we have works well allowing multiple people to get on the same space together.

I would love to see a 1,000-player team deathmatch, with teams of 50 or 100 players going against each other. Why not? As long as the play space is big enough. With game two it’s something we’ll try to explore, upping the player count to something that’s still reasonably possible and then seeing how that large-scale interaction works. Again, if it’s a systemic world, if it’s emergent, like a lot of the spaces I like creating, it’s easier to build. But these kinds of large-scale interactions excite me because no one’s really pursuing them. Everyone’s still happy with 20 or 30 or 100 players. Come on! It’s been 20 years already. Give me millions of players, please.

GamesBeat: A lot of game designers have said that that’s all they can see as being fun. Would that many players in a game be fun for the individual? The Call of Duty designers are perfectly happy with six-on-six.

Greene: Again, 100-player battle royale probably wasn’t seen as fun before it happened, and it turned out to be a lot of fun. I don’t think we can say something isn’t fun if we’ve never experienced it. I struggle with that kind of–it can never be fun if it’s over whatever number? Let’s try it. Maybe it’s fun and maybe it’s not.

I’m not trying to make games with millions of players. I’m just trying to create these shared social spaces for millions of players to have experiences together. Maybe they’re games. Maybe they’re concerts. Maybe they’re all kinds of things. But it’s more that you have large-scale interaction. But hell, bring on 1,000-player battle royale and see what happens. Bring on 1,000-player search and destroy. Look at the real world. You see now–paintball games used to be six-on-six, but now you have whole teams of hundreds of players going at each other in some of these massive paintball tournaments.

I don’t know. Any new technology scares the stalwarts, right? You saw it with that lovely ILM documentary, “Light and Dark,” about moving from puppetry to computer graphics. We can’t do it? Oh, shit, we can do it. Of course puppetry has now evolved into something even more special. It’s been forced to evolve because of other tech taking away the low-hanging fruit. It’s always an evolution. You should want to see it move forward, rather than just trying to trap it in a box.

GamesBeat: I remember games like World War II Online. They were trying to get 100,000 people or more into an MMO, so that they could replay historical battles. Would something like that be doable inside this kind of world?

Greene: Wouldn’t it be great? We could get 100,000 people all playing together. That would be great. The tech should hold up. But again, this is what game two and game three are intended to test and prove, to make sure that we have multiplayer, that we have interaction systems, that we have all these AI systems that work well together. By AI I mean bots in games, so you can control stuff. Having all this level of interaction and scale all working. As I said, Melba, Preface, it’s all open. Not open source technically right now, because that comes with certain responsibilities we’re not ready to commit to yet. We need time to work. But we’re still doing it with this open mentality, where nothing’s encrypted. It has to be built with the community. The internet was, and I think the metaverse has to be the same.

GamesBeat: In this kind of game world, does the concept of shards still exist?

Greene: No, because I don’t see servers. That’s the thing. I think it will be peer to peer. We’ll have a hybrid peer system, where you’ll have peers that handle–you could be one of these peers if you have a decent enough system, handling the high-level simulation for physics, weather, ballistics, these other heavy needed simulations. That sends data to lower-end devices. That’s how I see this working. We’ll have some kind of peer to peer system that will self-validate or self-auth rather than being reliant on servers.

I still think we’ll have a hybrid peer-server type of model that will hopefully be able to distribute across both users and a more commercial grade. But again, I don’t think–it can’t be based on servers, or else we’ll never get to hundreds of thousands of players. It just doesn’t work like that.

GamesBeat: Is it starting to look more like a decentralized blockchain infrastructure?

Greene: No. It’s decentralized in the sense of that word. I still think “federated” is better than “decentralized.” It achieves the same general goals. There was that interview I did a year ago with Nathan where he asked me about blockchain, and then the next day it was “PUBG guy making blockchain game!” That filled me with joy.

Blockchain or hashgraph or whatever, decentralized ledgers are useful in certain regards, especially when you’re trying to build a decentralized network. Whether we’ll use them, we don’t know. We’re years away from actively investigating that. It’s an interesting space, but I don’t see us using it in a similar way to how it’s been used so far. As a tech stack or a tech layer it’s interesting, but it’s not something I’m going to build games on. I don’t get that part. I’m building our own engine. It may incorporate some level of the tech as a layer to facilitate digital bookkeeping, but for me, that’s about the usefulness of it.

GamesBeat: Are you confident in the ability of a peer to peer system to handle something so large?

Greene: Just brash confidence, right? With reckless abandon I say yes. I think we’ve seen, with Bittorrent and blockchain, that decentralized peer to peer can be secure. There are some new blockchains that do this kind of self-auth stuff quite well. I’m relatively confident, as confident as I can be with the knowledge I have, that something will be there that can work.

Because we’re not building a game, so to speak – we’re building a world – then there’s certain–we don’t have to make it as performant, for example, as an FPS game. There are certain things we don’t need to ensure at that level. But then if you want to have an FPS game within our world, we’ll probably have to use a more known network protocol to enable a good experience there.

GamesBeat: What if the player is requesting a certain world? “You have a great wilderness world, but I want a city. Can you generate that for me?” Instead of getting a random world, can they wish for a certain kind of world?

Greene: With Preface, everyone gets the same world. With Artemis, everyone will get the same world. If you want to create your own world, the tech stack will be there for you to do that. Maybe we’ll provide a way where you can give us some money and we can create a world for you. I don’t know. This is 10 years away. But for me it’s always been like Minecraft. We’ll give you Minecraft survival. You can go there, explore, create, do things in the world using the tools we provide, but if you want to create your own world, you have to put it together yourself, host it from your own machine, rather than relying on us.

We’ll provide one layer, and experiences for lots of parts of the world, but you won’t be creating a new world when you press play locally. You’ll just be entering our world. Also, it may not be just our browser that you use to enter this world. Maybe someone has already created a new browser, better than the one we have, that allows you to do more in the world.

GamesBeat: Do you think that your world is going to be a contiguous world, an actual 3D planet, as opposed to something like–Second Life is this collection of places you can go, but it’s not the map of a world.

Greene: I would like our world to be contiguous. I would like that it seems to be the one world. But again, I don’t know. Ultimately I want to create a contiguous world. That’s what I would like to do. I would like something like this you see in the background, a massive world that’s there to explore. There’s lots of stuff to do. People can do whatever they want with it. Great. That’s the aim. Let’s talk again in a few years and see where it’s going. But that’s the aim, to provide a contiguous, unique 3D planet that allows you to spawn at various locations and create some stuff. It might have some urbanization. Early on it’ll probably have very little. But as we add more systems it should get more interesting.

GamesBeat: Would you get something like the actual physics of the Earth?

Greene: Why not? Exactly. Then maybe we have a more extreme world, or a more playful world. It should be easy sliders for me. That’s ultimately what we want to create with Melba. It should be that easy. We can just change a slider and the gravity changes. The world is created in real time, so if the data slightly changes, we should be able to do that.

GamesBeat: I think I know the answer to this, but others might be wondering. How do you build something this big without 10,000 game developers?

Greene: That was always the aim. When we sat down to do a 100 kilometer by 100 kilometer map initially, when I was still at Krafton, we discovered–okay, you need that many game devs to build that world, because it takes so much time. That’s why we tried to solve–how do you create a world in real time and generate it? That’s how we’re doing it. We already have the terrain part of that solved. We still have to figure out how you store persistent data in an efficient way, but at least we’ve solved the terrain generation part.

Now comes the gameplay and other systems. But since they’re always systemic, they’re pretty simple, especially in the real world. I hesitate to say I don’t see this as much of a problem, but I think we’re solving the bigger problems. The terrain was a big challenge. We’ve solved it in a pretty unique way, in a breakthrough way. There’s still a lot to do, a lot I don’t know, but I think the vision is clear. I’m confident about getting there.

GamesBeat: Financially, is your situation still pretty similar to what it was a year ago? You had your own money. You had money from a couple of companies.

Greene: We have funding to get us through launch and after. Of course we would like more money, but we prefer to make that from selling the game and using that to reinvest in the studio, rather than looking for another round. My aim with all of this, always, is to make sure the team can pursue the vision without having to worry about just pumping out products for sale. Whatever we choose to do moving forward, it’s always with that priority in mind. I have to give the team that safe space to dream, to be able to be psychologically safe. “This is a good place to work. We’re doing some good stuff.” We’ve achieved that pretty well over the last year. People feel good coming to work and excited about the project. I want to continue that. We need to sell games, but we’re pretty good right now.

GamesBeat: When you look down on the micro level of things like the cabin you had, it was pretty detailed in there. On that side, do you envision–do you have to have an army of creators making these small things that could be useful for players in this kind of world? How much work is that?

Greene: I’d love for our art director to give you a proper answer on this, but it’s more that the tools these days, for example Houdini, are allowing us to do a lot more variation on stuff like cabinets. Ultimately there will be some kind of blueprint that can generate multiple different variations. We have something like 300 variations of the cabin spawned across the world, because it’s relatively easy to do. It doesn’t take a lot of dev time. The cabins still look pretty good. With the variation they’re relatively believable.

It does take time. I’m not going to say it doesn’t take time. But I’m impressed by how far they’ve come in the last six months. When Petter, our producer, joined about nine months ago, he asked, “Where’s the build? Where can I play the game?” There weren’t many responses. Within a week he got a playable build up and running. Since then, the progress has been remarkable. We have a game that I get excited to start up, excited to run and try to find my way through it. I can’t wait to get it in the hands of more people.

GamesBeat: It sounded like one thing you were asking players to give feedback on was the level of detail in the world, if it was enough. Do you think you’ll have a difference in the quality of what you can generate compared to the quality they’d expect in single-player Unreal Engine 5 games?

Greene: I think it looks pretty good already. The forest landscapes–we still need some more detail, for sure. Especially the terrain level, to make it a bit smoother. But it’s keeping me happy. I’m pretty pleased with how it looks. The forests look natural enough. It’s still early days. We still have six months of work to focus down on the look and feel. But I’m pretty happy with what we have already. I think players should be excited to explore the world. There’s enough detail already that it doesn’t look bad. Let’s put it that way.

GamesBeat: The Flight Simulator people said that compared to 2020, the 2024 game has 4,000 times more detail in the landscape. That suggests a rate of progress they can continue to ride on. Is that something you can do? If players do demand it, is that a curve you can ride in some way?

Greene: We’re trying to build the engine in a very generic way, so that as new tech comes on stream, we should be able to update that part or add it in. It shouldn’t be much of a problem. The world we’re building in Prologue behind me, we’ve already gone through various iterations on the terrain uprezzing tech. We’ve already gotten it down to finer detail. As our agents improve, as the training improves, it will get better and better. As you’ve seen with a lot of AI – image generation, video generation – it will always improve. We’re building the engine with that in mind, that it will constantly be iterated. If a new thing comes online, we should be able to adopt it as quickly as possible.

If people want more detail, sure. I don’t know if you’ve played the playtest yet, the build, I’m pretty happy with how the world looks. It’s a bit rough still, but the forests look pretty good. I’m excited.

GamesBeat: Well, I’m still very impressed with the scope of the ambition here.

Greene: I try to be consistent with my madness, right?

GamesBeat: Would you have advice for people around sticking with their ambitions?

Greene: Just be stubborn. Or, well, no. Someone told me I’m not stubborn. I’m single-minded. I’m in a privileged position to be able to do this. I know the games space right now is not the most wonderful place to work. There’s been a shit-ton of layoffs. There’s this conglomeration of IP where studios are just being thrown out the door. We’re in a privileged place right now, that we can pursue this and have me in a position where I don’t have to worry about anything else other than pursuing it. But being single-minded about what you do–if someone tells me no, I look for a way around it. If you really believe and think it’s reasonable and possible, then you should pursue it.

There are always going to be people that tell you no. Like you said about game designers who’ve decided that games of 1,000 people are probably not going to be interesting. They said that about games of 100 people, and now those are some of the most popular games out there. If you’re sure about something, if you’re confident and optimistic, just pursue it. Be single-minded about it.

That’s not very wise stuff. That’s what everyone says. It’s hard, though. You’re going to get knocked down a lot. But it’s having that anger inside you, the spite inside you, to say, “I’m going to prove you wrong.” Just going and doing it. It takes a lot of work. We were lucky with battle royale. It took about three years to form a genre. Counter-Strike took a lot longer. DOTA took some time as well. Things take years to cement and become something. That’s the other thing to remember. It doesn’t happen overnight. It might seem like it does, but it took me a year and a half or two years to make sure battle royale was in a place where it was picked up by someone bigger and went somewhere crazy. It does take time. Don’t give up. Keep going.

GamesBeat: The metaverse seemed to inspire a lot of people, including you, some years ago. It’s gone out of fashion now. Do you still believe in the metaverse, or has your view of that changed?

Greene: I just don’t see the metaverse that everyone else is building. This idea that it’s an IP bubble–even in the interviews that have been going around, that the biggest challenge is the business to business. The metaverse isn’t controlled by companies. It’s not my metaverse and your metaverse and this metaverse and that metaverse. It’s the metaverse, I believe. That’s only achievable if someone builds an open-source platform or protocol that everyone can use. There’s no partnerships needed. It’s just there, like HTTP. We tried to monetize that with AOL and other things, but really the metaverse just has to be an open-source platform.

That’s what I’m trying to provide with Melba, which is just this open-source tool that creates digital places, much like HTTP generates web pages. That’s where I think the metaverse is. I haven’t gone off it. I’m still plugging forward toward it. I think that’s what it should be, rather than what everyone else is trying to build, which seems to be just a funnel to sell you skins.

I don’t think we should be thinking about what fits in the world. There’s always going to be a joker in a crazy costume running the ultramarathon. This world might have billboards put up because someone can afford to do it. This is a beautiful world. What people make of it? Well, we don’t know. But let’s see.