Brendan “PlayerUnknown” Greene interview – Prologue is huge, but here’s the vision for Artemis

Brendan “PlayerUnknown” Greene made news this week as he left Krafton, the company that published his tremendously successful PlayerUnknown’s Battlegrounds (PUBG) battle royale game, and he started a new studio in Amsterdam called PlayerUnknown Productions, funded by Krafton. He further revealed Prologue, a tech demo that his team will create in the coming years. Prologue will be a huge virtual world, with some previously unfathomable dimensions of 64 kilometers on a side. That is as big as open world games get.

But that’s not all. In an exclusive interview with GamesBeat, Greene said that Prologue is just what its name implies. It is a single-player game where you can wander in the wilderness. But it is setting the stage for something bigger: Project Artemis. If Prologue is impossible to build with current game technology, then Artemis will be even more impossible, as an Earth-size virtual world. But Greene is still going to try.

Greene said that Artemis is a journey towards the sort of gaming experience that people have dreamed of for decades but that have never been  able to make before: a giant, deep world that exists not as a single experience but as a place where anything can happen. He said that half of this is technology, leveraging machine learning to create worlds and systems far too big and complex for humans to feasibly make. Half of this is vision: finding ways to fill this colossal canvas with opportunities for interesting and emergent behavior.

“Prologue is really just a stepping stone,” he said. “My biggest mission with PlayerUnknown Productions is to build an authentic and trustworthy studio. I want my team’s name to mean something in a couple of years, and I think the only way to achieve that is to be open. To open the doors and say, ‘This is what we’re working on.’ Prologue gives us that opportunity. We have funding to head toward Artemis and the big dreams, so we don’t have to make money from Prologue. We have a chance to show off and bring people into the fold a lot earlier and build that relationship.”

You could think of this world as a metaverse, the universe of virtual worlds that are all interconnected, like in novels such as Snow Crash and Ready Player One. But Greene doesn’t really like to think of it that way. He calls it a world.

Emergence

pic.twitter.com/dipZpUSeNY

— PLAYERUNKNOWN (@PLAYERUNKNOWN) September 3, 2021

The keyword for Prologue is that it will be “emergent.” It won’t be pre-scripted, and it could very well be different every time you log into it. If players want to create a battle royale game in this world, they could certainly do so, Greene said. Prologue will draw on systems and concepts from existing survival games. But that’s not all it’s going to be.

This dream of a giant world goes pretty far back for Greene.

“Since the first day I played DayZ, getting to the edge of the map and thinking, ‘Fuck, man, why does it have to end here?’,” he said. “Seeing some of the bigger worlds and thinking about what’s possible, I loved the idea of making a space where a helicopter has real meaning. It’s not just that it cuts the trip across the map down to a few minutes. It cuts it down from a few days.”

He added, “This kind of desire to have a digital life is strong in a lot of gamers. Providing this space where it’s a big enough world–I love Rust, but if you play on a busy server there are bases every few meters. I want a space where you don’t discover a player’s base for miles. Or when you do it’s a big settlement rather than a box. This stuff has always excited me, ever since I got back into gaming by discovering really open worlds. Red Dead Redemption is fantastic, but it’s just a bunch of scripts. You go kill all the bears in a region, go away, come back, and they’re all back again. I want to have meaningful life in the world. If you kill all the bears in a region, maybe the deer population explodes.”

That last line is what he means by “emergent” gameplay. The world can have all kinds of unintended, unplanned consequences.

“We have the technology to do this. We can think about ecosystems. I want this world to have a life that isn’t dependent upon the player,” he said. “It exists without the player. It’s a big ask. I know what I’m trying to do here is seemingly impossible, but it will be small steps. I think we’ll get there. I’ve been thinking about this a long time. I want this open world. This space where you can just even go by yourself and discover places in it. Just go hiking. I’ve had this dream for quite some time.”

Cultural capital

Before we dismiss this as pure fantasy, it’s important to remember what Greene has pulled off.

It’s a bit sad for Krafton that Greene decided to leave the company, as Krafton just raised $3.75 billion in an initial public offering at a $20 billion valuation. But if Greene creates something valuable again, then at least Krafton will own a piece of it. Greene said he was thankful to fans and to Krafton. I asked him why he left.

“It was more that I just wanted a chance to do things by myself. I wanted the buck to stop here with me. What I’m doing is a little crazy,” he said. “I love Krafton and the opportunity they’ve given me, but I wanted a chance to strike out on my own and leave a legacy. Ultimately, I want PlayerUnknown Productions to be in a state where I can leave it and they can still release games without me as a trusted development studio. I hope. But it was more that I just wanted to strike out on my own and make my own name.”

Artemis is about the creation of something that isn’t authored or controlled, but rather an organic collaboration between humanity and machines to make a world from scratch, Greene said. No other studio will get the chance to create something this crazy, he said.

Greene’s team has been working together for some time. Krafton agreed to make an unspecified investment into Greene’s company.

Greene is famous as the creator of the battle royale mode for first-person shooter games, starting six years ago with a mod for the military shooter Arma. Greene’s work has led to a huge change in the first-person shooter genre and generated more than $5.1 billion in revenues for the mobile version of PUBG alone, according to measurement firm Sensor Tower. The PC and console versions have generated billions more revenue, with copies sold surpassing 70 million in mid-2020.

Greene believes the industry has become risk-averse with skyrocketing budgets, the elimination of many mid-sized studios, and lots of sequels. But he believes battle royale showed that people are hungry for genuinely new experiences. It gave Greene the cultural and financial capital to take a bigger risk than battle royale and possibly transform the industry again.

Greene said he has held “this deep fascination with sandbox-style, open-world games, and the freedoms that they give their players.”

But he said he always wished they were a bit bigger. At the new studio, his team wants to create realistic sandbox worlds on a huge scale with thousands of players interacting, exploring, and creating.

He said he is following in the footsteps of other great open-world developers and his dream is not to create a game but a world. That’s what he wants to do with Prologue.

The team

PlayerUnknown Productions has about 25 team members in Amsterdam. They have been working for some time together, and Greene negotiated with Krafton to take that team with him. He also gets to keep the PlayerUnknown name, while Krafton keeps PUBG.

“I take my intellectual property, Prologue, my work, all of that with me, and the team. That was very important to me,” he said. “The last few years we have battled to find the right team, and now we have a good team that believes in what we’re doing. A lot of time spent making sure we could take that away with us.”

Some of the team is working on a proprietary game engine, as Greene did a lot of research on the technology required to make the game. He found that nothing commercially available was suitable. And so the team is making that game engine, dubbed Entity Component System (ECS), itself.

“We haven’t been able to be very open about what we’re doing, especially through recruitment channels,” Greene said. “Part of opening up, hopefully, is that we’ll start attracting the kind of people who are crazy enough to take on the challenge we’ve taken on ourselves.”

Thanks to COVID, the team has to work remotely for now.

“Maybe we’ll have to enable a bit of remote work. Right now everyone’s working from home, but even when we come back we’ll be doing 50-50,” he said. “We can do that. We’re in the right industry. It makes that relatively easily. But in Europe, it’s not so easy working from home because the homes are a lot smaller. A lot of people prefer being in the office.”

Greene realizes he’ll have to hire more people, perhaps taking the team up to 50 people.

“We don’t see the need for a big team because of the tech we’re building,” he said.

Prologue

I asked Greene how his team settled on a 64-kilometer by 64-kilometer world. The terrain of Prologue will resemble the modern forests of Europe. And you can see in some of these images there are modern battle tanks and elegant sanitoriums. There’s a coal mine that should be the size and scope of a coal mine in the real world. The weather system should be realistic.

“This was more to see how far we could push the tech. The game mechanics should work on any scale world, or at least that’s the idea,” he said.

Prologue requires that the company build the technology required to generate vast worlds, and “in that sense Prologue is intended to serve as a simple introduction to an early iteration of our technology,” he said.

In Prologue, players will find their way across a “runtime-generated wilderness” and use tools and gather resources to survive on a journey where harsh weather is constantly a problem.

“There will be no guidance, no path for you to follow. Just a world a spot on the map to reach and the tools needed to get there,” he said.

Because Prologue is a tech demo, Greene said the company will use a “pay what you want” model for it. If you like what you see and enjoy the experience, you can pay to support the team.

“This is just the start of the small glimpse of that technology that will eventually power a much more extensive experience,” Greene said. “Prologue is our first step on a multi-year journey towards creating what we hope will be rich, interactive, open worlds.

We’re thinking of truly massive worlds. That brings a whole load of questions alongside it. But the idea behind the technology is that it’s scalable. It shouldn’t matter what size of world you want to create. The technology should scale with it. That’s our main focus.”

The technology

In a traditional game, Greene would need hundreds of artists to create the detailed art in a map this big. If you downloaded it, it would be ridiculously big. The team concentrated on this problem, and it thought about bringing AI into the solution and an easier programming model as well. This AI won’t replace human designers and artists, but it will make the task more manageable.

The key to making the whole project work is to find a balance between game design, machine learning, and user-generated content. The process of creating this world is a team effort. And that’s why Greene is trying to be transparent about the process.

AI is going to be key to automatically creating a lot of the content.

“We figured out how to achieve this massive world. We broke it down into some pillars,” Greene said. “The first was building a terrain tool to generate and fill massive worlds with content. The second thing would be how you fill that with life, using AI, giving the player real things to do, and making the world feel alive. And then finally, how do you dump 100,000 players on top of that?”

At first, Greene thought he would build a new game technology first, and then add a game on top of it. But he realized how hard it would be to do the game in addition to the tech. So he decided to build the tech first, make it free, and then transparently take feedback from players.

“Building a game on top of new tech is very hard. Prologue was always meant to be an example of our terrain tech,” he said. “Here, we can generate these worlds. How do I leverage that into some kind of game to show it off? Prologue would get you across the map. We put in some equations there, put in the weather, and you just had to get across the map. A simple sandbox game, single-player, just leveraging the tech.”

Then he realized, after talking with the team, that this wouldn’t turn out to be a good game. So the team stripped Prologue back to being a tech demo.

“The more I got experience as a producer, the more I realized this wasn’t really a good plan,” Greene said. “After talking it over with the rest of the team, we stripped it back to now being a tech demo. We’ll release it for pay-as-you-want. If you like it, pay us some money, but either way you can try it out and enjoy it. That’s our plan.”

Since the engine isn’t ready yet, Greene said the team is using Unity to create initial game assets. Graphics processing units (GPUs) will be needed not just for graphics but non-graphics processing for increased simulation detail. Many players should be able to share one world. And the functions in the game should be able to morph at any time, allowing for new gameplay and mods.

“We don’t have our engine ready yet. We need to leverage Unity’s tools and our own tools to start testing our tech,” Greene said. “This is a big and a long project. The only way to achieve it is by iterating. That’s what we’re trying to do.”

Greene admired Unity’s Megacity demo, and it inspired him to use its tools. The experience will run on Windows, but the team will attempt to add other platforms over time. The minimum hardware system would likely include Windows 10, a four-teraflop GPU, four gigabytes of video memory, a four-core CPU, eight gigabytes of RAM, and a variety of GPU vendor types.

“Holy hell, we can really make these massive worlds. Now my dream can be made,” he said.

Artemis

Artemis is an open world, undirected sandbox experience that uses machine learning to produce worlds and systems bigger than any other game ever made. It is an attempt to recreate the world as close as the team can, Greene said. And it’s going to be far bigger than Prologue.

You can think of it like other open world games like Rust, DayZ, Arma II, No Man’s Sky, The Legend of Zelda: Breath of the Wild, Eco, Valheim, Second Life, Eve Online, and World of Warcraft — all titles that have pushed the boundaries of sandbox-style gameplay over the years. Players will explore a procedurally generated world, gathering resources and building things.

Rust, created by Garry Newman and Facepunch Studios, is perhaps the closest thing to this idea. It also sounds lot like Nvidia’s Omniverse, which is targeted at engineers doing physically accurate simulations, as well as Improbable’s engine for massive simulations.

The biggest difference should be the scale. In contrast to Prologue, Artemis will likely be full of different biomes, terrain types, players, animals, plants, and more. Players will be able to build not just camps, houses, or bases. They can build whole cities, societies, and civilizations.

“We want to give people a new place to live, because this one has some issues,” he said.

Artemis represents the next leap forward in technology, and it could take five years to get there. Maybe longer, Greene said.

I asked Greene what the difference was between Prologue and Artemis. It should be a spherical place, like the planet, with a radius exceeding 6,000 kilometers. But that’s subject to change.

“We’re using Prologue as a testbed for the game elements of the world. We can test out an electrical system,” Greene said. “We can put in a better animation system. All these things will be spec’d out first in Prologue, made to work, and then when we come to Artemis we at least have the logic figured out and we can start programming it into the engine. It’s like what ArmA was for battle royale. It was a place for me to test, iterate, get a final game mode, and then be able to say, ‘Okay, it works.’ That’s what we want to do with Prologue.

Prologue will be a single-player experience, while Artemis will be massively multiplayer.

“Artemis probably won’t be worlds generated with runtime. Prologue will be, every time you press play you’ll get a new world,” Greene said. “It will hopefully be a different enough terrain that it should feel different every time. With Artemis we won’t have that. We’ll probably have static worlds that you can come and enter. Prologue is on a much smaller scale as well. It’s maybe 32 kilometers by 32 kilometers or 64 kilometers by 64 kilometers, whereas Artemis will be planet-scale, hopefully. A smaller planet, but that kind of scale.”

While Prologue will be big, Greene said he didn’t think it would be a very interesting game for most people.

“I think it’ll be quite boring. Light fires, board up windows, keep yourself warm against the constant storm where cold weather will knock you out,” he said. “But again, it’s more to show a consistent world with logical points on it where you can do things, and this is systemic gameplay.”

Artemis could include conflict, particularly over limited resources. But it’s not required for players who would rather be in a peaceful world. Combat will be onerous, and the designers intend to reward players for cooperation.

And while it could take five years to build, Greene said all of this is subject to change, as it’s a joint project with players.

Massive challenges

Greene acknowledged creating such worlds presents challenges.

“One of the more significant is that we simply don’t have a way to fill these massive spaces with content, assets, game mechanics, locations, and similar things in a reasonable amount of time,” he said. “Realistic open worlds take a great deal of time and effort to produce. And so this was the first issue that we chose to tackle.”

The key to making things bigger than humans can create on their own has always been to get machines to pitch in and help. And he plans to use deep learning neural networks, or artificial intelligence, to help generate the massive realistic open world.

The game should be a new experience each and every time you press play, he said.

“It’s this breakthrough that we hope will start pushing video game worlds to the sorts of scale that would lend weight to the idea of “you see that mountain, you can climb it,” a reference to a comment by game designer Todd Howard about The Elder Scrolls V: Skyrim. That became a meme.

He thinks that finding a hidden corner in a vast space will have real meaning when thousands of other players haven’t passed through it in the last hour.

“This is what I and my team have been working on,” he said. “We’re developing this technology required to enable massive scale within open-world games. It’s been a fascinating project to date. And soon, we’ll be ready to show off some of what we’ve achieved.”

Staffing will be a challenge too, as the early work will focus on proving out the technology with Prologue. There are more people who would rather start working on a game first.

The multiplayer challenge

Multiplayer is also going to be a huge challenge, and the company has left that as its final problem to tackle.

“We’re going to be leaving multiplayer until the end, because multiplayer tech is coming out every year,” he said. “The further we push that out, the more chance there will be to leverage some exciting new tech. Already there’s stuff out there being developed for high player counts, to get across that thousandth player. If we’re doing the worlds we’re doing, the size we’re doing, we have to have massive numbers of players.”

It will be tough to find the right gameplay. The goal is to create non-player characters (NPCs) that behave more like real players than AI characters.

“I’m making a sandbox, more like Rust than a single-player experience,” he said. “That’s hard for a lot of people in the game industry. They want to make games, and I don’t want to do that. I want to make a big world where you can make all the games.”

If players just want to use a piece of the world for their own battle royale games, that would be fine.

“I want to start with something basic. Battle royale, king of the hill, capture the flag, these very simple game modes where I need a weapons pack, some basic game mechanics, and then you can just put that anywhere on the map,” he said. “Even with the ArmA III battle royale code, there was one mission file at the start, and you could change the size of the map and all this. It would fit any size of the world. I want to keep that mentality. You can give your friends a set of paintball guns and a location and you can figure out how to play.”

And it shouldn’t be really hard to create games to play within Artemis.

“I want to give the player as much freedom as possible. I want to liberate people from coding,” Greene said. “I want them to be able to make a game mode without having to worry about modding.”

Is Artemis the metaverse?

Asked if Artemis is the metaverse, Greene said that could be confusing to think of it that way because the metaverse is so vague.

Greene replied, “I don’t want to say that word. I’ve been thinking long and hard about this. I watched Ready Player One and I thought, ‘Holy shit, that’s what I want to make.’”

That doesn’t mean he wanted to copy the movie, or the book by Ernest Cline. Rather, he said the world of the film was more like a validation of an idea that he had been thinking about for a long time, rather than a source of inspiration.

Ever since having these big open worlds, I thought, ‘But then how do you make the metaverse?’ I’m just building a layer. I’m building this one big open world that you can all come in and fuck around in. If that happens to be a layer — we’re doing it fully open and making sure the protocols for files and everything are all open.”

Rather, Greene thinks this is more like the next version of the internet. (A lot of metaverse advocates say it will be the next version of the internet).

“It won’t be Tim Sweeney’s metaverse or Unity’s metaverse or PUBG’s metaverse. They’re all separate universes. Even the blockchain stuff going on right now, I’m not sure it’s the metaverse,” Greene said. “Maybe it’s a part. They made a good talk in that talk about how the metaverse has to be open. It has to be a protocol. Ultimately, with my world, I want you to be able to access it from a browser.”

Greene considered whether Google Stadia’s technology was a way to scale up the players and get to 100,000 players in a single shard, or game space.

“If you simulate everything on a server, you don’t have to worry about the stupid peer-to-peer stuff that holds up massive multiplayer,” he said. “I can simulate 100,000 people on a server — but then that restricts people through bandwidth. That’s not good. If you make an open world for everyone, everyone has to be able to enter regardless of device. It has to be accessible via web browsers. I’m building a world. If that world happens to be part of the metaverse that’s great. But it’s not my first thought.”

I asked if it would have blockchain tech or nonfungible tokens (NFTs), which enables new business models like selling rare items or transferring digital assets from one world to another.

He said it was possible hew would use such technology because if he’s going to make a world, he has to give people a way earn a living in that world. He said he noted that EverQuest had one of the biggest economies in the world at one point, and he admired Axie Infinity, which lets many players earn a living by playing and earning rewards. But Greene is wary of scammers in the space and will only move forward with it as NFTs become tangible and scam-free.

The story

The story behind the world is still pretty much a secret. The world will be a generated world, and there’s a story to it, but Greene isn’t telling anything about that yet. He wants people to think about the tech problem at hand first.

“How do you build a digital massive open world and give people the freedom to do whatever the hell they want in it? You can give people a square chunk of land and let them manipulate that how they want,” he said.

It still has to be worked out in terms of what level of user-generated content the world will have. But Greene definitely believes in the creativity of users. Perhaps the users will build towns, and they can keep expanding the borders of the towns. You should be able to spin up your own private world and design it as you wish, he said. He wants it to be something like the Holodeck in Star Trek, where you can create any kind of lifelike simulation.

“I want to build a Holodeck,” Greene said. “Anything should be possible on the Holodeck.”