According to XDA-Developers, Brendan Greene—better known as PlayerUnknown—is launching his new survival game Prologue: Go Wayback! into early access on November 20. This marks his first major project since stepping away from PUBG development in 2019. The game serves as the first part of a three-game plan culminating in Project Artemis, a world-scale MMO that Greene believes requires machine learning technology. Prologue: Go Wayback! specifically uses ML to generate 2048×2048 pixel height maps locally on players’ PCs in about 60 seconds. The technology creates terrain using open-source data from agencies like ESA and NASA without requiring always-online connections.
How the magic works
Here’s the thing about Prologue: Go Wayback!—it’s basically a tech demo disguised as a survival game. Players navigate from a cabin to a weather tower across generated terrain, dealing only with weather and survival mechanics. No enemies, no multiplayer (at least for now). The real star is what Greene calls “Guided Generation,” where ML creates black-and-white height maps that determine where rivers and mountains go. Then traditional tools like Unreal Engine 5’s PCG system populate the world with trees and buildings.
What’s interesting is how different this is from typical procedural generation. Most games use algorithms to create random but consistent worlds. Greene’s approach uses machine learning trained on real geographical data to create more natural-looking terrain. And it all happens locally—no massive server farms required. That’s actually pretty clever when you think about it.
The grand plan
Now, why go through all this trouble for what sounds like a relatively simple survival game? Because Greene’s playing the long game. Prologue: Go Wayback! is essentially the training wheels for Project Artemis, his white whale—a world-scale MMO that would make No Man’s Sky look quaint. The technical challenge is mind-boggling: how do you store data for an entire planet-sized game world? You can’t. So the solution is generative worlds created locally on each player’s machine.
Greene’s argument makes sense from a technical perspective. If you’re building something truly massive, traditional server-based approaches just won’t cut it. The data storage requirements would be astronomical. But here’s my question: when has any developer actually delivered on the “world-scale MMO” promise? We’ve heard these ambitions before, and they usually end up scaled way back.
The ethical elephant
So let’s talk about the uncomfortable part. Machine learning and AI in games have become seriously controversial lately. We’ve seen backlash against games like Arc Raiders and Call of Duty for their AI use. Greene insists his approach is different—it’s using open-source data for terrain generation, not replacing artists. He even says his studio still employs environment artists to design scenes and create assets.
But then he drops this analogy: “Think of it like an orchestra. Instead of having someone playing a violin, they now just conduct the orchestra.” Okay, that sounds cool until you realize he’s basically saying the conductor is more important than the musicians. And that’s where I get nervous. When this technology becomes accessible, what stops big publishers from deciding they need fewer “violinists”?
Greene claims this tech will allow “smaller teams of artists to iterate on the world a lot quicker.” That sounds great for indie studios, but terrifying if you’re an artist at EA or Ubisoft. The line between “tool that helps artists” and “tool that replaces artists” gets pretty blurry pretty fast.
Wait and see
Look, I’ll give Greene credit where it’s due. He’s actually shipping a game (Prologue: Go Wayback! is real and playable), not just making lofty promises. His approach of local generation using open data is more ethical than many AI projects we’ve seen. And the technology itself is genuinely interesting—imagine being able to generate realistic terrain without needing massive server infrastructure.
But I can’t help feeling like we’re at a crossroads here. Machine learning in games could either become a powerful creative tool or another excuse for publishers to cut costs and jobs. Greene’s studio might use it responsibly, but will the big corporations? Probably not. So while Prologue: Go Wayback! looks like a fascinating experiment, it’s also a test case for where this technology might lead the entire industry. And that’s a conversation we all need to be having.
