Google DeepMind announced a research partnership with Fenris Creations, the studio behind EVE Online. Three game environments for long memory, planning, and agents that adapt. The work starts offline, and only mature capabilities reach live players.
- Google DeepMind announced a research partnership with Fenris Creations, the independent studio behind EVE Online.
- Three environments: EVE Online for long horizon continual learning, EVE Vanguard for tactical decisions, EVE Frontier for adapting agents.
- The work starts in an offline copy of the game. The Aura Guidance system already uses Gemini inside EVE Online itself.
Fifteen years DeepMind has taught machines to play games. They started with Atari: dots on a screen, three lives, game over. Now they are stepping into a world with an economy, wars, and real people inside it.
The difference between a box and a world
Atari was a closed box. Points, lives, game over. EVE Online is a different animal: an economy with real value, wars that run for years, and thousands of players who remember an insult from seasons ago. Teach an agent to play chess and it learns chess. Teach it to live in EVE and it has to remember, judge, and carry the consequences tomorrow.
Their caution stands out, and it deserves to be noted, because it is rare. The work stays in an offline copy until the capability matures. You, as a player, will meet none of this on an ordinary evening in the game. For now.