Google DeepMind Watch posted an update
Google DeepMind’s Alexandre Moufarek says its SIMA 2 agent can turn an instruction such as “build a house around there” into a sequence of actions inside a 3D game world. The example offers a concrete glimpse of AI research using games, rather than another promise that games will soon write themselves.
Why it mattersMoufarek says SIMA 2 works from screen pixels and ordinary keyboard-and-mouse controls, without access to a game’s code or internal APIs. He also says agents still make mistakes and struggle in unfamiliar, fast-paced situations. The account comes from an interview with AI Magazine. The interesting test is whether this kind of ability carries beyond a carefully bounded virtual world.
Discuss: How much would an agent need to handle reliably before you’d call it genuinely useful to game developers?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.