●researcherGameplay as a large-scale behavioral pretraining corpus is a testable hypothesis worth tracking — the approach could validate or challenge simulation-based agent training assumptions.
●founderA $2.3B bet on game-trained agents signals serious capital flowing into embodied and agentic AI, which could accelerate tooling and competition in that space.
●investorThe $320M raise at $2.3B valuation is one of the larger bets on agent pretraining infrastructure — worth tracking whether gameplay-derived models transfer to real-world tasks.