Frontier research lab building large action foundation models trained on billions of action-labeled gaming video clips for space-time understanding, action prediction, and world modeling that transfers to physical domains.
General Intuition is a frontier research lab founded in 2025 and based in New York, spun out of Medal, a platform where millions of gamers capture and share gameplay clips. The company's work centers on large action foundation models trained on billions of action-labeled video clips drawn from gaming environments spanning tens of thousands of games. The stated objective is a machine capable of space-time understanding, action prediction, and world modeling - what the company describes as a general intuition of reality - with learned capabilities that transfer from virtual to physical domains.
For robotics engineers, the relevant claim is the virtual-to-physical transfer: models trained on gameplay video are intended to perceive, anticipate, and improvise in ways that generalize beyond the games they were trained on. The technical stack sits at the intersection of embodied AI, spatiotemporal reasoning, video understanding, and transfer learning, with training data sourced at scale from the Medal clips ecosystem, which captures billions of gameplay moments per year. The company publishes no team size or funding figures.
General Intuition operates as a public-benefit corporation and states a commitment to collaborating with, rather than competing against, the gaming industry and creative professionals. Work is positioned as open-ended frontier research across both virtual and physical environments.
