FrontierJuly 26, 2026via MarkTechPost
Induction Labs Photon-1 Simulates Desktops, Plays Checkers, and Models Billiard Physics From One Pretraining Run
Why it matters
Induction Labs challenges the assumption that video-based AI agents require labeled action-frame pairs. Photon-1's unsupervised pretraining approach on raw video could unlock cheaper, faster agent training—a potential inflection point for embodied AI development.
Key signals
- Photon-1: 106B-parameter sparse MoE model
- Trains on raw video with NO action labels
- Demonstrates capability on desktop simulation, checkers, billiard physics
- Architecture: imagination models (foundation model for video understanding)
- Single pretraining run produces multi-task competency
- Challenges current paradigm: action-label dependency in video learning
The hook
106B parameters. Zero action labels. Induction Labs' Photon-1 just proved video models don't need supervised action data to learn world physics.
Most agents that learn from video need to know what action produced each frame. Induction Labs is arguing that this requirement is the bottleneck. Last week, they released imagination models, a foundation model architecture that pretrains on raw video with no action labels at all. Their test system …