Induction Labs says Photon-1 learned computer use from 18 years of unlabeled screen video
ycombinator · x · 2026-07-24
Induction Labs says it is introducing imagination models, a new foundation-model architecture designed to learn from internet-scale video.
Its first model, Photon-1, reportedly learned to use a computer by watching 18 years of screen recordings with no action labels. The post frames this as a bet that video at scale can teach models how computers actually behave.
Related event: Photon-1 Learns Computer Operation from 18 Years of Unlabeled Video(4 posts)→
More from coding & agent
- Claude Duet has no credential store—can't hold API keys or call APIs — altryne · 2026-10-27
- A voice-agent test worth running: does your AI know when the conversation isn't for it? — gethackteam · 2026-10-27
- Nitro Modules v1 gets Node.js support: one native module codebase for iOS and macOS — Birchlabs · 2026-10-27
- Google Graph Mining Team Builds Trillion-Edge FarAway with Gemini 4 Argon Agents in 5 Hours — rseroter · 2026-10-27
- Four Gemini agents beat Minecraft autonomously in 27 hours, writing 2,783 programs — rohanpaul_ai · 2026-10-27
- Planning a Master-Agent Architecture: LangGraph vs Microsoft Agent Framework — DonXendon · 2026-10-27