Photon-1 learns computer use from 18 years of unlabeled screen recordings
ycombinator · x · 2026-07-24
Y Combinator is amplifying a startup’s announcement of imagination models, a new foundation-model architecture trained on internet-scale video.
The first model, Photon-1, reportedly learned to use a computer by watching 18 years of screen-recording video without action labels, suggesting a new route to learning from passive observation rather than manual supervision.
More from Research
- A practical entity-normalization flow for clean knowledge graphs — Roger_M_Taylor · 2026-07-24
- Numerai podcast teases a breakdown of robust modeling and feature neutralization — matlabulous · 2026-07-24
- Turing Institute highlights AI methods for forecasting a changing climate — turinginst · 2026-07-24
- Google releases ATLAS 1.0, analyzing 15 million AI interactions across Gemini and AI Mode — jackclarkSF · 2026-07-24
- AI Security Institute red-teams monitor AI agents for rogue actions — HZoete · 2026-07-24
- Frontier-Bench debuts with 74 tasks and top agents scoring about 34% — giswqs · 2026-07-24