OpenAI Exec on Cheap Intelligence in Labs
LiamFedus · x · 2026-07-12
Liam Fedus says he loves the phrase "intelligence too cheap to meter" because in their lab, nearly every equipment will soon have 140+ IQ.
He gives an example of a materials synthesis lab: early reliance on technicians and scientists manually recording data from characterization devices; later switched to script auto-capture, which was scalable but not "smart", lacking experimental intent and context, missing anomalies and details. As AI becomes nearly "too cheap to meter", data collection becomes scalable while understanding experiment intent, viewing full records, and actively finding more valuable evidence from devices.
He likens this to the lab version of Rick and Morty's Butter Robot: the system's "purpose" is to watch the characterization device, and AI makes it truly meaningful.
Related event: OpenAI Exec Foresees AI Native Lab Equipment(2 posts)→
More from AGI Musings
- Superintelligence will be maximum good, not stupid or evil, argues Patterson — davidpattersonx · 2026-09-11
- Mathematician Daniel Litt Launches Problem Repo to Track Human vs AI Progress: 15 Problems, 1 Solved — littmath · 2026-09-11
- Should AI models be taught morality? Breakout incidents expose missing ethical training — Pfungus_ · 2026-09-11
- SoftBank's Masayoshi Son predicts 100 trillion self-replicating AIs: "humans' era as top life form is ending" — Puzzleheaded-King584 · 2026-09-11
- We are witnessing the unreasonable effectiveness of inference-time scaling — sqcai · 2026-09-11
- The AlphaFold lesson: AI-solved math may mean fewer mathematicians needed — kiki-le-koala · 2026-09-11