Neel Nanda: Losing Monitorable Chain of Thought Would Be a Safety Tragedy
NeelNanda5 · x · 2026-09-03
DeepMind interpretability researcher Neel Nanda pushed back on a growing view that keeping Chain of Thought monitorable doesn't matter because interpretability will save us, or CoT is already useless. He called that take 'total bullshit': CoT is our best current tool for AI safety and interpretability, and losing it would be a major tragedy — a pointed comment amid the industry shift toward latent-space reasoning.
More from AGI Musings
- Hiring at high-performance orgs shifts from quantity to quality as AI agents amplify output — andreisavu · 2026-10-22
- AI made the once-hard easy; the new challenge is finding discipline — pixlpa · 2026-10-22
- Grady Booch Slams Altman and Amodei: AGI Hype Is Irrational and Untrustworthy — Grady_Booch · 2026-10-22
- Don't call LLMs useless before throwing trillions of tokens at a problem — iruletheworldmo · 2026-10-22
- Crosby trains judge models to evaluate legal AI where no unit tests exist — graceisford · 2026-10-22
- BaseLabs open-sources mammoth-1 and elephant-1, top-ranked decision models — saranormous · 2026-10-22