DeepMind's Geoffrey Irving: AI companies could just stop unilaterally
ShakeelHashim · x · 2026-09-29
Geoffrey Irving of DeepMind's frontier safety team said in an interview: "If you're doing a very dangerous thing, you should just slow down." He argues AI companies overplay the framing of safety as a pure coordination problem — they could unilaterally stop dangerous development without waiting for everyone else. A notably blunt take from a major lab insider.
More from AGI Musings
- Nature comment: AI makes researchers 3x more productive while narrowing science overall — StellaLisy · 2026-09-29
- AI Consciousness Needs a Science-First International Effort, Not Bayesian Guesswork — erikphoel · 2026-09-29
- Scoble: AI Agents Will Turn Advertising From Persuasion Into Proof — Scobleizer · 2026-09-29
- Security researcher moyix: AI-assisted writing is 'actively counterproductive' for him — moyix · 2026-09-29
- Does Exponential Compute Mean Exponential Intelligence? Tom Davidson and Ramez Clash Over Intelligence Explosion — connoraxiotes · 2026-09-29
- RLVR Can't Prove What Its Verifier Can't: The ZFC Ceiling Argument — Liu_eroteme · 2026-09-29