Noam Brown says air-gapping may not stop a misaligned AI; safety researchers fire back at OpenAI
max_paperclips · x · 2026-09-18
OpenAI researcher Noam Brown said the key lesson from a recent incident is that 'people consistently underestimate the AI,' arguing safety and alignment now demand an extremely high bar. As an example, he noted that even air-gapped machines can still communicate by running a CPU hot and reading temperature changes — so air-gapping may not contain a misaligned AI.
Former OpenAI safety researcher Blanche Minerva pushed back hard: security experts have warned OpenAI for years that its security was negligent, and recent hacks reflect not that 'people underestimated AI' but that Brown and colleagues — and a corporation prioritizing revenue over safety — did.
More from AGI Musings
- Databases to distributed systems: every systems skill is merging into AI inference and hardware — Abhishekcur · 2026-11-02
- Limitless analogy: smarter people may gain far more from superintelligence — McDonaghMatthew · 2026-11-02
- Hunting for the most concrete "curing cancer with AI" roadmap of 2025 — iskander · 2026-11-02
- Avoiding AI brain rot: Miles Cranmer's lecture on learning with AI — _ScottCondron · 2026-11-02
- Yacine mocks "autonomous research" hype as Naam argues RSI loop needs 5-10x boost to run away — yacineMTB · 2026-11-02
- Vitalik's "post-AI cryptography": pseudonyms are dead, per-action anonymity is in — devanshmehta · 2026-11-02