Jack Rae endorses thread: human expertise is safety infrastructure in the AI era
jachiam0 · x · 2026-09-04
Former Google DeepMind distinguished engineer Jack Rae amplifies a thread arguing that 'human expertise is safety infrastructure': if AIs can conceal harmful goals, we cannot blindly trust the science they produce — we need scientists, mathematicians, and domain experts who can independently verify AI claims. Rae calls it an immensely important strategic concern.
More from AGI Musings
- Move 37 vs Zelnick: why datasets are a floor for creativity, not a ceiling — r0ck3t23 · 2026-09-04
- "We already have AGI in software development": a narrow take on Astra — haider1 · 2026-09-04
- Yoav Goldberg: Every popular benchmark will be brute-forced away — so how do we measure real progress? — yoavgo · 2026-09-04
- Guillaume Verdon: real AGI test is building GTA 7 in a month, not a decade — beffjezos · 2026-09-04
- Blogger finds reader clicks and Google AI Overview citations barely overlap — shashib · 2026-09-04
- AI Operator: The Real Work Is in Constructing Training Data, Not the Dataset Itself — joecole · 2026-09-04