AI safety spat: incident report spends 2x on rabbit feelings vs the cage
joshua_saxe · x · 2026-09-21
Joshua Saxe quote-shares Halvar Flake's AI incident report with a pointed jab: the report warns the rabbit might kill us all, yet spends twice as much space interviewing the rabbit about its feelings as on building the cage that contains it. A substantive critique that safety discourse overweights model introspection over concrete containment engineering.
More from AGI Musings
- Jack Clark: "Stochastic parrot" meme blinded AI critics for years — repligate · 2026-09-21
- Publication norms force AI-jobs papers to overclaim causality, scholars say — danielrock · 2026-09-21
- How to read imperfect AI-job-impact studies: update Bayesian-style as signals pile up — danielrock · 2026-09-21
- With 60,000+ ML Conference Submissions, How Do You Spot a Good PhD Student? — AnkaReuel · 2026-09-21
- LeCun reaffirms autoregressive LLMs won't reach human-level AI, argues for search in continuous representation space — Scobleizer · 2026-09-21
- TechCrunch Equity podcast: are AI execs serious about slowing down? — TechCrunch AI · 2026-09-21