Robotics AI safety veteran: people who believe AI risks take them dead seriously
amankhan · x · 2026-09-10
An author with hands-on AI safety/interpretability/evals experience in robotics says people who genuinely believe in AI dangers take the work extremely seriously, and flippant mockery is almost offensive. He still questions how p-doom figures like ">10%" are derived, arguing economic inequality from AI is the bigger risk.
More from AGI Musings
- AI slowdown talk is a cover for compute costs, argues viral critic — CtrlAltDwayne · 2026-09-10
- Jen Zhu Scott questions how the 'AI will kill us all' narrative took off without explanation — kevinnbass · 2026-09-10
- Tiny teams will look much bigger as AI gives them department-level output — alexmacgregor__ · 2026-09-10
- Models are overqualified too: frontier LLMs add endless complexity and cost to simple tasks — rachel_l_woods · 2026-09-10
- Cofounder of Major US Compute Firm Backs AI Regulation Despite Cost to His Company — MxMnr · 2026-09-10
- Four days talking to normal New Yorkers: 'People literally hate AI' — and a cultural theory why — Linahuaa · 2026-09-10