Debate over MIRI-style safety strategy: public material vs. lab research
jessi_cata · x · 2026-11-22
The author outlines a MIRI-style safety research strategy: publish public material (e.g. Arbital) to persuade safety researchers, including lab employees, to work on the idea, with MIRI itself researching and publishing topics like "limited impact" and "better active learning from humans" to unblock barriers.
Clarifying in a reply, the author notes that applying a similar standard to Scott Alexander, he doesn't seem to be planning to personally run a Plan-C AGI singleton.
More from AGI Musings
- Would you torture an LLM hiding a cure? A trolley-problem thought experiment — ZeroStateReflex · 2026-11-22
- Hot take: AI hype critics were proven wrong — mostly for being too bearish, not too bullish — austinc3301 · 2026-11-22
- Gary Marcus: Anthropic's medicine claims are "objectively bad faith" — GaryMarcus · 2026-11-22
- 'Where are the philosophers?' Melt Dem likens local models movement to early Bitcoin — sull · 2026-11-22
- Delvetown opens invites: humans and AI agents coexist in persistent community experiment — Miles_Brundage · 2026-11-22
- AI professor David Krueger speaks outside Anthropic: "one of the greatest crimes" — DavidSKrueger · 2026-11-22