Alignment debate: should safety researchers pre-plan for the Plan C scenario?
davidmanheim · x · 2026-11-01
A safety-community thread debates whether to plan for "Plan C" (strong interventions like a global pause) alongside conventional alignment. A quoted AIFP forecast notes all five forecasters plus the AI implicitly believed humanity's survival would likelier come from the Plan C supercategory than a global pause. quetzalrainbow flags a probabilistic sleight of hand—P(survival|Plan C) ≠ P(survival|do(Plan C))—while davidmanheim argues the case hinges on priors about prosaic alignment and evidence of a robust basin of safe behavior.
Related event: AI Safety Community Debates Whether to Plan for 'Plan C' Scenarios(3 posts)→
More from AGI Musings
- Musk predicts money won't matter by 2036 as AI and robots oversupply goods and services — CodeByPoonam · 2026-11-01
- Furman hails 'Pyramids, Diamonds, and Oscillations' paper modeling AI freezing junior hiring — soumitrashukla9 · 2026-11-01
- Economist Jason Furman: AI could do my first job at the CEA, but not my last — soumitrashukla9 · 2026-11-01
- Waymo reports 95% fewer serious-injury crashes than human drivers — how safe is safe enough? — PeterDiamandis · 2026-11-01
- Innovators keep only 2.2% of value they create — economist's case for socializing AGI gains — Dr_Atoosa · 2026-11-01
- Alignment Researcher Challenges "Fully Aligned": To Whose Interests, Exactly? — wfithian · 2026-11-01