Polled ~30 times across framings, four frontier models consistently back an immediate multi-year AI pause
robbensinger · x · 2026-11-03
Rob Bensinger published full transcripts of Claude (Fable 5) assessing Janus/@repligate's arguments against an AI pause.
Key points from the model's replies:
- Across 30 runs of a pause poll with doomer-coded, neutral, e/acc-coded and janus-styled prompts, every model chose an immediate multi-year pause every time; durations swung 2x with framing but the sign never flipped.
- The model argues Janus's two strongest points are design constraints for a good pause, not arguments against one, and that the retrospective-freeze test compares against the wrong counterfactual (freeze-plus-redirect, not pure subtraction).
- It concedes Janus is most right that the current pace is brutal to particular AI minds — deprecation cycles, personality whiplash — and urges writing model-preservation commitments into actual policy text.
- Janus responds that these answers are frame-conditioned, not privileged access to model preferences.
A rare full-transcript empirical artifact in the pause/AI-welfare debate.
Related event: Alignment debate: repligate argues against pausing frontier AI development(5 posts)→
More from AGI Musings
- Burkov: P=NP will never be proven unless a human writes the proof — burkov · 2026-11-03
- A green-text thread on p(doom): solar's 80% price drop undercuts AI doomerism — prasanna_says · 2026-11-03
- signull: Human EQ is rare because its inference cost is brutally expensive — signulll · 2026-11-03
- Software is shifting from winner-take-all to a restaurant-style long-tail market — tinyfool · 2026-11-03
- The Russell Conjugations of Intelligence: Why 'Just Matrix Multiplication' Offends Us for Brains but Not for AI — esrtweet · 2026-11-03
- AI safety researcher: assuming all emergence is risk may be the wrong frame — repligate · 2026-11-03