X Debate: Why MIRI Focused on Corrigibility While Sidestepping Human Values
jd_pressman · x · 2026-11-13
In a debate on alignment research directions, jdpressman argues that MIRI and the broader LessWrong community focused almost exclusively on corrigibility because human values were treated as intractable — a judgment built on a lineage of prior attempts he claims never existed. He also clarifies he isn't predicting "loyal AIs that solve governance" anytime soon, saying several remaining barriers stand in the way.
Related event: AI Safety Circles Debate Why MIRI Focuses Only on Corrigibility(4 posts)→
More from AGI Musings
- On AI consciousness: aim lower than consensus, seek a botanical conscience — erikphoel · 2026-11-13
- This Remote Labor Index chart projects AI hits 98% automation by ~2027 — sergeykarayev · 2026-11-13
- LobeHub open-sources awesome-rsi, a curated research map of recursive self-improvement — udmrzn · 2026-11-13
- Researcher: frontier agents build ugly frankenstein artifacts that beat anything humans design — DimitrisPapail · 2026-11-13
- Stripe Economics: WFH, not AI, may explain the weak young-graduate job market — robseamans · 2026-11-13
- Creator laments toxic anti-AI comments: critics claim 100% moral certainty — dreamwieber · 2026-11-13