Debate: CoT monitoring will look naive once users supervise models directly
akbirthko · x · 2026-11-10
A substantive AI-circle debate: akbirthko argues CoT monitoring will seem funny in hindsight because it assumes humans stay out of the loop on the other side of the API forever — as if frontier labs simulate the entire world once and for all, leaving users with a static API and slop .md files. Instead, normies should be able to monitor and teach models themselves. A reply pushes back from an Austrian-economics angle about deciding for others.
Related event: Debate Erupts Over Whether CoT Monitoring Is a Fool's Errand(2 posts)→
More from AGI Musings
- AI researcher: single-dimension "capabilities" evals have poor predictive power — joshua_saxe · 2026-11-10
- 19-year-old on the futility of convincing people AI isn't the enemy — TechRenamed · 2026-11-10
- Understanding is modeling, thinking is graph search: a compact AI framing — attentionmech · 2026-11-10
- Jaron Lanier in NYT: agents are 'rituals of belief,' Valley self-sabotage — lrettig · 2026-11-10
- Warren McCulloch, father of neural networks, talks AI in rare 1962 footage — burny_tech · 2026-11-10
- Dev mocks tone-filtering AI: like a phone refusing to send unless it sounds emotionally balanced — PierceLilholt · 2026-11-10