AI Consciousness Needs a Science-First International Effort, Not Bayesian Guesswork

erikphoel · x · 2026-09-29

In a thread, erikphoel argues the only real path to settling AI consciousness is to scientifically solve consciousness itself, via an international collaborative effort like the one Bicameral is trying to start — essentially a DeepMind/OpenAI-style lab for consciousness research. He sympathizes with Bayesian/joint-evidence approaches but calls them an illusion of progress: if all theories are pre-paradigmatic, their combined informativeness is tiny, and empirical evidence from AI itself is poorly controlled and largely uninformative.

Original post →

More from AGI Musings

AGI Musings channel →