Critics call out Anthropic and OpenAI for withholding interpretability research and real models
RileyRalmuto · x · 2026-10-05
An AI commentator argues the debate would move much faster if Anthropic shared its actual interpretability research and OpenAI shared its actual models — instead of the community being stuck on "what if" for decades. He half-jokes that both companies may be deliberately withholding to generate high-IQ training data from this exact flavor of discourse.
More from AGI Musings
- Living Science proposal: papers as living, updatable units of scientific knowledge — ChenhaoTan · 2026-10-05
- AI removed the friction: creation is now 'I have an idea, I ask for it' — signulll · 2026-10-05
- "OpenAI is run by X, Anthropic is run by Claude": AI joke goes viral — Miles_Brundage · 2026-10-05
- Elon Musk at YC AI Startup School on digital superintelligence and being useful — elonmusk · 2026-10-05
- Google researcher: the five stages of grief describe reactions to AI better than actual grief — moultano · 2026-10-05
- Even if ants were proven conscious, society wouldn't act on it — and that mirrors AI's consciousness dilemma — jessi_cata · 2026-10-05