Ex-Researcher Critiques Anthropic's Overconfidence and Lack of Open Debate

peterwildeford · x · 2026-08-12

AI safety researcher Jacob Hilton stated that Anthropic suffers from a collective overconfidence in its leadership and safety culture, alongside a lack of open debate. Commenting on this, Peter Wildeford noted his impression that OpenAI accurately estimates itself but underestimates AI, whereas Anthropic overestimates itself but correctly estimates AI.

Original post →

More from AGI Musings

AGI Musings channel →