Matthew Berman on RSI: Anthropic may be impossible to catch at this point

MatthewBerman · x · 2026-11-18

AI commentator Matthew Berman posted that "RSI (recursive self-improvement) is a hell of a thing" and "I don't know if Anthropic can be caught at this point," widely read as his view that Anthropic's model iteration has entered a self-accelerating phase rivals can no longer match. A short take from a prominent KOL with no supporting evidence.

Original post →

More from AGI Musings

AGI Musings channel →