mark_k: RSI is as poorly defined as AGI and will arrive gradually, not overnight

mark_k · x · 2026-11-19

Reacting to renewed RSI talk, markk argues "recursive self-improvement" is poorly defined like AGI: it spans many levels from mere AI-research assistance to full automation, will happen gradually rather than as a binary switch. The quoted post predicts Anthropic, OpenAI, xAI, Google and Meta will each reach RSI within a few years.

Original post →

More from AGI Musings

AGI Musings channel →