Google and Meta Are Back, RSI Is Talked About Openly, and Linear Extrapolation No Longer Works

haider1 · x · 2026-09-10

haider1 argues that regardless of whether we actually cross the human-level threshold, the signals are unmistakable: Google and Meta are making a comeback, recursive self-improvement (RSI) is now being discussed explicitly, and continual learning wouldn't surprise him anymore. His takeaway: things are moving too fast for normal extrapolation to work.

Original post →

More from AGI Musings

AGI Musings channel →