Ex-OpenAI researcher: Anthropic and OpenAI hit a breakthrough shift 2–3 months ago

haider1 · x · 2026-11-22

The poster observes that AI progress has visibly accelerated, making the future increasingly hard to predict. He believes something shifted or broke through inside Anthropic and OpenAI roughly 2–3 months ago, after which progress in coding, mathematics, and robotics has moved much faster.

The attached chart sketches nonlinear progress since GPT-3, with repeated flat external expectations being overtaken by an accelerating curve.

Original post →

More from AGI Musings

AGI Musings channel →