We overestimate the gap between average-human difficulty and human limits — and misjudge AI

S_Conradi · x · 2026-11-22

AndreaM argues humans tend to overestimate the gap between 'average difficulty for humans' and 'the limits of human capabilities,' so we downplay machines that match average humans and are surprised only by those hitting human limits — a systematic misjudgment of AI ability.

Original post →

More from AGI Musings

AGI Musings channel →