Why Replacing Feynman Is Harder Than Replacing Einstein in the AI Era
CatAstro_Piyush · x · 2026-11-16
A thought-provoking thread argues that AI swarms can hill-climb complex research problems, but within their capability limits every hill looks flat—they don't feel human cognitive friction. Left to teach, AI either over-explains the obvious or steps over conceptual chasms where humans get stuck. A great human teacher calibrates in real time to a student's confusion. Hence replacing Feynman is harder than replacing Einstein—and with an avalanche of AI-generated discoveries coming, we shouldn't accept a world where our own understanding is sidelined.
More from AGI Musings
- Do Frontier Weights Deprecate? The Inverse-U Debate on Context vs. Internalization — agarwl_ · 2026-11-16
- Noah Smith: Both TypeSafe AI Founders, Now Worth Billions, Spent Years 'Bumming Around' — paulnovosad · 2026-11-16
- Bindu Reddy: China lacks Blackwells, US lacks power — an AI slowdown looks inevitable — bindureddy · 2026-11-16
- The jagged frontier: fully automated coding, but still writing key docs by hand — Zergylord · 2026-11-16
- Top mathematician Tsimerman says he's quit pure math to work on AI safety — tak3sh8 · 2026-11-16
- Critics Say Effective Altruism Spends Too Little on AI Risk Messaging — AndyMasley · 2026-11-16