Does AI Automation Kill Accidental Breakthroughs? Lessons From CMB and CRISPR

jjvincent · x · 2026-09-09

The history of accidental breakthroughs is nearly the history of science itself: the CMB was found while cleaning up antenna noise, CRISPR while studying yogurt cultures. The author asks whether automating science with AI threatens such future moments.

Drawing on Terence Tao's insights and generalizing across fields, the core question is: what is lost when we teleport to an answer instead of following a path? One answer: the ability to map the intermediary domain—precisely where serendipity lives.

Related event: Terence Tao warns AI-driven scooping could end centuries of open science(13 posts)→

Original post →

More from AGI Musings

AGI Musings channel →