Manning: AI absorbed 2,000 years of knowledge but still learns far worse than humans

chrmanning · x · 2026-09-17

Stanford professor Christopher Manning argues today's LLMs look impressive largely because they've absorbed 2,000+ years of human knowledge into one giant neural net, but that's exploiting human learning rather than discovering knowledge. The secret of human intelligence is learning speed — a few examples suffice for humans, where current AI still falls far short.

Original post →

More from AGI Musings

AGI Musings channel →