Human Go skill improved after open-source AI, not AlphaGo: seeing moves isn't enough

thesephist · x · 2026-09-04

An interesting observation: human Go playing strength seems to have improved not right after AlphaGo beat top players, but after open-source Go engines were released.

The author's explanation: just watching AI moves wasn't enough for humans to learn — open-source versions let players look under the hood and actually understand the strategy. A thought-provoking data point on how open-sourcing AI enables real human learning.

Original post →

More from AGI Musings

AGI Musings channel →