LLMs are weirdly good at math and coding, and that still feels surprising

paul_cal · x · 2026-07-21

The post highlights a recurring AI oddity: people are less impressed that computers are good at math than that LLMs are good at math and coding, even though that is a surprisingly specific and unusual capability profile for a computer system. The attached screenshot frames it as a joke, but the underlying point is about how strange the current division of labor between LLMs and traditional software really is.

Original post →

More from AGI Musings

AGI Musings channel →