Taking AI Seriously: Borretti traces how programmer AI advice keeps retreating

zetalyrae · x · 2026-11-21

Fernando Borretti's long-form essay Taking AI Seriously charts how programmers' advice on integrating AI keeps retreating as capabilities advance — from "tab completion needs human repo-wide context" to "agents produce slop, review everything" to today's "code was never the hard part." He likens it to a losing state's war propaganda, with victory tales creeping ever closer to the capital. The core failure mode, he argues, is not taking AI seriously: the term "AGI-pilled" conflates short timelines (weakly evidenced) with transformative consequences (easy to believe), and serious thinking requires separating the two.

Original post →

More from AGI Musings

AGI Musings channel →