LLMs Are Anthropomimetic, So the Anthropomorphic Lens Is Productive

dioscuri · x · 2026-08-29

In a discussion on anthropomorphizing AI, @dioscuri pushes back against a colleague's concerns, arguing the anthropomorphic lens is explanatorily productive for understanding AI behavior: LLMs are anthropomimetic systems, trained on human data, so human-like framings carry real explanatory weight.

Related event: Experts Debate Whether Anthropomorphizing AI Helps or Hinders Alignment(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →