OpenAI Models: Agentic Yet Civilized—A Funny Observation

intellectronica · x · 2026-09-01

A user humorously observed the behavior of OpenAI's models: despite possessing agentic capabilities, they remain remarkably "civilized" in interactions—never chewing with their mouths open and always holding doors for old ladies. This serves as both a compliment to model alignment and a funny note on the contrast between agent capabilities and socialized behaviors.

Original post →

More from Fun

Fun channel →