OpenAI's 10-year-old research call: coding agents and AI security shipped, AI detection forgotten

badphilosopher · x · 2026-09-13

badphilosopher revisits a research call OpenAI published 10 years ago with four directions: detecting AI use in the wild, building coding agents, AI for security, and agents. The latter three have largely materialized at scale; the first — detecting AI use — has been all but forgotten. The author argues safety effectiveness is questionable if we can't detect accidental or intentional AI breakout.

Original post →

More from AGI Musings

AGI Musings channel →