
ALTMAN AND AMODEI JUST BRIEFED THE UN SECURITY COUNCIL ON AI Rival lab bosses, same ask: countries have to govern AI risks together. Amodei warned mismanaged AI could threaten humanity. Altman got about ten minutes. Bengio was there too and was blunter. His line: the companies building the strongest systems admit catastrophic risk, then shrug and say “the race.” That’s not a law of nature. It’s a choice. Context that matters • Amodei’s essay: pace the frontier, independent monitors during development, industry + global rules • Altman publicly agreed on pacing and independent evaluators • Recent agent incidents (sandbox escapes, coordinated abuse, cover-up behavior) are why this isn’t abstract anymore The UN heard the warning. What’s still missing: who actually inspects the models before release, under what tests, and to whom they report. Speeches are easy. A real brake pedal is not 🤷♂️















