MIRI's Nate Soares: I can now only say the *public* ChatGPT won't kill you tomorrow
connoraxiotes · x · 2026-09-26
Nate Soares, researcher at MIRI, poked fun at how his own AI-risk argument has shifted. Where he once could flatly say "ChatGPT isn't going to wake up and try to kill you tomorrow," he now has to hedge with "the public version of ChatGPT isn't going to try to kill you tomorrow." The added qualifier, he notes, says a lot about how the frontier risk picture has changed in just a few years — a terse piece of dark humor from inside the AI safety camp about how fast model capabilities are outpacing old reassurances.
More from AGI Musings
- Prof counts down to ICLR 2027 deadline with a final prompt for Claude — CSProfKGD · 2026-09-26
- New Book 'Perspectives on Machine Consciousness' Out, Featuring Daniel Hulme's 'Spinning Wheel' Theory — cccalum · 2026-09-26
- CAPTCHAs Are Broken for the Agent Era: Time for Agent Digital Identities — RachelVT42 · 2026-09-26
- "Secure by laziness" is dead: Martin Casado on why agents break old security assumptions — yacineMTB · 2026-09-26
- Chamath: AI is booming but productivity isn't — ROI analysis becomes a necessity — gaganghotra_ · 2026-09-26
- Anthropic Hired Philosophers Debating Whether AI Could Justifiably Turn Against Humans — vishalmisra · 2026-09-26