A Potentially Historic AI Breakthrough

When it comes to AI safety, the news lately has been a little bit dreary.

It goes beyond the title with direct discussion of detectors, nick, detector, including: Or at least, I should say, the headlines have been.

1. 01We've been living through a new wave of panic and concern, driven in part by the reality that models are getting more capable and closer to what many view as key inflection points
2. 02I, like many people, noticed it when OpenAI's Rune posted, when Nick told me that he had some interpretability breakthroughs, he was, uh, underselling it
3. 03Interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building