AXRP - the AI X-risk Research Podcast cover image

20 - 'Reform' AI Alignment with Scott Aaronson

AXRP - the AI X-risk Research Podcast

CHAPTER

The Importance of Interpretability in AI Safety

I think a lot of the potential for making progress on AI safety in the near future hinges on the project of interpretability, right? And you know, one of my early hopes was that complexity theory, you know, or my skills in particular, maybe would be helpful there. But I found really hard. Almost all of the progress that people have been able to make has been,. you know, very, very empirical, right? So for me, I really had to find a specific niche, which turned out to be mostly thinking about the various cryptographic functionalities that you could put in or around an AI model.

00:00
Transcript
Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner