The Gradient: Perspectives on AI cover image

Jeremie Harris: Realistic Alignment and AI Policy

The Gradient: Perspectives on AI

CHAPTER

The Power of Smiley Faces

There's this kind of interesting anthropic paper, which you probably seen as well on. They found that you could just ask a language model to do app in a slightly less biased way. But I don't particularly take that as addressing the more central issue,. Look, at some point, the representations contained in the models are going to be such that it has an incentive to seek power and execute against that incentive. That's why I think the smiley face is actually a meaningful like, like this is a disa deception, not exception but it gives us false comfort.

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