The Gradient: Perspectives on AI cover image

Jeremie Harris: Realistic Alignment and AI Policy

The Gradient: Perspectives on AI

CHAPTER

The Paradigm Shift: How AI Is a Dangerously Creative Solution

I think there have been studies on how reinforcement learning from human feedback can decrease risk of hallucination, but also decreases some capabilities that these models have. And so sometimes what you want is something that is just more if you're willing to use the term creative but just kind of better at putting things together that you might not have thought about before. But again, this is just another case of there's something fundamentally different here. 100% I think this directly connects to risk in the sense that you simply don't get the intelligence without the risk.

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