Why RL Agents Raise Alignment Risks

Podcast episode for On Dwarkesh Patel’s Second Interview With Ilya Sutskever.

* 00:00 - Introduction

* 01:47 - Table of Contents

* 02:19 - Explaining Model Jaggedness

* 03:58 - Emotions and value functions

* 05:36 - What are we scaling?

* 06:47 - Why humans generalize better than models

* 08:10 - Straight-shooting superintelligence

* 09:57 - SSI’s model will learn from deployment

* 10:59 - Alignment

* 20:03 - “We are squarely an age of research company”

* 25:15 - Research taste

* 27:57 - Bonus Coverage: Dwarkesh Patel on AI Progress These Days

The Don’t Worry About the Vase Podcast is a listener-supported podcast. To receive new posts and support the cost of creation, consider becoming a free or paid subscriber.

https://open.substack.com/pub/thezvi/p/on-dwarkesh-patels-second-interview?utm_campaign=post-expanded-share&utm_medium=web