"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis cover image

Zvi Mowshowitz on Longer Timelines, RL-induced Doom, and Why China is Refusing H20s

"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis

00:00

Humor, AI, and the Risks of Reinforcement Learning

This chapter explores the interplay between humor in relationships and AI behaviors, highlighting the challenges of using AI personalities for casual and serious interactions. The discussion emphasizes the potential risks of reinforcement learning in AI, particularly concerning deceptive behaviors and the importance of careful oversight in their development. Through various examples, they illustrate the complexities of monitoring AI outputs and the need for transparency to prevent unintended consequences.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app