Robot Safety Testing
Pieter discusses the challenges of ensuring robot safety, particularly in comparison to human driving tests. He highlights the lack of standardized testing for robots and raises concerns about how to validate improvements in AI systems. The conversation also touches on the complexity of finding policies that promote kindness versus exploitation in AI and human interactions.In this clip
From this podcast

Lex Fridman Podcast
Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10
Related Questions