Double-Header Holiday Podcast on AI & Auto Safety

Holiday weekend listening: a two-parter with my friends from the Center for Auto Safety

As the US takes a long weekend for the Independence Day holiday, here is a double-header of discussions with my friends at the Center for Auto Safety on AI and auto safety. Each episode is about a half hour long.

Part 1: https://rss.com/podcasts/autosafety/2087377/

A general discussion of AI and auto safety. We touch on Embodied AI, the benefits and challenges of end-to-end machine learning, LLMs faking sincerity, the ELIZA effect, and an illustrative personal story about why naming your kid a “Jr.” is not doing them any favors.

Part 2: https://rss.com/podcasts/autosafety/2095084/

A theme from my upcoming new book is that you need to have literacy in four areas to have any hope of understanding automated vehicle safety: safety engineering, cybersecurity, machine learning, and human/computer interaction. I also discuss what I think about using LLMs for safety, issues with driving safety beyond avoiding crashes, and other topics related to robot drivers.