Episode 255: Beloved Child of the House (Susanna Clarke's "Piranesi")
Very Bad Wizards
Chat GBT
Ruth Ben-Ghiat: I think a little moron canny, at least in my experience of it. The way these language learning models work is that they're literally just trying to predict the next word from the previous word. It's like if you ever played the game of auto completing on your text messages,. You know, you could make it say some really weird things. And so, it could be that that ends up being really dangerous and bad. But, but I find it harder to get into that mindset when I can see just see so clearly what happened.
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.
