Episode 67: Ellie Pavlick on the Cutting Edge of AI

Does ChatGPT really understand anything? Or is it faking it? Hell, do we understand how it works? Ellie Pavlick joins Vasant Dhar in episode 67 of Brave New World to share her thoughts on where AI has reached — and where it’s going.

Useful resources:
1. Ellie Pavlick at Brown and Google Scholar.
2. Symbols and grounding in large language models — Ellie Pavlick.
3. Sam Bowman on ChatGPT & Controlling AI — Episode 58 of Brave New World.
4. The Nature of Intelligence — Episode 7 of Brave New World (w Yann le Cunn).
5. Music, Mind, and Meaning — Marvin Minsky.
6. Emergent Abilities of Large Language Models — Jason Wei et al.
7. Hugging Face.
8. EleutherAI.

Check out Vasant Dhar’s newsletter on Substack. Subscription is free!