On self-consciousness, sentience claims about AI, and grounding in LLMs.
This interview is incredible. I have been studying referentiality for over 20 years now and this lecture connected a lot of dots. I love the resurgence of realism in the context of linguistic and cognitive scientific analysis of LLMs. Can you interview someone who can explore the process of vector representation of words? I want to know a lot more about that. So cool. Keep doing the good work.
I would love a follow up episode that goes into some of the questions you both touched on towards the end. There's a whole series you could don with Prof. Millière.
I really enjoyed this conversation!