Neither Nor Podcast Notes - lydgate/mindmeld GitHub Wiki

Episode 4: Intuition versus Rationality or why Language Models make Bad Philosophers

Part 4 in a series of interviews on the book I'm working on, Neither/Nor. In this episode, Isabela Granic and I explore whether AI language models can ever be helpful for developing a useful philosophy for our current times. We explore the myriad ways human thought differs from the processes and outputs of language models. The philosopher, Arthur Schopenhauer, figures centrally throughout our discussion, as we apply his insights about intuition versus rationality. I'm particularly compelled by how Shopenhauer writes about perception, which is immediate and the basis of intuition, versus rationality, which is encapsulated in language, sequential, and unfolds in time. We end with considering the dangers inherent in language, which reifies our concepts and keeps us separate from our lived experience.

Links we discussed:

Clerestory by Bryan Kam • Infrequent updates at Substack • All my work plus exclusive content at Patreon

My website. Occasional updates on my Substack. Frequent updates on my Patreon. See also the InterIntellect for events.

Notes if you want to edit:

  • I have a bunch of ums and uhs when I'm particularly struggling to understand or cohere / summarize what I am understanding. Sections that are particularly troublesome which you could delete entirely without missing any real meaning in the episode:
  • 11:17 - 13:00
  • 16:44 - 16:58
  • 47:35 - 48:25
  • there's loud noise in the background, I think it's traffic on my side; particularly loud around 50:28 for about 5 min