B40 Reading Circle - On Feral Library Card Catalogs

About

How do LLMs actually function? How should we understand them? What are the real risks associated with their adoption and how does our understanding of what they are influence how we choose to respond to them?

Link to essay: https://www.science.org/stoken/author-tokens/ST-2495/full?referrer=https%3A%2F%2Fbactra.org%2F Shorter version (if you're in a rush): https://bactra.org/weblog/feral-library-card-catalogs.html#ngopnikism

Questions:

  1. How does the essay define "cultural technology"? Do you agree with this framework?
  2. Do you believe large language models capable of "thought" or something functionally analogous to it?
  3. What are the biggest risks of AI? How do we actually address those risks?
  4. How much does the "lossy" nature of large language models limit its usefulness to human beings?
Discussion