Skip to content

Comment on A gentle introduction to automated reasoningparent

Comments

How and why do LLMs get to eliminate symbolic/logic based reasoning?

I'm thinking about how LLMs like chatGPT did a complete end-run around the more historical attempts of formulating a grammar, translating English into some kind of logical form, using that to query a database, etc etc.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.