Skip to content

Comment on LLMs and self-referentialityparent

Comments

Fact checking is harder. A RAG can only check what's in the database, so you have to know what to know beforehand.

A global database of facts would make easier for AI to stay factual

Also ironically it would also make it easier to align AI to do things like consistently censor or distort some political facts

AboutSource Built by g1lg1l

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