Hacker News new | ask | show | jobs
by joseda-hg 316 days ago
The problem is the non deterministic part

An LLM can generate many valid reasons for doing X, including the right one, but you can't know which one of those it will give you