Hacker News new | ask | show | jobs
by jqpabc123 56 days ago
Our analysis shows that current LLMs are unreliable delegates:

Who knew that a tool that relies on probability could make such a mess?