LLMs can’t really evaluate things. They’re far too suggestible and can always be... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

yunwal on April 9, 2025 | parent | context | favorite | on: The Agent2Agent Protocol (A2A)

LLMs can’t really evaluate things. They’re far too suggestible and can always be broken with the right prompt no matter how many layers you apply.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact