Skip to content

Comment on The security paradox of local LLMsparent

Comments

It involved an LLM reading a reddit comment or github comment and immediately executing.

right, so you shouldn't give the LLM access to execute arbitrary commands without review.

AboutSource Built by g1lg1l

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