Skip to content

Comment on The security paradox of local LLMs

Comments

If you are writing and deploying code without reviewing it then yeah, but all of this should be caught easily in the review step. No way my eyes are going to glaze over some random header logic, “eval”, or anything that looks obfuscated.

I don’t really think this matters at all in the local vs frontier model discussion.

AboutSource Built by g1lg1l

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