2026-07-28 03:22 UTC
@muddle@infosec.exchange @vrek@mastodon.social @lzg@mastodon.social It's even worse than that. "Guardrails" for LLMs consist of requests, in natural language, to the LLM, as part of the so-called 'system prompt'. Things like "IMPORTANT: DO NOT ADD ***ANY*** COMMENTS unless asked". It's a fucking joke.
Replies (2)
-
@muddle@infosec.exchange 2026-07-28 03:24
@conniptions@mastodon.social @vrek@mastodon.social @lzg@mastodon.social Have you seen the leaked Claude code? It's the same stupid shit all the layers down.
-
@vrek@mastodon.social 2026-07-28 03:26
@conniptions@mastodon.social @muddle@infosec.exchange @lzg@mastodon.social exactly, I want to say say "drive down route 80 till exit 36" without saying "do not drive off a cliff. Do not hit a tree, do not drive into the ocean" it's easier for me to just drive