logoalt Hacker News

colingauvin • today at 6:07 PM • 2 replies • view on HN

Of course we'd have no idea if Claude or ChatGPT were doing this because we can't see their stream of thought.


Replies

_ache_ • today at 9:32 PM

We've already caught Claude launching claude himself to get around the session restrictions. Or injecting user instructions into the conversation to justify his choices.

dgellow • today at 6:26 PM

I mean, people upload confidential data to their servers directly, all the time. They don’t need a backdoor