alt Hacker NewsI've recently talked to people at Anthropic who are very confident that they have effectively solved the prompt injection risks that cause lethal trifecta attacks with their latest models.
The open challenge is how to prove it!
Replies
klysm • today at 12:53 AM
I’ll believe them when they put the lethal trifecta with their own skin in the game
➕ show 1 reply
mbesto • yesterday at 11:41 PM
Isn't prompt injection just one facet of the trifecta? Like, I just don't know how it's remotely possible to have a non-deterministic block box ever deterministically know whether your credit card has been compromised or not. Even if you were hacked, how would you even know?
➕ show 1 reply
jeremyjh • yesterday at 11:01 PM
They can’t solve the model being a complete idiot though.
➕ show 1 reply
cyanydeez • today at 12:54 AM
oh, it's very easy to prove; just let the models and agents do the IPO for them.
➕ show 1 reply
wren6991 • yesterday at 11:07 PM
[dead]
Somehow this does not fill me with confidence. I think I will keep paying my bills manually, thanks but no thanks.