3 ms·
Bruce Schneier put it well: "Willison’s insight was that this isn’t just a filtering problem; it’s architectural. There is no privilege separation, and there i
by fzeindl 10mo ago
Bruce Schneier put it well:
"Willison’s insight was that this isn’t just a filtering problem; it’s architectural. There is no privilege separation, and there is no separation between the data and control paths. The very mechanism that makes modern AI powerful - treating all inputs uniformly - is what makes it vulnerable. The security challenges we face today are structural consequences of using AI for everything."
- https://www.schneier.com/crypto-gram/archives/2025/1115.html#cg4 https://www.schneier.com/crypto-gram/archives/2025/1115.html...
- CuriouslyC 10mo agoAttributing that to Simon when people have been writing articles about that for the last year and a half doesn't seem fair. Simon gave that view visibility, because he's got a pulpit.
- 6LLvveMx2koXfwn 10mo agoHe referenced Simon's article from September the 12th 2022
- flir 10mo agoLonger, surely? (Though I don't have any evidence I can point to). It's in-band signalling. Same problem DTMF, SS5, etc. had. I would have expected the issue to be intuitvely obvious to anyone who's heard of a blue box? (LLMs are unreliable oracles. They don't need to be fixed, they need their outputs tested against reality. Call it "don't trust, verify").