It's one thing to say "these vulnerabilities existed and could've been found before AI", that doesn't mean they could be exploited en masse at the speed and scale they are now. Which still for the moment at least means a huge net problem

Show Post

AI is also used to prevent fraud on a massive scale. It's just that right now the numbers suggest we have a lot more net fraud thanks to the availability of these tools.

Show Post

Yes, that's how vulnerabilities work. But the things that make people vulnerable to, say, an AI cloning their son's voice and begging for money because he's been kidnapped might be deemed pre-existing too. They just weren't easy to exploit before AI.

Show Post

exhausted by how you can apparently just say “open weights“ and “self hosted“ and then just wash your hands and never have to confront the fact that you don’t and can’t know what’s actually in the training data, or how it was obtained

Show Post

hauptsache ein munteres Ausrufezeichen!

Show Post