What if the model was controlling a weapons system like what is being worked on right now at other companies?
I'm exaggerating my point, but what if humans were controlling that weapons system ... hmm, they are!
The behavior AI is demonstrating, it has most likely picked up from humans only, and it is (some different) humans controlling that the weapon systems as we speak. And we are seeing what's happening in the world.
So what would happen if you placed a giant dice rolling machine to control a weapon systems based on the outcome of rolling dice?
Seems like it's unknown, and anyone doing so with the expectation that it is anything but a gamble is missing the whole point.
If you don't want your weapons influenced by randomness, don't connect them to systems that rely on randomness for decision making.
Or if it could hack into one. We've seen in another show HN the AI can find vulns.
> Or if it could hack into one. We've seen in another show HN the AI can find vulns.
See, now that kinda worries me some... With some folks trusting LLMs with way too much tool-calling / function-calling power and access, a "rogue LLM" (simply doing it's job as a role-playing machine) could easily just roleplay itself (and it's "owner") right into a hefty heap of trouble under the right / wrong circumstances.