

"Using AI to murder with no oversight or >accountability is also fine, but only when it’s not >Americans.”
I read the whole statement. That was not what they said. They said it might be a future application of the technology, but right now the technology isn’t ready or safe enough for this.
Before LLMs, there were already machine learning algorithms and software to do automatic targeting. I don’t understand why we would even want to employ these tools trained to generically do anything okayish vs a highly specialized tool for this kind of application. The most exciting and useful applications of AI are highly specialized and trained applications: breast cancer screening, alpha fold, automatic targeting, etc. Why the fuck would you push to use Claude for this anyway?!

What do you think the drones are for?