Thats kinda the point, they’re talking about misaligned AI’s. Im worried about the misaligned humans in charge of them. They’re plenty dangerous because humans are operating them and creating them. Like a gun, not dangerous at all until someone picks it up with bad intentions, or lack of precautions.

I don’t think computers only do what you program them to do anymore.

Based on what? I have seen no evidence to the contrary

emergent behaviour and microscopic differences in huge amounts of parallel computations

Emergent behavior meaning what?

Which “microscopic” differences?

Hey, Socrates

look for “emergent behaviour in computers” and specifically the non-programmed behaviour

look for “floating point non determinism” to see how the same program can produce different results

keep in mind you asked me on what I base that I think computers don’t only do what they are programmed to do

You know, acting like a dickhead doesn’t really strengthen your argument.

I’ll speak to someone less pompous about this

yeah ask someone else

Not one of those so-called AI systems is an autonomous entity. Every one of these machine learning systems could be regulated, slowed down, or turned off. If a responsible human being was there to do it. Those CEOs simply refuse to take responsibility.