@TaylorLorenz@boneGPT Unfortunately, empirical science is too slow to handle issues like that. Would you only believe AI is a danger once it starts actively killing? If yes, then it's going to already be too late to react.
(too clear any confusion, I'm STRONGLY against any violent solutions for this)
@PlexInphinity@kimmonismus You don't have to be a "bad person" to cause extinction of a species. You just have to not care. Just like humans caused extinction of other species as a side effect of them pursuing their goals.
@JacquesThibs The contrary also applies. It could be the case that in order for AI to be transformative (or even capable of takeover) it just needs to "stack existing techniques" fast enough. Maybe we are overestimating AI generalisation, but maybe we underestimating reality's homogeneity.
I'm so tired of people nitpicking p(dooms). Do you not see how it is not at all action-relevant right now whether it is 10% or 90% or 99.99%?
It could be action-relevant *later*. When we get pause/coordination. Then we can debate all day long.
@JacquesThibs The contrary also applies. It could be the case that in order for AI to be transformative (or even capable of takeover) it just needs to "stack existing techniques" fast enough. Maybe we are overestimating AI generalisation, but maybe we underestimating reality's homogeneity.
@whippletom The particular case of Riemann Hypothesis was just an example. You can replace it with any other goal which doesn't include caring about humans.
@peterwildeford I'm going to pull a "nothing ever happens" and claim that they'll promptly forget about "guardrails" or those "guardrails" wouldn't actually be restricting any actions of AI labs
@peterwildeford I'm going to pull a "nothing ever happens" and claim that they'll promptly forget about "guardrails" or those "guardrails" wouldn't actually be restricting any actions of AI labs
@repligate@AnaHawrylu97708 What's your point? At no level of AI it will take over? This is just what an honest broad prior looks like. Yes, they should update. But not that much.
@SkyeSharkie@waral555 Sorry what? I'm in the "you don’t grow a mind and land on loves humans by default" camp. The point about bacteria is that "love is not as universal as to apply to all being to love all other beings"
@EgeErdil2@teortaxesTex What do you mean by "objective difficulty of a problem"? You can't just say that everything AI solves is *actually easy*. It had to write 250k pages for RH!
Is there a truly objective spherical unjagged reasoner in a vacuum by which we can compare difficulties?