I find most speculative discourse on AI really bothersome.

"but what if we create a machine and it doesn't value things that are good for people?"

we already created machines that don't value things that are good for people. They are called corporations, families, and states, and most of the people using that kind of rhetoric are active supporters of them

@tidsguiden It's great that you are lucky enough to have one of the few families that actively undermine the vast power they could wield over you. Most are not youth liberationist though. Most tend to be very authoritarian and obsessed with "disciplining" and "controlling" children rather then letting them explore and learn. There are so many parents who think it is okay to hit children even though they would never tolerate that kind of behavior towards adults (despite those adults being less defenseless than children)


No, I do legit count myself lucky.

And I'm not gonna, I do seem like an outlier.


@forAll52 "No no, you don't understand.

What if we create a machine and it doesn't value things that are good for people and it affects ME??"

I will admit though, the number of ways that AI could end human life before we realize what's happening does put it in a category all of its own. The gray goo scenario in particular is pretty frightening.

@sandrockcstm I would have thought that was more boringly called "greenhouse gas emissions"

@forAll52 It's closer to the borg more than anything. Basically you'd have an extinction event in a matter of weeks by an unstoppable, invisible force that literally devours organic life.

Even if we did stop it the damage would be catastrophic.

@sandrockcstm A lot of the worst case scenarios about AI's depend on them being able to have almost complete knowledge about how the world works. Basically they implicitly assume P=NP, something that is not settled math yet.

@forAll52 Right, and as our good friend @dankwraith often points out, AI are pretty dumb and can't tell the difference between an orange and an orange umbrella most of the time.

The question is one of crossing the tipping point. Because of how fast computers are, when we reach the point of NP, it could be too late.

It's a pandora's box. We don't know what'll happen.

@sandrockcstm @dankwraith It might very well be impossible to ever reach NP. certainly modern cryptographers and everyone who uses the internet or online banking pretty much is making that bet by default anyways

@forAll52 @dankwraith Yeah, we can't spend too much time obsessing about it. At some point we all have to live our lives. But it's good to stay informed and be aware of the possibility, if for no other reason than to preempt it by calling for ethical restraint in AI development.

@sandrockcstm @dankwraith Well, it certainly shouldn't happen under capitalism. Mostly because even the dumb ai we currently have is being used for actively bad purposes

@forAll52 so corporations are the real paperclip maximizers? hmm that scans.

Sign in to participate in the conversation
Serenity Laboratories

Welcome to the Serenity Laboratories public Mastodon instance. This is a stable instance with high standards of curation and moderation. Please read those before registering. The short version is: No Nazis. No Fascists. No bigotry. Listen and be excellent to other people.