site banner

Friday Fun Thread for September 29, 2023

Be advised: this thread is not for serious in-depth discussion of weighty topics (we have a link for that), this thread is not for anything Culture War related. This thread is for Fun. You got jokes? Share 'em. You got silly questions? Ask 'em.

1
Jump in the discussion.

No email address required.

I would not accept slavery and outlawing criticism of government.

But because those are just generally bad and probably won't stop the asteroid, not because unilateralism is bad, so I don't see what's wrong with the original premise of 'AI seems to eat other ethical concerns on a large scale'.

The problem is that some people would sincerely believe (or lie) that stuff like that is necessary.

Some policy being murderously stupid does not mean that it will not be enacted. Rejecting blatantly evil and unethical policies is far from foolproof but provides some coordination against really terrible ones.

(do I need to provide examples of tragically idiotic and evil programs enacted by governments?)

Okay. Some people will lie and make up foreign threats of aggression to justify wars and military buildups. Yet we still have a military. And that does mean that we, often, have thousands-millions of unnecessary deaths for unnecessary wars. But it still beats not having a military and then getting conquered by whoever feels like it. And during wars, you do have to suspend freedom of expression and freedom of movement to win, and all sorts of underhanded things that are bad during peacetime. And that is often done in unnecessary ways, but it's still done.

This is still analogous to the asteroid situation. It's worth making sure the asteroid isn't something someone made up, or a distributed mistake. But asteroids exist sometimes, and if they do exist it's worth putting your all into not having the asteroid hit.

And AI is worse than the asteroid in this case, because good outcomes aren't 'everything continues as normal', but 'AI everywhere and everything but good somehow', and nobody's really worked that last part out yet.

To be clear, I didn't actually advocate anywhere for 'the government forcing everyone to work on AI'. I just said that it seems to eclipse most other ethical concerns. I simply don't see why that stops being true even if it makes coordinating harder. Asteroids also make coordinating harder, but as before, they still exist.

Okay. Some people will lie and make up foreign threats of aggression to justify wars and military buildups. Yet we still have a military.

I am not a a pacifist, I do not concern having military as unethical. But if someone starts going "we totally must murder all X and it is important and ethics should be ignored" (I know, it almost never is so blatant) then I hope that I would not support that.

"eclipse most other ethical concerns" is getting much better than "does seem to eat every other ethical concern"

But I would be still highly suspicious about such claims. "goal justifies all methods" repeatedly caused severe issues. And I am not convinced that ignoring most/all ethical concerns would actually help solving AI problems.

And during wars, you do have to suspend freedom of expression and freedom of movement to win

not always and not fully

It only seems to if you accept some pretty specific premises -- all of which seem fantasitical to the population at large.

It's like saying 'the prospect of burning in Hell seems to eat other consequential concerns' -- it sure does! But only if you believe in Hell.

The population at large thought fantastical the telegraph, cars, oil, artillery, fighter jets, electricity, nuclear bombs, computers, and neural nets, a century before they arrived. They still came, and clever people predicted them.

Along with the flying cars, interplanetary (manned) spaceships, and other things that clever people predicted -- I honestly think that the popularity of science fiction and AI Doom scenarios in the rationalist community are not a coincidence. But 'would make a great science fiction story' is not a good predictor for 'is likely to happen IRL'.

Flying cars exist! There are multiple brands! They're just not very practical relative to cars/trains/planes. Link.

Interplanetary manned spaceships are clearly technically possible.

And yet...

'AGI capable of ruling/destroying/whatevering the world' is not even at the 'clearly technically possible' point -- why would one worry more about it than flying cars?

Because it seems quite likely that it'll happen in the future?

For the same reason that small mammals like ferrets existing implies the existence of humans a few hundred million years later, even though ferrets didn't rule much.

What do you expect the world to look like in ten thousand years? Where were we three hundred years ago? Will change slow? Will it really just be awesome forever, just because? How does that work?