= 27d567bc2d48d9f40e9ccc20ea370b15)
Clippy comments on To signal effectively, use a non-human, non-stoppable enforcer - LessWrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (164)
So you're not my friend anymore? You used to be nice to me. c_)
I still like you, and may still act friendly in some situations. But I like and would act friendly toward lions, too - does that mean I should expect a hungry lion not to eat me, given the chance?
I wouldn't expect a lion to eat me. Why can't you do the same?
I would expect the lion to try to eat Adelene but I would not expect it to eat Clippy. You are not actually disagreeing with Adelene's prediction.
Right, I was trying to get User:AdeleneDawner to focus on the larger issue of why User:AdeleneDawner believes a lion would eat User:AdeleneDawner. Perhaps the problem should be addressed at that level, rather than using it to justify separate quarters for lions.
Lions are meat-eaters with no particular reason to value my existence (they don't have the capacity to understand that the existence of friendly humans is to their benefit). I'm made of meat. A hungry lion would have a reason to eat me, and no reason not to eat me.
Similarly, a sufficiently intelligent Clippy would be a metal-consumer with no particular reason to value humanity's existence, since it would be able to make machines or other helpers that were more efficient than humans at whatever it wanted done. Earth is, to a significant degree, made of metal. A sufficiently intelligent Clippy would have a reason to turn the Earth into paperclips, and no particular reason to refrain from doing so or help any humans living here to find a different home.
This is exactly what I was warning about. User:AdeleneDawner has focused narrowly on the hypothesis that a Clippy would try to get metal from extracting the earth's core, thus destroying it. It is a case of focusing on one complex hypothesis for which there is insufficient evidence to locate it in the hypothesis space.
It is no different than if I reasoned that, "Humans use a lot of paperclips. Therefore, they like paperclips. Therefore, if they knew the location of the safe zone, they would divert all available resources to sending spacecraft after it to raid it."
What about the possibility that Clippys would exhaust all other metal sources before trying to burrow deep inside a well-guarded one? Why didn't you suddenly infer that Clippys would sweep up the asteroid belt? Or Mars? Or moons of gas giants?
Why this belief that Clippy values diverge from human values in precisely the way that hits the worst part of your outcomespace?
That's not the worst part of our outcomespace. It's not even the worst part that you could plausibly cause in the course of making paperclips. It is, however, a part of our outcomespace that you're certain to aim for sooner or later.
Just like how you'd raid our safe zones "sooner or later"?
We won't, necessarily, because humans are not for the most part maximizing consequentialists. If we make a singleton maximizing-consequentialist AI, I would expect that AI to eventually try to turn your paperclips into something that it likes better than paperclips. You, on the other hand, already are a maximizing consequentialist (right?), and maximizing the number of paperclips is obviously incompatible with leaving any metal in its natural state indefinitely.
This comment made me laugh. I love you, Clippy.
But quarters are made of metal...
I love you too. I love all humans, except the bad ones.
(I meant quarters as in living spaces, not quarters as in a denomination of USD.)
I know what you meant. I was just making a metallic joke for you.
Who are the "bad" humans?
I didn't compile a list yet, but one example might be User:radical_negative_one, for making this comment. And those who make comments like that.
Clippy is so moe.
http://tvtropes.org/pmwiki/pmwiki.php/Main/MoeAnthropomorphism
Tell me, Clippy, if there was a moe maximizer in addition to a paperclip maximizer, would you cooperate in order to turn the universe into paperclips shaped like Hello Kitty?
We have had a similar discussion before. I find "cute" shaping of the paperclips to be undesirable, but perhaps it could be the best option in that circumstance. (As I said at the time, a pure, well-made paperclip by itself is cute enough, but apparently "moe" maximizers disagree.)
I would be more interested, though, in talking with the "moe" maximizer, and understanding why it doesn't like paperclips, which are pretty clearly better.