Hacker News new | ask | show | jobs
by k8tte 3986 days ago
one week Hawking warns about skynet-style AI. next week he insist of us go hunting it. good troll?
4 comments

It strikes me that the two are very different things. ET != AI.
Funny you say that:

The Dominant Life Form in the Cosmos Is Probably Superintelligent Robots

“If they were interested in us, we probably wouldn’t be here,” said Schneider. “My gut feeling is their goals and incentives are so different from ours, they’re not going to want to contact us.”

That’s a welcome divergence from Steven Hawking’s claim that advanced aliens might be nomads, looking to strip resources from whatever planets they can, and that all efforts to contact said aliens may end in our own demise.

“I’d have to agree with Susan on them not being interested in us at all,” Shostak said. We're just too simplistic, too irrelevant. “You don’t spend a whole lot of time hanging out reading books with your goldfish. On the other hand, you don’t really want to kill the goldfish, either.”

http://motherboard.vice.com/read/the-dominant-life-form-in-t...

EDIT - HN discussion for that article, 212 days ago: https://news.ycombinator.com/item?id=8773778

Hah. I suppose it depends how you classify "Artificial Intelligence". It was my understanding that Hawking was concerned about humans creating skynet-style AI ourselves, rather than discovering AI created by an Extra-terrestrial.
In just a few hundred years, human civilization will be expanding at something on the order of the speed of light, unless we manage to blow ourselves up first.

How is that not interesting?

SAI != ET.

The idea of SAI is that it progresses at a significantly faster rate than biological processes can. For example, if humans and chimps were on a "staircase of evolution" and were one step apart; SAI would be one step above us a mere hour after it is turned on. After 2 hours it would be multiple steps above us. So the theory goes, anyway.

This means that evil aliens would be unlikely more dangerous than evil SAI.

Not that he's suggesting we announce our presence to aliens, merely look for them.

A minor note on word usage here: usually the word "theory" is reserved for explanations with a good track record rather than wild speculation.
A timescale of hours is not required for most of the AI-doom arguments; if it took a year, that would have a similar outcome.
It's definitely not a requirement, it's just an example of how it could rapidly progress. Within the scope of the argument 1 hour is as plausible as 1 year - the main point is that it happens shockingly fast and continues to happen even faster.
Humans being humans, one year is not nearly as shocking, which could contribute to people actually considering the possibility.
thanks for pointing this out, it made sense in my head in the lines of: super civilization creates AI, AI outlives civilization, humans find trace of super civilization, eventually establish contact with the AI
Well, they might not be the same thing or they might - or perhaps more likely of all, some kind of hybrid society with AI and natural intelligences.
He wants us to to look for ET's, but not reveals ourselves to them: to see but not be seen. So just like in the AI question, his stance is to be cautious.
I think Hawking would ultimately like ET's help in the coming war against AI.
Either that or early warning?