News for hackers. Stuff that matters.
Re:Artificial intelligence risk research by andreyf on Saturday December 12, 2015 @03:46PM (#10724071)
I think it will be obvious once HGI is sufficiently understood, in a way that will not need more than a paragraph, if that.

The fear and idea of somehow creating an HGI without understanding how to control it came from The Terminator. It's pure fantasy.

Re:Artificial intelligence risk research by jerf on Saturday December 12, 2015 @08:42PM (#10725056)
That doesn't even rise to the level of proof-by-disbelief; it's proof by hypothetical future disbelief. You've given no more reason to be confident in your hypothetical than the concerns about hypothetical self-improving AIs.

You're failing to sell this argument to someone generally inclined to agree with you about the outcome.

Re:Artificial intelligence risk research by andreyf on Saturday December 12, 2015 @09:57PM (#10725221)
What is proof-by-disbelief? I'm not trying to prove that HGI is impossible, nor do I think it's impossible.

However, until one proposes a concrete methodology for modeling the world which includes the method that is doing the modeling (self-awareness), a decision making process, and a way to take action autonomously and without intervention, it's very hard to talk generally about preventing catastrophic bugs.

Re:Artificial intelligence risk research by eli_gottlieb on Saturday December 12, 2015 @05:10PM (#10724357)
Just because we will understand, in principle, how to control it, doesn't mean we'll engineer that control particularly well.
Re:Artificial intelligence risk research by robertk on Saturday December 12, 2015 @05:00PM (#10724331)
Factually disputed. The concerns raised in this area originated in the 90s, conceived prior to release of the first Terminator film.
Re:Artificial intelligence risk research by andreyf on Saturday December 12, 2015 @09:47PM (#10725200)
The first Terminator was released in 1984. The idea existed before it, of course.