News for hackers. Stuff that matters.
Re:The Last Invention of Man by apsec112 on Friday October 06, 2017 @07:37PM (#15421265)
Rebuttal to that essay by the Machine Intelligence Research Institute (AI safety group):

https://intelligence.org/2017/01/13/response-to-ceglowski-on...

Re:The Last Invention of Man by crooked-v on Friday October 06, 2017 @08:00PM (#15421360)
That write-up seems to mainly try to shrug off the "what is intelligence" issue by virtue of "maximization", while ignoring that when dealing with human intelligence, selecting for maximization of a single field gets you autistic savants, not Lex Luthor-grade manipulative geniuses.
Re:The Last Invention of Man by SiVal on Friday October 06, 2017 @09:15PM (#15421663)
My own rebuttal to that unimpressive Maciej argument is that he merely listed many ways in which something that seemed as though it ought to work didn't work out at all as expected, so we can safely assume that something as challenging as "superintelligence" won't work out either, and is therefore no danger.

But the argument that there are countless ways for plans to go awry and instead cause something unexpected to happen is hardly reassuring evidence that there is, therefore, "nothing to worry about".