4 ms·

Rebuttal to that essay by the Machine Intelligence Research Institute (AI safety group): https://intelligence.org/2017/01/13/response-to-ceglowski-on-superinte

• crooked-v 9y ago
  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.
• SiVal 9y ago
  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".