4 ms·
The arrogance of the subtext that machine intelligence is nothing like human intelligence and therefore inherently inferior and nothing to be afraid of is more
by numinary1 11y ago
The arrogance of the subtext that machine intelligence is nothing like human intelligence and therefore inherently inferior and nothing to be afraid of is more disturbing than the prospect of artificial intelligence gone awry as it reminds us how far awry natural intelligence goes routinely.
- argonaut 11y agoI find it equally arrogant to believe that humans will achieve strong AI any time in the near future.
- cybertronic 11y agoThe Manhattan project or the Apollo program were even more arrogant.
- argonaut 11y agoSo you're agreeing with me that the person I was replying to was wrong?
- Rapzid 11y agoI'm not convinced we are any closer to creating human-level AI than we were before we invented computers..
- maaku 11y agoOkay. It was arrogant of Alexander the Great to think he could conquer Persia. It was arrogant of Kennedy/NASA to think they could go to the Moon within a decade. What's your point?
- TheLogothete 11y agoMachine intelligence is nothing like human intelligence and I thank the author for writing this article. It's a glorified calculator. An input-output program. It is not "intelligent". In fact, most research into artificial intelligence, in the sense that we think about in sci-fi is stopped. For more than decade, approaching on two. It is stopped because nobody got anywhere with it, despite lots of money and lots of people trying to get it off the ground. What is dubbed "artificial intelligence" nowadays is some statistical learning or other mathematical models which find optimal solutions. Its got nothing to do with intelligence. But it is easier to sell to executives.
- maaku 11y agoThe field of artificial general intelligence does exist. AI risk people for the most part aren't concerned over the latest machine learning buzz, but rather the potential of an AGI.
- TheLogothete 11y agoYes, that's what I was talking about. It does exist, in a much smaller form than in the 80s and 90s.
- argonaut 11y agoThere is no AGI research that actually produces anything that actually works. It's all about abstract concepts like value systems and the like, none of which is used for ML that produces real results.