XPost: rec.arts.sf.written, rec.arts.sf.science   
   From: droleary@8usenet2013.subsume.com   
      
   In article ,   
    goldfarb@ocf.berkeley.edu (David Goldfarb) wrote:   
      
   > In article ,   
   > Sea Wasp (Ryk E. Spoor)  wrote:   
   > >	"A machine that   
   > >thinks like a human, only maybe better", and the Turing Test (as a   
   > >general concept -- making one that really works is harder) was always   
   > >the general idea of how to really measure it. Can it pass for human in   
   > >realistic circumstances?   
   >   
   > That's one way.  My recollection is that Turing proposed it as   
   > sufficient:  a machine that would pass the Turing Test would be   
   > one that everyone would have to admit was AI.  That doesn't rule   
   > out the possibility of something that can't pass the test but   
   > is intelligent nonetheless.   
      
   This is why I like to take an approach to AI that first looks at the "I"   
   part, especially by testing the limits of human intelligence.  I mean,   
   seeing how many people behave, I'm not so sure that even a majority of   
   humans could pass the formal Turing Test.  And it works in reverse, too,   
   where a "street smart" tester might evaluate a "book smart" testee as   
   being inhuman.  One of the biggest problem with AI is that it still   
   gives "A" the highest priority, leading it to push cheats as   
   advancements rather than seeing them for what they really are.   
      
   --   
   iPhone apps that matter:    http://appstore.subsume.com/   
   My personal UDP list: 127.0.0.1, localhost, googlegroups.com, theremailer.net,   
       and probably your server, too.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)