XPost: rec.arts.sf.written, rec.arts.sf.science   
   From: YourName@YourISP.com   
      
   In article , Gutless Umbrella   
   Carrying Sissy  wrote:   
      
   > Doc O'Leary  wrote in   
   > news:droleary-FC7285.11271224012014@news.eternal-september.org:   
   >   
   > > In article ,   
   > >  Greg Goss  wrote:   
   > >   
   > >> "Michael F. Stemper"  wrote:   
   > >>   
   > >> >On 01/23/2014 06:17 PM, Your Name wrote:   
   > >>   
   > >> >> There was an article in the newspaper not long ago about the   
   > >> >> experts working with robots etc. saying that getting to a   
   > >> >> point of having actual computer intelligence is stil decades   
   > >> >> away.   
   > >> >   
   > >> >It always has been. Kind of like fusion.   
   > >>   
   > >> Well, a lot of the skills that were defined as part of "AI" get   
   > >> redefined out once they're met.   
   > >>   
   > >> Expert Systems?   
   > >> Basic language translation?   
   > >> Grandmaster-level chess?   
   > >> Driving a car on public roads?   
   > >>   
   > >> Nope.  Those aren't AI after all.   
   > >   
   > > They'd be AI if they were solved with AI.  Instead, they are   
   > > solved by clever humans using computers as tools.  Things like   
   > > Watson and self-driving cars are, as currently implemented,   
   > > cheats.   
   >   
   > That doesn't really address Greg's point, which is shifting goal   
   > posts. If "xxx is part of AI" is changed once xxx is possible, then   
   > AI is, by definition, never, ever possible.   
   >   
   >   
   > > As I have posted in the AI newsgroups, you can easily set up a   
   > > standard for an AI driver.  Run this experiment: allow a person   
   > > to drive a car remotely using only a video camera, measuring   
   > > their performance as you reduce the resolution of said camera.   
   > > Then you pick a "minimum acceptable" point for their ability to   
   > > navigate the environment.  That's your Turing Test for AI   
   > > drivers, getting around with only a crappy web cam as your   
   > > input.   
   > >   
   > > Compare that to how Google and others are solving the problem   
   > > and it is easy to state unequivocally that AI is *not* being   
   > > used.   
   >   
   > That, too, seems a matter of redefining AI to keep it from ever   
   > happening.   
      
   Intelligence isn't the ability to do only one thing well, so a computer   
   being able to play chess at "Grand-master-level" is not "artificial   
   intelligence" ... it's simply a computer that has been programmed to   
   play chess well, something that's been posibile since the days of the   
   Apple II, VIC20, and before.   
      
   A computer being able to play chess at "Grand-master-level" may or may   
   ont be considered one tiny step on the way to "artificial   
   intelligence".   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)