XPost: comp.ai.alife, comp.ai.games   
   From: joe@burgershack.com   
      
   Ray Dillinger wrote:   
   > wtkiii@hotmail.com wrote:   
   >   
   >>    Since people don't really understand how the mind works,   
   >>programming an AI is difficult and the past half century or so of   
   >>research has been unsuccessful.   For games though, there might be   
   >>little tricks for making the character appear intelligent.  I would   
   >>think along those lines and not waste time on AI for game characters.   
   >   
   >   
   > I don't think I agree that it's been "unsuccessful." We've learned a   
   > lot.   
   >   
   > We have better search techniques, better algorithms for machine   
   > vision, better methods of spatial representation, and scads and   
   > scads of useful and workable heuristics for finding "good"   
   > solutions to intractable problems - even if they're not necessarily   
   > the "best" solutions.   
   >   
   > We've learned a lot about language and parsing, a lot about efficient   
   > methods of doing mathematical regressions, about the mathematical   
   > properties of intractable problems, etc.   
   >   
   > We've been able to make a lot of successful applications, too: from   
   > chess and poker playing games to useful (though limited) natural-   
   > language interfaces and systems for searching vast amounts of   
   > NL text for writing about particular subjects, to expert systems   
   > that have proven incredibly useful especially in industrial   
   > control applications.  None of them is as smart as a person, nor   
   > smart in exactly the same *way* as a person, but they can do things   
   > that we need done, so they're useful.   
   >   
   > Various people have held forth goals and achieved them.  Various   
   > others have contented themselves with pointing at the goals that   
   > no one has achieved yet.  Every time something becomes an acceptable   
   > engineering technique, we quit calling it "AI".  But that doesn't   
   > change the fact that a lot of successful engineering techniques   
   > come from AI research.   
   >   
   > And people are still making new discoveries....   
   >   
   > 					Bear   
   >   
      
   If the quest to land 'a man on the moon' had been equally as (un)successful as   
   the quest to build an artificial system that exhibited human levels of   
   intelligence, all of our space flights would have:   
      
   - sat inert on the launch pad   
   - fizzled out or exploded somewhere in flight   
   - crash landed on asteroids and wrong planets   
   - stopped communicating   
   - been lost in space   
      
   Over and over and over...   
      
   Worse yet, each outcome would have occurred almost randomly, with no apparent   
   progress toward achieving the original mission except that we might know   
   something about where NOT to point the rocket.   
      
   IMO, the level of performance of today's AI systems is not measurably better   
   than it was 40 years ago.  Yes, we've discovered some interesting (or   
   entertaining) byproducts.  But if the Apollo missions had made a comparable   
   lack   
   of progress as has AI, nobody would be lauding the merits of its derivative   
   technology.  NASA would be rightly embarrassed by their zero percent success   
   rate and their unquantifiable rate of progress.  And like the good engineers   
   NASA is/was, they'd quietly go back to their drawing boards.   
      
   Unfortunately, I think AI still has most of its journey yet to go, and it's   
   still not clear whether the rocket is going in the right direction.   
      
       Randy   
      
   --   
   Randy Crawford   http://www.ruf.rice.edu/~rand   rand AT rice DOT edu   
      
   [ comp.ai is moderated.  To submit, just post and be patient, or if ]   
   [ that fails mail your article to , and ]   
   [ ask your news administrator to fix the problems with your system. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)