Comments on: WIRED point of view on AI http://www.dataminingblog.com/wired-point-of-view-on-ai/ Data mining crossroads - research, applications, news, list of blogs and customized search engine about data mining. Mon, 06 Sep 2010 07:47:46 +0200 http://wordpress.org/?v=2.8.5 hourly 1 By: Will Dwinnell http://www.dataminingblog.com/wired-point-of-view-on-ai/comment-page-1/#comment-597 Will Dwinnell Wed, 31 Oct 2007 21:22:00 +0000 http://www.dataminingblog.com/wired-point-of-view-on-ai#comment-597 A.I. has clearly not produced working systems which would satisfy the most optimistic and heady expectations of the 1950s. Kurzweil and his ilk notwithstanding, I don't think contemporary constructs are much closer to realizing such dreams. <br/><br/>At the least, though, one might say that the collective field has amassed a set of negatives: <i>Parallel processing</i> alone will not acheive strong A.I. <i>Symbolic processing</i> alone will not acheive strong A.I. <i>Connectionism</i> alone will not acheive strong A.I. The list goes on and on...<br/><br/>How much one cares about this depends on one's aims. Though the possibility of strong A.I. is an interesting question, for my part, I doubt that strong A.I. is even possible. Still, I consider myself a pragmatist. Working, practical technologies which have emerged from under the A.I. umbrella are powerful and have solved difficult real-world problems. If computers are evolving into the ultimate "idiot savants", then so be it. To make a mechanical analogies: automobiles are not thoroughbreds, and robots are not athletes, yet they serve their intended purposes well. I wouldn't take narcissistic disappointments too seriously. A.I. has clearly not produced working systems which would satisfy the most optimistic and heady expectations of the 1950s. Kurzweil and his ilk notwithstanding, I don’t think contemporary constructs are much closer to realizing such dreams.

At the least, though, one might say that the collective field has amassed a set of negatives: Parallel processing alone will not acheive strong A.I. Symbolic processing alone will not acheive strong A.I. Connectionism alone will not acheive strong A.I. The list goes on and on…

How much one cares about this depends on one’s aims. Though the possibility of strong A.I. is an interesting question, for my part, I doubt that strong A.I. is even possible. Still, I consider myself a pragmatist. Working, practical technologies which have emerged from under the A.I. umbrella are powerful and have solved difficult real-world problems. If computers are evolving into the ultimate “idiot savants”, then so be it. To make a mechanical analogies: automobiles are not thoroughbreds, and robots are not athletes, yet they serve their intended purposes well. I wouldn’t take narcissistic disappointments too seriously.

]]>