mje at mich.com
Mon Feb 22 07:51:44 EST 1999
Malcolm McMahon wrote:
> What we've actually seen is the phylosophers move the goalposts of
> "Intelligence" each time AI advanced. Playing chess used to be regarded
> as an excellent test of intelligence until computers starting to beat
> people at it. The Turing Test started to look a little shakey when Eliza
> came along and had to be tightened up.
The problem with this argument is that what you describe was never the case. Chess
has long been realized to be a problem with a solution- not a game. Being such, it
only awaited fast enough computers. If we had much, much faster computers, chess
could be solved through a brute-force solution, trying every possible outcome.
And the "Turing test" has never been a proof of intelligence. As you note,
Weitzbaum showed that a very small and crude program like "Eliza" could fool a lot
of people without anything that would be called intelligence.
Michael Edelman http://www.mich.com/~mje
Telescope guide: http://www.mich.com/~mje/scope.html
Folding Kayaks: http://www.mich.com/~mje/kayak.html
customer at email.usps.gov consumerline at ftc.gov
admin at loopback $LOGIN at localhost $LOGNAME at localhost $USER at localhost
$USER@$HOST -h1024 at localhost root at mailloop.com
president at whitehouse.gov
More information about the Neur-sci
Send comments to us at biosci-help [At] net.bio.net