Path: news.net.uni-c.dk!newsfeeds.net.uni-c.dk!newsfeed1.uni2.dk!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newshub2.home.com!news.home.com!news.mindspring.net!not-for-mail From: J Thomas Newsgroups: comp.ai.neural-nets,comp.lang.apl,comp.lang.awk,comp.lang.beta,comp.lang.cobol,comp.lang.dylan,comp.lang.forth Subject: Re: Einstein's Riddle Date: Wed, 14 Mar 2001 00:46:35 -0600 Organization: MindSpring Enterprises Lines: 47 Message-ID: <3AAF13CA.C7EA3113@ix.netcom.com> References: <3AACB567.A59B8497@Azonic.co.nz> <3AACE6CF.7F05484D@ieee.org> <0W8r6.178$fo5.14165@news.get2net.dk> <3AAD60F3.120F284A@ieee.org> <3AAE371A.2F9F596F@brazee.net> <98m43a$fe2$1@localhost.localdomain> <3AAEAD1A.BCDE11DB@ix.netcom.com> <98mugg$2mj$1@news.igs.net> NNTP-Posting-Host: 3f.27.04.7d Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Server-Date: 14 Mar 2001 06:45:20 GMT X-Mailer: Mozilla 4.76 [en] (Win98; U) X-Accept-Language: en Xref: news.net.uni-c.dk comp.ai.neural-nets:67565 comp.lang.apl:29390 comp.lang.awk:17156 comp.lang.beta:12758 comp.lang.cobol:102631 comp.lang.dylan:24179 comp.lang.forth:78568 donald tees wrote: > "J Thomas" wrote in message > > aph@redhat.invalid wrote: > > > In comp.lang.forth Howard Brazee wrote: > > > : But the trouble with defining whether or not we have AI is that > > > : there is no solid arrival point. > > > Sure there is: the Turing Test [1]. That's why it was invented. > > But the Turing Test only checks whether the program can imitate the > > particular forms of stupidity common to human beings. It doesn't > > work as an intelligence test. > Sure it does. If you're doing the Turing Test, and you ask what is 1355693147 * 25190678237 and you get a quick correct answer, you can conclude that it probably isn't human. One thing needed to pass the Turing Test is to make the kind of logic mistakes that humans make. The Turing Test isn't about a program that's good at finding solutions to problems, or a program that's good at redefining problems to make them easier to solve. The Turing Test is about a program that's good at imitating stupid humans. > If you take that line, then the only logical endpoint > is that there is no such thing as intelligence (which may be true). I doubt that there's a unitary intelligence. Different brains are good at solving different problems. We won't be ready to understand the intelligence of oak trees until we get a feel for what problems they have to solve. If understanding the problems of oak trees and their solutions turns out not to be in our repertoire then we may never notice their intelligence. The point of AI shouldn't be to imitate humans. We already have a lot of humans who're good at doing that. My thought is that since at present computers have as their sole ecological niche to serve humans, useful AI would involve predicting what humans will want well enough to be ready to give it to them when they first need it, before they think to ask for it.