home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!edcastle!aiai!jeff
- From: jeff@aiai.ed.ac.uk (Jeff Dalton)
- Newsgroups: comp.ai
- Subject: Re: Reply to Ginsberg, Blenko, Whedon RE Loebner Test
- Message-ID: <7897@skye.ed.ac.uk>
- Date: 10 Nov 92 18:30:25 GMT
- References: <1992Nov6.202339.19045@cs.ucla.edu>
- Sender: news@aiai.ed.ac.uk
- Organization: AIAI, University of Edinburgh, Scotland
- Lines: 16
-
- In article <1992Nov6.202339.19045@cs.ucla.edu> colby@oahu.cs.ucla.edu (Kenneth Colby) writes:
- >No program could pass a completely unrestricted Turing test anymore than a
- >human could. No single human has all the knowledge of all the domains
- >of all the humans in the world.
-
- Now wait a minute! The idea of the TT is "can it be distinguished
- from a human?". Humans therefore ought to be able to pass it.
- Moreover, since humans do not have all the world's knowledge at
- their command, machines whose performance was indistinguishable
- from a human's wouldn't have to either.
-
- (By "indistinguishable" I mean only that, based only on what comes
- over the teletype, our best efforts to reliably pick (think Star Trek)
- out the real humans fail.)
-
- -- jd
-