hdunne@amethyst.UUCP (10/08/87)
Since there's been some discussion on the Turing test, I'd like to point out that the Turing test is not sacred. A lot of AI people disagree that it proves anything. If I can't figure out how a stage magician pulls a rabbit out of a hat, does that mean magic exists? Besides, a lot of people have been fooled by relatively simple programs such as Elisa, which wasn't even designed to pass the Turing test but merely to be a parody of a certain school of psychology. - Hugh Dunne
mojo@reed.UUCP (10/10/87)
Someone asked about the Turing test; given the direction of discussion here,
it seems like it might be a good thing to have posted. In a nutshell, the
Turing test consists of a setup with a human, a computer, and an observer
hooked up to both of them through some sort of electronic network (or some
such--the point is the observer can't tell which is which). The computer and
the human both try to persuade the observer that they are the human; if the
computer can do it, it's intelligent.
I don't agree with this in an idealistic sense, but it sounds like a pretty
good acid test.
"A variable is a thing which varies." -- my Poli Sci prof
Nathan Tenny
{tektronix || ogcvax}!reed!mojo