Newsgroups: comp.ai,comp.ai.philosophy,comp.ai.games,comp.ai.nat-lang
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.sprintlink.net!howland.reston.ans.net!agate!library.ucla.edu!csulb.edu!csus.edu!netcom.com!jqb
From: jqb@netcom.com (Jim Balter)
Subject: Re: Loebner Prize $2000 and a medal
Message-ID: <jqbD6ss4B.Fss@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
References: <3ls4nh$qkn@hopper.acm.org> <vlsi_libD6Kx7q.BGo@netcom.com> <jqbD6LA71.8Dp@netcom.com> <3lvi06$icc@mycroft.rand.org>
Date: Mon, 10 Apr 1995 02:26:35 GMT
Lines: 25
Sender: jqb@netcom15.netcom.com
Xref: glinda.oz.cs.cmu.edu comp.ai:28933 comp.ai.philosophy:26721 comp.ai.games:879 comp.ai.nat-lang:3194

In article <3lvi06$icc@mycroft.rand.org>, Jim Gillogly  <jim@acm.org> wrote:
>In article <jqbD6LA71.8Dp@netcom.com>, Jim Balter <jqb@netcom.com> wrote:
>>To the degree that the point of the TT is to separate "intelligence" from
>>it's implementation, this goes the wrong direction.  Yes, AI programs
>>*can* do 2D processing, but is that *necessary* for intelligence?
>>Loebner's new rules display a willful ignorance of the whole point of the
>>exercise.
>
>Spot on.  This reminds me of the sequential redefinition of "success" in
>chess programming during the 60s and 70s -- the AI critics were happy to
>move the intelligence milestones further whenever they got close.
>Programs could be considered intelligent with respect to the chess task when
>
>- they could beat a human.  No, I meant when
>- they could beat a USCF-rated human.  No, I meant when
>- they could beat <me>.  [Greenblatt program arrives here.]  No, I meant when
>- they could beat a master.  No, I meant when
>- they could beat a grandmaster.  No, I meant
>- the world champion.  In a match.  At tournament time controls.  No, I meant
>- chess-playing ability isn't really germane to AI at all.

But chess playing ability is just a matter of TECHnology, Jim. :-)
-- 
<J Q B>

