From newshub.ccs.yorku.ca!torn!cs.utexas.edu!uunet!haven.umd.edu!darwin.sura.net!Sirius.dfn.de!unibwh.unibw-hamburg.de!unibwh!p_misiak Wed Sep 16 21:22:09 EDT 1992
Article 6818 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca comp.ai:4223 comp.robotics:2120 comp.ai.philosophy:6818
Newsgroups: comp.ai,comp.robotics,comp.ai.philosophy
Path: newshub.ccs.yorku.ca!torn!cs.utexas.edu!uunet!haven.umd.edu!darwin.sura.net!Sirius.dfn.de!unibwh.unibw-hamburg.de!unibwh!p_misiak
>From: p_misiak@unibwh.unibw-hamburg.de (Carlo Misiak)
Subject: Re: Turing Indistinguishability is a Scientific Criterion
In-Reply-To: sef@sef-pmax.slisp.cs.cmu.edu's message of 7 Sep 92 15:07:24 GMT
Message-ID: <P_MISIAK.92Sep8153708@grafix.unibwh.unibw-hamburg.de>
Sender: news@unibw-hamburg.de
Organization: University of Federal Armed Forces Hamburg
References: <Bu7rCD.CMG.1@cs.cmu.edu>
Date: Tue, 8 Sep 1992 15:37:08 GMT
Lines: 45

In article <Bu7rCD.CMG.1@cs.cmu.edu> sef@sef-pmax.slisp.cs.cmu.edu writes:


       From: ward@sun17.vlsi.uwaterloo.ca (Paul Ward)

       If I can push a nail into the wall with my own hands, then why would I build
       a hammer?

       If I can pull a plough, then why would I make a tractor?

       To get into the domain of computers; if I can perform 50 Gigaflops then
       why would I need a Cray?

       One does not make machines that merely duplicate human function - they must
       do it better.

   Nonsense!  We make machines all the time that do jobs that a person could
   do, but doesn't want to.  Consider automatic pilots, automatic door
   openers, automatic elevators, home dishwashers, or even dial-operated
   telephone exchanges.  One could argue that a human can do each of these
   tasks better than the equivalent machine, but the human would get bored,
   want to be paid a lot, go on strike, etc.  So a machine that does these
   tasks WELL ENOUGH is valuable, even if it's not nearly as good as a human.
   Then the humans can spend their time doing things that humans like better
   (or, in certain economic systems, they can starve).

   -- Scott

No nonsense ! If you evaluate performance with a function that keeps an eye on
costs ( as I understand is done in certain economic systems ), you will
probably come to the conclusion that any machine that actually is employed is
better than a human. BTW, I do not believe that a machine that will have a
chance to meet the TTT-criterion currently propagated by Harnad will ever be
built in order to merely duplicate human-function - no return on investment,
since there is the strong assumption that this kind of machine would like to
spend its time doing things that TTT-proven machines like better.



 
--
Carlo Misiak

*** All that we C or Scheme is but a mind in the machine *** (remember POE) ***



