From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!bronze!chalmers Tue Jan 21 09:26:34 EST 1992
Article 2821 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!bronze!chalmers
>From: chalmers@bronze.ucs.indiana.edu (David Chalmers)
Subject: Combinatorial explosion
Message-ID: <1992Jan17.082347.15868@bronze.ucs.indiana.edu>
Organization: Indiana University
Date: Fri, 17 Jan 92 08:23:47 GMT
Lines: 28

In my opinion, combinatorial explosion may be one of the most
overrated "problems" with AI.  One frequently hears claim along
the lines that AI would be trivial if it weren't for this problem
(someone said that the essence of intelligence is beating
combinatorial explosion).  Hence I'd like to propose a 
thought-experiment.  Imagine that we have computers that run
infinitely fast -- or at least, say, 10^10^10 times faster than
current computers, with arbitrarily large amounts of memory.  The
question is: what would the state of AI be?

i.e. what kinds of things would be trivial, what would still be
hard?  Would full intelligence have been achieved by now (assuming
35 years effort)?  Obviously some things would be utterly trivial,
chess-playing for example.  However, I suspect that most
aspects of intelligence wouldn't be vastly more tractable than
they are now (obviously any search space could be searched
instantly, but you still need good criteria for the search).  It
seems to me that the real bottleneck is in developing algorithms
themselves, not in speed.  I may be wrong about this, however, so
I'd be very interested to see people's thoughts about what would
and wouldn't be vastly simplified by these conditions.  (Perception?
Language processing?  Various kinds of reasoning?  Action in the
world?  Learning?)

-- 
Dave Chalmers                            (dave@cogsci.indiana.edu)      
Center for Research on Concepts and Cognition, Indiana University.
"It is not the least charm of a theory that it is refutable."


