From: Byrne Hobart (sometimesfunnyalwaysright@gmail.com)
Date: Thu Jul 12 2007 - 10:05:03 MDT
I might be defining AI away from being meaningful, but to what extent could
an AI project consider itself successful if it was neither implemented in a
human brain nor in a computer but as a collection of memes, institutions,
and protocols that yielded humanlike results from the contributions of
numerous human and nonhuman entities? If so, the stock market qualifies: it
passes the Turing test in the specialized domain of "How much cash would you
hand over for the following set of future cash flows?"
To a significant extent, a single human acting in the stock market is not a
thinker as much as a compiler, taking basic information on fundamentals or
price movements and outputting buy or sell signals in a pretty
easily-automated (
http://www.newyorker.com/reporting/2007/07/02/070702fa_fact_cassidy?printable=true)
way. If the market is, roughly, lots of automata, with feedback-driven
biases, is it a solid analogy for AI.
I presume this has been considered before -- can anyone point me to a good
discussion/essay/anything on AI being implemented with institutions rather
than machines?
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:58 MDT