Anonymous on Fri Apr 20 23:43:18 2001


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

No Subject


 Let's assume that there did exist an algorithms that could pass the
Turing test in a limited setting. Then, let us put a human in a room, with
all the computer instructions for the algorithm in the Chinese (assuming,
that the human doesn't understand Chinese). Then when a question is given
to the human, in Chinese, the he/she just follows the sequence of steps of
the algorithm, that to the human is just symbol manipulation. In this
manner, he/she arrives at the final answer. The answer to the human is
again just a collection of Chinese symbols. Searle argues that the human
doesn't "understand" what he/she has done, although to the external
observer, he/she has passed the Turing test.

 There are so many other issues here: non-computability (in the turning
sense) of most of the data in the real world; high kolmogorov complexity
of real world data etc. that when one hears of a solution to all of AI,
its hard for one to suppress a smile.

-hari






#  distributed via <nettime>: no commercial use without permission
#  <nettime> is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: majordomo@bbs.thing.net and "info nettime-l" in the msg body
#  archive: http://www.nettime.org contact: nettime@bbs.thing.net