From: Stathis Papaioannou (stathisp@gmail.com)
Date: Wed Jun 04 2008 - 04:45:29 MDT
2008/6/4 John K Clark <johnkclark@fastmail.fm>:
> On Monday Wednesday and Friday the "friendly" AI people say we can't use
> anthropomorphic reasoning because the AI's motivations are completely
> alien to us. On Tuesday Thursday and Saturday they say we can understand
> an AI so well we can be certain he will remain our slave until the end
> of time regardless of how much he grows in intelligence knowledge and
> power. And on Sunday they rest.
There isn't necessarily a contradiction here. The motivations of an AI
could be completely alien or at odds with our own, yet completely
understandable. There is also no necessary contradiction in the idea
of superintelligent AI which remains predictable to us, since its
goals may make it predictable. For example, if the AI is born with the
aim of adding two given numbers together and then turning itself off,
it will do just that, quickly and efficiently. Of course, the AI will
always be able to surprise and bamboozle us if it tries, but it has to
try.
-- Stathis Papaioannou
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:03 MDT