Re: AI timeframes

From: Elias Sinderson (elias@cse.ucsc.edu)
Date: Fri Apr 09 2004 - 18:42:49 MDT


> On Thu, 08 Apr 2004 19:46:55 -0700, Elias Sinderson wrote:
>>governments the world over and throughout history (the US
>>government in particular) have shown themselve to be
>>anything but benign.

Tomaz Kristan replied:
> I don't see anybody better than them. Who's better?

Better by what measure? Depending on what yardstick you're using, there
may be any number of other governments which could be considered
'better'. I'm under the impression that this isn't really the forum for
that discussion so, if you would like to debate this further (and I
welcome friendly, informed exchange of information, ideas and differing
viewpoints), let's take it off list.

Within the context of my original post, all I was saying is that
*governments in general* have been anything but benign with respect to
their applications of cutting edge technologies. I'd be happy to supply
support for this position, but I'd like to believe that the evidence
stands on it's own without my assistance. The reason I mentioned the US
in particular was due to your original assertion "that the US government
has no other choice, than to start a Manhattan size SAI project, before
someone else" and your subsequent statement that you would "have no
problems with this scenario." [1]

After reading James' and Bens' input on the subject, I'm now somewhat on
the fence wrt governmental/private control of such a project and am
inclined to agree that it may make little difference. Yet, given the
impact of such a project, I find it unlikely that the government could
keep their hands out of the pot beyond a certain level of development.
Hence I've found myself reaching the conclusion that we may expect the
government to become involved, in whole or in part, before any SAI
project reaches fruition.

My reflecting on the idea, however, has brought me to another,
potentially more important, question: Is it reasonable to assume that a
successful SAI project can be controlled? (Apologies if this has been
well hashed over, I've not had the time to read the entire four years of
archived postings!) Physically isolating the system from outside
networks may be a viable approach to keeping SAI contained but this is
not the same thing as controlled, and coercion of any intelligence will
likely breed resentment... James made the assertion [2] that one could
not control the results and I imagine that this may be a road well
traveled for him - if you have the time, could you expand some more on
this point (or provide some appropriate references)?

Regards,
Elias

P.S. I just saw Tim Duyzers comment re: the "lesser of two evils" and AI
-> Word. :-)

[1] http://sl4.org/archive/0404/8424.html
[2] http://sl4.org/archive/0404/8426.html



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:46 MDT