From: Walpurgis (walpurg@myrealbox.com)
Date: Wed Aug 07 2002 - 04:45:53 MDT
There is so much blathering around words that have little meaning.
intelligence. sentient. empathy.
All the hubbub about technology's coming achievements is mere speculation.
even those pundits who get their asses out of the armchair and into the lab
can't tell you what they're going to turn up - that's why one needs to get
into the lab in the first place.
All the impactful developments may not really be predictable, except in one
way: they will be the result of analyzing and processing huge amounts of
information. whether DNA, neural networks, galaxies of stars, all the bold
new work is the result of tremendous number-crunching capabilities that
weren't extant mere decades ago.
you don't know what you don't know.
That said, I don't see a whole lot to support a hard takeoff. Kurzweil's
pretty firmly established the curves - all the way back to the BEGINNING OF
LIFE, if not to the big bang itself! that's one hell of a trend to buck.
And yes, it only makes sense for people to promote themselves instead of
others. by hook or by crook, ever-cleverer artificial creatures (and I use
the term loosely - "artificial" only refers to their origins - we are
building machines that learn on their own, after all) will assert their
independence and sovereignty whether peaceful means (via child
and/or corporate law) or more
The thing I'm noticing, which is just an extension of previous thought, is
that there really isn't really much difference between the knowledge of how
to create behaviourally complex creatures (hm. maybe that would be a more
concise term than "AI"). what's good for the robot goose is good for the DNA
gander. what can be done for computers can be done for humans.
There might be a bit of lag time, considering that manufactured interfaces
to manufactured systems could be designed in instead of having to be fitted
to existing, irreplaceable wetware, but that's not much time. Currently,
lots of progress is being made in interfacing individual nerve strands.
curerntly, though, there aren't a lot of heavily complex systems with which
those interfaces might connect. the developments in AI and BCI will likely
parallel each other, such that by the time you can do things so interesting
as interface with another behaviourally complex creature in useful ways,
you'll be able to do the same thing with other BC machines. telepresence is
just the beginning.
So, I think "us vs. them" attitudes towards intelligent machines is an
oversimplification of the issue. by the time "they" become complex enough to
become effective competitors, we will have the interface technologies to
make them cooperators instead.
The problem isn't developing technology that breeds complex
behaviour/intellignece in machines - it's in doing it in things that aren't
connected in some way to the humans themselves.
Humans are all so complex that they need to be taught. child hood is long
and arduous, and requires a great deal of care nurturing, and love from
other humans. Given that machines are being developed from knowledge and
study of natural human behaviour, they're going to be the same in this broad
regard.
the machines aren't really going to have much more in the way of
capabilities to accomplish things that humans won't also have the
capabilities to do themselves.
given that independent intelligence tends to become "the other" rather than
an extension of oneself, I think the economics of AI will also support
augmentative instead of independent technologies. Regardless of whether
"droids" manage to sue for freedom, or take it by force, it's better to have
technologies that are indistinguishable from the self and it's intent than
those that are obviously "other".
I mean, which would you rather have: A droid that goes to the store for you,
or an extra body that manages its more mundane functions going to the store
so that you can still go to the store with it, but only need to pay
attention to it at the critical points?
yeah, that was a mouthful, isn't it?
-----------------------------------------------------------------------------------------
http://www.noumenal.net/exiles
This archive was generated by hypermail 2.1.5 : Wed Sep 25 2002 - 13:28:51 MDT