From: Eugene.Leitl@lrz.uni-muenchen.de
Date: Sun May 20 2001 - 08:19:17 MDT
Anders Sandberg wrote:
>
> Building houses with nano is most likely best done with convergent
> assembly. Grow the basic building baterials in vats, then assemble them
Do you still need to build houses when you have nano? Such capabilities
as described do not arise overnight. A few tango rounds of the world
dancing with their precursors, and their precursors precursors might
make a lot of difference cumulatively. We're habitually waffling about
shrinking prediction horizonts here, but we sometimes have trouble to
eat our own dogfood.
I've just written a private email to a guy off wear-hard, trying to sketch
out the next 30 years in computing, and I realized the particular set of
scenarios described was starting to unravel in front of my eyes. The possible
future trajectory funnel is fanning up exponentially, or even
superexponentially. I still sent it, but what I should have said instead
was: "next 30 years in computing? Wish I had a clue". It's all heavy
haze, the farther you look, the less you can tell for certain.
That's a Nissan at the neighbour's garage, that's probably a refinery
over there, it's a bit hard to tell, and there... Um, yeah that could
be a mountain. Or a cloud. Whatever.
We're far from being omniscient, and the system noise present at each
step plus divergence drive the funneling up of future trajectories.
I think it's really more constructive and more fun to predict the future
by inventing it. Unfortunately, few people would wish to pump megabucks
into somethiang with a very uncertain ROI, decade-delayed at that. It
pays much better not even investing in a better mousetrap (or, as one Jewish
businessman told me 15 years ago, invent something <insert the functionality
of Propecia and Viagra here>, and you would be swimming in cash), just do
insurance, or real estate. Even investing in tulip bulbs would have
worked until a year ago.
Which kinda takes all the wind out of the wild dreamer's sails.
> into the rough building, then let a second stage of slower nano-growth
> do the integration. The parts that need only to be bulk are quickly
> made, while the complexity on the different scales (macro and nano) are
> provided by engineering on the same scale.
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:07:44 MST