From: Damien Broderick (d.broderick@english.unimelb.edu.au)
Date: Sat Jan 08 2000 - 18:39:07 MST
At 06:37 PM 7/01/00 -0800, Gina wrote:
> I forwarded your post to a few appropriate channels.
Thanx, N-grrl. Add this to the fwds?
Thanx also to those who've responded so far, although there's been the
predictable drift away from economic realities to the plausibility of
competing tech paths to an assembler. Here's an extra bit I've written to
amplify one ref in the previous post. (BTW, for those who've never seen THE
SPIKE v. 1.0, there's already an ample account of the Drextech approach as
sketched in everything from Engines through to Nanosystems. Trouble is,
skeptical assessors of my book have been scathing abt Eric - for reasons
that escape me - and demand [i] more real-world examples of progress and
[ii] more detailed economic analysis of the path to assemblers.)
The superscripts will drop out, alas:
============
Corporations such as Geron (in cancer, stem cell and telomere research) and
Zyvex and Ntech (in micro- and nanotech realms) work on both immediate and
long-term studies into topics that might yield truly flabbergasting
outcomes eventually and in the meantime could bring more limited products
to market
Consider Zyvex, a developmental engineering and directed research startup
company founded in April 1997 in the Telecom Corridor of Richardson, Texas.
Its explicitly declared goal is `to build one of the key pieces of
molecular nanotechnology; the assembler.' They acknowledge an intellectual
debt to Drexler in forming this goal (while recommending close attention to
such refereed journals as Nanotechnology, Surface Science, Journal of
Vacuum Science and Technology, and Journal of Physical Chemistry), and
state that `our first product will be some type of assembler, which others
can buy and set to work to make useful things for the market they are
already familiar with. We don't expect to go conquer all known markets and
make all possible products by ourselves. The assembler will be enough for
starters.' The company was not seeking outside funds in early 2000, and
announces `a 5-10 year time horizon for its first revenue'.
What kind of assembler do they have in mind? Certainly not a handheld
miracle machine, or an AI-controlled Genie. Rather, they envisage:
a system of unspecified size (possibly quite large initially--say one by
two by two meters), capable of manufacturing materials or arbitrary
structures with atomic precision, getting nearly every atom in the desired
place. An error rate of 1 in 106 would allow meaningful structures to be
built, and is probably a reasonable goal for the first generation. A more
mature system would have error rates in the 1 in 1012 range or better.
A Zyvex assembler might be regarded as a child of a current Atomic Force
Microscope (or Probe), powered and controlled from the outside, but ideally
able to copy itself and even produce an improved version (again, naturally,
following instructions provided by engineers). To handle useful amounts of
raw materials and produce saleable quantities of goods, huge numbers of
assembler manipulators will be governed by a central control processor.
Zyvex expect to build assemblers with a million positional devices, each
doing 1000 tiny tasks a second. That's a billion molecular building blocks
put into exact place every second, for each chunky assembler--still not a
great deal in visible terms. The second generation might graduate up to
1015 building units a second (that's a million billion), and the third
generation commercial assembler would process 1024 molecular Legos a second.
Their game-plan for reaching even the earliest of these mints is
cautious. They'll start with current microelectromechanical machinery (MEM)
technology. A MEM assembler, has the advantage, as a testbed, that what
it's compiling will actually be visible under a scanning or even an optical
microscope. Next might follow a fullerene or nanotube assembler, handling
and chemically bonding pre-grown microscopic carbon tubes. Trust me, it
will not be building a diamond spaceship out of old potato peelings,
however. This humble forerunner of the mature mint will put together very
strong structural beams, welding carbon tubes into rods or pipes stronger
and lighter than steel. Zyvex have in mind a device that is far from
perfect but can function without the hard vacuum or difficult temperatures
needed in more classy designs. So it won't build diamond or diamondoid, but
will make useful products, and compile them cheaply. `All we need,' Zyvex's
promotional website confesses, `is a suitable reaction, a suitable
catalyst, a suitable positioning device, and some luck getting the catalyst
stuck onto the end of the positioning device just right. Does that sound
hard?' They are honest, and wry. `Yes.'
Even so, by 1999, the company had already announced the launch of Zybot
Mark 1, a MEM micromanipulator with two little grabbers able to move with
five degrees of freedom, `holding MEMS Precision Instruments tweezers'. The
Zybot will make more elaborate MEM gadgets, and so on--all the way down,
perhaps.
So you can start to see how the program comes together.
======================
Damien Broderick
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:26:09 MST