anopresenceanopresenceI spent a good deal of my professional life working for an outfit that specialized in exploring other planets. It was found, after some trial and error, that this is best done by machines, controlled remotely by people.
Controlling machines remotely, is called, you guessed it, "remote control." But there is a fancier word for it, "Telepresence." Telepresence is what you want when you are mixing up radioactive chemicals. Telepresence is what you want to have when you're dealing with unpredictable viruses. Telepresence is handy when you don't have air, water, bathrooms and convenience stores right there in front of you. You want to be there and you don't. At the same time.
A very useful aspect of nanotechnology is nanopresence. The ability to be at some very small but important place, building, predicting, and controlling the outcomes of events in that place. I will assume the nanoevangelists have done their job and have made the benefits of nanopresence obvious to everyone; better surgery, small appliances, five pound jet engines, self-dusting furniture, fantastic voyage, inner space, etc. Getting there is a bootstrap process. Rather than suddenly jumping to nanothings in a global nano infestation, consider:
Nanopresence requires Micropresence.
Micropresence requires Millipresence.
Millipresence requires Centipresence.
and so on. You get the picture.
Keeping the Workshop Clean
Power's of two make sense as a reduction strategy because the whole
shop doesn't disappear in one evening, and digital people like powers of
two. It doesn't take very many powers of two to get small quickly. Watchmakers
are already accomplished at Centipresence and VSLI people understand Micropresence.
One problem is that as things get smaller, phenomena that we neglect at
one scale can become more important. Surface tension, viscosity, dirt,
static repulsion and other little problems get in our way as we scale our
systems ever smaller. When we reach Millipresence, the lights go out and
we have to shorten the wavelengths. Short light can damage things. So we
have to go in small, incremental, reproducible steps to get there. And
we have to award and encourage the people who are trying, in small steps,
to accomplish just that. We have to set goals, with realistic examples
and big cash prizes. "Come on down. You're the grand prize winner today."
Another issue is whether we want to take the whole machine shop down with us, or just the thing that the machine shop built. We really have no choice but to take the whole shop with us. That means doing some serious thinking about what the best shop is. Eventually it starts looking more like chemistry than machines, and that will change the way we build, the processes we rely on, the hammers, chisels, lathes, mills, lithographs and so forth.
We will have to employ as many people in the scaled machine shops as we do in the full sized ones. So each scaling step requires the maintenance of a fleet of machines and a fleet of staff, in one long building the size of a Ford assembly line. You can't fire the staff from the previous level when you reach the new level. You have to keep them on to fix things.
Cutting to the Chase
My first choice, this great example, is the eye hand combination complete
with the holy grail of biology, the opposable thumb. Not replacing my eye
and my hand. No. Instead my eye, and my hand reach into successively smaller
places and build things at that level. Eventually my eye and my hand see
and act at a nano level through whatever abstractions are necessary. This
isn't going to be done overnight, no it will take nine years. Nine years
after the president says, "We're going there." The technology chosen for
shrinkage should be the technology that allows sensing and effecting to
be built and refined at successively smaller levels. Because if you do
that, you can keep going. We can't just shrink the machine shop. We have
to shrink the people too. We do that by choosing that as the first example.
We have to learn and adapt at every new level, like when the lights go
out at Millipresence. We have to keep humans in the driver's seat. Computers
and AI have taught us that.
Side Problems:
Sound that is not significant at our scale can break little machines.
The uncertainty principle also governs how big we can make parts. We can't
simultaneously know how big the are and where they are once they reach
a critical smallness. This affects assembly as well as operation and maintenance
of our nanomachines. We shouldn't get discouraged. We already routinely
operate two dimensional microcircuits which, if you count electrons as
the moving parts, are pretty darn small.
Costing It Out:
To do this right will require hiring all the planners, designers, fabricators
and administrative staff (PDFAS) to run a general purpose machine shop,
electronics section and foundry. We will focus on the self sustaining machine
community necessary to build and connect an autonomous robot arm, a video
camera eye, and a set of tools to be used by the robot.
To build a robot at our scale requires about 10 people, if we allow off the shelf fasteners, wire, electronics and the like to be used. But we can use off the shelf stuff, cause we're the ones doing the shrinking. What this really suggests is that to do nanotechnology right, we need to get an entire industrial complex and economy running so we can get tiny off the shelf stuff. At least the shipping costs will be low! I would estimate that this will swell our minimal PDFAS staff to around 1000 people. Insulation makers, screw thread makers, rivet makers, video camera shrinkers, washer makers etc., etc., etc.
So if anyone comes and asks for money for nanotechnology, we just ask them how many employees they think they will need. For every generation of shrinkage we estimate they need 1000 people. To go from watchmaker's centipresence, which is cheating, to millipresence is 1000, to go from milli to micropresence is three generations so 3000 people will be needed. To go from micro to nanopresence is three more, so 3000. Also we cheated and started at centipresence, even though all the Swiss watchmakers are gone, so add 2000 there.
So to get from current scale to nanoscale will require 9000 people,
working concurrently.
Assuming a person costs $50,000 with overhead and they each use their
own cost in tooling, that makes $100,000 per person. So for about a billion
dollars we can have the first nanoeyehand exploring inner space, crawling
around in Johnny's brain, scraping away the stroke he got from smoking
crack. Not too bad. Less than a moon shot. About the same as a new space
shuttle. Only twice the size of the outfit I worked for. Piece of cake.
© 1996 L. Van Warren, All Rights Reserved
For further reading: (First two links courtesy of a fish@kzin.cs.utah.edu (Russ Fish))
Cornell
National Nanofabrication Facility