People who know me know that I've had a deep interest in nanotechnology
for a lot
of years. There was a Scientific American article in 1989 describing
that got me interested, and soon I read Engines of Creation
and started reading and posting on sci.nanotech, going to conferences,
and occasionally even writing software
. So I'm
down with the whole molecular manufacturing thing. One of the
particularly interesting ideas associated with nanotechnology is the
Imagine a robot with enough smarts and dexterity that it can build a
copy of itself. Call this a "capable robot". Start with one
capable robot, tell it to make more of itself, go away for a few
months, come back, and you have an army of them. Not surprisingly,
this has economic implications.
Those economic implications depend strongly upon the cost of the input
materials that the robot uses for self-replication. If the robot is
only snapping together two complex subassemblies, then there are no
implications. If the robot digs up ore and smelts it into metal, makes
its own lathes and drill presses and nuts and bolts, and purifies its
own silicon from rocks and sand, then the implications are big.
The implications are big because the input materials are cheap. The
robot does not have to participate in a complex interwoven economy to
get its subassemblies, it makes them itself. The price of a
freshly-minted robot drops to nearly the same as an equal weight of
rocks and sand. You can give the robot to anybody and they can
instruct the robot to make more robots.
The price of a capable robot is infinite today (completely
unavailable), and some day it will be billions of dollars, and a year
or two later it will be falling precipitously and it will level off
near the cost of the raw materials to make a new robot.
If capable robots were distributed to every village, they could change
life on Earth. Poverty could be wiped out. Diseases could be cured. We
could live in a robot-labor-powered paradise. It has been argued that we
might easily end up in a robot-labor-powered hell of perpetual
unemployment. So the distribution of capable robots is important to
understanding how they would reshape the world.
So there are some interesting questions.
- Is this "capable robot" idea possible at all? In the short term, it's more
efficient to manufacture stuff in stages (one guy smelts ore, another
guy builds a metal lathe, and a third guy builds a building). The
capable robot trades away economic efficiency in favor of economic
- We are currently on a trajectory of increasing
concentration of wealth; the gap between rich and poor is growing
while the middle class is shrinking. Can capable robots turn that
around? What mechanisms and policies can we put in place today so that
when robots arrive, the poor and unemployed aren't left behind?
- How do these questions and answers change when we replace robots
(presumably built with technology available today) with mature
nanotechnology? The cost of inefficiency declines.
As far as the thing about not leaving the poor and unemployed in the
dust, Marshall Brain has suggested that everybody
get a $25000 minimum wage just for breathing. It's interesting, and
even affordable. I have what I think may be a better idea.
Suppose Acme Robots (or a real robot company) offers robot
purchasing savings accounts. I open my account and plunk down twenty
bucks. Then I go about my business for ten or fifteen years. My money
collects interest at some reasonable rate, and one day the price of a
capable robot decreases to the amount in my account. On that day, Acme
Robots ships me a capable robot (or just tells one of their robots to
walk to my house). In the developing world, the initial deposit might
be shared among all the families in a village, or a group of villages.
When the robot arrives, one of its jobs is to make more robots.
Could cheap robots help to discover new medical treatments, or find
the next new galaxy or solve the next mathematical grand challenge?
These are more computer questions than robot questions
since the challenges involved (aside from the physical
working-in-the-lab part) are primarily cerebral. So: Could
advances in computer hardware and software trigger advances in
scientific and medical progress? I think so. Big fast computers enable
simulation. They enable literature database searches. They can sift
through mountains of data to find correlations that might be too
subtle for human minds to notice.