From: hal@finney.org
Date: Sat Jan 06 2001 - 09:39:32 MST
Steve Van Sickle, <sjvan@csd.uwm.edu>, writes:
> I've spoken with some of the top researchers about this...it is *very
> hard* to design wild replicators, and of very limited use. I can see,
> myself, only two uses for a "wild" replicator: offensive military, and
> possibly terraforming when you are unable to deliver more than a few
> milligrams of payload (say a distant start system). If anyone can figure
> out some other use for *self-replicating nanoscale devices* that cannot be
> done with some other architecture, I'd like to hear it.
I agree, this talk of how "one mistake with nanotechnology will
wipe us out" ignores the wide variety of implementations and design
approaches which will fall under the general category of nanotech.
Nanotech encompasses a very broad range of manufacturing technologies.
What they have in common is that they are working precisely at the
atomic or molecular scale. Within that scheme there are millions of
possible approaches.
Drexler emphasizes that the pick-and-place robotic arm which many people
view as the canonical approach to nanotech manufacturing is actually
likely to play a very small role. Far greater efficiency will result from
specialized equipment which makes specific products. Nanotech "assembly
lines" can perform repeated operations to raw materials carried along
conveyor belts and transportation pallets. The entire work flow will
be optimally structured for whatever product the factory is producing.
You could have a nanotech factory that just made micro-scale gears,
for example. If that factory makes a mistake, it's not going to wipe
us out, it's just going to make some bad gears. And similarly for all
the other kinds of products we will need in a nanotech world.
Not only is such a manufacturing methodology more efficient, it is likely
to be far easier to design than some kind of general-purpose eat-anything,
build-anything magical genie machine. Furthermore it integrates into
the existing economy more easily and allows incremental development of
new products.
In such a nanotech world, there is no way that one mistake is going
to wipe all of us out. Yes, there will be dangerous products, just as
there are today. But they don't destroy the world.
Now, of course we can imagine the design of products which could,
potentially destroy the world - omnivorous (or biovorous as Robert Freitas
describes it) gray goo replicators. This is a problem recognized since
the earliest days, and one we have debated many times. At a minimum,
it is clear that the design of such dangerous devices is not going to
be easy and will require a long-term, intense effort.
I don't think it is fair to associate the entire field of nanotechnology
with the dangers posed by this specific problem. It's something that
we have to watch out for, yes. But there is far more to nanotech than
gray goo devices.
Hal
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:04:35 MST