Re: Bootstrapping to nanotech

Hal Finney (hal@rain.org)
Thu, 4 Sep 1997 12:08:33 -0700


A somewhat late follow-up to an earlier message about the Feynman Grand
Prize for nanotech from the Foresight Institute.

From: Hagbard Celine <hagbard@ix.netcom.com>
> Hal Finney wrote:
> > I suspect that there is an implicit assumption of a particular path
> > towards nanotech where these devices would in fact represent an inter-
> > mediate point. The fact that 32 copies are needed presents a hint.
> > Specifically, an atomic force microscope enhanced with interchangeable
> > tips which can catalyze certain bonds might work. You'd probably only
> > be able to build one or a few copies of the nano device, so 32 would
> > perhaps be within reach.
>
> Foresight is steeped in Drexler -- he's the friggin' Chairman for pete's
> sake! Why would they do that?
> [...]
> I don't have the science background to assign a sensibility to either
> case, but I do know that Drexler's works are persuasive to a layman. Can
> you point me to discussions concerning likely route technologies other
> than the Drexlerian biotech route?

I don't see Drexler as being committed to the biological or self-assembly
routes to nanotech. Rather, I think the consensus today is that
the STM/AFM approach is the most productive. Look at the most recent
issue of the journal Nanotechnology, table of contents at:

http://www.ioppublishing.com/EJ/Info/bin/toc/na0803A/

Virtually every article has to do with probe microscopes of various
types.

I think this is unfortunate. While these devices offer quick payoffs
and we are learning some interesting things, I don't see how they will
really work as a path to nanotech. It is going to be so difficult and
so tedious to gradually learn to arrange 5 atoms, then 10, then 100,
500, 1000, etc. But we won't have anything really useful until we can
arrange millions of atoms. And even then we'll be lucky to have as many
as the 32 copies required for the prize.

The biotech route seems to offer more prospects for useful intermediate
results. That's what you need if industry is going to fund the research
in a big way. Biochips, and other integrated devices containing a mix
of biological and micromechanical technology, are potentially a very
important technology. Biological sensors, manipulators, artificial
tissues of various kinds, all have immediate economic payback. As we
gain more skill in utilizing and modifying cell organelles and creating
artificial variants, the line between biotech and nanotech will gradually
blur.

>From the beginning, with biotech we should be able to have mass production
of the devices, at least in quantities of thousands when we are using
lithographic techniques, up to billions and beyond when we use pure
biological and chemical techniques for construction. With probe microscopy
techniques, it is hard to see any useful payoff until they manage to build
a full-blown self-replicating assembler, and set it to work making copies
of itself.

I don't think STM research is wrong or wasted, because they are getting
data about the behavior of atoms in systems which are too large for
pure QM simulations to work. These data will feed into the next generation
of molecular design tools and make them more accurate. But I don't see
STM's as a path to nanotech.

Hal