Re: SciAm: nano and cryonics

From: Robert J. Bradbury (bradbury@aeiveos.com)
Date: Sat Aug 18 2001 - 07:57:08 MDT


On Sat, 18 Aug 2001, Eugene Leitl wrote:

> Sterical hindrance or sterical constraint are much harped upon items of
> every chemical undergrad, and are also mentioned in highschool chemistry.

Are you saying that they aren't thinking it through at a deep enough
level? Sounds like they need to read Nanomedicine Section 3.5 on
Molecular Receptor Engineering, esp. 3.5.5 regarding the limits
of atomic positional accuracy.

> Yes, but you're not trying to make deep cages, and you're not trying to
> stick stuff to a surface.

You can assemble deep cages from the inside out and sticking a methyl
group onto a molecule *is* sticking it on the surface. I *will*
admit there are going to be some very difficult, perhaps even impossible
assembly problems without a programmable nanoassembler. But the
phase space of what you can assemble by tacking things onto each
other or assembling things as nested hierarchies would seem to
be some reasonable subset of what you can assemble atom by atom.

> Chemists do random polymer libraries now, too.

I was unaware of that. Must make a pretty gooey mess.

> They're molecules, not machines, though. Very different from biochemistry.

Machines are subset of all the possible sets molecules. Nature
has produced "machines". Polymerases, ion channels and the
proteosome are some that immediately come to mind. I think
you are splitting hairs if you consider enzymes to not be machines.

> Not your typical computer scientist. They understand bits, but not atoms.

I was thinking the A-life people.

> A few years back it really looked liked we were getting an open source
> virtual sandbox to study dynamics of small and not so small molecules and
> assemblies of such interactively. Sadly, the efforts fizzled, so we're now
> limited to the efforts of a few academic researchers, in toto a few warm
> bodies, and with a biology, not nano bias.

I think that is because there is so much money in it due to the pharmaceutical
industry. As soon as you develop some skill set in the area you get sucked
off into a commercial effort.

> Computers are plenty fast now, even if you're not talking about a cluster.
> We really need something painless, which lets people tweak atoms
> interactively, and see what that stuff does, and interacts. Once we've
> made a few interesting things in machina, we can start trying to make
> them, and to see whether they would work in reality.

I think Goddard is on the right track with this. And if Nano@Home
picks up some steam and feeds into what Goddard et al are doing with
the simulations then I think your vision may be realized. One thing
is for sure -- with a multi-teraflop machine coming online nanomachine
simulations will be doable. If the simulations say they should work
thats going to up the pressure on the synthesists to actually build them
to confirm what the simulations predict.

Robert



This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:40:10 MDT