[extropy-chat] Re: extropy-chat Digest, Vol 2, Issue 51

[Chris Phoenix]

On Monday, Eugen Leitl wrote:
> On Sun, Nov 16, 2003 at 08:29:48PM -0600, Damien Broderick, channelling 
> Dr Vijoleta Braach-Maksvytis, wrote:

> > < I go for number 1 - the assembler still resides in the fiction domain and
> > the science needs to be cracked before it will proceed. 

> I would rather agree with this statement, adding a footnote: that it's > not only science
> (a library of experimentally validated machine-phase reactions) but also
> technology (implementation of a nanolithoprinter) utilizing such > reactions to
> deposit a rich set of structures, with a sufficiently high processivity for it
> to make copies of itself, and a number of other devices.

It's postings like this that make me wish I could keep reading this
list.  Remember I won't read any replies to this post unless you email
them to me.

Developing an MNT fabricator will certainly require both science and
technology.  The question is, will it require major unplannable
breakthroughs, or can it be done with planned research?  

What do we need?  
1) A set of mechanochemical tool tips and motions for making diamondoid.
2) A mechanical design for a sufficiently stiff and programmable
nanoscale diamondoid manipulator/fabricator.  (Not a full assembler with
onboard computer--just a robot.)
3) A bootstrapping plan, which may be as simple as putting the tool tips
on an SPM, or may involve building a NEMS or self-assembled
proto-assembler.
4) All sorts of picky little closure issues, like how to maneuver the
parts into place once you've fabricated them, and how to move feedstock
and nothing else through the assembler wall.
  And to make it useful,
5) A CAD program that can handle MNT designs, including peta-component
integrated designs.
6) A design for a nanofactory integrating lots of fabricators with
appropriate control, power, convergent assembly, etc.

1) is almost certainly doable by planned research.
2) is pretty straightforward once we know what motions we need and have
some questions from 4) answered.
3) requires a lot of engineering creativity and lab work, but it's not
at all obvious that it requires more than that.  A big question is
whether we can do 1) with 3DOF positioners.
4) requires lots of engineering.  There's always the chance of some
little problem being a huge pain and requiring real creativity to solve,
but we can't know till we try.
5) we can do in rudimentary form today, but better programs will be
hugely rewarded; if MNT is plausible any time in the next 20 years, we
should start designing the CAD system today.
6) we can probably wait on a bit, but we definitely want it before the
first fabricator is developed.

> A scientist is not an engineer, typically. The set of reactions are worthless
> all by itself, if there's no design, and no viable bootstrap pathway from here to
> there. 

By this argument, each piece of the project (reactions, design, and
bootstrap) is useless, and this seems to imply that we shouldn't work on
any of them ever.  The way out of this is to start working on whatever
will best decide the issue and/or will best advance the project.

The set of reactions would be useful right now today, for at least two
reasons.  First, some people are claiming that MNT is impossible because
the reactions are impossible.  If they are wrong, we need to know it,
and having a set of reactions would go a long way toward deciding the
question.  Second, the set of reactions would both inform and inspire
the rest of the project.

The proposal we have now is detailed enough to evaluate for
plausibility.  So far, no one has pointed out any problem in the
calculations.  Several smart and/or prominent scientists have an uneasy
feeling about some of the theory or results, but that's very far from a
guarantee that the proposal is fundamentally wrong.  And if the proposal
is basically right, then MNT 1) can almost certainly be done in a
decade, and 2) will be very much worth doing relative to the expected
state of technology in 2010 or even 2015.  

We do not have nearly enough information to be confident that no country
or company will try and succeed by 2015.  The question of whether MNT
will exist in 2015 or not is 1) very unsettled; 2) incredibly urgent. 
If the answer is "no", the only way to learn that is to start doing MNT
research until we discover the contradiction.  If the answer is "yes",
then maybe it's safer not to do the research and delay it as long as
possible, but currently we think otherwise.

Chris

--
Chris Phoenix                                  cphoenix at CRNano.org
Director of Research
Center for Responsible Nanotechnology          http://CRNano.org