Re: Genetic transition to posthumanism

From: Eugene.Leitl@lrz.uni-muenchen.de
Date: Tue May 01 2001 - 18:25:49 MDT


CurtAdams@aol.com wrote:

> No, I hadn't heard of buckysprings. I think it indicates how far we are
> that a spring is cutting-edge.

They're not just springs. They're linear actuators. They protract when
you apply an electrical potential, due to Coulomb forces, and pull
back due to attraction of (weak, short-range) van der Waals forces
between the concentric graphite sheets, within few 10 ns. That's
right, there's not much mass to move, so even little forces acting
over short periods of time move stuff *fast*.

Think of them as telescoping tubes with a regulation knob. By
tweaking the knob and thus applying the voltage, you can
regulate the protraction amplitude. Very fast, very elegant.

The main point is that a common chemical you can produce in kg
quantities in a statistical process contains components which
can be used as nontrivial parts for a molecular machine. I sneered
at C60. Silly molecule. No good for nothing but structure
fetishists, who drool over anything even slightly sexy
geometrically. I become interested, when it exibited a
rich chemistry. I become rather interested, when buckytubes entered
the scene. I'm now getting more and more impressed by graphenes
by the year, and since recently by the month. This class of molecules
has such a richness to it, it might be enough. Reactivity dependant
on curvature, resulting in being useful for highly inert ducts.
Mechanical properties to rival diamond. Thermal conductivity to
give diamond a good whacking. Ballistic electron transport over micron
distances. Metal, semiconductor or isolator, dependent on chirality
and doping. Can tolerate absurdly high current densities and field
strenghts without going up in wisps of smoke. Transistor action
demostrated in T-junctions. Processing of nanotubes by current
pulses demonstrated. Mechanical manipulation of nanotubes by
proximal probe demonstrated. A certain company from Texas does nifty
6DOF manipulators for buckys. Buckys are superconducting.
Buckys are highly sensitive to adsorbed material. Buckys sing,
dance, and throw confetti. Buck nekkid buckys win the sexiest
molecule of the decade award.

And I'm not tracking the field that closely. That's just a random
list.

So it's good for structural material, logics, ducts, mechanosynthesis.
Did I already mention that I was highly impressed by graphenes?
 
> I think that's very premature, given that we can do *neither* universal,

Of course it's premature, I wouldn't be saying that in a paper.
It's just my personal, private opinion. Nevertheless, if I was
in the nanotechnology field, I know what I'd be working on right
now.

> nor even broad, mechanosynthesis *and* that all mechanosynthesis
> requires decidedly macro machinery.

Of course they're macro, we're in an early bootstrap phase yet.
The next step would seem to use the smallest MEMS you can get,
and hybridise it with invidual molecules, and try what you could
build next. It is important to realize that there is no sharp
boundary between MEMS and NEMS, they are a continuum. We already
have MEMS parts which resonate mechanically at quite a few MHz.
  
> We do need mpsrn for the nanoSanta which will solve all our problems.

I've stopped beliving in Santa (he was called Ded Moroz back
then) when I was a small kid. Why should I start believing
into him now?

> Nifty new computers would be nifty new computers but they won't
> build houses for everybody on earth.

You're of course aware that the Bradbury/Forste/Leitl faction
would recommend you to not become too attached to the ground
under your feet. We'll try to be very civil, but. Ahem. Excuse
me, but you're standing on my planet. Would you kindly move
aside?
 
> We have created computers which can do some very limited parts
> of human thoughts, such as voice recognition. For this reason,

Very lousy voice recognition. Very very bad machine vision. In fact
I need reliable machine vision right now for a project of mine, and
I can't get it. It still bungles, and requires manual intervention.
It sure saves lots of manual works, but it needs a monkey in
attendance.

> given Moore's law, I think it's *possible* that human-equivalent
> AI could be created in 20 years or so. I wouldn't bet the bank on

Given that the state of the art in software development doesn't
move visibly, and that the complexity bareer still exists, I
don't see how anybody could hope to sit down, and explicitly
write down an AI.

Why are you focusing on a smart entity to create smartness? You
don't need that. Evolution is pretty damn stupid, and it did
produce us. Weeeel, sure, we're not thaaaat smart...

> it though. I wouldn't bet the bank *against* it either.

If I was a betting person, I actually would.

> This differs from proximate nanotech or significantly mentally
> enhanced people, which I would bet the bank against.

I would also bet the bank against boosted people, but molecular
circuitry is not very far away. Drextech, now that's lots more
difficult.
 
> I know you're joking, but there's nothing defeatist in saying progress
> will go on for some time to come, that we personally won't be
> obsolete in the next few years, or that nanaSanta's not on our

Next few years? Next few decades, at least.

> doorstep. That's my concept of practical optimism; there are a
> lot of great things under development; things can improve over the
> next 20 years and more after that. It's counterproductive, however,
> to *expect* very improbable miracles; it'll throw off your discount
> rate. You'll end up doing suboptimal stuff.

Um, that's me, 'gene here. You're probably talking to somebody
else.
 
> If we had machine-phase nanosynthesis or computers that could
> write good papers I'd be singing a different tune. Someday we
> probably will. But today we don't.

How about next Friday, 12:03?



This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:07:25 MST