[extropy-chat] Nanotech educations

[Adrian Tymes]

> If we want the private sector to stay clueless, and
> the President's 
> Technology Council to stay misinformed, while we
> wait and see which 
> government catches on first and spends a billion to
> develop a 
> nanofactory, then our current strategy is very
> effective.

One can't get 100% odds on these things, but I
suspect the odds favor the American government doing
this first, given what they're already funding.  But
this will be once someone comes up with a solid way to
make a nanofactory for US$1,000,000,000.00 or so,
having formed the proposal enough that one could make
useful budgetary projections.

> A liter product
> would have 10^18 of 
> your voxels.  I doubt AutoCAD could handle that.

It could, and I understand some people do use it for
10^18 voxels.  But from what I've seen of that kind of
usage, managing that many elements at a time is so
ungainly that there's much room for a better solution.
One could say the entire field of bioinformatics was
set up to address a very similar problem.

I think we might be agreeing on the idea here, just
using different words to describe it.

> But writing useful CAD software could take five
> years.

What all needs to be done to write "useful" CAD
software?  As had been pointed out, efforts towards
this are being made; I'd view these as components,
while the rest of the components (and assembly into
the final software package) would wait for the actual
hardware.

> * Succeeding or failing on their own merits:  I
> agree that their 
> technical success is not determined by their
> desirability.  However, 
> their economic success is.  If you can build a
> single Merkle "assembler" 
> for $10^7, it's worth doing.

*nods*  But the proposals will have to be evaluated.
It's likely that any such proposal which has a serious
chance of resulting in MNT by itself (plus everything
that's already been accomplished) for a mere $10 mil
will get the funding.  I don't see any such proposals
at this time, though I do see a bunch of research to
create that which could let those proposals be
formed.

> * Similar end result: If *any* technology is able to
> achieve the MNT end 
> result in the next few years, then we have to
> consider that the end 
> result is imminent.

Logic error: just because something could happen
doesn't mean it automatically will happen.  Certain
techs could lead to MNT in the next few years.
Doesn't mean they will, and while we can alter the
odds somewhat, nothing we can do will give them 100%
or 0% success rates.

> At that point, we can stop
> talking about the 
> details of the technology (while funding a crash
> project to develop it) 

Part of the crash project would be to discuss, as in
iron out and nail down, the details of the technology.
In fact, that'd possibly be the most important part.

> Can you tell me about the architectures you've
> thought of?

Sure.  I think I may have mentioned them on this list
a short while back.

#1: Start with two probes, rigged to be able to hold
and release atoms.  One is the "stage", onto which
the product is assembled; at start, it holds one atom
that'll wind up on the edge of the final product.  The
other is the "manipulator", which slides along a rod
to one of a series of storage bins, each holding bulk
feedstock of different elements.  A laser or electron
beam ablates one atom from the corner of the stock,
which the manipulator latches onto before sliding back
to the stage.  The stage moves and rotates to position
the product relative to the manipulator, such that the
atom is in approximately the right place (within an
angstrom or so, anyway), then the manipulator releases
its atom with a pulse of electricity, freeing it to
latch onto the product.  The manipulator then slides
back to pick up the next atom, et cetera.

#2: Take single atom thin layers of material, possibly
created by deposition on an evaporatable substrate.
Press these layers onto the product being made, and
fire laser pulses at the substrate to knock off single
atoms in the desired position.  (Substrate chemistry
makes sure it does not bond to the product beneath -
perhaps the substrate evaporates up while the layer
evaporates down.)  This one is not as well thought
out; the substrate would need to be identified, as
well as the precise means of making layers (though the
latter may be doable today).  It's mostly a way to
convert those techniques for making small-radius dots
and lines into something that could actually build
real products.

> >> I think this will delay MNT.  And make it happen
> >> more abruptly and 
> >> proliferate less controllably when it does
> arrive.
> > 
> > The former is bad; the latter, frankly, is
> arguably a
> > good thing.
> 
> I arguably dis/agree with you.  We don't know
> whether it's a good thing. 

Which is why I said "arguably".

> If you think that nano-anarchy is survivable, then
> I'd definitely argue 
> with you.

I declare that there appears to be a nonzero, even
majority (though not 100%), chance that the human race
(for very broad, inclusive definitions of "human")
would survive for at least a hundred years (or
"forever barring other forces", but I'm trying to be
specific here) if a general purpose molecular
assembler and replicator capable of doing everything
described in Nanosystems were to be developed this
year and appear, without any legislative controls that
do not apply to most types of physical property, for
sale at under $100/gram at retail hardware stores in
most industrial nations, and software to control said
replicators from WiFi-enabled PDAs (like, say, Palm's
Tungsten C model) freely distributed on the Internet
(as available as, say, major distributions of the
Linux operating system, including bandwidth
requirements), by year's end.

Long winded, but before I explain my reasoning behind
that conclusion, does that meet your definition of
nano-anarchy?

> Maybe we can also agree that nano-oppression is a
> near-vertical slippery 
> slope.  And one that we're arguably likely to step
> on.

You mean oppression by those using nanotechnology
against those who don't?  I'd agree there's a
significant danger that that will happen, somewhere.
I wouldn't agree that it's likely that fate would
befall a large fraction of humanity, whether in the
industrialized nations that could produce MNT (since
the oppressed would also know how, and soon enough
gain their own) or in the rest of the world (since
oppressive governments tend to waste resources like
no one's business, meaning that they are unlikely to
be able to truly tap the exponential power of
assemblers - a king having assemblers wouldn't help
his grunts that much if the king keeps the assemblers
all to himself - thus it won't upset the power balance
that much, thus it wouldn't really be
"nano"-oppression).

We're just talking the about availability of MNT
here, not the intelligence to effectively clamp down
on its use.

> >> Identify ways we can hurt ourselves with it, and
> >> take whatever actions 
> >> are necessary to avoid the worst dangers.
> > 
> > I definitely disagree, then, based on the results
> to
> > date of attempts to do so.  But only directly with
> the
> > taking action part of it.
> 
> Hm... Which attempts are you thinking of?  I'm
> thinking of various 
> anti-proliferation and pro-communication and
> pro-stability measures that 
> have been taken to reduce the nuclear threat.  Do
> you think the ABM 
> treaty did not make things more stable?

I'm thinking that practical efforts, in today's
political climate, would wind up being closer to the
stem cell funding ban than to the ABM treaty.  And
the funding ban has made things less stable, by
retarding research into biodefenses while countries
more amenable to supplying terrorists (and not cutting
them off when their initial job's done, anyway) made
great strides towards developing bioweapons.

Besides, the ABM treaty was a deal between the only
two significant sources of nuclear weapons at the
time.  There are how many labs researching nanotech
these days?  And not all of them will respond to
international treaties - indeed, a few of them are
probably acting in deliberate defiance of local laws.

> Someone got their stories *really* mixed up, to the
> point of 
> irresponsibility or even lying.

No, the sources you gave are consistent with what I
heard.  They weren't talking about the rat story, just
the fish.  From:

>
http://www.eurekalert.org/pub_releases/2004-03/acs-ob031904.php

> In a controlled laboratory study, the researcher
> exposed nine juvenile largemouth bass — confined to
> 10-Liter aquaria — to a form of water-soluble
> buckyball (C60) at a dose of 0.5 parts per million.

What they don't say is how this "exposure" happened:
direct injection into the fish throats.  I recall
hearing concerns that the injection completely
blocked the throat; correlating to the description
above, it sounds like they were concerned the
buckyballs weren't the only thing in that injection,
or that the balls formed a semi-solid despite the low
dose and water solubility.

> The fish researcher reported her pilot-study
> findings at a scientific 
> meeting.  The press picked up on it.  She was widely
> castigated for 
> having talked about it at all.  I think this was
> inappropriate; what, 
> she shouldn't have talked to colleagues?  Or they
> should have kept the 
> press out of the meeting?

The latter, if either.  It was the press that altered
the details so that what they said failed to match
reality, no?  Of course, barring the press is itself
rarely a good solution...although one wonders if
gross misrepresentation of the facts like this might
be cause to suspend journalistic privleges, or at
least incur libel suits.

> Is that the main lesson they learned?  That they
> dodged a bullet by 
> avoiding legislation?

Main?  No.  But it was among the results.

> They might realize that they're playing 
> with new kinds of chemicals, and some fraction of
> them *will* be 
> significantly toxic, maybe even with new mechanisms,
> and they should be 
> thinking about how to deal with that reality.

Oh, they realize that.  That's why the study was done
in the first place.  But there's a wide difference
between identifying and dealing rationally with
dangers, and slapping on the Precautionary Principle
to ban an entire industry.

> Then you'll have already seen my counter-argument
> coming.  Even if the 
> exact nature of the threat is not known, if we have
> good reason to think 
> that a threat exists, we should take action to avoid
> it.

But what action?  It's a cliche among politicians to
pass useless laws because the public demands they "do
something" about certain problems, even when the
politicians have no idea what the problem is or how to
effectively address it.  In too many cases, these laws
do more harm than good.  (And if you need proof of
that, I'll step aside and let any of this list's
Libertarian members chime in.)

> And the most likely and appropriate response to the 
> initial studies is... more studies.  Not a big risk.

Ah...that's different from what I thought you were
arguing for.  Yes, more studies are a good idea.  I
thought you were arguing for passing laws right now to
ban certain activities or the like.

> OK, let's hash this out.  Advanced cheap avionics
> within 10 years: 80% 
> certain.  Arms race based on this: 50% likely.  Arms
> race will go 
> unstable: I'm guessing another 80%.  Overall chance
> of geopolitical 
> instability from this one chain of reasoning: 32%.
> These percentages 
> are based on my thinking about this for years.  My
> main uncertainty is 
> the 50% chance of arms race.

Only 32%, and I'd say you're being generous with the
chance of arms race and unstable arms race.  (Who
would we have an arms race with, for starters?  The
whole world's playing technological catch up to the US
military as it is.)  And an unstable arms race would
not necessarily invoke WWIII.  And given the US
tendency towards smart weapons, if WWIII does seem
imminent, what's the chance one side or the other will
simply assassinate the leaders to encourage regime
change (and suceed at it), thereby making total
casualties under 1,000?  And, and, and...

> So I'm 
> predicting at least 3 million statistical deaths
> from just one tiny 
> consequence of molecular manufacturing.

Balanced against how many lives saved due to improved
medicine?  And lost from other consequences?

> Now, are you arguing that the best strategy is to
> ignore what I just 
> said, rather than working to verify or criticize my
> numbers?

I question the applicability of the numbers.  They
only seem to capture a tiny portion of the big
picture.  But it is the big picture that matters: will
a certain action cause more deaths than its absence?

(By contrast, my "50%" above is based on comparisons
to history, which have the big picture available.  All
that has happened, has happened.)

> Can we at least agree on the value of making
> contingency plans?  So that 
> if events start validating a model which predicts
> disaster, we know what 
> we need to tweak to avoid it?

So long as these plans allow for reversal if they
turn out to be cause unforseen problems or to have
been invoked in error, or at least have minimal
negative consequences in said cases, yes.  For
example, setting up orbital and lunar colonies as
backup locations for humanity, and getting them
self-sufficient, just in case the grey goo scenario
actually happens (however unlikely it is).

> > Advisable *at this time*, no, based on
> > an analysis of previous instabilities/vicious
> cycles
> > caused by premature efforts to do so.
> 
> Examples?  And note that I'm not proposing immediate
> action; I'm 
> proposing immediate study.

The examples would be of immediate actions.  Again,
immediate studies are good.