Re: One NanoDream, Deconstructed

Dr. Rich Artym (
Tue, 1 Oct 1996 02:58:16 +0100

In message <>, Robin Hanson writes:

> At some point there will be a device which
> -- can disassemble generic stuff in the environment into feedstock,

Not quite. Disassembly and assembly don't have to be co-located in
either time or space, though they could be.

> -- scan any (house-or-smaller-sized) object for its atomic-level design,

There is no upper limit on size: you don't need to enclose the item
being constructed. Nor do you have to scan the whole item before you
start assembling a replica, though it might help in some circumstances.

> -- sythesize a plan for constructing such an object from feedstock,

That may be the way to do it in some cases, but it's not the only way.
"Rip up and try again" strategies such as used in circuit-board track
laying could also be used. Also, inevitably the experience of building
a lot of different things will build up a database of component plans
and heuristics for tackling such tasks, so "synthesizing a plan" may
turn out to be a Lego-fitting exercise.

> -- retool itself and construct the object with a low error rate, and

Again, the "retooling" will probably entail selecting appropriate nano-
factory components from a library, rather than doing it from scratch
for each new job. Items in the library will be those that experience
has shown *can* be manufactured with a low error rate, no doubt. The
error rate in the end product will I expect be raised through iteration
and constraint checking at the point of assembly. Unlike in wafer
fabrication, yield will not be limited by having to get it right on
first attempt.

> -- repair itself, or break slowly enough to replace itself anew often.

Yes indeed. One can expect to find a range of monitoring mechanisms to
detect errors, isolate them, and replace the faulty units.

> If such a device were smaller than a house,

Size not relevant.

> and could replace itself
> in less than a year

Drexler's estimates suggest that the likely timescale is minutes or hours.

> using stuff available from an acre of land,

There's no restriction on source of feedstock. Tunnel down, tunnel
along, set up transcontinental feedstock pipelines if you want, or
don't bother with continuous tunnels at all --- just send your emmissary
out to where the feedstock is and bring it back to you, or else maybe
manufacture the necessary part where the feedstock is and then have it
transported back. There's nothing that says you can't crawl around
within the Earth's crust to your heart's content; certainly a lot of
nice challanges there. In any event, "one acre" becomes irrelevant.
That notion belongs to a time when property was defined by access to
the *surface* of land and to a time when such restrictions could be
enforced; neither of these are valid in a nano world.

> and if
> copyright on the design were not strictly enforced,

It's unenforceable. And anyway, why should anyone want to enforce it?
This is "old world" thinking, based on the premise that current-day
economics will survive unchanged into a nano age. That belongs to
another thread already in progress, so I won't say more here.

> but if property
> rights to such acres of land were strictly and cheaply enforced, then

A hell of a lot of if's. Property rights? Enforcement? Cheap?

> it would seem that within a hundred self-copy times, most everyone who
> owned an acre of land should have such a device, obtained at the cost
> of a small fraction of their acre. And it seems they could then, if
> they chose, soon have an acre full of devices for which the design
> copyright was not strictly enforced. And they might so choose if
> public designs were not much worse than private designs, and if local
> production were near as efficient as distant specialized production.
> Given all this, some folks seem to want to infer that at some point
> most living humans will fit the scenario of living off their own small
> plot of land, using basically free designs.

This is a figment built upon a mirage built upon a misconception. The
main misconception is, I think, that you see all the other institutions
of the 20th century staying more or less unaffected by the progressive
development of nano capabilities. I find this extremely unlikely. As
the capability is put into use in more and more areas, it is likely to
cause *extreme* uproar, panic and histeria among people that suddenly
find they live in a shifting world of horrendous freedom, and the whole
fabric of civilization is going to be given a severe shaking. I don't
know where it will lead, because people are unpredictable. However, I
know where the technology is leading, and that's very, very far away
from the terribly limited scenario that you describe above.

> If I've understood this right, I can challenge the conclusion by
> pointing to all the assumptions that may not hold.

You haven't, so you can't. (:-)

> I do no have much
> doubt that eventually such a device will be possible to construct.
> However, by that time many other changes may happen as well. The sun
> may dim due to intervening solar collectors, or the O2 and CO2 may be
> stripped from the atmosphere. An upload or other population explosion
> may mean that the ratio of population to acres is far less than one.

There's a causal inconsistency here. In all the scenarios I've read,
you need nano to give you control over matter before the above situations
can arise. It's also inconsistent with the estimates from Foresight.

> Or wealth may be very unequally distributed, so few can afford an
> acre.

Wealth? Distribution? If you want an acre, go and get it --- there's
lots going spare, and if everyone does it (and they will because everyone
wants more) then counter-measures are unenforceable because, presumably,
the will of the people is what matters, no? (;-)

> Copyright on these many devices may in fact be strictly enforced, by a
> combination of self-destruction upon scan attempts

Whatever is the point of so doing? This sort of argument is still
wedded to the product-is-wealth mentality. Free yourself! (:-)

> and draconian
> punishments for violators.

Unenforceable, and no point in enforcing it.

> Public domain designs of most sorts may be
> quite inferior to private designs, or generic one-size-fits-all
> devices may be rather inferior in efficiency to specialized devices,
> even given shipping costs.

Or public domain designs may be greatly superior to private designs,
since there will probably be *vastly* greater resources in the public
domain than in the private. It's already true now in some walks of
life: eg. using a proprietary C compiler is asking for trouble, as
the support you get from a private company is terrible compared to
the support of thousands of GNU developers. Anyway, what's this
difference between public and private that we're talking about? In
a nano world where individuals and families can have the same kind
of resources and capabilities as international corporations, these
differences become more or less irrelevant.

> Property rights to acres of land may not
> be cheaply enforced, but require payment of substantial taxes or PPL
> fees.

Think 3D, in the Earth's crust, in its oceans, and beyond the clouds.
Yeah, it'll be war for the choice cuts, but I'm heading out of here
anyway. :-)

> To pay such fees, one may have to devote a substantial fraction
> of ones resources to production at an advanced level - public domain
> defense system designs may be nowhere near up to the task.

Why? Public-domain systems ought to be far, far more capable than
anything that a remnant of government or state can muster. After all,
who will want to continue to *work* for government or state? What
would be the point?

[Robin, I must confess that I'm not sticking to the straight and
narrow in much of the above, just because I'm having so much fun
responding to your points because the world you see is so terribly
*limited*. You need to transcend a little bit quicker. :-)]

> Finally,
> most people may prefer more to live be closer to other people, in
> dense cities,


> than to have their own isolated plot of gadgets.

Yuck. It's not like that at all. Far more like living in your
own world amid your loved ones and friends and defended by your
own animates. And it *will* be your own world. The need for
protective containment will see to that.

> Now folks may have plausible arguments for each of these issues.
> And I'd like to hear them. But what are the chances that *all* of
> these issues go the way of this NanoDream?

Collosal at present, because nobody has yet come up with any way of
actually stopping it from happening, so it *will* happen because that's
the direction in which the technology is heading.


###########  Dr. Rich Artym  ================  PGP public key available
# galacta #  Email   :
# ->demon #  Web     :
# ->ampr  #  AMPR    : rich@g7exm[.uk] BBS:GB7MSW
# ->NTS   #  Fun     : Unix, X, TCP/IP, kernel, O-O, C++, SoftEng, Nano
###########  More fun: Regional IP Coordinator Hertfordshire + N.London