Re: Leading nanotech experts put 'grey goo' in perspective

From: Chris Phoenix (cphoenix_at_best.com)
Date: 06/16/04 

Date: 16 Jun 2004 03:20:12 GMT

Jim Logajan wrote:

> iph1954@msn.com (Mike Treder) wrote:
>
>>A paper published today in the journal Nanotechnology...
>
> "Accordingly, the construction of anything resembling a dangerous self-
> replicating nanomachine can and should be prohibited."
> Ugh. Any really useful "self-replicating nanomachine" is likely to
> "resemble a dangerous" one - it's the nature of the beast.

Maybe "prohibited" wasn't the best word... but I don't think the
recommendation goes much beyond the Foresight guidelines. The whole
point of the paper was that there really is a substantial difference
between dangerous replicators and efficient exponential manufacturing
systems--regardless of scale. I'll answer your technical questions
below, using information from the Nanofactory paper which was referenced
from the IoP paper. Perhaps you should have read more before you attacked.
http://www.jetpress.org/volume13/Nanofactory.htm

> .... But the first step toward MNT WILL
> require the need for exponential growth, and tiny self-replicating devices
> are still the best way to attain that curve.

If a computer controlled robot can build its mechanics but not its
computer, is it self-replicating? Not by my definition. During the
bootstrap process, the tiniest devices would not include any computer.
They'd just build more robot arms. As the required control bandwidth
increased, simple onboard computers could be added--but they'd be too
small to hold the full blueprint, which would still be streamed in from
a bigger external computer.

> The opinion piece simply
> doesn't attempt to compute the size of the nanofactory in terms of number
> of base pieces or how the first one is to be built or how fast it can be
> built. A "desktop-size" factory is HUGE (really really huge - seriously)
> relative to the molecular scale - no matter how you slice it.

No, the opinion piece doesn't do this, because it doesn't need to,
because the Nanofactory paper (which it cites) goes into great detail on
this. A desktop-size (half cubic meter--OK, maybe more like
appliance-size) factory would require 144 quadrillion robot arms.

> Without some
> relatively TINY self-replicating device available to boot-strap the
> construction of the first nanofactory, the nanofactory approach is a
> non-starter (unless you are willing to wait millions of years for
> nanofactory number one to roll out of the lab!)

The nanofactory paper has a whole section on bootstrapping.

> Also, can a nanofactory build a copy of itself or not? If all I have access
> to is a nanofactory, solar energy, and soil, am I self-sufficient or not?

Not. But you would be self-sufficient if you also had blueprints for a
chemical processing plant, and enough purified chemical feedstock to
make one (plus the solar cells).

You could, in theory, design a robot chassis that could load
nanofactories and chemical plants on itself, travel until it had loaded
enough raw materials, then tell the nanofactory to build another robot,
nanofactory, and chemical plant. I do not think this makes the
nanofactory self-replicating. It does indicate that self-replicating
systems can be built, but we always knew that. The nanofactory may make
it easier--but not easy! And still pointless.

One thing I don't know, and would like to know, is how small such a
robot/chemical plant/nanofactory combination can be made. The
non-computer portion of the nanofactory, of course, can be quite small.
  The limiting factor will be either the computer (with full blueprint
storage) or the chemical plant.

> If I'm not, why not? Will it be because I and other people like me are not
> to be trusted? Will there be an elite, select few, who have access to full
> nanotechnology and decide who is trustworthy and who isn't? Just how will the
> prohibitions be enforced? How will violations be detected? What remaining
> freedoms will be sacrificed?

Very good questions. CRN doesn't have an answer for them. We don't
even know what the workable options are. We would like people to start
wondering about them and working to answer them.

>>Mike Treder, Executive Director of CRN, said, "We hope that this
>>article will advance the discussion of the actual implications of
>>molecular manufacturing. There is no need for panic, but there are
>>urgent concerns that must be addressed before the technology arrives."
>
> There are no reasons to panic simply because no one is in a position to
> build either a nanofactory or even the simplest non-self-replicating
> assembler - and wont be for many years.

How many? I say it might be as few as five years to a nanofactory,
given enough billions of funding, and could be less than ten years with
probably well under a billion total. That's soon enough that we should
be very worried about the lack of policy. And that assumes there's not
already a secret program underway.

Chris 

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