[extropy-chat] Social Implications of Nanotech

Chris Phoenix cphoenix at best.com
Tue Nov 11 19:59:38 UTC 2003


Robin Hanson wrote:
> Would the mildest versions
> of nanotech really, by themselves, induce superhuman AI?  It is 
> not obvious.

I assume the word "nanotech" here means MNT.  And MNT becomes
interesting when it becomes capable of exponential manufacturing.

A "mildest version" might be a Merkle-type "assembler".  I have argued
that this could lead rapidly to a nanofactory capable of making
monolithic kg products: http://www.jetpress.org/volume13/Nanofactory.htm

Does this count as "mild"?  It's still limited to diamondoid, and still
requires human design to build products.  But such a system could
certainly build billions of tightly networked computers--probably for
just a few dollars.  (Could build trillions, but I'm not sure about
cooling them.)

> >Sure, a couple of centuries worth of hitherto progress rolled into a
> >month, or a
> >couple of days. Accelerating up to a rate of 3 kYears of progress within
> >24 hours
> 
> The mildest versions of nanotech don't seem capable of inducing such rapid
> change.  Again, the point is to try to be as clear as possible about what
> assumptions lead to what conclusions.

We don't know how powerful the algorithms might get.  If you put enough
crunch power in a box, and run an evolutionary system on it, you might
indeed get a system that could be amazingly efficient at discovering new
science and technology.  But there's a lot of supposition here.

> >What does it take to produce a good meta-designer? A robust morphogenetic
> >code, an evolutionary system, a good nanoscale simulator, and lots of
> >computronium to run the above. As embarrassingly parallel as they come.
> >And that's about the only metainvention you need to make.
> 
> And how damn hard is it to have "lots" of "robust" "good" items like these?

Are you questioning the ability to build such a system fault-tolerant? 
It's not hard to make a very reliable system on top of a very unreliable
system.  Takes only one level of abstraction (e.g. TCP/IP).  And you can
play this game as often as necessary to make it appear that you have
billions of 100% reliable computers.  It might cost you some speed and
energy, but it's doable.

> I meant that open source counts as a market form.  We have some insight into
> it, and some open questions remain, as with all market forms.

Interesting.  I'd like to float a definition here.  An "unlimited-sum
transaction" is one in which the benefit to one (or both) of the parties
is much higher than the cost, and is not correlated with the cost. 
Copying a text file off a web site is an example.  The bandwidth is
trivial.  The benefit will vary widely depending on who reads the file.

I argue that open source is based on, and enabled by, unlimited-sum
transactions.  And, that commercial trading is incapable of dealing
adequately with unlimited-sum transactions.  
http://CRNano.org/systems.htm

Note that economic theory probably can deal with unlimited-sum
transactions.  If economic theory can deal with war, which is also
outside the scope of commercial trading, then it can probably deal with
open source.  But commercial trading can't deal with open source. 

Chris

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



More information about the extropy-chat mailing list