From: Robert J. Bradbury (bradbury@aeiveos.com)
Date: Mon Dec 02 2002 - 21:03:07 MST
On Mon, 2 Dec 2002, Ramez Naam wrote:
> Okay okay. Maybe I'm out of date. The safety concerns I've posted
> deal specifically with self-replicating assemblers.
But we already have self-replicating assemblers. Every single living
microbe (not viruses) matches this description. In fact there are
more copies of microbial self-replicating assemblers in or on your
body than there are copies of your own cells making up your body
(documented in Nanomedicine VI).
Given that situation (that we manage to survive in a sea of self-replicating
assemblers) you have to make the conversation much more precise -- at least
I hope you would see this.
> I have far fewer problems with highly specific non-replicating
> nano-devices. (I hesitate to call these assemblers, though.)
Why? A ribosome is an assembler that assembles a large variety
of protein based structures (on which all of nature operates).
If you want to be specific then you need to state a "universal
assembler". It isn't clear that a "universal assembler" can
exist. It isn't clear whether it is possible for an assembler
to assemble both diamondoid (largely C based) and sapphiroid
(largely Al2O3 based). They may require two very distinct assembler
architectures.
> And again, I feel like the non-replicating model results only in
> incremental improvement of current manufacturing techniques.
You have to get this Mez -- we already have self-replicating
nanotech. You use it every time you eat a slice of cheese,
a container of yogurt, consume a bottle of wine or beer, or
wash your clothes in detergents equipped with "enzymes"
(derived from bacteria). Many of your prescription drugs
are also produced using this technology.
The self-replicating model is here now -- we actively use
it in our existing manufacturing processes. The question
is whether there are better ways to control it (there are
but I'm not at liberty to discuss them).
> I don't feel that it achieves the potential for vast social change that
> Drexler envisioned in _Engines of Creation_.
EOC is a very early view of what is possible. People are working
on re-writing it. Exponential assembly achieves much of the social
change that EOC envisioned without the dangers of self-replication.
(because broadcast architectures can operate one-to-many).
> That's fine by me - those social changes include very scary
> risks as well,
Granted, but there are paths that allow the risks to generally
be avoided.
> and big parts of Eric's reasoning on the economic leveling impact of
> self-replicating nano have always seemed a bit dubious.
I would disagree -- if you can have exponential growth of the benefits
of Nano without the risks of true self-replication then one might be
able to steer a path of having the benefits without the hazards.
That produces the economic leveling effect. It also lays the foundation
for a subsequent set of problems, documented to my mind in Robin
Hanson's "If Uploads Come First" paper.
Its not over until its over.
Robert
This archive was generated by hypermail 2.1.5 : Wed Jan 15 2003 - 17:58:32 MST