Re: Homework, was Re: Goo You--was Re: Nuke weapon/reactor/waste horror story links?

From: S.J. Van Sickle (sjvan@csd.uwm.edu)
Date: Wed Jan 10 2001 - 13:49:02 MST


On Tue, 9 Jan 2001, John Marlow wrote:

> I'm reading it now. The science is good, obviously,
> and I'm not all the way through it, but holes appear
> in the first few paragraphs. For example--the comment
> that a stealth buildup would take 20 mo to accomplish
> and that a more rapid buildup would be detected,
> thereby enabling defenses to be deployed.
>
> Yah, right.
>
> WHAT defenses? The basic problem, of course, is that
> you really need a nanodefense--and that you can't
> possibly develop one until (as Drexler has pointed
> out) the tech is already here.

The difference being that there is a wide variety of possible molecular
nanosystems, of which self-replicating goo is only a small subset, and
likely the most difficult to design. It is possible (but by no means
certain) to have non-self-replicating nanodefenses in place well before
anyone is capable of building a nanoreplicator. Read towards the end of
Rob's paper, where he makes a strong but admittedly not entirely
convincing argument that prepositioned defenses can trump replicators.

The rub in this argument is the time lag between the first nanosystems and
the first wild nanoreplicators. If we have to rely on more or less
conventional design (CAD systems, shape description languages,
standardized part systems, design compilers, etc) the the time lag could
be many years. If, on the other hand, MNT results rapidly in strong
super-human AI, then the lag could be only hours. AI is the important
factor (IMHO), not nanotech itself.

It is also not quite true to say that nano is the *only* defense against
nano. Rob and I discussed it, and came to the conclusion that, for
localized infestations, napalm would work just dandy. Inelegant, though,
and rather tough on the neighbors.

> But, as I say, I'm still looking it over. Homework is
> what I'm doing; that's why I'm here.

Good for you! May I also suggest Chapters 13 and 14 of Drexler's
"Nanosystems". They concern overall architecture, and are a little less
technical than the previous chapters. This may give a little
perspective beyond the "self replicating little critters" paradigm.
Hint: neither the word "replicating" or "self-replicating" appear in
Nanosystem's index. Pages 66 and 67 of "Nanomedicine" also have a good
discussion of the varieties of replication, and their relative benefits
and difficulties.

steve



This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:04:44 MST