NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/sci/nanotech/795

Internet Message Format | 1993-01-27 | 3.9 KB

Path: sparky!uunet!news.univie.ac.at!scsing.switch.ch!univ-lyon1.fr!ghost.dsi.unimi.it!rpi!uwm.edu!linac!att!cbnewsi!cbnewsh!att-out!rutgers!igor.rutgers.edu!planchet.rutgers.edu!nanotech
From: [e-mail] (Malcolm McMahon)
Newsgroups: sci.nanotech
Subject: Re: Surviving
Message-ID: <[e-mail]>
Date: 27 Jan 93 04:53:53 GMT
Sender: [e-mail]
Organization: Computing Services, University of Warwick, UK
Lines: 61
Approved: [e-mail]

Yes, I that is what I understand by the term "hothouse" too. But
consider, if nano-technology is to produce significant amounts of
product of any sort it must do something utterly beyond any machine we
have devised to date. It must reproduce itself. In many important ways
this immediately makes it more like a horse than a car.

Now it may well make sense to use a heterogeneous set of machines
rather than one general purpose machine. Yet to do the job we must
have a minimum set of these machines capable of reproduction. It does
not matter whether the reproductive unit is one generalised assembler
or a whole set of assemblers, molecular mills and so on. Someone has
to have a comp ete reproductive unit and that person will be the one
essentially in control of the manufacturing process. That person will
be the one who implements changes in product.

Making a machine reproduce itself is an enormous challenge. To me it
seems overwhelmingly the most demanding thing these machines are
likely to be called on to do If they can reproduce themselves they are
already most of the way to being able  to do any other likely job.
When the product is actually being manufactured there may indeed, be a
specialised array of machine to make it but this is essentially just
the final stage. Firstly enough reproductive units must accumulate to
create this final production array in a reasonable time.

What environment do assemblers actually need? One basic requirement is
that the manipulators must have a eutactic environment: In other
words, no molecules admitted except on a leash. This implies more than
vacuum. It implies a whole structured environment with molecular
conveyors of some sort to deliver feed stock and fuel and remove
desired and incidental products. You are not just going to be able to
dump naked assemblers into a fluid mix of feed stocks. As I see it,
replicators, of necessity, are going to have to build their own
environment as they go along, just as our own cells do.

Hence my egg idea. The eggshell provides a barrier within which
eutactic conditions apply. Feed stock and fuel enters in a rigidly
controlled way, molecule by molecule carried in by molecular mills
that are the equivalent of the molecular transport channels in a cell
wall. Once admitted (and generally attached to a carrier molecule) the
molecules are carried to where they are required in an orderlye
fashion, delivered to the right place at the right time. As new
elements are created during the process of reproduction the network of
conveyors may be extended to supply them. There are two options.
Having generated a complete copy of all its active components
internally the egg must either grow or fission. If your end product is
a macroscopic artefact rather than just a chemical growth seems the
most sensible option (although I am concerned about the relatively
lower surface area of larger eggs). To manufacture your end product
all the assemblers are going to need to have a fixed spatial
relationship so why not let the replicators hang onto their parents in
the first place rather than being allowed to scatter into the fluid medium?

To put it simply the environment available in the kind of vat system
that Eric Drexler envisaged in "Engines" is not good enough and
requires that the replicators generate their own micro-environment
within it. That being done global environmental requirements are
relaxed and you hothouse begins to look like a jam-jar.



Malcolm McMahon