[extropy-chat] Computing Power: Moore's Law keeps going and goingand going

Robert Bradbury robert.bradbury at gmail.com
Fri Jan 27 15:58:29 UTC 2006


On 1/26/06, Brett Paatsch <bpaatsch at bigpond.net.au> wrote:

The unit of biology is the cell not the atom. Its the cell that does the
> replicating and any particular human body we have experience of to date has
> developed from cells, not atoms.
>

Let me state this very clearly -- self-replication is *NOT* necessary for
"real nanotechnology" (at least in my book).  Josh Hall wrote a paper 5+
years ago demonstrating quite clearly that nanotechnology assembly lines
could be very fast.  Put another way a specific dedicated assembly process
can be much more efficient than a general purpose assembly process.  This
relates to the "broadcast" architecture for nanorobotic assembly methods --
you do not have to replicate the subcomponents which store and retreive the
information (DNA & RNA in cells).  All that is required is the receiver for
the instruction scheme.  This corresponds someplace between SIMD and MIMD in
computer architectures -- but not general purpose self-replicating
assemblers.

While some people view self-replication as a goal to be achieved (i.e. 'we
have created life') I tend to view it as misdirected energy which could be
better utilized designing nanoscale parts and/or nanofactories for them.  We
already have self-replicating, relatively general purpose assembly systems
(bacteria) which we have been and continue to engineer for various purposes.

The advantage a "cell" has over "soup" is that it allows for reactant
concentration and speeds up the rates at which molecules can be broken down
and/or reassembled.  That becomes largely irrelevant when nanofactories are
fed relatively pure small molecule reactant streams.  (Why do you think I
wrote the paper about the assembly of an aircraft carrier from an oil
pipeline???).

 DNA does not replicate on its own. Ribosomes don't work outside of their
> cellular environments.
>

DNA is replicated using a reactant pool (esp. Primers, DNA bases  & ATP) and
a single molecular machine (one of the various DNA polymerases).  This is
the entire basis of PCR amplification (replication) of DNA.  DNA replication
does not require a cell.  DNA polymerases are molecular machines that I
would guess are around 10-20nm in size (still smaller than current scale
lithography).  Extracellular production of proteins from RNA is also
feasible (I believe you can buy kits for this from the molecuar biology
companies).  Whether all of the components on the kits are produced de novo
or whether some are isolated from cells I do not know.  But the chemistry
and biochemistry for the production of all of the necessary molecular
machines for these processes is well known.  It is simply a cost issue of
whether it is cheaper to do the reactant synthesis from cheap materials or
separate the already built reactants from biological systems.

The *generally* in that sentence is one hell of a hard working word.
>

Actually, it is "biological systems" that may be the point of confusion.  I
was *not* refering to self-replicating biological systems which probably
require a minimal size of 200-300nm, more probably 500+nm (I'll believe
nanobacteria exist when I see the genome sequence for one deposited in a
database).  What I was refering to was biological disassembly or assembly
systems (everything from single enzymes to enzyme complexes to enzyme
systems).  Those I believe fall into the 5-20nm range.  There may be a few
that are somewhat larger but I doubt many are larger than 100nm (though
obviously the protein chain could be longer).

But there are feedback mechanisms in place such that glycolysis operates
> sometimes and gluconeogenesis other times and various GLUT receptors on
> various types of cells are sensing different levels of blood glucose.
> [snip]
>

Yes of course.  I was talking about systems that would circulate *only*
small molecules such as glucose.  Both the kidney and dialysis machines are
examples of systems which allow selective transport of various molecules
from one side of a membranes to another (look up their wikipedia entries or
see [1]).  In fact you could get the same effect by inhibiting the
transporters in the kidney which reabsorb glucose.  The reason that
diabetics have sweet urine is that the blood sugar glucose levels are so
high that they overwhelm the reuptake transporters [2].

But rather than taking a drug to inhibit glucose reuptake it seemed much
cooler to plug yourself into something which uses the glucose for a useful
purpose.  Damien is right about the connection problems of course.  But this
is something that people are working on very intensely because they want to
monitor blood glucose levels and be able to have a computer administer
insulin as required.  There are different routes being pursued towards this
which involve implanted and external solutions which obviously have
different advantages and disadvantages.

Robert

1. http://arbl.cvmbs.colostate.edu/hbooks/molecules/hexose_xport.html

2. http://science.uwaterloo.ca/~mpalmer/MetabolismNotes/page-14.7.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20060127/5cb89847/attachment.html>


More information about the extropy-chat mailing list