[extropy-chat] Dark matter and ET

[Eugen Leitl]

On Wed, Jul 20, 2005 at 02:59:01AM -0700, The Avantguardian wrote:

> Ok, but that is like comparing the space shuttle to
> some wrenches and saying that the space shuttle is
> wasteful because it is bigger than the wrenches. None
> of the "nanoparts" shown do 5% of what the virus does.

No, the virus as depicted doesn't do anything. It needs
a very large volume of water in order to be able to dock
to a lipid bilayer membrane of a monkey cell, and inject 
its contents into it. So you need a whole animal cell,
in even more water, for the virus to do its task. You
need about 1-2 cubic micrometer for a bacterial cell, 
plus medium, for a self-contained self-rep capable 
systems, it's about 2-3 orders of magnitude larger
assembly than the virus itself.

The parts depicted have much higher functionality 
concentration, as they don't need to encode self-assembly
instructions inside the substrate thus diluting the 
desired gadget concentration, as proteins do. They also don't
need solvent to operate (they hate solvent, in fact), 
they're covalently crosslinked so that they're structurally
stiff and can operate at much higher temperatures, etc.

In human design, you have an assembly of exchangeble parts with clean
interfaces, and control code which can be loaded into a new machine
when the old one has degraded beyond repair, and can be recycled
en bloc. Of course, you could also
integrate autorepair along with diagnostics into the part, as biology
does, and actually achieve sustained operation.

The only sustained cell line in biology are the gametes, which
is a shabby sort of immortality (there's also an intensive
selection process occuring there, so very few cells actually
can be considered immortal, and you can't tell which of them
are that in advance). 
 
> > 
> > It is pretty obvious that a functional, addressable
> > 1-bit computational 
> > element can be constructed roughly within the cubic
> > hydrated bilayer 
> > volume. 
> 
> The bases of DNA stack at approx 3.4 Angstrom (.34
> nanometers) from one another. Each base is two bits.
> (dibit?) 

You can't read DNA in dry state, and in fact need
associated proteins and enzymes in order to handle it
without breaking. Also, I already said that it's a pretty
dense form of storage (lattice defects in a crystal can
be much denser though, and of course we could use compression --
try running bzip2 on a human chromosome sequence, you'll
be surprised; there are of course better compression
algorithms especially for organism genomes).

Storage doesn't compute, though. If I want computation,
I need a cell to be able to communicate its state at least
to its direct neighbours, which needs signalling. You also
need a way to power it, and to remove the heat, even if it's
a reversible spintronics device, operating largely in ballistic
regime.
 
> > The storage density of RNA within the virus capsid
> > is sure considerably
> > denser, but it can't compute on its own. It needs
> > ridiculously large volumes of 
> > water with biomachinery to even unfold. 
> 
> True but won't your bilayer need some form of I/O Bus
> to compute? RNA/DNA by itself is more like a harddrive

I've already factored a 3d crossbar of single-walled
nanotube as addressing elements into the estimate of 
a 1-bit computational element.

No cooling yet, though one could use helium through
dedicated bucky lumen (gas molecule collision screw with
electron transport, though, so the whole assembly should
be in UHV and preferrably at a deep temperature -- if
there's an optimal temperature regime for computation
it might be well below 300 K, and have very little direct
thermal blackbody radiation as a signature).

> than a whole computer and pretty useless without the
> rest. But why is that less efficient than a bilayer?

The task of the bilayer is to form compartments, with
good lateral diffusion of building blocks, allowing
their modification and replenishment from integrated
and associated molecular machines (proteins). You need
that in order to be able to form cell organelles, and cells,
and you need cells in order to form an anisotropic excitable
computation medium.

The engineer attempting to build computronium for
AI and emulation of human minds doesn't have to deal with
the evolutionary baggage, and the need to factor self-maintenance
within the system.  As a result, the design is radically
different, and drastically more performant.

I could go on, but my work break is almost over.

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a>
______________________________________________________________
ICBM: 48.07100, 11.36820            http://www.leitl.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20050720/f2331c0f/attachment.bin>