From: ghio@temp0181.myriad.ml.org (Matthew Ghio)
Date: Sun, 18 Jan 1998 00:49:10 +0800
To: cypherpunks@cyberpass.net
Subject: Re: Nanotechnology
In-Reply-To: <199801162110.WAA17429@basement.replay.com>
Message-ID: <199801171645.LAA01243@myriad>
MIME-Version: 1.0
Content-Type: text/plain



nobody@replay.com (Anonymous) wrote:

> I think people here seriously underestimate how long nanotech will take.
> Suppose you have a machine which takes atoms and deposits them onto a
> surface, building it up layer by layer.  This is possible with today's
> technology, but it's a slow process.
> 
> Consider this: If someone builds a replicator which takes a week to make
> a copy of itself, every person on this entire planet could have one within
> eight months.  Think about it.


Let's consider a rough estimate of whether this could actually work.

Suppose you have a desktop fab with an ion gun that can put out 1 amp of
current, housed in a 10-centimeter cube, and you want to build another
cube.  Assuming an average ionization energy of 10 eV and supposing this
achieved an energy efficiency of 10% that'd be 100 watts.

1 coulomb/second = 6.24 x 10^18 atoms/sec.  Assuming an atomic spacing
of roughly 10^-10 meters, at that rate it'd take you about 18.5 days to
construct one side of the cube with a thickness of 1mm, or almost four
months to build another cube.  I'd guesstimate another 4 months to build
the electronics for it.  And then you'd need some way to assemble the
pieces.

So roughly 8 months just to build one copy, rather than to achieve world
domination. :)

Still, unless I'm way off on these estimates, it's within the right
ballpark.  If the cubes were 1 cm, you could make copies in less than a
day, assuming it didn't get too hot at that power level.  Just figure
out how you're going to feed the electricity and raw materials into all
those little things...

This sort of replicator is not what is usually considered nanotechnology,
but if it actually worked, such a device could become quite popular.