X-Message-Number: 25122
Date: Sun, 21 Nov 2004 08:58:21 -0500
From: Thomas Donaldson <>
Subject: CryoNet #25114 - #25120

To the various Uploaders or those desiring Uploading:

In the very broadest theoretical sense, uploading is possible, However
that statement obscures the gross practical difficulties in doing so.
First, you'll need to produce a computer able to grow new connections
between its "neurons" and also produce and destroy its "neurons". The
memories in our brains consist of our set of neural connections. The
"neurons" themselves would act as small computer, individually not 
with much power. 

I know of computer hardware which allows the designer of a computer
to cut various connections. It's not that far from making hardware
in which a program running on it could change the connections. As 
for creating new "neurons", I know of no such hardware available at
present. However in one major sense discussion of these two problems
fails to discuss one important issue: that of scale. We may well be
able to imitate some very stupid class of worms (insects would be
too smart). Scale makes a big difference, and that problem isn't
so easily solved by any method (at least) known to me. At one time
someone on Cryonet suggested radio connections, for instance. Even
1 million neurons would then require that every neuron be able to
communicate with every other. Yes, you can separate the neurons
so that each could have a straight path to the other; this gets
you into additional problems, since neurons don't send only one
kind of message to one another. (Think of dopamine, serotonin, and
many other chemical transmitters, plus the physical connections 
between neurons. Could you make your signals more complex to match
the complexity of nerve signalling? In theory. And here is another
problem: the kinds of faults and mistakes which happen in our 
brain would not happen in this "computer brain". Therefore the means
to deal with such incidents will have to be quite different ---
which will ultimately make your "computer brain" behave differently
from our own brains. 

Basically, once the number of neurons becomes large enough, growing
new connections becomes a lot more efficient than starting with every
neuron connected to every other. This may  even tell us something
about why our brains have the traits that they have.

And of course our brain is a highly parallel system, something 
ultimately not producible with only one processor. 

If anyone out in Cryonet land believes that these problems can be
solved for human brains BEFORE THEY DIE OF OLD AGE, I'd like to
hear from them.

             Best wishes and long long life to all,

                   Thomas Donaldson

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=25122