X-Message-Number: 9168
Date: Sun, 15 Feb 1998 13:33:38 -0800
From: "Joseph J. Strout" <>
Subject: Re: CryoNet #9163 (hardware for uploading)

In Cryomsg #9163, Thomas Donaldson <> writes:

>I do not see the problem as nearly so well defined. First of all, it's all
>very well to claim that we can build devices of "hard" materials which
>can change their connectivity, but another thing entirely when you start
>seriously thinking how it might be done.

Certainly no harder than imagining nanobots which can resuscitate a frozen
brain.  For example, you could have all your processing units on the inside
of a sphere, signaling to each other via light; since light beams can cross
without interference, any unit could be connected to any other at any time.
This is a ridiculous approach, of course, but as a proof of concept it
works just fine.  In reality, of course, a neuron can't connect to just any
other; only to those in its physical neighborhood, or those within a
neighborhood to which it projects.  (And these gross, long-range projection
patterns probably do not change in adults.)  So the problem is much simpler.

But really, what's the point of arguing this?  Manufactured brains will be
made out of something; you have your prediction, I have mine.  I'll bet you
a dollar (plus 10% per year interest) that they won't be made out of
proteins and lipids.  Now why don't we just wait and see who collects?

>Furthermore, you're not going to
>escape having to deal with biological materials, if only finding a way to
>read out our memories from our (biological) brains.

Yes, but I believe I already sketched out how this could be done: by
imaging frozen sections.  Frozen tissue is relatively easy to work with.
Warm tissue is not.  And though I can perhaps imagine nanomachines building
a frozen brain, that isn't really what you want in your head -- you want a
warm one.

>For that matter, the ability to read memories from a damaged brain comes
>awfully close to an ability to repair it. Just think about that, a bit, in
>comparison with what we can do now.

I think not, for the reason mentioned above: you can get all the
information you want out of a frozen brain.  Perhaps this comes close to
putting information into (i.e., repairing) a frozen brain.  But that still
leaves the big jump from a frozen brain to a warm wet one.

>(Yes, our brains
>contain one or more neural nets, but they work differently from any of the
>neural nets so far made as devices by computer engineers).

The functional properties of networks of spiking neurons (as opposed to the
more common analog neural networks) appear to be very similar (in important
ways) to that of biological neurons -- see the recent work of Wolfgang
Maas, for example.  But that's a different topic...

>It even happens to be true that a lot of work is going on right now to
>find out how to make brains regrow connections.

Yes, I don't suggest that we couldn't grow brains or grow new connections
in old brains.  I only suggest that we can't tell these connections where
to form with enough specificity to recreate a previous, specific pattern of
connectivity.

>Or is it that you just don't like squishy things?

I like them just fine.  Future engineers will build out of whatever
materials are most convenient, and I refuse to tie their hands by requiring
that they use (or don't use) a particular class of materials.  Personally,
I believe that materials originally useful because they evolved to
self-replicate in an ocean will not necessarily be the most useful for
making a configurable information processor.  You believe that the very
same materials will turn out to be the best choice for both applications.
But we can settle our bet sometime next century.

>It seems to me that before we draw such conclusions we
>need at least one example of a real, advanced (not just capable of acts
>a paramecium can do) electronic brain.

One could say the same about the machines purported to cure frostbite.

> Without
>doing that the real brains we have will inevitably lose in comparison to
>the purely theoretical electronic brains some people think up: for no
>other reason than that imaginary constructs inevitably end up superior
>to reality. Dreams are often so much better than reality.

This is exactly my feeling about nanotechnology.  And that's why I'm
comforted to find that full-scale, brain-fixing Nanotechnology may not be
necessary to revive cryonics patients.  Uploading appears to require
improvements in a variety of fields, including microtechnology, computer
engineering, and basic neuroscience -- but it doesn't require
nanotechnology.

>What you really want, Joe, is an ORGANIC COMPUTER.

Maybe, but who are we to say?  I am not a 21st century engineer, and
neither are you.  We can with confidence say that an organic brain would do
the job, if we can just figure out how to build one.  And we can say that
we know how to build electronic (and perhaps optical and quantum)
computers, which could probably do the job.  The clever folks of the next
century will make one approach or the other work, and maybe even both.  I
figure you'll owe me about $13K then (but perhaps I'll owe you!).

And that's basically what it comes down to.  I can't show you a working
brain emulator, and you can't show me how to restart a frozen brain.  We
can speculate about which is easier, but if our speculations differ, we'll
just have to wait and see.

Best regards,
-- Joe

,------------------------------------------------------------------.
|    Joseph J. Strout           Department of Neuroscience, UCSD   |
|               http://www-acs.ucsd.edu/~jstrout/  |
`------------------------------------------------------------------'
        [ Help stop spam: http://www.imc.org/ube-sol.html ]

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=9168