X-Message-Number: 1105
Date: Thu, 6 Aug 92 11:32:25 EDT
From:  (Perry E. Metzger)
Subject: CRYONICS: more on LHe temps

>From:  (Keith Lofstrom)
>>     Drexler et al have
>>pretty much shown that thermal jiggling on the atomic scale (not the
>>electron scale) isn't an issue.

>Where is this shown?  I have seen a few of Eric's studies, as regards
>thermal effects on rod logic and assembler machines.  Could you
>direct me to a paper on sensors, or long distance, high bandwidth
>communication?  In the macroscopic world, these are two areas where
>cooling is used to improve performance.  Is it inconcievable that the
>nanotechnological equivalents would stoop to using these
>macrotechnological techniques?

I find it quite concievable that nanotechnological equivalents would
stoop to using these techniques. What I'm denying is the need to get
to low temperatures to do CELL REPAIR, a very specific task that
doesn't need long distance high bandwidth communication (99% of damage
requires coordination only to within a submilimeter distance range,
and 90% of the rest requires communication of no more than 10 or 15
centimeters of distance, with none of it ever even concievably needing
to go more than two meters or so, the maximal length of the body in

>Page 15, EOC, (1986 Doubleday first edition) has a paragraph on thermal
>effects on assemblers.  Chapters 7 and 8 of EOC discuss room temperature
>cell repair.  Chapter 9, "A Door into the Future", has a section called
>"Reversing Biostasis", page 136 to 138.  This section describes repairing
>a biostasis patient at LN2 temperature.  Nothing much changes between the
>77K Eric assumes and the <4K I am assuming except the amount of thermal
>noise, and the choice of working fluid, if any.  Wet chemistry is stopped.

Yes, thats quite true. 

>This section includes such sentences as:

>   "Small devices examine molecules and report their structures and positions
>    to a larger computer within the cell."

Very true. You refer to long distance communications. Most somatic
cells are hardly in the realm where "long distance communications"
becomes an issue. Indeed, the number of bits needed to completely
characterize the average cell must of needs be less than a few billion
given the structure of DNA; one can transmit such data using rod logic
very rapidly; certainly even being grotesquely pessimistic (by many
orders of magnitude) in the tens of seconds range. Tell me, where do
you envision the need for your "high bandwidth long distance"
communication within a somatic cell?

>There is NO natural analog to this process.  Certainly a molecule can be 
>identified with an enzyme surface in a "blind" fashion - just as a book
>could be "read" by comparing each page to all possible permutations of a
>page, then computing which comparisons made the best "fit".  That is a heck
>of a slow way to work.

We are talking about structures that are very very small. Your cells
operate very rapidly by such techniques when needed. Unless you can
give me some DATA showing that the situation isn't analagous, I'm
inclined to disagree.

>In an old patient there will be quite an accumulation
>of damaged - but still functional - proteins that can't just be blindly
>discarded.  These molecular structures will have to be examined and
>modeled ad hoc, to determine which healthy structure should replace

Why? What makes you think this is needed? I'd say that odds are we
won't need to do this, as I can't think of an example; give me some
evidence for why you think it would be needed, preferably by
describing a real protein that we might want to model in situ that we
couldn't recognize through normal techniques and couldn't simply
discard of.

>Some of these molecules will have internal structure not directly accessable
>to enzymatic pattern matching at the surface.

So, just the protein apart and put it back together again afterwards.

>Those internal structures 
>may express themselves only at room temperature.  Do we heat the patient up
>with internally flawed molecules?

No, just take it apart.

>>Sorry. Your cells work fine operating by touch. 

>Mine don't.  The damn things fail after only 3 billion seconds or so. They
>don't compile any of the popular programming languages.  They eat any old
>crap I put in my bloodstream, and sometimes choke on it.  They won't run
>revision 2.0 DNA.  The diagnostic readout is broken.  They cost too much.
>Pretty shoddy - I'm getting mine fixed at the earliest opportunity  ;-)

We aren't talking about future modifications. We are talking about
cell repair.

>Seriously, we are moving from a paradigm of random transport, enzyme matching,
>and wet chemistry, to a one of directed repair, central control, and 
>observation and modeling in a rigid framework.  THE SAME PROCESSES DO NOT

I'm a computer scientist by trade. No central control paradigm is
workable for the body repair problem other than on the grossest of
levels. Most of the work is going to go on at the cellular level. the
fact that we are in a rigid system in which water chemistry doesn't
work doesn't mean that all of the natural paradigms need be abandoned.
In any case, however, all you've done is give platitudes; give some
specific examples of problems IN CELL REPAIR we can't somehow solve
at 130 degrees k that we could somehow solve at 4 degrees k.

>>Why should we believe your claim that we have to go colder in the
>>absense of any evidence at all?

>Why, Perry, you can believe anything you want.  I am putting forth the idea 
>that certain measurements become easier the colder you get.

Doubtless, but you haven't given us an example of a measurement we
would NEED to make for cell repair that we would need to get that cold
to make.

You may very well be right; it may very well be necessary to go that
cold. No one has built one of the things yet. However, before making
highly confident sounding pronouncements that we are going to have to
go to liquid helium temperatures, a concrete example or two would be

>Perhaps they
>are easy enough at 77K, and therefore do not justify the approximately
>1 megajoule needed to cool a 70Kg patient to 4K, or 3 megajoules to 0.001K
>(at Carnot efficiency).  I would think that a 20x or a 70,000x improvement
>in measurement accuracy, with a corresponding 8000x to a 3e14x improvement
>in 3D position modeling, would justify a nickle or a dime's worth of

The issue I brought up wasn't energy expenditure but engineering
complexity. Trying to design nanomachines to operate at temperatures
that cold is going to be a bitch; most nanomachines are going to be
designed, for obvious reasons, to work at room temperature.
Redesigning for liquid helium, where only the most exotic of
structures will operate, is going to be a super-bitch. I'm sure that
if its absolutely needed, it will be done, but just as people don't
bother making kitchenware to submicrometer precision even though we
have the capacity, mostly because its expensive and no one would care,
its very possible that we won't bother cooling to LHe because we don't
need to do it and no one will care.

>What I am having trouble understanding is where all the hostility is coming

Not hostility; an intellectual difference of opinion.

>Perry, you and I are part of a small band of brothers facing a sea of
>superstitious, antagonistic moral midgets who will try to smash either of us
>if given the chance.  I may someday help with your suspension or revival, as
>you may help with mine.  We need each other.  Perhaps I am misinterpreting
>an "East Coast - West Coast" language difference here; if so, let's get it
>worked out.  Let's save our energies for battling the "Monash Menace" :-),
>and BUILDING this stuff, and not fritter it away on doctrinal minutiae.

Oh, I agree that all this is true. I even agree that cooling to LHe
temps MAY be needed. What I don't agree with is your seeming
insistance that it WILL bee needed rather than that it MIGHT be needed
without evidence. Just think of me as a colleague asking you to
radically tighten up a paper before submitting it for publication.


Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=1105