From: Billy Brown (bbrown@transcient.com)
Date: Sat Feb 26 2000 - 16:48:19 MST
Eugene Leitl wrote:
> Let me try again. I have a couple of motley colored designer gummi
> bears, which I drop into a big beaker of cold water.
<snip>
> Now I bring the solution to the boil. Soon after I have a homogenous
> gelatine solution, all traces about the gummibear's original shape and
> coloration erased. Now time stops in the beaker. Freeeeeze. 256 years
> later, time resumes, and the system is given advanced (but not
> Omega-grade) future intelligences for analysis and reconstruction.
>
> Which information is still present about the shape and distribution of
> coloration of the original gummi bears? Would the type of technique
> used for reconstruction influence the result a lot?
Well, a good reconstruction could tell you the exact mass of each flavor of
gummy bear that was originally present, but that's about it. This seems to
be a very close analogy to the case of a brain being dissolved in acid.
However:
Suppose, instead, that we subject the gummy bears to a different series of
insults. First we expose them to a series of reactants that break down the
dyes that give them their color. Then we cool them to 0 C, neatly dice them
into lots of tiny gummy bits, and preserve the resulting pile of discolored
fragments in liquid nitrogen. Can we reconstruct the original shape and
color of the gummy bears?
If we don't know that it started out as a pile of gummy bears, probably not.
If, however, we know what happened, it isn't even all that hard. You can
tell the color of a gummy bear fragment by recognizing the reaction products
left over from its de-coloring, and you can fit the pieces back together
based on their shape.
Now, I think that the case of cryopreservation is a lot closer to my example
than yours. Typical slides of preserved tissue show a bewildering jumble of
mixed-up parts, not a homogeneous sludge. In this kind of situation the
quality of your repair techniques has a very large impact on your ability to
do a reconstruction.
> This is not about pessimism at all. You aren't an engineer, are you?
I'm a software engineer - and yes, that is a very different thing than being
a programmer. But that's another thread.
> If I use a true random generator (or a pseudorandom generator unknown
> to you) to scramble the picture, information is lost. Period.
No kidding. I've said that several times myself. There's no need to keep
beating a dead horse.
> You are surely aware that the position you question are held by the
> only specialists in the matter there are? There are no better ones
> yet. What are your credentials in the matter? Programmer, aren't you?
Ah, the argument from authority. Does this mean you don't have actual data
on hand?
There are no specialists on brain reconstruction. Neurobiology,
conventional medicine and cryopreservation are all only tangentially
related. Asking a specialist in one of these fields about reconstruction
techniques is like asking a paint-and-canvas artist about digital image
processing. Heck, most 'experts' still seem to think that clinical death
and the capacity for spontaneous recovery are meaningful yardsticks here.
The only group that even tries to understand the real situation are the few
pioneers who do cryopreservation research for the cryonics organizations,
and they haven't had the time, funding or manpower to create the kind of
well-supported body of experience that would allow them to become a real
specialty. Besides, the last I heard they were cautiously optimistic except
when they're asking for money.
Meanwhile, the people whose opinions I most respect in this regard
unanimously hold the same position I do. The only difference between my
claim and the comments that Drexler, Merkle et. al. make is that I left out
the minimal qualifiers that scientists habitually attach to any statement of
fact.
> So we assume a best case as default. No further research needs to be
> done, because the information Is Still In There, Somehow.
No, we take past experience as our guide. Finding ways to computationally
unscramble mangled structures is so common that we should expect it to
happen as a matter of course whenever someone decides to devote some real
effort to any given problem. If you don't think it can be done in this
particular case, you need to actually show that the necessary information
does not exist.
Which brings me back to the point you've been ignoring. The fact that a
structure 'appears' scrambled says nothing about its actual information
content. The fact that cremation and long-term decomposition cause
irreversible damage is easy to demonstrate, but the same is not true of less
severe insults that leave substantial amounts of structure intact. So, do
you have any actual experimental evidence regarding the degree of
information loss caused by various forms of trauma? Failing that, do you
have a good logical argument for believing that the information loss is
catastrophic?
> > 2) Every repair scheme I've ever heard of relies on massive
> computing power
>
> ^^^^^^^^^^^^^^^^^^^^^^^
> > and very limited intelligence. This is totally unrealistic. We should
> ^^^^^^^^^^^^^^^^^^^^^^^^^
> This sentence makes no sense. It is also pure fabrication.
You need to calm down and start actually reading me before you reply. Look
at Drexler's scenario in Engines of Creation for a typical example of what
I'm talking about. You've got a bazillion MIPS of computing power being
used to run something that sounds like a mating between a molecular CAD
program and an expert system. I understand why he didn't want to spend
another 50 pages getting the reader fully up to speed on what is really
possible, but his brief references to engineering AI go right past most
people without ever making an impression.
Certainly, there are people around who know better. Drexler is one of them,
and so are most of the other big names in nanotech. But most people (even a
lot of extropian and transhumanist types) are still thinking in terms of a
program that shuffles atoms using algorithms that any good programmer could
implement.
That is not even remotely realistic. What they should be imagining is
something more like the FBI working over an especially important crime
scene, only with even more layers of specialized knowledge and processing.
The degree of expertise that can be applied to solving each particular
sub-problem (membrane reassembly, oxidation reversal, etc.) will grow
steadily over time, and it is difficult to say exactly where it will reach
its limits.
> I don't see many problems in getting frozen, a number of precautions
> assuming. I see a lot more problems in staying frozen, mostly because
> of above attitudes. "What, me worry?" hits too close to the mark for
> comfort.
What on Earth does an argument about the quality of current suspensions have
to do with the long-term stability of cryonics organizations?
> > nonsense. Most cryonicists do in fact end up being suspended when they
die.
> > So, people shouldn't use this as an excuse for dismissing cryonics - if
you
>
> Nobody is dismissing anything. You once again missed what this is
> all about.
Actually, if you look back on this thread, you'll find that my original
remarks were directed against a post that argued this very point -
retroman@turbont.net was claiming that the odds of not getting suspended at
all were high enough to constitute a valid reason for not bothering to sign
up.
Billy Brown
bbrown@transcient.com
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:27:03 MST