Joys of Flesh

John K Clark (
Sat, 14 Sep 1996 20:46:42 -0700 (PDT)


On Thu, 12 Sep 1996 Eugene Leitl <> Wrote:

>A DRAM cell needs 1 Transistor + 1 capacitor

In A CMOS transistor the voltage on the gate controls the current through
the device, the rod logic equivalent is the interlock, the movement of a
rod is dependent on the position of a control rod. Drexler's interlocks
are just two pyridazine rings, only 20 atoms for both. A CMOS logic gate
must have a large capacitor and that slows the signal down, it takes eons
for the capacitor to charge up, A rod logic gate does not need anything like
a capacitor so it has much less signal delay. Drexler is designing for
10^17 logic elements per cubic centimeter, I can't see why he can't make it.

>>in 1988 when Moravec wrote his book it was
>>reasonable to think that if you multiplied 10 bits
>>by the number of synapses in the brain you could
>>get a good estimate of the storage capacity of the
>>brain. It is no longer a reasonable assumption


Because after the January 28 1994 issue of Science by Madison and Schuman
we know that the brain has a huge amount of redundancy, LTP spreads out.
If you change the strength of one synapse you also change the strength of
thousands of other ones, so they can't hold independent data. It's not
like new values are being recalculated, the synapses are being crudely
changed by the diffusion of nitric oxide produced by the original synapse.

>Can you expand a bit which energetic gains can be expected
>from them? [reversible logic circuits]

Consider a computer with the smallest possible memory, just one bit. The
computer could be in two states, zero and one. Now record something into the
computers memory, for example one. You have reduced the states the machine
can be in, from 2 to 1 in this case, and because the entropy of an object is
the logarithm of the number of ways the parts of the object can be rearranged
without changing its macroscopic attributes, that means you have reduced the
Entropy of the machine too.

According to the second law of Thermodynamics you can locally reduce the
Entropy of something but it takes energy to do it. The absolute minimum
energy it takes to erase one bit of information and record something different
in its place is ln(2)kT , k is Boltzmann's constant 1.381 X10^-23 J/K, and
T is the temperature of the computer in degrees Kelvin. This is not a lot of
energy by everyday standards, but it is free energy that must be dissipated
as heat if you want to erase one bit of information. With reversible
computing, that is where the output uniquely determines the input, nothing is
erased in computation so you don't have this energy loss and a logical
operation can be performed with an amount of energy that is arbitrarily close
to zero.

None of this is an important consideration to chip designers, today's logic
circuits are so huge, and they deal in such a tiny amount of information,
that the effect is utterly trivial. Chip designers can safely ignore this,
Nanotechnology engineers can not. In the early days there were some
embarrassing incidences where people thought they were designing a nano
computer, but when you looked closely at it and the amount of heat the thing
would give off if operated at the design speed, you realized that what they
were really designing was a first rate high explosive. They do a lot better

>>There is not the slightest doubt that quantum
>>cryptography works, and not just in the lab.
>>Recently two banks in Switzerland exchanged

>Yes, but what does it offer, and which costs? Which advances
>does it have in relation to vanilla cryptography? What is
>with triggered (by stimulated emission) photon message
>spoofing, can it get detected in principle?

A message can be detected but NOT without you knowing it was detected, the
idea is to use Quantum Cryptography to send a One Time Pad of numbers to the
person you want to talk to. With a random One Time Pad you can generate a
perfect encryption scheme, even if to everybody's surprise it turns out that
P=NP, even if your opponent has a quantum computer, even if your opponent has
a computer of infinite power, he can never read your message. Despite its
perfect security it has severe practical limitations, how do you distribute
the One Time Pad to the person you want to talk with? You can't send the pad
electronically, if your electronic channel is not secure then an eavesdropper
can tap your line and get a copy of the pad, if your channel is already
secure then you don't need the pad. The only secure method is to physically
hand a disk with the pad on it to the person you want to talk to and then
hand him another one when that one gets used up. That's not practical in most
cases, certainly not for Crypto Anarchy in an economy based on anonymous
electronic money.

Public Key Cryptography solves the key distribution problem but at a (small)
price, it is breakable in theory but probably not in practice. In a year or
two it will be expensive but possible to break a 512 bit key, a 1024 bit key
would be astronomically more difficult. Many users of PGP have a 2048 bit key,
even with Nanotechnology there is not enough matter in the observable universe
to make a machine that can break a key of that size before the heat death of
the universe, or the Omega point whichever comes first. We can be absolutely
certain that a 2048 key will never be broken.. unless..

Unless somebody finds a way to improve existing factoring algorithms
by a lot. An improvement of a measly few hundred trillion percent would by
useless, I said a lot. This doesn't seem very likely, but it has not been
proven to be impossible. There is one other possibility.

Unless Quantum Computers are possible and somebody actually manages to build
one. (Such a machine would revolutionize the world so much that breaking your
PGP key would probably be the last thing on your mind, but let's ignore that
for now.) Would this mean the end of Crypto Anarchy? No, because we can use
quantum mechanics on our side too, we can use Quantum Cryptography, and unlike
Quantum Computers we know this will work because it's already been done.
This is how:

I send you a bunch of photons, each photon is polarized in one of 4
directions, horizontal, vertical, left-diagonal and right diagonal, - | \ / .
In this example I send you 10 photons polarized as follows
| | / - - \ - | - / .

You have a polarization detector, you can set your detector to measure the
horizontal and vertical photons (+) OR you can set it to measure the
left-diagonal and right diagonal photons (x). The laws of physics do not
allow you to measure one photon both ways, because measuring one destroys all
information about the other.

You set your detector at random, let's say you set it to find rectilinear
photons and let's say you have guessed correctly and it really is a
rectilinear photon. If you can detect the photon after it passes through your
polarized material, you will correctly deduce that it is a horizontal photon.
If you can not detect a photon after it hits your polarizing material you
will correctly deduce that the photon is vertical.

What if you guessed incorrectly when you set your detector, what if you set
it to detect a rectilinear photon but I send you a diagonal polarized photon?
Then the photon will hit your polarizing material at a 45 degree angle so
there is a 50% chance the photon will get through, a 50% chance it will not.
In other words you get a random result.

I send you 10 photons polarized as follows | | / - - \ - | - /
At random you set your polarization detector as follows x + + x x x + x + +
So you might claim the photons were polarized as follows / | - \ / \ - / - |

Now you tell me over an insecure channel how you set your detector for each
photon, Big Brother is free to listen in, it won't help him. I tell you over
the same channel which settings on you polarization detector were correct, in
this example settings number 2,6,7 and 9 were correct * | * * * \ - * - *
We only use those readings and junk the others, and we agree that horizontal
and right diagonal photon means 1, and vertical and left diagonal means 0.
So we have sent the number 0011 and we can be as certain as we want to be
that there has been no eavesdropping.

An eavesdropper can not know what type of photon is being sent, and just like
you must guess what direction to set his polarization detector. He will be
wrong 50% of the time and when he is he will change the polarization of the
photon and give himself away. We compare N bits in the string of numbers
sent over an insecure channel, if there are no discrepancies then there is
only one chance in 2^N that somebody is eavesdropping, so we can use the
remaining bits as a one time pad. As I said this has already been done and
messages have been sent about 35 miles in this way. I learned about this
stuff mostly from Bruce Schneir's wonderful book "Applied Cryptography".

>Your other post on QC has been very illuminating. Do you
>think these theoretical values are achievable in reality,
>without assuming unrealistically high demands on fabrication
>precision (structure geometry, alignment, etc?).

I honestly don't know. A few years ago I would have laughed at the idea
because it seemed that such a machine would need to operate with absolute
precision and be completely isolated from the environment, and that is a
ridiculous idea. When they started to find quantum error correcting codes I
stopped laughing. Sometimes I think that at the last minute we'll find some
law of Physics that will prevent a bizarrely counter intuitive object like a
Quantum Computer from existing in the real world. On the other hand, the
results from the last couple of years seem to point in the opposite direction.
We may know in a few years if this is really going to work or if we'll just
have to settle for old fashioned Nanotechnology.

Just how bizarre would this gadget be? If a Quantum Computer can produce
meaningful output information from examining the quantum mechanical state of
its memory, then there is no reason you couldn't run the machine in reverse,
feed in information and change the quantum mechanical state of the memory in
a meaningful way. In other words, a quantum mechanical matter manipulator that
works at the atomic level or smaller. Yes I know its crazy but don't blame me,
I didn't invent Quantum Mechanics.

>>Just one cubic nanometer of diamond contains
>>exactly 176 carbon atoms.

>Yes, but how many operations/s can you gain from these atoms?
>How many bits can be stored in this volume? One bit?
>(probably much less).

Nanotechnology by definition means the ability to control atoms at the atomic
scale, if you can't store 176 bits in 176 atoms then you don't have
Nanotechnology. I should add that there is no theoretical reason that you
can't store more than one bit of information in one atom, a lot more, but
Drexler is an old fashioned conservative sort and he makes no use of this
fact in any of his designs.

How big would a rod logic nanocomputer need to be to equal the power of a
modern mainframe? Drexler figures it would need to have about 150 billion
atoms, a cube 1000 nanometers on a edge would do it, that's about as big as a
mitochondrion. You could fit well over a thousand of such mainframes into a
cell of average size.

>Apropos picotech, see the post on femtotech I forwarded to
> the >H list ;)

I remember hearing speculation about femtomachines that operated at a scale
of 10^-15m . The idea was to build things with strangelet quark structures.
Unfortunately I don't see how this could work, objects would not be rigid,
and you can't build machines with a liquid. There is little more I can say on
that subject because unlike nanotechnology, building things at this scale
would require a scientific breakthrough and they are inherently unpredictable.
By the way, it's not very clear to anybody if a scientific breakthrough is
needed to make a Quantum Computer or not.

>Mechanosynthesis has not been shown to work.

We can do chemistry, and not just physics on atoms one at a time. In the
June 16 1995 issue of Science it's shown that if electrons of the correct
energy are shot at an atom from the tip of a scanning tunneling microscope
the atom will resonate and the resulting vibration will break the chemical
bond. The procedure is somewhat faster than expected and does not require any
exotic conditions such as very low temperature. J.W.Lyding, one of the authors
of this report, is quoted as saying " We'd like to make small, electronic
devices on the nanometer scale".

>This is a problem. The defect threshold in the clone
>structure must not have _any_ negative impact on positioning
>tip accuracy.

Some error is acceptable because it would not be cumulative, such is the
advantage of working in the digital domain, and Nanotechnology is certainly
digital. You may need to build a rod 7 carbon atoms long, but you will never
need to worry about a rod 7.01764513 carbon atoms long.

>The machine needs to build structures from the outside,
>not from within.

Why? Neurons are machines and they can do it, but I like the idea of pre
wiring lots of connections in 3 dimensions and then changing their weight.

John K Clark

Version: 2.6.i