> > Mark Gubrud wrote about the construction of a brain-equivalent computer.
On Wed, 29 Mar 2000, Eliezer S. Yudkowsky wrote:
> I don't know if you're aware of this, but the atoms within the brain do
> not remain constant. I forget what the exact rate of turnover is, but
> something like 50% of the atoms get replaced every six months. So
> whatever "we" are, it's clearly not the atoms.
This is an oversimplification of the details. Given the complex physical
structure of the brain, there is a variability in the turnover rates.
The proteins in the liver turnover at something like a 1% per hour
rate (intracellularly). That goes up to ~2%/hour when people are on
caloric restriction. In the brain, you probably get similar rates
of turnover in the proteins around the nucleus. Proteins and structures
(e.g. mitochondria) in the axons or at the terminals would turnover
on the order of days because it takes them that long to get transported
up and down the axons. I'm unsure of whether there is any proteosome
activity (breaking down proteins into amino acids) in the regions of
the synapses (Anders might comment). The extracellular matrix and
the lipids in the axons and cell body turn over much more slowly
if at all.
It may not be strictly true that the atoms get replaced (if Eliezer
has references on this I'd like to see them). I doubt for example
that many of the atoms in the DNA get replaced. It is certainly true
that they get recycled and there may be some flow of the basic building
blocks into and out of the brain.
> If you scan in one neuron at a time, destroy the old neuron, and replace
> it with a robotic interface to equivalent computations taking place
> elsewhere, you can migrate, one neuron at a time, to the new substrate,
> without ever losing consciousness. In this case, would it not be
> possible to say that the new individual is unambiguously "you", with the
> change of physical substrate as irrelevant as the turnover among your
> constituent amino acids?
This is the Moravec/Minsky replacement approach introduced in the
1980s. Being conservative it should be noted that it is not clear
that this is entirely feasible. I've recently confirmed that
nanobots are of the order of the same size as synapses and Nanomedicine
suggests that they should be able to sense and modulate neurotransmitters.
They should also be able to monitor the action potential spikes in
the neurons. However, until further work is done on this and someone
comes up with a real design for doing it, this remains a hypothetical
The following is a great quote, so I thank E.Y. for sharing it:
> "But I am not an object. I am not a noun, I am an adjective. I am the
> way matter behaves when it is organized in a John K Clark-ish way. At
> the present time only one chunk of matter in the universe behaves that
> way; someday that could change."
> -- John K Clark
The emphasis in the statement should be on the "could".
My personal opinion is that it should be "will likely".
This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:06:46 MDT