Re: [sl4] Uploading

From: Stuart Armstrong (dragondreaming@googlemail.com)
Date: Fri Dec 04 2009 - 03:38:08 MST


>> > I assume that you are referring to the no clone theorem, which says
>> > that it is impossible to perfectly duplicate a given quantum state.
>> > However, this ultimate level of copying fidelity cannot be fundamental
>> > to identity,
>
>> "Cannot" is a very strong word.
>
> Yes but no stronger than accuracy demands.
>
>> "Very unlikely" is more apporpriate.
>
> Show me how it could possibly be untrue.

First of all, I don't have to. You are claiming an impossibility
theorem - quantum effects have no significant effect on consciousness,
with probability one - so the proof here is for you to supply. All I'm
claiming is that there is not enough evidence for probability one.

But I'll still try and suggest how it could be untrue, anyway.

Second, if AI is possible on traditional computers (which I believe,
with probability 0.999), and quantum computing is possible
(probability 0.99999), then you can implement an AI on a quantum
computer. You can even do so in such a way that the data of
consiousness, the connections to use, is stored in supperimposed form.
In this AI, erasing the quantum information would kill it. So there is
nothing wrong in principle with an consciouss dependent on quantum
effects.

Thirdly, appealing to chaos theory, you can get great macroscopic
stability in the outcomes of certain differential equations. It
therefore seems pretty likely that you could implement the AI above in
such a way that it is resilient to macroscopic shocks. Even if that is
not possible, simple redundancy would accomplish the same thing.

Fourthly, our brain is a mess, cobbled together according to what
worked at the time, and there is no such thing as "turning our brains
off" or erasing the quantum effects and starting it up again.

So to sumarise - quantum dependent resilient consciousness is
possible, our brain has structures close to the quantum scale,
evolution would have used quantum effects if these were easily
accessible to it, and we have not seen what erasing these qunatum
effects does to the brain. So a retreat from perfect certainty seems
justified.

Stuart



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:05 MDT