Re: Human mind Turing computable according to Eliezer

From: Anne Marie Tobias (
Date: Fri Oct 15 2004 - 04:07:42 MDT

Hello wei,

I think you're having a different conversation...

Oh, and the answer to your question is to do a high speed comparison
between your current
visual field and the visual cortex of the brains in storage. Since there
is a lovely one to one
correspondence between visual formating and the neural activation, you
should be able to
find the brain in question with all due haste.


Wei Dai wrote:

>On Thu, Oct 07, 2004 at 12:39:19AM -0400, Eliezer Yudkowsky wrote:
>>Mitchell Porter wrote:
>>>Look, do you agree that blue patches, loud noises, etc., exist, or not?
>>>If yes, explain what's blue about a bunch of cortical neurons. If not,
>>>explain how I can interpret my experience without blueness, loudness,
>>>etc. In my world, these things are *epistemically* basic, whether or
>>>not they're ontologically basic.
>>I'm replying to Mitchell about this off-list; we'll see how that goes.
>Why the off-list reply?
>Let me try this approach to the issue, and see if it's helpful. Consider
>the following thought experiment. You're a disembodied brain living in a
>jar, with remote control of a telepresence robot. You just heard that the
>storage warehouse housing your jar is on fire, and rush the robot back
>to save yourself. The bad news is there are thousands of brains in the
>warehouse, and you don't know which one is yours. The good news is there's
>a brain scanning machine with resolution down to individual neurons, and
>you can try to use it to find out which brain is you. Question is, how to
>do it?
>How do you take a bunch of neurons and figure out what it's experiencing?
>Would the answer to this question also answer Mitchell's question, or is
>he asking something else?
>Is this question worth trying to answer? A mind with full access to its
>self-state presumably wouldn't have this problem with determining which
>brain is itself. It would just directly compare the scanner output with
>its own self-inspection. Assuming that we'll get this feature soon,
>is there some other scenario where the answer would be useful?

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:49 MDT