From: Toby Weston (email@example.com)
Date: Wed Mar 04 2009 - 16:32:21 MST
I was not linking levels of processing power with conciousness.
I was maintaining that there could be a dumb procedural non recursive program that appears conscious through brute force... But is not. Like a rock.
I am. A rock is not.
On 4 Mar 2009, at 16:33, Matt Mahoney <firstname.lastname@example.org> wrote:
--- On Wed, 3/4/09, Toby Weston <email@example.com> wrote:
If there is no detectable difference, then there is no
sense in asking the question.
Not ok. This Dennet, Blckmore take on consciousness is
lacking in that it does't help me understand the origin
of subjective experience, whether my motives are purely
philosophical or engineering oriented.
In fact the statement that conciousness does not add
anything is itself pointless: it helps understanding not one
Perhaps the following will aid in understanding. If I hard wired your brain to connect the "2+2" neurons to your "5" neurons, you would not ask "what is 2+2?" You would ask "why does 2+2=5?"
A super inteligent, but not conscious, AI recieves an
upload. The company that owns the AI demos the newly
uploaded person to it's family. It passes all sorts of
tests - Turing style and White box - the audience concluded
that this really is uncle Bert. It does not have a body, it
is not a zombie, it is a set of responses that has fooled
the room. It may require 10x the processing power of a human
brain to run. In this example consciousness is a performance
optimization allowing complex behaviour to run on smaller
Uncle Bert is not conscious - this is not irrelevant for
people considering the procedure.
So lets say that the upload service figures out an optimization that allows Bert to be simulated and pass all the tests as before with 1/10 the computing power. Does that make it conscious? Or are you saying that's impossible because currently the brain does 10% of your thinking and your consciousness (by some mysterious process) does the other 90%?
Or is it possible that 2+2=4?
-- Matt Mahoney, firstname.lastname@example.org
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:04 MDT