Re: [sl4] The Jaguar Supercomputer

From: Matt Paul (lizardblue@gmail.com)
Date: Mon Nov 23 2009 - 18:41:26 MST


Ok, this is probably gonna get me banned...

I've been following SL4 for a while now. The discussions are
certainly intellectually stimulating in a "university" sense, but what
I still don't get is what exactly the perceived value of the AI you
guys discuss is beyond normal scientific desire to understand. I don't
see the practical and prudent value of a machine that acts like a
human brain. Fascinating and cool certainly, but I don't see the
actual benefits to mankind. I do see many potential problems for
mankind though...

Rather than flame me for these statements, please answer my question.
I honestly am trying to understand the subject better.

Lizardblue

On Nov 23, 2009, at 5:58 PM, Matt Mahoney <matmahoney@yahoo.com> wrote:

> Mike Dougherty wrote:
> > How much complexity is in the genetic space for the development of
> a brain? ex: total for all genetic encoding minus the parts about
> the bloody viscera and redundant so-called junk-DNA =?= less than
> 10^9 Like you said, nature just needs *a* brain - not an exact
> replica of a particular brain.
>
> That's for a baby brain. To model an adult brain, you have to add to
> that all the information learned since then.
>
>
> -- Matt Mahoney, matmahoney@yahoo.com
>
>
> From: Mike Dougherty <msd001@gmail.com>
> To: sl4@sl4.org
> Sent: Mon, November 23, 2009 6:15:30 PM
> Subject: Re: [sl4] The Jaguar Supercomputer
>
> On Mon, Nov 23, 2009 at 11:20 AM, Matt Mahoney
> <matmahoney@yahoo.com> wrote:
> When I say that the brain has 10^9 bits of memory, I mean its
> Kolmogorov complexity. There are 2^(10^9) possible brains you can
> distinguish by their behavior. (It happens to take 10^15 synapses to
> achieve that, however). So the Kolmogorov complexity of the desired
> outputs for any traning set also has to be at least 10^9 bits, or
> else there would be some brains that can't be distinguished.
>
> If your goal is to produce *a* brain (say, to pass the Turing test),
> and not a copy of some particular brain, then I suppose you could
> get by with less.
>
>
> How much complexity is in the genetic space for the development of a
> brain? ex: total for all genetic encoding minus the parts about the
> bloody viscera and redundant so-called junk-DNA =?= less than 10^9
> Like you said, nature just needs *a* brain - not an exact replica of
> a particular brain.



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