[ExI] Human memory was Looking for a word

[Keith Henson]

On Fri, Jun 25, 2010 at 12:31 AM,  Ross Evans <ross.evans11 at gmail.com> wrote:
>
> On Thu, Jun 24, 2010 at 11:31 PM, Keith Henson <hkeithhenson at gmail.com>wrote:

snip

>> Considered as a computer, human brains tend to be long in processing
>> power and short on memory.

> This isn't really true, and it demonstrates how using a digital computer as
> an analogus framework for the human brain, can lead to erroneous
> conclusions.

Always happy for the state of my knowledge to be improved.

> The human brain both processes and stores vast amounts of
> information.

As I understand it, to simulate human brain processing would take tens
of TF while a lifetime of information at a few bits per second adds up
to less than 200 M bytes.

It's been well over a decade since I had a computer with less than a
200 M byte disk.  The current desktop computer might run a few GF and
a and a few hundred GB of disk.

So to match this a ten TF brain simulation should have access to ten
TB of storage rather than 200 M byte.

> The reason why the brain is not so wired that the average
> person can recite pi to a few thousands digits, is simply the lack of the
> necessary evolutionary pressure. To proffer a computing analogy that is
> germane, the hardware of the brain is very poorly utilised by its software.

How would you suggest better utilizing the hardware?

Keith