From: "Nick Arnett" <[EMAIL PROTECTED]>

> > Well, what do you mean by processing at the same bit rate then? It's
> > been a long time since I made the calculations that came up with 2020-5
> > as the time when our computers are powerful enough to simulate a mind,
> > but I definitely took into account having to handle the many
> > connections between neurons.
>
> Still, having the connections isn't enough, as I'm sure we'd all agree.
But
> do we really know what more is required?  To draw on my earlier metaphor,
> how much effort would have to go into creating appropriate data structures
> and paths?  Can we even know until we do it?

The real problem is creating the framework for a brain in which data can be
accepted and processed. Really, it's only as hard as creating the brain in a
baby, which does seem quite difficult but in reality is probably only
regulated by a few hundred genes at most. These would control the growth,
direction and suicide of various neurones according to an array of trophic
factors and cell surface proteins. It's a dynamic process, sure, but it's
certainly not mysterious and I think that given another couple of decades
we'll have it figured out fairly well.

In other words, it's not the structure that matters as much as the rules and
guidelines that are invoked in creating that structure.

This is of course assuming that the human brain is the only way in which to
create an artificial intelligence that can communicate with humans.

Adrian

========================================================================
Adrian Hon                          |  www.vavatch.co.uk
Generation Mars and New Mars        |  www.genmars.com - www.newmars.com
Astrobiology: The Living Universe   |  www.ibiblio.org/astrobiology

Reply via email to