[ExI] Kurzweil's latest

[Anders Sandberg]

On 02/05/2013 06:59, Kelly Anderson wrote:
> On Mon, Apr 29, 2013 at 3:51 AM, Gordon <gts_2000 at yahoo.com 
> <mailto:gts_2000 at yahoo.com>> wrote:
>
>     We need to understand the brain before we can hope to duplicate
>     one. It's pretty naive, I think, to suppose that we can create one
>     even before we know how they work.
>
>
> Agreed. Have you read Kurzweil's most recent book? What do you think 
> of it?

I read it. I think it is surprisingly mainstream: many neuroscientists 
(computational  or not) believe roughly in something similar. What many 
do not believe in, and I think Ray does, is that the system is as clean 
as described. In particular, he is a "cortical chauvinist", believing 
that the cortex is mostly what matters - but the subcortical systems is 
what sets our goals, style of learning and evaluations.

His philosophy of mind chapter is light - the recent email discussions 
has, despite much heat, delved into many issues he just saunters past. I 
was also dismayed by his very simplistic take on how to set AI motivations.

All in all, my view is that it is a readable popular introduction to 
connectionist style models of the brain.

-- 
Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20130502/e1fd1d56/attachment.html>