[ExI] The symbol grounding problem in strong AI

[Eugen Leitl]

On Fri, Dec 18, 2009 at 10:41:37AM +0000, BillK wrote:

> Yes, but.....
> See:

Seen. Tilera actually ships better product.

> <http://news.cnet.com/8301-30685_3-10407818-264.html>
> December 2, 2009
> Intel hopes 48-core chip will solve new challenges
> 
> "The machine will be capable of understanding the world around them
> much as humans do," Rattner said. "They will see and hear and probably
> speak and do a number of other things that resemble human-like
> capabilities, and will demand as a result very (powerful) computing
> capability."

To say with JK or P&T: BULLLLLSHIT. It's not a problem of
a particular chipset. Architecture, yes, but with enough scale
you can kill everything with COTS stuff or just ASICs. 
The question is how much money you want to spend. Electricity
is quite expensive.
 
> The Tera-scale project doesn't fundamentally address one of the big
> challenges in today's computing industry, though: getting multicore
> chips to run today's computing jobs that are often designed to run as

Can't be done. Multithreaded is a dead end. Shared-nothing asynchronous
message passing is the only thing that this universe allows. It's
not just a good idea, it's The Law.

> a single thread of instructions rather than independent tasks running
> in parallel.

MPI goes back three decades. Most scientific code today is exactly
"independent tasks running in parallel". In order to make TOP500 you
have to be able to.

> ------------------
> 
> Tera-scale Computing Research Program
> <http://techresearch.intel.com/articles/Tera-Scale/1421.htm>
> 
> >
> > Positive feedback enhancement runaway by *real* personal computers
> >  is obvious and a non-problem?
> >
> 
> 
> Different problem.  *That* problem happens after computers achieve
> self awareness and the ability to self=improve.

It isn't AI if it isn't human-equivalent across the board. "Self awareness"
is meaningless, every embedded today could host a self model, and a meta self
model. Insects do it. Nothing to it.

You don't have AI until the thing can do everything you do. Tomorrow,
it will do better. And half a day after better still. Even today's paltry
analog/digital hybrids can do a day's worth of dynamics in a wall clock second.

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE