[ExI] The symbol grounding problem in strong AI

[Eugen Leitl]

On Fri, Dec 18, 2009 at 09:58:34AM +0000, BillK wrote:

> Because it hasn't been done yet.

Nothing is done until it's done.
 
> And I think that our present computing systems will probably have

Our present capabilities are quite impressive. A 300 or 400 mm
wafer is a lot of real estate. Especially if you approach ~nm
device geometries. What you do with that potential is of course
up to you.

> great difficulty in achieving it. (Though they might achieve
> reasonable simulations of it).

So chess computers don't play chess, they only simulate playing chess.
 
> Once systems with thousands of processors and multi-streamed logic
> become common, we're in a different ball-game. But it will take new

They are common. They are called clusters. We're in meganode country
now.

> programming techniques and new software to fully utilize these
> systems.

You don't need anything beyond MPI, a lot of cores on a 3d lattice 
signalling mesh and a decent morphogenetic darwinian system. I could
have said that 20 years ago. Wait, I did.
 
> This problem is one of those that will be overtaken by circumstances.
> Once it has been done, everyone will say it is obvious and a
> non-problem.

Positive feedback enhancement runaway by *real* personal computers
is obvious and a non-problem?

-- 
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