[FoRK] Scalability of intelligence

Eugen Leitl eugen at leitl.org
Thu Aug 5 02:46:48 PDT 2004


On Wed, Aug 04, 2004 at 10:29:45PM -0700, J. Andrew Rogers wrote:

> Things like human language seem to be around a=24 level algorithm 
> spaces.

It doesn't require intelligence to increase intelligence. Darwinian evolution
isn't an intelligent process, yet it produced us. Of course, given hints like

	http://today.uci.edu/news/release_detail.asp?key=1187

it should be possible to figure out the architecture of the neuroanatomy, and
reverse-engineer it.
 
> I should add that I generally do not believe a genuine "hard take-off" 
> is possible in the absence of proper general purpose MNT, and I think 

You don't need general purpose MNT, only means of fabbing molecular
electronics in bulk cheaply enough to prototype in low submegabuck range
(preferrably, a lot cheaper, as it increases the number of hardware
developers).

Current fabs would do, if there was enough money to burn on building
WSI-scale cellular architectures, including an interwafer mesh 
signalling fabric.

> it currently looks like we'll have general purpose AI before we have 
> general purpose MNT (obviously arguable).

If you have enough crunch to do human-grade AI, letting them figuring out machine-phase
chemistry, viable self-rep designs and a bootstrap route largely in machina
is very small beer in comparison.

People could do it with today's technology, but it would be a very
large-scale effort, and take several decades.

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a>
______________________________________________________________
ICBM: 48.07078, 11.61144            http://www.leitl.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE
http://moleculardevices.org         http://nanomachines.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
Url : http://lair.xent.com/pipermail/fork/attachments/20040805/c826c938/attachment.pgp


More information about the FoRK mailing list