From: Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Tue Jul 17 2001 - 05:12:46 MDT
On Mon, 16 Jul 2001, Samantha Atkins wrote:
> add to that at this time. I hope someone talks you out of this
> as I can scarcely think of a more crippled way to attempt to
> write and maintains something as sophisticated as an SI.
What many people don't seem to understand, that an SI seed is going to be
a number crunching project, of the embarrassingly parallel kind, and a
rather impressive hardware hunger even considered decade-away hardware.
Thinking doesn't come for free. Thinking good enough to be able to rewrite
yourself is anything but shallow thinking. Theorem proving and chess
playing is cheap, this is not.
What I see is a high-processivity crunch engine (to be eventually
implemented as an ASIC), more or less regular data access pattern, a high
memory throughput (taxing even embedded memory), lots of nodes, low
latency/high-bandwidth interconnect, very tight local coupling, less tight
global coupling.
And of course it's going to be darwinian, and of course you'll need a
high-speed reality emulator for behaviour evaluation, and several of them
critters coevolving in a box.
Monkeys are smart enough to do that, barely. It is not a trivial project.
-- Eugen* Leitl leitl
______________________________________________________________
ICBMTO : N48 10'07'' E011 33'53'' http://www.lrz.de/~ui22204
57F9CFD3: ED90 0433 EB74 E4A9 537F CFF5 86E7 629B 57F9 CFD3
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:08:50 MST