RE: Review of Novamente

From: Ben Goertzel (ben@goertzel.org)
Date: Sun May 19 2002 - 20:08:10 MDT


This conversation seems to go around and around in circles....

I think one way to make more progress might be if you would to the
following.

Write a 2-page document summarizing, as clearly and in language *as
theory-independent as possible*, the high-level structures and dynamics that
you think any AGI system has to support.

Vague things like "memory" are not very useful, since obviously I think
Novamente supports "memory" very well.

On the other hand I find the DGI paper is not so useful as a guide in this
regard either, because it mixes up

a) your view of the structures and dynamics that you think need to be
supported
b) your particular theories about how these structures and dynamics come out
of lower-level phenomena

Of course, some such mixing-up is unavoidable, but I think you could
probably make a much briefer "high-level requirements document" that
contained less of b) and more of a).

This would be very interesting to me and probably some others on this list,
and I can't imagine it would take much more time than typing long e-mail
responses ;>

ben g



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:39 MDT