From: Ben Goertzel (ben@goertzel.org)
Date: Wed Feb 01 2006 - 10:23:31 MST
On 1/29/06, P K <kpete1@hotmail.com> wrote:
> There seems to be some talk about AGI programmers having to be good
> learners. We seem to be forgetting one little detail; AGI hasn't been
> INVENTED yet. AGI programmers would have to invent AGI using new and
> creative approaches.
I have a different perspective, since I believe I have a valid design
for an AGI.
I don't need supergeniuses to help me build and teach an AGI according
to my design, merely very smart people with a mastery of C++
programming and a reasonably strong grounding in a few areas such as
textbook narrow-AI, algorithms and data structures, evolutionary
learning, dynamical systems theory, probability theory and cognitive
science. This set of requirements is not that easy to fulfill in
practice, but it's not so hard as the requirements being discussed on
this list.
-- Ben
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT