From: Benjamin Goertzel (ben@goertzel.org)
Date: Sat Jul 14 2007 - 20:19:36 MDT
There are many specific things that a person matching the description
you mention could usefully do to move us toward an understanding of
AGI, Friendly AI and the path to Singularity.
A question is whether the person is interested in doing "math theory
stuff" or "computer science experimentation stuff"? I would have
different suggestions regarding the person's taste in this regard.
(Of course, the difference btw these categories is not crisp, but,
doing theory in support of an implementational experiment is different
from doing implementation of simple test software in support of
exploring a theory...)
There is work regarding trying to build a prototype probabilistic
self-modifying AI system.... Then there is work regarding trying to
create a formal theory of such systems and their behavior.... But
describing the work that needs to be done in either of these areas
would get technical...
-- Ben G
On 7/14/07, Emil Gilliam <emil@emilgilliam.com> wrote:
> Is there anything specific that an extremely motivated and talented
> researcher, not currently working for SIAI or the recipient of one of
> its upcoming research grants in the near future, could work on that
> would be of highest expected utility towards the eventual building of
> a Friendly AI?
>
> Suppose (to give a hypothetical example of personal importance) that
> this person is a PhD student whose previous education is in vision and
> imaging systems, and can master all the math in Jaynes, Pearl, Hutter,
> Mitchell, etc. quickly. If "PhD student" puts too many restrictions on
> this person's time and freedom to possibly produce any SIAI-relevant
> research, suppose this person is an independent researcher.
>
> - Emil
>
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:58 MDT