From: Stephen Reed (reed@cyc.com)
Date: Mon May 20 2002 - 12:35:31 MDT
On Mon, 20 May 2002, Eliezer S. Yudkowsky wrote:
> Cyc has a viewpoint pretty distant from my own, so I may have to use Ben
> as an interpreter to say anything about it...
Fortunately, I think that I understand your writings and see how to bridge
the gap between where Cyc is and where Seed AI starts. You may appreciate
my perspective that the gap between Cyc and Seed AI is smaller than the
gap between starting from scratch and Seed AI, although you hold that our
approach is fundamentally flawed.
Regarding Seed AI, I find your writings to be mostly in line with
thinking that I have held since the early seventies, and which motivated
me to stay in touch with the AI field and utimately to work at Cycorp. I
was looking for a tool by which a software system could use knowledge of
algorithms to improve its own behavior - as the starting point for AI.
I especially value your writings on Friendly AI and the programming
required to assure it. Cyc has a vocabulary for agents and goals that I
should extend in the directions you describe. I can introduce a
vocabulary and programming support for influence diagrams and bayesian
decision theory (or at least put on my to-do list), for that time when Cyc
initiates its own activities.
Regards,
-Steve
--
===========================================================
Stephen L. Reed phone: 512.342.4036
Cycorp, Suite 100 fax: 512.342.4040
3721 Executive Center Drive email: reed@cyc.com
Austin, TX 78731 web: http://www.cyc.com
download OpenCyc at http://www.opencyc.org
===========================================================
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:39 MDT