From: Ben Goertzel (ben@goertzel.org)
Date: Wed Feb 22 2006 - 15:01:42 MST
On 2/22/06, Eliezer S. Yudkowsky <sentience@pobox.com> wrote:
> Ben Goertzel wrote:
> >
> > Based on crude considerations, it would seem SIAI is nowhere near the
> > most advanced group on the path toward an AGI implementation. On the
> > other hand, it's of course possible that those of us who are "further
> > along" all have wrong ideas (though I doubt it!) and SIAI will come up
> > with the right idea in 2008 or whenever and then proceed rapidly
> > toward the end goal.
>
> A disingenuous argument, Ben; many AGI projects are farther along the
> path to AI implementation than Novamente. In fact, many AGI projects
> have completed implementation. Then they didn't work, as is the prior
> for AGI projects, because the people who built them rushed ahead with
> vague ideas of which they were very proud even though they had no
> concrete idea of how intelligence would magically flow from their AI.
> Don't boast of how many lines of code your project has. Demonstrate
> your understanding of intelligence. Only then does your partially
> completed code matter more than the already-completed code of Eurisko or
> SOAR. I understand the impulse to show off one's own strong points, but
> Cyc might as well boast of more funding.
Sure, I basically agree with your comments, which is why I started my
paragraph with "Based on crude considerations...." ;-)
However, less crude considerations also indicate to me that SIAI is
not that far along the path to a practical AGI design. There is also
the fact that there is very little that looks useful for practical AGI
work in any of the SIAI publications (while there is a lot that looks
useful for Friendliness theory, for example).
But anyway, predicting who is going to make a huge breakthrough is
really really hard, and not the most useful or fascinating topic for
discussion...
Ben
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT