Reason wrote:
>
> Depends on how you want to define "well understood". I'd substitute "highest
> level design spec mostly completed" as a minimum. That gives you the all
> important end goal and a navigable pathway towards modularizing development.
> Then the project is just partitioned and subpartitioned, etc, until you end
> up with things that can be coded.
I think that trying to code a living mind is likely to be just a little
bit more complicated than that. If there are aspects that can be isolated
and clearly defined, enclosed within the surface tension of an API bubble,
then that means there are aspects that can be forked off to a loosely
associated programmer on the SIAI team, perhaps a part-time volunteer who
is nonetheless a part of a closed-source project. I don't think you get
into the area of things that can be open-sourced until you descend out of
the cognitive-science stratosphere. I think that coding AI will require
continuous creativity and deep understanding from everyone on the team,
and that the problem will expand as we try to solve it.
-- -- -- -- --
Eliezer S. Yudkowsky http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:39:44 MDT