COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Dear Colleagues:
C00006 ENDMK
C⊗;
Dear Colleagues:

	As you all have heard, I would like to be replaced as Director
of the Artificial Intelligence Laboratory.  This letter is to make
explicit my reasons and the considerations that I think should guide
a search for a successor.

	My basic reason is that I am finding one of the most important parts
of the job increasingly uncongenial, and I hope someone can be found
who will do it better.  This is evaluating the research of the
research associates and faculty in the Lab, comparing it with work
done elsewhere, deciding whom to keep and whom to let go, and
praising this work to ARPA and other funding agencies.  The reason
I find it uncongenial is that the number of people in AI and the
number of subfields has greatly increased, while simultaneously
I have wanted to concentrate my attention in the areas in which I
am directly active.  It seems to me that if the Stanford AI Lab is
to remain one of the world leaders, this job just has to be done
better than I want to do it.

	The other administrative parts of the job are not onerous
given the help of Les Earnest.
While ARPA gave us a hard time recently, and may do so again, at the
present moment AI seems to be in good repute there.

	I want to concentrate my attention of mathematical theory
of computation, the formalization of knowledge, proof-checking and
other aspects of formal reasoning.  I will continue my interest
in the operation and improvement of the AI Lab's time-sharing
system and would be willing to continue in charge of it.

	The new Director should be someone who can be appointed as
a full professor, does good work in AI, has good taste in AI,
and has a natural inclination to maintain an acquaintance with
who is doing what in the field.  (I seriously considered acceding
to Jerry Feldman's ambition to be Director of the Lab, but eventually
concluded that I didn't think he would lead the Lab in directions
that would result in good research).  At present I have two other
candidates in mind both of whom are well known to the Department.

	I am in no special rush, and I suppose I could continue
indefinitely if no suitable candidate could be found.