OFFICE MEMORANDUM e STANFORD UNIVERSITY © OFFICE MEMORANDUM © STANFORD UNIVERSITY © OFFICE MEMORANDUM

To

FROM

SUBJECT:

Date: February 20, 1973

Tuesday
File

Joshua Lederberg

Last Friday Tom and Ed and I had more or less definitely concluded

that we ought to narrow the focus of our further effort at SUMEX to
concentrate on biochemical applications. This was on the basis of some
discouragement whether we would ever be able to justify the more
comprehensive PDP-10 based program. However, King and Baker were out here
today and Baker is very keen to see us persevere and he gave us a much
more optimistic picture of how the study section viewed our application
than we had gotten before. He thinks we should go for broke in trying

to organize some reasonably effective national resource in the field of
artificial intelligence applied to medicine. And we just might make it.

We were negotiating on the basis of being able to use about 50% of such

a resource under local control and about 50% for distribution to other
people around the country. The users would be further classified in

terms of collaborators who will get some further support besides computing
time and those who newly come on to the service. There was not very much
that we could do but accept this offer as it is coupled with implied
Support expressed very strongly to find some way to maintain the continuity
of our research efforts of the order of $180,000 to $200,000 a year as a
supplement to our DENDRAL project in the event that the larger framework
is not at this time accepted. There is some tension as to whether to

pull the study section to a cliff and demand that they jump, on some all
and on basis, as opposed to giving them a range of options which they are
very likely to take only fractionally. The usual problem of linkage in
negotiation!

There just seem to be some real understanding of the problems of negotiating
as complicated a system as this and of the importance of having some very
close-in discussion about the style with which it will eventually be managed.

I promised Baker that I would have a memorandum for him on the long-range
implications of GCMS technology for health purposes which he might use for
program justification purposes. In addition, we have to provide a range of
pegs which are not too visible but which can be pointed out at option in

a course of the study section's deliberations that give them the possibility
of segmented approval of our request but do not push them into it.

I mentioned the case of ethanolamine accumulation and the implication to that
this has that at autopsy it may be necessary to freeze tissues for later
biochemical purification. We have a tricky problem that there are a number

of qualified AI oriented users at Stanford but we probably have to leave it
the the advisory committee to select which ones will be able to qualify.

On the other hand, we have to encourage a number of others outside to submit
independent applications. We had too much of a real time emphasis for the
local community before and obviously this is not the kind of thing that can
be offered over the wire.

WNONVYOWIW FDiddO © ALISYZAINN GYOINVLS @© WNGNVSOWAW FADdIddO © ALISATAINN GYOINVLS © WNANVYOWAW 3D1ddO © ALISUYZAINN GYOINVLS ©