Ss UW Cc») € +d

June ?6, 1°36.
ti.L. Royden

G.E. Forsythe

Requested Minsky appointment

I have forwarded an evaiuation of Minsky which McCarthy
has given me. I hope that with this in hand you can take
another good look at Minsky.

I have been looking at some of Minsky's publications
myseli. 1 £ind them to be excellent contributions, far more
than expository in aature. It is true that the nomber of
tangible results is less than with a candidate for 2 proiessor -
ship in mathematics. I think the reason is perfectly clear
the structure of the field of computing and artificial inteil-
igence is far less ciear, and a research man hes far less on
which to build. Nevertheless, what is being done is signif-
secant and important.

The aim of artificial intelligence is to deve Lop comp:-ters
as an extension of the human intellect, i:st as other types of
machines are an extension of human muscles. The first stage
ok this has been to see what cognitive activity digital computers
could be programmed to carry ont unassisted. This direction
was partiy the result of heman optimism, and partiy dictated
by the reiative inaccessability of present computers to humane

tinman beings are far too slow to work as a team with the
full tine of a microsecond comp ter,

This first stage has proved very diifiewit. Le wil
probably pay off eventually, but the pay-off is not a rapid
one .

However, & second stage of research is appearing -- the
use of a competer together with a human as a team for prob iemn-
solviug. This direction is now being made possible by tech-
nological developments. The pay-offs are coming considerably
sooner than those to be expected from machines on their own.
lt ig a priori evident that a heman being with a computer can
solve many problems better than a human being withor t a commter,
provided the total system is organized cight.
There is no question that Minsky is a leading candidate
in this whole area of artificial intelligence. Whatever is
said by his evaluators, no other serious candidate is put
ahead of him. In my opinion, there is no question of the
importance and value of the field of artificial intelligence.
This is why 1 am strongly urging the appointment of Minsky.

It may be asked why we need Minsky when we have McCarthy.
The ceason is simply that we are working McCarthy very hard
right now. He is covering a wide field of compiiter activity.
The whole character of the enormous future of Stanford's and
SLAC's comprting activities is receiving its technical
direction from McCarthy. This is a big responsibility to add
to the responsibilities of setting up a research and educational
program in artificial intelligence and other non-mimerical
uses of computers. We simply cannot ask McCarthy to spread
himself this thin and expect a program to go forward.

Stanford has the greatest momentum in improving its
program in computation of any university in the nation. This
momentum must be centinued, if we are aot to lose our position
and even ovr present leaders.

GEF -dcs