COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	rota[f88,jmc]		Reply to Rota
C00003 00003	∂18-Oct-88  1545	JMC 	Rota article   
C00008 ENDMK
C⊗;
rota[f88,jmc]		Reply to Rota

There is the usual error of regarding Japan as having a single policy.
Only a small fraction of Japanese eggs are in the Fifth Generation
basket.  Some Japanese, including some involved in the Fifth Generation
Project, advised the Inamori Foundation to award its Kyoto prize
to an American working in AI.

Whom did the chess programmers get to closely guard their site?
How does Rota relate his leaked information about the 1983 secret
meeting with the public fact that one of the recent programs won
the 1988 Pennsylvania championship tournament?
∂18-Oct-88  1545	JMC 	Rota article   
To:   barwise@CSLI.Stanford.EDU  
It just arrived today, and I've just read it.  Are you sure you
want to publish it?  It's rather bad history, journalism and
amateur philosophy.  There is one aspect that puzzles me.  It
is dated December 26, 1985.  Subsequent to that date, Rota got
Los Alamos to sponsor a meeting on AI and a special issue of
Daedalus devoted to AI.  Daedalus is the publication of the
American Academy of Arts and Sciences.  Rota took part both in
preparatory meetings in Boston and in the meeting at Los Alamos.
The special issue came out at the beginning of this year.  Does
he really have nothing to add to his 1985 paper based on all these
meetings he attended, and the special issue he helped edit?

If you still want me to write a few pages of reply to it, I would
spend one page on listing some errors, and the rest on a de novo
treatment of Mathematics and AI, but the latter wouldn't be based
on his random remarks.

Incidentally, I would remark that if the Notices wants to go in
for science journalism, then it has to employ a ``fact checker''
like the New Yorker and other respectable publishers.  I remember
being asked a large number of detailed questions by a New Yorker
fact checker when they published a Jeremy Bernstein profile of
Marvin Minsky.

A few errors:

Minsky's thesis was 1954, not 1953.  His adviser was Albert
Tucker, not Lefschetz, and the readers included John Tukey, John
von Neumann and Lloyd Shapley.  It was 400 pages not 1,000 pages.
It cost me a long distance call to verify these facts, but it
would have only been an intra-M.I.T. call for Rota.
Lefschetz was my adviser, but my thesis was 23 pages and on
differential equations.  I mention it, because Lefschetz, who
was Department Chairman, may in later years have confused the
two of us.  He wouldn't have been the only distinguished scientist
to have done so.

The characterization of hackers is an exaggeration, and I'd bet
that the role of Feynman is misremembered.  A call to Danny Hillis,
the head of Thinking Machines and who can and does write rather well
would establish some facts.  For example, Hillis has an article in
the Daedalus issue that Rota instigated and helped edit.
  Hillis is young, but not as young
as Rota says, and some of the people active in Thinking Machines
are older than Rota.

He might mention that Jack Schwartz is now Director of DARPA's
Information Science and Technology Office and hasn't been exactly
generous to AI.

Maybe Rota sent you the wrong manuscript.  Surely he must think
all the effort he put into the Daedalus issue taught him something.