perm filename CASTI.1[LET,JMC] blob sn#862520 filedate 1988-10-19 generic text, type C, neo UTF8
COMMENT āŠ—   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Dear Dr. Casti:
C00005 ENDMK
CāŠ—;
Dear Dr. Casti:

	I read your book, especially the chapter on AI.  The
enclosed notes were made as I went along, so that some of the
criticisms might have been obviated by later material.  However,
my impression is that hardly any were obviated.

	The book reads very well.  That's how I happened to get
through it.  It was much easier to read than what I should have
been reading.  However, you will see that my criticisms are
rather sharp.

	The biggest problem was with chapter on cognitive science.

	First of all, the errors in your description of formal
systems are so numerous and serious that your reader will
get only the delusion of understanding and will be a sucker
for false arguments if he thinks he understood.  However,
many of them could fixed by consulting a random PhD in
mathematical logic.

	Second, you express various opinions and animadversions,
perhaps typical of the circles you move in, without either
arguing for them or citing sources.  The reader is simply
invited to share your prejudices about who are bad guys.

	Third, I disagee with many of your conclusions and
arguments.  The notes don't contain any exposition of my
own views.  If you care, you might start with my article
in the 1988 issue of Daedalus devoted to AI.

	By the way, please send me the technical supplement.