COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	4 "My own interests in artificial intelligence are biased toward its potential
C00006 ENDMK
C⊗;
4 "My own interests in artificial intelligence are biased toward its potential
for counteracting the dehumanizing influence of natural science,
for suggesting solutions to many traditional problems in the philosophy of
mind, and for illuminating the hidden complexities of human thinking and
personal psychology." The "counteracting" part suggests that beliefs can
properly be adopted for their effects.  This suggest deviations from
honesty in the direction of wishful thinking, but maybe I'm oversensitive
about what is, to this point, a mere possibility.

16 It's not important in itself, but the explanation for why most computers
today are binary is speculative and incorrect.  Many of the earlier computers
were non-binary, and this was justified as being closer to human usage, and
the Russian ternary computers were justified as representing information most
compactly.  Binary computers won the competition, because it is easier to
make a computer fast and reliable if it is binary.  It is more natural
for programming also.  For example, in binary computers, masks and the
data masked are the same kind of entity, whereas in decimal computers,
the masks are still essentially binary.  The tendency to rationalize is
regrettable.

25 One understands why one wouldn't want the program to print the same
belief over and over, but what psychological mechanism is supposed to
correspond to the program's behavior of putting the regnant belief in
the background once it has been expressed.

ch2 Colby's neurotic program is an excellent example.

46 "Serious ..." a rather ad hoc criterion for serious neurotic
conflict.
68 2:21am

126 If two programs call each other recursively, does that make
the organization heterarchical?

153 Can Schank primitives represent "color it red", "destroy it",
"damage her in some way physically or psychologically", "weaken
it", "cut it",

Richard Stark's proof of decidability of knowledge calc. essential idea:
"If S*p appears in q and p' is subordinate to p and q' is formed be
replacing every S*p by S*p', then q' is subordinate to q.

175 Wilks's criticism of deduction.

299 Marquise of O.