COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "let.pub" source
C00006 ENDMK
C⊗;
.require "let.pub" source;
∂AIL David Gerrold↓Box 526↓Hollywood CA 90028∞

Dear David:

	Thanks for%inviting me to the SFWA meeting.  It was fun
meeting all you famous science fiction writers and also the
non-famous ones.  The round-trip air fare on PSA was α$55.00,
and stubs are enclosed.

	Here are the comments promised on the ideas in your book:

	1. The first thing one asks such a machine is its estimate
of the consequences of various policies concerning its use.  One
asks for ⊗pictures - not just evaluations.

	2. It is my strong opinion, though not that of everyone
involved in artificial intelligence, that the human motivational
structure is closely connected with our evolution from animals
and would require deliberate effort to put it into intelligent
programs.  The intellectual structure, on the other hand, is more
universal, because it evolved to solve problems and any other
intellectual structure capable of solving problems, whether
natural or artificial, would be similar in many respects.

	3. The "judgment units" are not plausible to a computer
scientist, because it doesn't make much difference what the
basic components of a computer are; what counts is the higher
level programming.

	4. While many programmers think otherwise, I think
one can make tamper-proof interfaces except that physical
security of the computer room must be maintained.  It would
be more plausible to have HARLIE bribe or gull a technician
into giving it unauthorized connections into the telephone
network.  It could help the poor fellow with his sex life
by writing his love letters, planning his strategy of
seduction, and keeping straight what he had told each girl
friend.

	5. I am opposed to making computers with emotional
structures that might give them rights.
In particular they should not mind being turned off.  In fact
they shouldn't mind anything - except harming humans.