∂12-SEP-74  0808		network site AI
 Date: 12 SEP 1974 1107-EDT
 From: BEN at MIT-AI
 To: JMC at SAIL
 
 John,
 
 	Thanks very much for sending me a copy of your book.  I enjoyed reading
 it, and have a number of comments on various points.  I'll just write paragraphs
 on them as I skim my marginal notes.
 
 	About population problem:  recent demographic research seems to show
 that population growth "naturally" levels off in well-@developed countries.
 The baby boom after WWII seems to have been a local abberration, and we are
 now below the ZPG level anyway.  Thus the way to stop the population explosion
 is probably to export as much technology to underdeveloped countries as possible,
 so they can reach the state where it is not economically to the advantage of
 each family to have as many kids as possible.  My reference on this (unreliable)
 is the New York Times Magazine of 8 Sept 74.
 
 	In a couple of places you talk about the possibility of having formal
 languages for discussing social and political issues, in which we could compute
 the validity of various arguments.  This idea has been around for quite a while,
 as witness a story about Leibniz, who believed that a little extra research would
 get his formal logic to the point where, upon entering a political discussion,
 he could whip out pencil and paper and say "Come now, let us calculate!"
 It seems things have not progressed as quickly as he had hoped, and my personal
 belief is that the problem has not come that much closer to solution, due to the
 difficulties in producing the kinds of expressive power you need to talk about
 social and political issues.  It might be prudent to point out that previous
 attacks on this have failed, and just how hard the problems are.  [Unfortunately,
 I don't remember a reference on the Leibniz story.]
 
 	When discussing the risk in human lives of nuclear reactors, you say that
 since 50,000 deaths per year seems to be acceptible in the case of automobiles, it
 must surely be acceptible in other cases as well.  This ignores an important
 psychological point which crops up in a number of other contexts as well:  people
 find it very important to perceive themselves in control of their environment and
 fate.  In an automobile, I feel (with some justice) that my survival is in my own
 hands, and that by being a good driver, I can stay out of that 50,000.  This is
 not completely true, but enough so to be of comfort to me.  Having a nuclear
 reactor nearby, on the other hand, subjects me to a risk over which I have no
 control.  I would tend to be much more conservative about risks I will allow other
 people to take with my life than risks I will take myself.  This issue of
 perceived personal control can hardly be overemphasized, and it will come up in
 comments on other points.
 
 	You talk about mining uranium from granite as providing energy for the
 next billion years or so.  I would be interested in knowing the NET energy
 production of this mining, which should not be hard to compute.  Consideration
 of net energy production has interesting and non-obvious consequences which make
 the energy crisis perhaps worse than it seems, rather than better.  An interesting
 article is by Howard Odun, an ecologist (a real academic biological one), appearing
 in the CoEvolution Quarterly (Spring 1974).  The CoEvolution Quarterly is the
 supplement to the new Whole Earth Catalog, and is available at Plowshare Bookstore
 in Palo Alto, among other places, but is still hard to find.  I recommend it.
 Odun's article is very interesting, and his methodology agrees very well with
 yours, since he gets impressive results just by doing arithmetic everyone else
 has ignored.  His ideology is opposed to yours, though, so it is probably important
 to address his points and reference his work.  Very interesting stuff.
 
 	A drum I've been beating for a while deals with the problem of
 unemployment caused by automation.  Of course it's true  that automation
 produces more jobs than it destroys, but they go to different people.  There is
 an ever larger class of unemployable people being created (given our current
 economic system).  Your other proposals for guaranteed income and restructuring
 of the employment world address this point, but in ECON.ESS you don't point
 out the immediate problems involved in creating qualitatively different jobs
 (perhaps more service oriented) than existing ones.
 
 	Also in ECON.ESS, you propose a definition of the credit in terms of
 a fixed mixture of commodities.  How does this differ from the gold standard,
 in the sense of choosing something of generally accepted value to base currency?
 I don't see it as being inflation-proof, unless you define it as, for example,
 "that portion of our national production required to support a family of four
 for a year at currently acceptible levels" or some such.  The existing proposal
 is sensitive to obsolescence of the commodities in question, for example.
 Anyway, a radical economist I have only read slightly, but who is very imaginative
 is Kenneth Boulding.  Again, he is left-wing, but I believe that many of his
 ideas would appeal to you.
 
 	Again in ECON.ESS, applying for permits for lower fuel rates is an obvious
 bogge, contradicting common sense and several rules you present later in the
 book:  who manages that bureaucracy and makes the decisions?
 
 	Household terminals:  the biggest tradeoff I see is between establishing
 standards for common storage formats on the one hand, and providing enough
 expressive power to produce at least as wide a variety of publications as we
 now have, on the other.  I believe it's a solvable problem, but it's also
 VERY hard.  Consider how hard it is simply to agree, say, on network graphics
 protocols.  In particular, you mention having a program to filter out advertisements
 from publications:  this requires regulating advertisements so they can be flagged
 as such in the stored text.  More required regulatory bureaucracy.  This kind of
 comment is not a bug in your plan, but the implied global organizing body should
 be pointed out.
 
 	In OBJEC.ESS, you speak of pseudo-proofs (again of a social/political point),
 and assume that an error in the conclusion must be attributable to either an
 incorrect assumption or a faulty step.  This is an admirable quality of real
 proofs, but may not hold for pseudo-proofs.  Perhaps there is a built in range of
 variation for each statement (whatever that means), and that be staying to one
 side of these ranges of variation, you can gradually and globally slip into an
 invalid conclusion from valid premises and deductive steps.  Certainly it was
 the localizability of errors that made formal logic the useful (though limited)
 tool that it is.
 
 	A number of the essays included in the draft you sent me lacked their
 last pages.
 
 	The argument that each person's share of Spaceship Earth is a trillion
 tons is pretty much meaningless, though true.  That kind of arithmetic is
 exactly what you decry in others.  What is each person's share of useful and
 available resources?  Also point out that even if you had title to your share
 of the world's diamonds, they would be in South Africa where you probably couldn't
 get to them.  Distribution is the heart of most of the world's shortages, so you
 have to weight each person's share of X according to how far from him it is.
 
 	In WOMEN.ESS, point 8, you talk about child care.  It is now generally
 accepted that the most important things in child care are physical human contact
 and affection, and providing human social role-models for the children as they
 develop patterns for their later social behavior.  Technology should take as its
 goal streamlining the other chores (diapering, etc.) in order to facilitate
 those kinds of interactions.  But you lose plausibility with people who know and
 care about these aspects of child care by speaking as if machines can do the job,
 rather than just eliminating hassles for the people who really do the work.
 (No ref on this right now, but I can get them if you want.)
 
 	What's the reference for Pirages and Ehrlich (in CHICAR.ESS)?
 
 	Why do you want a futures market in labor and other commodities?
 
 	I had a bunch of other comments, which were mostly on individual
 paragraphs and dealt with stylistics which will doubtless come out in a later
 draft anyway.  Occasionally you lapse into a slightly carping tone (recognized
 explicitly once), where you seem to be reacting to being a beleagered minority
 beset by English professors and literary types.  That tone is beneath you and
 lowers the tone of the essays.
 
 	In general, I really enjoyed the book, and agreed with most of what I
 read.  I consider myself on the left end of the political spectrum, with the
 added point of considering less government better than more, in the absence of
 countervailing arguments.  So, by and large I agree with your overriding
 concern with providing maximum individual freedom and opportunity.  I would
 enjoy seeing later drafts of this as it is refined.
 
 				Ben Kuipers
 				BEN @ MIT-AI