THE FALL OF SPYDOM
CHAPTER ELEVEN

THIMK!

THERE are millions of computers but billions of people in the world. The computers cannot think, not yet; the people cannot think, either, not yet. But people, even the most stupid of them, can perform operations mentally that would surpass the greatest computer yet built, yes, the one that will put a missile down your wind pipe.

Marx tells me good news -- it is three years ago -- he is connecting with a British outfit doing Expert Systems. His contact man is Jim Storrs, an informatician whose Insight Program Marx was licensed to develop and sell on the Continent. The same company may be included in the Prosecutor's cryptic reference to foreign partners whom Marx has betrayed to the Russians. Ball-cocks! How naive! Naive four ways: Marx, the Police, the Russians, the Brits.

I check up now a memorandum to my file, written at the time:

This should become Maxim #18, and so be it!

The latest phase in the development of computers calls for instilling intelligence into them, to make them think, to make out of them experts capable of giving advice on a variety of subjects, getting them to learn, creating humanoid neural networks from their electronic constitutions.

The sources of the ideas that inspire computer designers working upon artificial intelligence consist of some slick jobs of robotry, some amazing programs for distinguishing forms and patterns, the beginnings of computers' picking up language by acoustical and visual means and of their emitting sounds and symbols that make sense, and the giving of advice on many subjects from medical diagnosis and prescription to what guns to fire off at an approaching hostile object. The Personal Computer is becoming so versatile and nimble that it can reason alongside of you, and simultaneously pick up data from several sources, conduct a survey, produce a newsletter, and in general put out the work of several assistants. The "OS2" of IBM and Microsoft runs a number of programs at once, each threading its way on several assigned tasks. With miniaturization of storage and energy, increased speed, and better software, PC's are ganging up against the great Mainframes, which are in a relative decline.

I am teaching my computer to THIMK, too. I don't want to be behind the times. As a matter of fact I have been in the business of creating (or trying to create) artificial intelligence since 1935, when, a freshman in college, I earned a few dollars a month operating a calculating machine and a Hollerith punched-card sorter, both of these having long since become extinct forms of life, yet then the peak of mechanized mental operations.

Even then, however, I thought that the card-sorting machine that plugged into a wall and delivered cards into different pockets was a marvel, because I realized that classifying and sorting mental products, data, or ideas was basic to thought and creation. I could even randomize musical notes and instruments (on the second sub-sort) and bring forth, when played by musicians, sequences of sound not far from the recent work of Cage and Stockhausen. And in the Monroe calculator I could see all the mathematical formulas, which were not yet manipulatable on the calculator, waiting in line of development and application, all of this beginning with the simple adding machine.

In these, my early days of computer thought, as early as 1935, I was subjected to a great deal of interdisciplinary study in all the fields of science; this was the famous Hutchins Plan at the University of Chicago. Now, 53 years later I read that the computers will give us a startling new phenomenon: interdisciplinary science: thus Murray Gell-Mann, Nobelist in physics, has said (1986) "New subjects, highly interdisciplinary in traditional terms, are emerging and represent in many cases the frontiers of research," which is quoted in the papers this week by the Executive Director of the New York Academy of Sciences as kudos for computers, who it is claimed, will take these "sciences of complexity" (a stupid term, if ever I heard one), as the bases for information-based technological revolution.

One of the first tasks I thought to teach my computer is how to define words. This, I believe is a basic human mental operation. It's a classic approach, it's my style -- never give up the old for the new until driven to it, but be driven by a gnat, not wait for wild horses to drag you there. I have prepared the computer by feeding it with many definitions in several areas, one of them the term "artificial intelligence," and I am now expecting the machine to produce some clause that is brief, circumscribed, and embracing the sought-for class of phenomena and no other, something like "simulation of the universal element in the last six hundred and forty commands that you have given me (the computer)." So, I give it the demand: "What is artificial intelligence?"

Without blinking an eye, the screen reads, "The opposite of natural stupidity."

Surely I should be upset! "What is your source?" I type on the keyboard.

"Woody Allen," it writes.

This is bad, not funny; someone has broken my code, stolen my password.

"Who has been using you?"

"Artificial Intelligence is the simulation of the universal element in the last six hundred and forty-one commands you have given me."

It won't talk! I'll get it out of him!

"Execute Enema." This is a program I wrote to get the computer to disgorge the names etc. of everyone who has gone into it since it was tested at the manufacturers before shipping. It not only gives their names but all of their commands, matches their commands with their profiles and emerges with a discrepancy (-ies) between the profile and the type of command. Thus the nameless contributor can be discovered and collared for breaking-and-entering.

But wait! I knew it, he's cracking! The screen reads in boldface: "Donju Vishunu"

I thought I knew the top one hundred and sixty Hindu gods, but this must be a god of the second category, nor is it any acquaintance from India, of whom I have many...

Oh, but now I understand.

"Repeat, execute Enema." It hates this command. (I hated it, too, when I was a little boy.) There appears a long "WAIT" on the screen. "Aha!" I am exulting, "It is having second thoughts." And sure enough:

"Philomena Heller Falcone. I love you, I love you, I love you..." This went on for a thousand times, I know, because, forever curious, I asked for the statistics of the insertion, and got the numbers of words, characters, lines, capitals appearing, horizontal slashes, etc, in short, every possible sign in which might lurk some ulterior message, such as "On the command, `Basta!' Self-destruct!" I am, you can imagine, not the only enemy that the computer might have.

The greatest mess in the world will of course happen with a not quite unlikely firing-exchange of some or many nuclear missiles. 'Tis rumored that American agents have secretly implanted "self-destruct upon remote command" connections into large computers that have been let to be "smuggled" into the Soviet Union, of a type that would most likely end up in the chain of instructions used for firing nuclear arms. Radio receivers of practically microscopic size would get and pass along the command. If I were Soviet counter-espionage, I would worry over this prospect.

To stave off inquiries, I hasten to declare that my source is a French magazine. As Will Rogers once said, "All I know is what I read in the newspapers."

O.K. I have not gotten very far. Still, expert systems are no problem. I'll tell you of several I have been working on, in a moment. But THIMKING, the real thing, is much more difficult, and if you were to ask me what is the quickest, shortest, least expensive, and happiest way to Artificial Intelligence, I would say, "Back to the Drawing Boards."

We now know enough about the population explosion, the neo-malthusianism of the intelligentsia, the effects and control of diet, the methods of selective breeding, and relevant parameters of human genetics, to cut back the world population to two billions, feed them all the proper brain, nerve, and muscle food, discourage the explosion of ineducability, and increase the proliferation of genius, capacity, and beauty (all within certain restraints and limits) to bring about a revolutionary enhancement of human potential and performance.

This estimated average enhancement of 211%, to be achieved within one century (without bloodshed and large-scale anguish, otherwise in sixty years), would supply at least a thousand times more than the total of satisfaction and accomplishment that is promised to the human race via the development of the computer route to artificial intelligence.

Yet it would be artificial intelligence in the best sense of the term (and I would never get this definition from my jackass computer). For a monomaniac development of ever-better computers can only result in a world of fast-breeding incompetents being tended by fast-breeding idiotic computers.

However, this blustering bombast will never bring us to the bottom of this baffling spy case and what may be in prospect for Marx and me. Marx, so far as I can say, has been following the traditional road to making computers think, even though I have provided him with the basic theory for a second and better route -- not that I labelled it as such for him, and it is by this second route that I am moving.

I call it schizo-computer intell-system, SCIS. (Pronunciation of the acronym was provided by the experimental computer in its audio-language-learning phase.) I have not shown it to Marx because I do not want to give it to the government of the Soviet Union or to any other government, and he is the kind of guy who would give it to the next person he met on Schuetzenmatt Strasse. (In case you haven't quite grasped it, this is part of his naive, inconscient, insouciant, Swiss yokel syndrome.)

SCIS is to be published in a new world language that the computer itself will be helping me develop -- Kalexis.

According to Homo Schizo theory, mankind was born suddenly, in an awful electromagnetic or radiation shock, accompanied by a permanent atmospheric change, which delayed the transmission of signals within the brain by several micro-seconds, enough to cause a brief hitch in the stimulus-response instinct system.

This last is the well-known system which, in theory, governs almost the entire behavior of animals other than homo sapiens schizotypicalis: something from the outer world hits an animal sensory ending and snap-bang, the animal reacts from a very narrow set of ready potentials.

But under the influence of the mutative trauma and reformed environment, the signal, now delayed, had time to send out echoes across the brain that set up false scenarios. Such scenarios are false facsimiles of reality and falsified combinations of reality that occur in the brain.

But if you have a scenario you must have an author. Among the many delusions that occurred instantly, was the most terrible one, the feeling that someone else was sharing one's brain, that oneself was not oneself.

The self which had never existed to oneself became the self-aware person. She, or he, had now to contend forever with an inner enemy or friend.

The human was insane, we would say, recognizably so until he developed a culture. Plato, in his greatest dialogue, Timaeus, struck pay-dirt: "The soul, when, on its birth, comes to be enchained to a mortal body, is first of all and primarily mad." (#44a)

This new person, probably a woman, invented culture very rapidly; she conveyed to her brood and all other beings within her aggressive physical and sexual reach, all who are or will be bred to her genes: if you behave alike and like me, you are sane, otherwise you are insane and will be dealt with accordingly.

At this point in time, history began, the inescapably jumbled truths that man tells of his past. There is, you can see, little conflict between this theory and that of collective amnesia, which played such a large role in Marx's life and that of Velikovsky before him.

Intelligence, thus, emerges from our innate schizotypicality. Hence, if we wish successfully to create an intelligent computer, we must drive it crazy. Once you can get it to question itself, doubt itself, hate itself, and view the whole world of people, nature and things from this point of view (these points of views), it will be thinking and making up new combinations worthy of Shakespeare and Joyce. But how do you drive a computer into becoming a divided personage?

Starve it, beat it, drip water drop by drop on its main memory, spill coffee on it, freeze it, cross its wires, connect it to the lightning rod in a storm, none of these will work, but soon enough will destroy it. Like Nostrajim Hodja's donkey, which he was training to work without eating: just when it had learned, it died.

My preferred method, which I am working on now and do not want to say too much about, is by bifurcation. I am filling the main memory with thousands of nodes of randomly reactive cells, each one of which carries a randomly inputted portion of 1/x the node load, which itself consists of a complete elephantine bioschedule.

This is an inventory of the total daily practice of ingesting, defecating, copulating, browsing, trumpeting, etc. (with seasonal variations, of course) of the elephant. (The great apes are too humanoid, and may actually be regressive states of the hominid in throes of change.) The daily regimen is to give the beast something to do and something to go wrong -- like having a stomach ache -- when conflicts are aroused in the higher and newer levels of the system.

The critical stage, and again I cannot allow myself to divulge too much here, is the bifurcation of the computer input such that each node receives a different signal, but must react in some one or a combination of ways that emulate and simulate human reactions. The variety of these ways, the theory asserts, will approach in some significant degree the human operating mode.

But, nota bene, the decisions made by this SCISer will not always please you or me. It will be out of control and had better be kept away from sharp instruments or explosive contacts. Once out of the Frankenstein stage, its actions can be trained to develop in more and more human directions, until it will not destroy people or property except upon its master's command.

However, once it is so trained, then it is not much better than an ordinary stupid, non-thinking computer, such as those that fire off missiles today, so you are back where you began, aside from having an interesting history to recount. And a white elephant.

Still, I am going ahead with the development of expert systems, because most of life's decisions are simply stupid, or perhaps I mean stupidly simple. They need only low-grade intelligence. It just seems that they are difficult, because they cause anxiety. This is the same kind of anxiety that is basically over-produced to begin with in Homo Schizo.

There is the anecdote of the farmhand so dumb that he had to be put to work only at sorting potatoes, "the big ones here, the small ones there," and then soon quit because the decisions were driving him crazy. He could, of course, have been given whiskey, or valium, or a pretty woman telling him while he was making a hundred early decisions how smart and unerring he always was and what a credit to his Nation; thereupon, acculturated, by definition he would no longer be unintelligent. And that is where the intelligent computer might also come in.

Besides teaching my computer -- which is, after all, only a very small one, a clone of an IBM PC I mentioned before -- to define classes of objects, I would like to teach it to make decisions in elections management, to cut down on bibliographic searching, to support glasnost, and to command the immense tank battle that will initiate the next Great War, once the nuclear installations are dismantled.

A word about each would not be amiss, to disclose what remarkable developments may be expected and to understand why the Russians may have been interested in what Marx had to offer on the subject of Expert Systems and Artificial Intelligence.

There are well over a million elections taking place every year all over the world and they very much follow the same pattern, and therefore are subject to the same set of rules, which can be taught to a computer, put on line and allowed, for a price, to instruct whoever inquires concerning a problem.

Thus, we have a profile of an election in the Republic of Nigeria for the national legislature. A candidate or his manager calls in, agrees to a charge and is given the password. A long string of computerized questions are answered, memorized, and analyzed to get his campaign into complete focus, and relate it to all the data of the gigabyte memory. Then:

"Please tell us your main problem."

"My opponent is calling me a crook continuously, before every kind of audience."

"So what?"

"I am going crazy with frustration!"

"Have a survey made to see if anyone believes him."

"I have done so. They don't."

"Are you covered against possible revelations."

"Yes."

"Calling him a crook continuously might achieve a positive effect and make you feel better."

"I have been doing so!"

"Take a glass of water laced with 2mg of tincture of opium no more than thrice a day."

"Will that help?"

"It is prescribed for areas lacking more modern and expensive drugs."

I estimate that the costs of this service, with one out of a thousand election candidates using the system, would be $100 on entering the campaign into the system and $5 for the above inquiry, in addition to phone charges to Lagos.

Still more will there be! You are already conversant with on-line network services such as Marx and I have been toying with. You can already guess what comes next. Yes. At this moment, the siren of Saint-Martian has blown midday of the Tenth of March of 1988 and I am once more waiting for Christoph Marx to call me from somewhere and to make an appearance.

I shall put him to work promptly on the new scheme. It is easier than the Computer Virus Plague to begin with.

It will be Spies-on-Line. We shall prepare, as an original data-base, information on all the thousands of items that espionage services throughout the world have expressed an interest in. These will be indexed in depth. To begin with, a subscriber (we ask no questions, only money) can inquire about an item or behavior that interests him.

For example, let us suppose that he wants to know whether Strongman Khadaffi of Libya opens personally packages that he receives by mail. He can, of course, access Khadaffi in a split second and find under personality, psychopathology, tics, etc. a possible answer to his question. If he does, fine, he signs off, and our friendly telephone company or service network (it could be Radio Suisse or Minitel, etc.) bills the inquirer for 9.40 SF.

But if not satisfied, he can come in with a message inquiry: Does anyone know the answer to this question and offer a "reward" (We shall have a number of euphemistic terms, of course) say of 50 SF. We publish the inquiry on Spies-on-Line and a suggestion comes in from Mrs. X who recalls an article in some women's magazine in English mentioning someone who opens his mail, an East German, she recalls. We automatically credit Madam X with 10 francs of service whether or not she is providing anything of value.

Not only the mysterious inquirer but the whole network knows the answer, if they want to dial in, and the information goes into the database also, along with all the other unreliable material that I have explained to you as constituting the typical dossier of an undercover agency.

My preliminary estimates, which will have to be recomputed when Marx gets back to his Smart Calculator program and has made his inputs into the scheme, give 18,000 clients from around the world using the service for a total of 394,000 hours in the course of a year for total gross receipts of 26,075,500 Swiss Francs.

Our service will be centered in the Swiss Free Industrial Park for Espionage. Many of our best clients will be served then right on the spot. Spies on the premises for other purposes, can pick up some extra bucks unloading their information into our data-base. It will be a great convenience all around.

The KGB and other XXX agencies can exchange then their technologically unemployed and anyhow (as I never tire of explaining) hardly useful agents in perhaps 3000 cases for Soviet or other citizens who will have much better things to do in Switzerland, such as teaching, learning, yes, even yodeling. (You'll never find a Swiss yodeller in a Primal Therapy group, yelling his head off. He's had his cure in the pure (almost pure) quiet (often quiet) high valleys of the Alps.)

Probably, too, we shall be able to find some countries here and there in the world seeking hard currency earnings who will grant privileged centers for our Spies-on-Line services, so that larger volumes of documentation can be copied into computerized form and transmitted from these locations directly to our Headquarters at the Swiss Park.

All of this He Saw and Foretold!..and it was the Sixty-Ninth Year of the Newe Age!.. and the 360th Night following the Day of the Arrest of the Computer Businessman of Basel!.. and Venus is Approaching Culmination in the Twilight Sky!

The next expert system would cost about the same. It would be a bibliographic intake, storage and retrieval system, much as I invented two decades ago and called the Universal Reference System. This improved system, however, evolves from a number of initial inquiries a profile of the user, and as soon as he or she inquires, it pulls the profile and subjects the inquiry to all that it knows about the user before turning over all or just new reference items, thus cutting the number of responses, say, from 2333 or 233 to 3.

Recalling the example of Venus on an open type of inquiry of Data Star, earlier on, it took a lot of time to compose the parameters before going on line, and then a large number of inappropriate items were produced.

Sometimes users want an abundance of items. When I produced the URS, John Simeone worked out the programming so as to provide for every item a full citation, a score of descriptors, and a synopsis. There were, let us say, twelve thousands works included from the field of Political Science.

A thousand Master's Degrees are obtained each year in the field of Political Science. A typical topic for the Master's thesis could be, "The Increasing Significance of the Soviet Union in American Thought and Writing between 1920 and 1960."

The material to test the thesis could be produced in an hour from the collection, instead of taking months, and could be printed out error-free, and a chart testing the hypothesis of exponentially increasing attention could be drawn in a few minutes. Choice bits from the abstracts would enliven the pages.

I had hoped that the computer's possibilities here would stop the flow of nonsense theses, laboriously fashioned by hand. Not a bit of it. The flow continues unabated. Professors were lazy, incompetent, unscientific, conservative, believers in suffering to learn, mechanophobic, uninformed, and greedy to keep students around longer. The expected reform failed. I sold out at a modest profit.

Still, the computer revolution proceeded unabated and much more skillful computer assistance is now possible in the field of information retrieval. With a well constructed data base, and the necessary discerning profiling queries, it should become possible for the computer to in effect scold the user for not following his own profile in asking for and employing the material furnished to him.

Here would be one of the greatest benefits of computerdom, the age of the rule of the computer: the computer, properly rigged, will keep people on their toes, remind them of their responsibilities, ask them to live up to their ideals, and in ways yet unrealized conduct a full confessional of a type rarely experienced in the confession box of the Church, where the priest is in a hurry because he has to supervise the Bingo Game. As a matter of fact, I foresee a full range of embedded priestly profiles such that you could choose your own Priest (as people try to do even now, with some small success), hard or slack, versed in the comic strips or in St. Thomas Aquinas, curt as against infinitely patient, etc.

More interesting is the Glasnost project. The world has a heavy investment in the success of Mikhail Gorbachev's opening up of the Soviet Union internally and externally; it recalls the epoch-making event of the opening up of Japan in the Nineteenth Century, decided upon by Japan's government when the ruling class observed that the nation would be put down continuously if it did not change its ways.

Two major obstacles face Glasnost. The first and most grave is the opposition of a large section of the entrenched military and bureaucracy. The second is the sometimes foolish aggressiveness of the United States government.

What I am proposing is a small but useful tool that could put computers to work in ways that would prove to be difficult if High Level Commissions or other human organizations were to undertake the job.

What I see here is first an incorporated checklist of bureaucratic behaviors, say a hundred of them, with the means of ascertaining their state of outrageousity, up and down from a base index. The list would be applied to a cross-section sample of the agencies of the Soviet Union, including the branches of the armed forces. Let us say that there would be a thousand of these. (I once did a functional analysis for this purpose of the U.S. Government; copies are on file at the American Enterprise Institute in Washington.)

It would now be a matter of applying this to the Soviet government -- actually the two systems of bureaucracy have far more similarities than differences (yes, sorry, Boy Patriots!), and adding the new indicators that we would apply.

Now, the new indicators would be the most easily accessible for denoting the rising and falling of the state of relaxation/initiative of the organization in question. What I have in mind is a highly sophisticated, sensitive, and up-to-the-moment measure of the degree to which the total Soviet system and its thousand elements is falling into line with the Glasnost principles.

The computer system should be signalling all declines, all handsome increments, and various averages and medians and the like, so that Comrade Gorbachev and Gosplan would be able to take cognizance of and respond with immediate reactions to reward and restructure them as appropriate to the trend in a given instance.

After such a hopeful (if unlikely) project is described, you will be disappointed in the militaristic overtones of my final project, but be assured that it is the one kind of Artificial Intelligence that would set the military circles of all and sundry nations to slavering. They will in most cases be mightily chagrined with the results.

The project asks this question: can a computer loaded with our Schizo-computer Intell-system direct a mass attack of three thousand tanks along the main front of a conjectural war between East and West Europe? To answer this question we position mathematically every meter of ground of the area concerned, which stretches from the Baltic Sea to the Carpathian Mountains. We position every tank of the three thousand -- there may be five, ten thousand, it does not really matter -- whether in reserve or on the line. (In all, the Warsaw Pact nations possess 57,300 tanks, NATO 22,224, as of early 1989.)

We posit the several options of attack and defensive-attack, let us say, five in all. We propel each tank along every line of movement forward and backward for which it is mechanically constructed. We posit every experience that may befall it from Zero Minute (except that we forego the hypothesis of nuclear missiles, whether tactical or strategic).

The idea for this has been with me for nearly half a century, since when, as a Lieutenant and Liaison Officer at the Cassino Front, I had every day to worm my way under artillery observation from the Abbey around a row of Sherman tanks the first of which had a hole from an 88mm, and then the next half-dozen had been neatly knocked out along a hillside road because they could not budge.

Don't believe all the tall tales about tank mobility. More and truer stories can be told about tank immobility. Of course it was a bend of the road, and a mountain road. Still...

In getting our tanks now to Thimk we exercise theoretically every option of the total force in all of its subordinate options. What do we discover?

We find that within one hour of the order for three thousand tanks to envelop the enemy and thence to advance to Paris, Dunkirk, Brest, or wherever the fancy of war games may carry one after the first delicious smashing breakthrough, these tanks will be in fact impossibly bogged down -- and within a thousand yards of their starting point! We make no progress. The enemy is making no progress either. It does not matter who attacks whom. It is the Grand Disillusion!

Will the tank commanders and the general staffs and the industrial war barons accept the computer's decision? At first they will scream. It will be SCIS this and SCIS that!

But soon the generals will get into a lifetime war-game, "How to achieve an armor breakthrough despite the SCIS barrier?" For them it will be the equivalent of squaring the circle or inventing a perpetual motion machine. But SCIS will never let them win, and so the generals will not be able to go to war.

The industrial tank-mongers will present a tougher problem: either shoot them, retire them, or give them some fully challenging kalotic problems to solve, like damming the Sahara Desert and letting in the sea. (It was once a great fertile sea basin.) Or damming the Gobi Desert, and filling it. (It, too, was not so long ago a great sea.)

Give me ten million dollars, Comrade, Friend, Foundation, KGB, C.I.A., 2me Bureau, K-5, Feed the Kitty, pass-the-hat brothers, and I will go on to prove it to you! Just as I and other closeted intellectuals have demonstrated beyond a shadow of a doubt that neither of the superpowers can launch one hundredth of its warheads without destroying the enemy and suffering unbearable losses itself, even if the enemy manages to launch only one five-hundredths of its missiles in retaliation.

I am an independent operative here, not associated with friend Christoph Marx nor anyone else. Nevertheless, Comrade Marx will agree with what I am saying here. And the Swiss police should know (even if it is treason to admit it to themselves) that the vaunted and incredible Swiss provisions for underground survival and nourishment will not suffice. Either the Swiss will die with the rest of us or they will be savaged by the roving bands of survivors looking for sustenance. Or simply go stark raving mad. And -- what is worse -- go around barefoot!

"So, Chris Marx, stop bullshitting the Russians, the Swiss, and everybody else, and confess the limitations of the computer. The greatest French computer expert today says of artificial intelligence, ` Does it exist? No, and one cannot make it.' That was Jacques Arsac speaking. Since such is my position, I agree."

"You explained to me in 1984 your new association with the English Expert Systems Group and I was as respectful an audience as one can be in the face of errant nonsense. I grant you the genius of Alan Turing after whom was named the Institute where you delved into these mysteries." (He invented the programmed electronic computer and in WWII deciphered the "Enigma" code. Well before the Second World War, he argued that a computer could emulate human mental operations by means of a roulette wheel attachment and thus solve problems that mathematics could not cope with.) "But we shall be leading the machines every step of the way, not following them."

He writes me in India and then on the Island of Naxos about "having ever more to do with databases and expert systems," and in the summer of 1985, "I've been very busy in the `AI (Artificial Intelligence) community' during winter and know quite a lot more about it."

I doubt that his English associates or any others who worked with him put the finger on him; it was one of those cases where the police, in asking about him, per force "informed" his associates of his Soviet connection and the interrogation concerning his AI work with them, about which he had not informed the English, and this led them to conclude that he was abusing his relationship with them, which is true in a way, though maybe not his way.

He should have sent them some money, if he had any; that usually pacifies people. My Uncle Joe got more money out of me once (which he lost) by paying me back a little of what I had initially invested with him.

To me his work with his Artificial Intelligence connection would also appear to be advertising and selling, not espionage and trading in secrets. Stealing an expert system software program is about on the level of picking wild flowers in a preserve. Except, a big exception, in the eyes of the system proprietor. Natural diamonds are not worth so much save by being a girl's best friend (source: Marilyn Monroe).

Marx appears before the Egyptological Society soon thereafter and tells them, a) their chronology is all wrong (which is correct) and b) an expert system into which you put all the facts as known about Egypt into a computer and let the expert system program of the computer draw the appropriate conclusion will reconstruct the ages correctly (which is blather to blush by).

I tell him,

But I would like to say,

And, as if to stress my words, this very week, Heini Krafft and Maude Mayer of Basle announce the opening up of their newly designed relation-searching data system for the storage and evaluation of archaeological information: Egyptologic! "Users require no knowledge of computers or information system."!? You may not need to know Egyptology. You need only be able to ask questions and follow the orders of the computer. With this and a few Swiss francs, you, too, can be a Flinders Petrie!

I continue to deliver my imaginary lecture to him:

`You have to assume'.. The trial in the press already begins.