Leaders and Computers
In the June of 1957 there were two men, Dwight D. Eisenhower and Nikita S. Khruschev, whose positions and ideas cast over-lapping shadows on their two societies. Eisenhower had been re-elected as president in the previous fall in the midst of the Suez and Hungarian crises. He was, by this time, secure in the affections of his countrymen. He wasn't thought to be brilliant, but there was absolute confidence, even among most Democrats, in his good will and basic common sense. The famous Eisenhower smile moved millions to feel, however unreasonably, that, as long as he was there, they wouldn't be nuked.
Comrade Khruschev had just established himself in the supreme position of power. The Communist Party Presidium had been giving him trouble. His enemies within it had tried to depose him, and perhaps even have him executed. Such things happened. It hadn't been long since Lavrenti Beria, the ambitious NKVD boss, had been shot as a British spy.
Khruschev had avoided that fate by convening the full Central Committee of the party, in which his friends out- numbered his enemies. When he won, Khruschev established a precedent. His enemies weren't executed. They were given jobs that carried a humorous message. Molotov, the veteran foreign minister and most worldly of all Soviet officials, was made ambassador to Outer Mongolia. Another opponent was put to teaching in a kindergarten, and so on. The people of Russia and the other republics weren't yet ready to love Comrade Khruschev, but he was thought to be warm, human, and a bit of a peasant. In fact, as a former farmer and miner, he had less formal education than any other leader of a major power. Regardless of that, it was obvious to anyone that he had a million ideas, boundless energy, and few inhibitions.
Khruschev believed that the Soviet Union existed precariously in a world which was implacably opposed to socialism, and which would take any excuse to destroy it with nuclear weapons. He had discovered that it was useful to have such weapons oneself. In the Suez crisis of the previous autumn, the British and French had given up their colonial pretensions very quickly when he threatened to destroy their countries with nuclear-tipped missiles. He was hopeful of soon having missiles that could reach the United States.
On the many days on which he thought that the rivalry between his country and the United States was more economic than military, Khruschev was confident that his planned economy, which had practically eliminated waste, could beat a disorganized and helter-skelter capitalist one.
Khruschev believed, quite sensibly, that food mattered more than anything else. As soon as his people began to eat better than the Americans, their resulting superior morale would enable them to triumph in all other fields. To eat better, it was necessary to greatly increase livestock production. To do that, it was necessary to produce great quantities of feed-corn. It wasn't for nothing that he was known as a kukuruznik (one gone crazy over corn).
Versatile and bright as Khruschev might be, he had no technical or mathematical training, nor was he any sort of theoretician. He hardly knew that there was such a thing as information theory, and computers seemed much less important than either tanks or corn.
President Eisenhower wasn't a kukuruznik, but he had grown up amid corn fields. Like his opposite number, he believed that the competition for world leadership was as much economic as military. Indeed, he had gone much further in that direction than any other general anywhere who had become a head of state. He had cut the military drastically, and his own service, the army, had suffered the most. The reason, he said, was simple. A bloated military could bankrupt the country and lose the cold war just as surely as one that was too weak. Generals had resigned and life-long friendships had been broken, but the 'New Look' army had become a reality.
Mr. Eisenhower hardly knew more about computers than Comrade Khruschev. But, quite apart from the great structural differences between the two economies, it was part of Eisenhower's temperament to run a looser ship and delegate more authority. If someone were to come to Khruschev with a new invention that had nothing to do with agriculture, the military, or heavy industry, that someone would be met with an immediate demand to either demonstrate the usefulness of his invention or depart.
If the same man had come to Eisenhower, the president would have referred him to an expert, and would have given the expert authority to fund the project if it seemed promising. He didn't think that he had to himself understand everything the government did.
It was the American military that supported the early computer projects, the ones that broke new ground but didn't pay. The man at the top, even in the midst of dismantling divisions, saw to it that there was money for experiments and adventures of a new sort. It was that attitude on the part of the president which ultimately made it possible for some forward-looking military officers to invest heavily in computer-aided research. The resulting progress in computer theory and practice was just beginning to have military applications on a scale which the president himself could scarcely have imagined.
At DRI and other such institutes there were many computer people who felt almost as if they were part of a crusade. They believed that computers would, in time, be capable of doing anything that could be done by persons, and then some. They were anxious to get on with the big projects, bigger than nuclear exchanges between super-powers.
The people in DRI who weren't part of the crusade viewed the crusaders as anything from eccentric lunatics to self- promotors who had rather cynically latched on to a good thing. But it was implicitly admitted, even by the people who disliked computers and their slaves, that they were there to stay.
There was really no organized effort, at any level, to put the crusaders in their place as long as they were willing to apply their talents to Red v. Blue.
Bruce Hammond was a quiet crusader, but a convinced and determined one. He, and most of the rest of Complab, felt that an essential part of their mission was the training and development of a corps of computer people whose influence would eventually be felt virtually everywhere.
Tom Williams was far from the first such potential ambassador of computing, and an extremely informal but highly effective training program was already in place before his arrival. He was encouraged to produce little programs almost immediately, and was soon running them.
With machine time costing thousands of dollars an hour, Tom would wait in the wings for a gap when someone finished earlier than expected. He would then dive into the programmer's seat and hand the operator his punched paper tape to be read into the machine. The operator would bear with him for anything from one to four minutes, and then boot him off. If he was lucky, he would learn in that time why his program wouldn't run, and thus be in a position to correct it. The chief operator, Ted Blitz, once said to him,
"With the other programmers, this is called de-bugging. Your programs are too shitty to even have bugs."
Ted didn't exactly smile when he said it, but it wasn't altogether an unpleasant remark. If they had been in the military, Ted would have been a senior non-com and Tom a young officer. Ted knew that Tom would eventually take his place with the other programmers, but Ted wanted to take the present opportunity to share a few truths with him.
Tom's programs gradually got bigger and better. One directed the computer to make a graph with the attached electric typewriter, using the letter 'x' to mark the various points. It was a fairly mundane operation, and the programmer would ordinarily have kept track of the column that was in front of the strike position by adding and subtracting movements of the carriage since the last 'x'. Tom was impressed with the need to save programming steps, and he found it more economical, in that respect, to simply backspace 65 times after each strike, so that he would always be at the left margin of the paper. When no one was looking, he discovered that, if the typewriter was already at the left margin, a backspace only caused it to make a peculiar noise and give a little jump.
Ted happened to be the operator on the morning that Tom ran the program. When Ted looked behind him to see the Flexowriter backspacing furiously, but with futility, he burst out,
"What the fuck is going on?"
When Tom explained, Ted replied, with studied patience,
"You may save a couple of steps, but you'll wear out the fucking Flexowriter."
With that, Tom was booted off.
Tom's other programs were less whimsical, and it was right after a not-so-simple one actually ran that he had his first face-to-face meeting with the director of DRI, Mac Hollins. It appeared that Hollins was on a routine visit to Complab, and had taken that opportunity to acquaint himself with the newest recruit.
Tom rose from his desk to shake hands, but Hollins motioned him to sit down while he himself sat on the table. He asked Tom how he was managing for living arrangements, and, when told that he had taken a room in the home of an old Italian lady in Bethesda, Hollins nodded approvingly. That seemed to be indicative of the approved life-style for a young man just arrived in Washington. Tom then mentioned Hollins' paper, and they soon got into a discussion of it.
Tom was used to talking to professors about their published papers. With Art Burks, one could say anything one liked. With most of the others, one was expected to have some objections, but to find the position and the arguments, if not quite convincing, at least powerful. It was the same with Hollins, and, in the end, Tom was, in fact, pretty well convinced. He then sidled up to the question which most concerned him, but which had seemed improper to ask directly. The culture of DRI was such that one didn't ask questions like,
"Will there be war?,"
and, still less, ones like,
"Will we all be killed?"
Indeed, one didn't ask the former sort of question exactly because it led so easily to the latter. Instead of doing that, Tom asked Hollins whether he thought he could assign probabilities to future wars. That was kosher enough, particularly since a probability would be expressed only as a number between zero and one. Numbers were automatically sanitized, and never had blood dripping from them. Hollins replied,
"Oh yes. Given the incidents of the last ten years, the probability curve peaks in just about five years with a figure of something like point eight eight."
"Of course, there's the issue of the reliability of a probability figure, or the probability of a probability. With nuclear weapons, that might be a little lower than in your studies of the past."
Hollins replied, as if it were no concern of his,
"Yes, I dare say. Not much work has been done on the matter of assessing attributions of probability in a systematic way."
"I might prefer to bet on picking an ace out of a pile of three cards, two of which are aces, to predicting that it will rain, even if I think the probability of rain is greater than two thirds."
Hollins understood immediately.
"Yes. You know that the probability of an ace is two thirds, and you only guess that the probability of rain is more than that. It wouldn't be unreasonable to choose to rely on the known probability even if you suspect it of being lower than the other."
The upshot was that the whole business of predicting nuclear war in the future was less solidly grounded than, say, a prediction of the second world war five years in advance would have been. It wasn't a big sliver of hope, but it was something.
Hollins then seemed to speak rather impulsively,
"You know, we have a few special projects floating around that we never seem to get to in the normal course of events. You might be interested in one of them. You'd still be here in Complab, of course, but you could think of the project as constituting something like a third of your work. Of course, we'd have to talk with Sam Harris about it."
Tom agreed. He had no idea what those special projects might be, but he knew that suggestions from almost any big boss really constituted orders. Sam Harris knew it, too. Hollins said to him,
"I don't want to steal your bright young men almost as soon as you get them, Sam, but this might turn out to be rather important."
Sam took it with good grace. It was then arranged that Tom would spend roughly half his time assisting the director in a special project. He was told that he didn't have to divide each day exactly in half, but that, over the course of weeks and months, he should try to preserve a rough balance between the two kinds of work.
Hollins didn't explain the nature of the project, but said only,
"You might spend a few days learning as much as you can about the history of the Cold War up to the present, and then come in to see me."
As Tom left the meeting with Harris and Hollins, he realized that he was in a tricky positon. They had agreed to share him, but each might expect so much that he would end up working twelve or more hours a day instead of eight.
On the other hand, Tom had never intended to stay at DRI permanently. He had a fellowship waiting for him back at the University of Michigan, and he might stay only until September, perhaps coming back to DRI the next summer. Or, if things at DRI looked interesting, he could stay longer, perhaps taking a year off from graduate school. Anyhow, since he intended ultimately to be a professional philosopher, he really had nothing to lose at DRI. He could take chances the others couldn't.
With Hollins' instructions in mind, Tom took the shuttle back to the headquarters building and went to the DRI main library. The building was a converted mansion, and the large panelled room that was the library looked as if it had hardly been altered. It was empty, and it seemed the quietest place imaginable.
Tom found a whole section devoted to the Cold War, and settled down at an antique table with a selection of books and unpublished materials. As he read about the Korean War (World War 2.6 according to the comedian Mort Sahl), the Suez crisis, and all the lesser incidents, that atmosphere of placidity helped him to put things into a rosier perspective. After all, these things had all happened without bringing about Red v. Blue with nukes. Hollins might easily be wrong.
On his next visit to the library, Tom discovered that there was a librarian, a peculiar man named Maurice Munson. He introduced himself as a historian who had taught at Johns Hopkins, and was thus highly qualified to run what amounted to a historical library. He seemed a bit young to be retired, and it sounded as if he had run into academic trouble somewhere along the line. It was well known that Weston Harrison occasionally hired broken-down academics who aroused his sympathies.
Munson was large, grumpy, and forbidding-looking. He also had a number of odd mannerisms. He would rise from his desk to walk to the window, but would pause mid-way to rotate slowly through 360 degrees with his arms combatatively akimbo. After looking out for a minute or two, never seeming to like what he saw, he would return to his desk. The return journey included another slow rotation, and then, just before sitting, he would spit into the wastebasket.
Fortunately, Munson went for extended coffee breaks and prolonged visits to the head, spending more time outside the library than inside it. Tom found it hard to concentrate on a book when Munson was in the room, going through his routine approximately every five minutes. Even if Tom sat facing away from him, it was hard not to occasionally look over his shoulder to see if Munson was again rotating and spitting.
One day, as Tom was about to leave, Munson came up rather aggressively, looked at his books, and asked,
"Are you working for that shit, Hollins?"
That was startling. DRI was certainly a loose organization, but casually referring to the director as a 'shit' seemed a bit much. Tom was just trying to decide what to say when Munson continued, more affably this time,
"You're just like the other one. He's gone now. You'll soon be gone too."
Tom left immediately, thereby confirming at least part of Munson's prediction. When he later saw Goldstein, he asked about Munson. Goldstein replied,
"He's certifiably loony. Weston is pretty good about recruiting smart people, but he sometimes mistakes pure eccentricity and craziness for intelligence. Or it may be that he really doesn't know the difference."
"Munson casually refers to Hollins as a shit."
"Yes. That's a standing joke. Munson was also a historian, and he's insanely jealous of Hollins."
"Does Hollins know about it?"
"Probably. They should fire Munson, of course. But I guess no one wants to confront him. I bet he keeps a pistol in his brief case."
"He told me that there was someone else doing a project for Hollins who apparently got fired."
"I never heard of anything like that. Probably just another figment of Munson's imagination. More amusing is the fact that he thinks that all of our security people are Russian spies."
"My God! I'm surprised that that hasn't got him fired."
"It doesn't make you a security risk to think that the spy- chasers are spies. And, anyhow, there aren't any secrets in the library."
By this time, Tom was convinced that Munson was truly crazy. On the other hand, he wondered if there were not a grain of truth in what Munson had said about his alleged predecessor.
When Tom stopped in to see Hollins a few days later, the latter asked,
"What have you learned?"
Tom had expected something of the sort, and replied,
"It looks to me is if there have been two adventurers, Stalin and John Foster Dulles. Fortunately, Stalin died just as Dulles was coming into power."
"There was a short overlap which could have been disastrous. But I would agree, basically. Dulles actually makes an explicit policy of moving to the brink of total war to gain limited ends. Stalin did that, in effect, although he didn't, like Dulles, claim that it was the morally virtuous thing to do."
Since Dulles was still secretary of state, it was interesting to learn that Hollins had reservations about him. Tom replied,
"Stalin did it because he wanted to dominate the world. I don't know why Dulles does it."
"There's both an unkind and a kind interpretation of Dulles. The unkind one is that his policy saves money, keeps taxes relatively low, and keeps the Republican Party in office. The Truman-Acheson policy of defending the perimeter of the free world at every point was very expensive. It's cheaper to let some of the defences lapse and threaten massive nuclear retaliation against Red if he steps over the line. And, of course, we have to keep going to the brink to make that threat credible."
"What's the kind interpretation?"
"It's just a little different slant. The assumption is that our people are unwilling to forego luxuries like getting new cars every few years in order to continue to pay the cost of a safer but more expensive containment defence. The massive retaliation-brinkmanship one is then the only defence that remains. So we threaten to turn all of Russia including Siberia into a great mound of radioactive rubble within the space of thirty minutes."
"What do you think?"
"I don't know. But it hardly matters. Dulles and Eisenhower are setting the policy and giving the orders. Our job is to see how they can be executed most efficiently. Is there anything else that strikes you particularly?"
"Well, the other thing is that, after the respite of the Malenkov period, the Russians are getting worse again. They're beginning to match Dulles, dangerous move for dangerous move."
That was evidently what Hollins had been waiting to hear. He smiled and said,
"I think that's the conclusion any intelligent and reasonably objective observer would have to come to. And, of course, it fits into my general theory of a series of incidents of increasing severity leading to war."
Hollins seemed to be much more comfortable with that idea than Tom was, but he continued,
"The trouble is that my theory is only accurate plus or minus a couple of years. We need something that will delve deeper into the actual causes of war and tell us when and why one side or the other will attack."
"Can the military and the CIA detect build-ups in advance?"
"Possibly. We might get a warning a few hours or days in advance. But it likely wouldn't be enough to allow us to hit them before they hit us."
"And your long-range predictions wouldn't be accurate enough to justify a strike?"
"President Eisenhower isn't going to order a nuclear attack just because I think there'll be war in a few years. We need something that at least claims to be able to predict the month of an attack, and which has the prestige of a computer simulation behind it. That's something the people at the top table will find it worthwhile to debate, and our job is to send them things they can debate."
"Is this something DRI is doing?"
"No. We're predicting the outcome of a hypothetical exchange under various conditions. We're not predicting the date of an enemy attack. That's the project I have in mind for you."
Tom was totally amazed and shocked, and supposed that Hollins must be joking. The latter smiled and replied,
"I don't mean that you can do it next week all by yourself. But you can map out the steps that would have to be taken and outline the process. That's the sort of thing philosophers are good at."
"But I'm not a political scientist or historian, and I can't read Russian. Isn't that something for the Russian Research Center at Harvard, or a place like that?"
"Those sorts of people don't know anything about computers, they aren't scientific, and they've never heard of operations research."
"Okay. I can imagine getting to the point where we might use our techniques to throw light on whether it's rational for the Soviet leaders to take a certain action, but that wouldn't tell us whether they'll actually carry it out."
"It's one of my strongest beliefs that politicians and statesmen are among the most rational of people. Getting power, in almost any system, is a winnowing process that selects for them. But you have to remember that their primary goal is always to stay in power. If you factor that in, you can assume that they'll do what it's rational for them to do."
"And that depends on what they think it's rational for us to do in response to various moves that they might make."
"Yes, if they think it's rational for us to launch a nuclear attack in response to a move they're contemplating on Berlin, it's rational for them to either cancel the move or cover it with a nuclear attack of their own."
"So I'll be in the position of trying to simulate their simulation of us, given what we think they know about us."
"As far as that kind of information goes, you can gets lots of help. Your position, and my recommendation, will get you to any of the Pentagon intelligence agencies, and to the CIA. They may not all tell you all they know, but they'll tell you most of what they know on this subject. And that's in addition to the extensive reports that come to us in the ordinary way of things."
There was really only one thing to say.
"Okay, I'll get to work."
"Tell me whenever you need anything, and drop by whenever you have anything to talk about."
It did occur to Tom that his predecessor, if there had been one, could easily have come to grief on a project such as this one.