Indeed, von Neumann had consciously modeled MANIAC after the human brain. “I propose to store everything that has to be remembered by the machine, in these memory organs,” von Neumann wrote, including “the coded, logical instructions which define the problem and control the functioning of the machine.” In this way, MANIAC became the world’s first modern stored-program computer. Von Neumann’s friend and colleague Edward Teller saw great promise in the computer and used MANIAC to perform calculations for the hydrogen bomb.
After two and a half years of work, the team at Princeton tested MANIAC against von Neumann’s own brain. Initially, von Neumann was able to compute numbers in his head faster than the machine. But as his assistants entered more and more complicated computational requests, von Neumann finally did what human beings do: he erred. The computer did not. It was a revelatory moment in the history of defense science. A machine had just outperformed a brain the Pentagon relied on, one of the greatest minds in the world.
The Pentagon’s strategy for nuclear deterrence in the 1950s was based on a notion called mutual assured destruction, or MAD. This was the proposition that neither the Soviets nor the Americans would be willing to launch a nuclear attack against the other because that action would ensure a reciprocal action and ultimately guarantee both sides’ demise. At RAND, analysts began applying the Prisoner’s Dilemma strategy to a nuclear launch, keeping in mind that the driving principle of the dilemma was distrust. This led a RAND analyst named Albert Wohlstetter to start poking holes in the notion that MAD offered security. The way Wohlstetter saw it, MAD most definitely did not. He argued that if one side figured out a way to decapitate the other in a so-called “first strike,” it might be tempted to launch an unprovoked attack to ensure its superiority. The only solution, said Wohlstetter, was to develop a new nuclear strategy whereby the United States had more nuclear weapons in more hardened missile silos secreted around the American countryside than the Soviets could decapitate in a preemptive strike. Wohlstetter’s famous theory became known as “second strike.” U.S. policy regarding second strike deterrence took on the acronym NUTS, for nuclear utilization target selection.
President Eisenhower began to see the madness of it all. The year after Castle Bravo, the Soviets successfully tested their own deliverable hydrogen bomb. If something wasn’t done to stop it, the arms race would only continue to escalate. Speaking to his cabinet, Eisenhower wondered if it was possible to put an end to nuclear weapons tests. He launched his administration’s first investigation into the possibility of stopping nuclear science in its tracks. His vision was short-lived. After a month of study and discussion, the State Department, the Atomic Energy Commission, the CIA, and the Department of Defense were all unanimous in their opposition to ending nuclear tests. Atmospheric nuclear weapons tests must continue, they all said. The safety and security of the country depended on more nuclear weapons and more nuclear weapons tests. The president’s advisors instead encouraged him to focus his attention on strengthening a national effort to protect civilians in the event of a Soviet nuclear attack, an unpopular program called civil defense. This job fell to the Federal Civil Defense Administration, a three-year-old agency with headquarters in Washington, D.C.
The plan for civil defense in the mid-1950s was to have people prepare to live underground for a period of time after a nuclear attack. An effort to build a national network of underground bunkers had been moving forward in fits and starts. The president’s advisors told him that his endorsement would boost morale. But the very idea of promoting civil defense put Eisenhower in an intractable bind. Ever since he had been shown the fallout map from the Castle Bravo bomb, Eisenhower knew how implausible a civil defense program was—how many tens of millions of Americans were destined to die in the first few hours of a nuclear attack. The idea that there was safety to be found in a civilian underground bunker program was apocryphal. One needed to look no further than what had happened to the men in the Station 70 bunker. Station 70 was a windowless bunker carefully constructed of three-foot-thick concrete walls with steel doors, buried under ten feet of dirt and sand. It was surrounded by a moat and had a secondary blast buttress wall. And even with a 10,000 factor of shielding, the radiation nearly killed the men inside; they barely made it off Enyu alive. After taking cover in the bunker’s urinal for eleven hours, the men were ultimately evacuated from the death zone by two Army helicopters in a carefully orchestrated military operation. The helicopter pilots were part of a ten-thousand–man task force, with unlimited access to state-of-the-art rescue and communication equipment. The rescue teams had fewer than one dozen rescue operations to perform, the majority of which had been rehearsed. Castle Bravo was a highly organized scientific test. In a real nuclear attack, there would be carnage and mayhem. Each person would be on his or her own.
To be caught outside, en route to a civil defense shelter, even forty miles away from ground zero, would be life threatening. The bomb blast and shock waves would rupture lungs, shred eardrums, and cause organs to rupture and bleed. Debris—uprooted trees, sheets of metal, broken glass, electrical wires, wood, rocks, pipes, poles—everything would be ripped apart and hurled through the air at speeds of up to 150 miles per hour. How, in good conscience, could the president urge the public to support a program he knew was more than likely going to kill so many of them?
Paradoxically, in the event of a Soviet nuclear attack, there was a fully formed plan in place to keep the president and his cabinet alive. An executive branch version of the Station 70 bunker had recently been completed six miles north of Camp David, just over the Pennsylvania state line. This underground command center, called the Raven Rock Mountain Complex, was buried inside a mountain of granite, giving the president protection equivalent to that of walls a thousand feet thick. The Raven Rock complex, also called Site R, had been designed to withstand a direct hit from a 15-megaton bomb. The idea of an underground presidential bunker was first conceived by U.S. Army military intelligence (G-2) during postwar examination of the underground bunker complexes of the Third Reich. The survival of so many of the Nazi high command in Berlin was predicated on the underground engineering skills of a few top Nazi scientists, including Franz Xaver Dorsch, Walter Schieber, and Georg Rickhey, all three of whom were hired by the U.S. Army to work on secret U.S. underground engineering projects after the war, as part of Operation Paperclip.
Plans for Raven Rock were first drawn up in 1948, including some by Rickhey. Work began shortly after the Russians detonated their own atomic bomb, known in the West as Joe-1, in August 1949, and by 1950, construction crews with top secret clearances were working around the clock to build the first underground presidential bunker and command post. Site R was a three-story complex with living quarters for the president and his advisors, a hospital, chapel, barbershop, library, and water reservoir. By the time the bunker was finished, in 1954, the costs had reached $1 billion (roughly $9 billion in 2015).
In the event of a nuclear strike, the president would be helicoptered from the White House lawn to the landing pad at Raven Rock, a trip that would take roughly thirty-five minutes. But the prospect of retreating underground in the event of a nuclear strike made President Eisenhower despondent. To his cabinet he expressed his view of what governance would be like after a nuclear attack: “Government which goes on with some kind of continuity will be like a one-eyed man in the land of the blind.”
While the president lived with his conundrum, the civil defense program grew. The details of the Castle Bravo test remained classified, as did the existence of the Raven Rock command center, leaving the public in the dark as to the implausibility of civil defense. Nuclear tests continued unabated, in Nevada and in the Marshall Islands. But the press attention created by the Castle Bravo fallout debate began to generate strong negative responses to the viability of civil defense.
In February 1955 the Senate Armed Services Committee opened a federal investigation into what civil defense really meant for the American p
eople. The investigating committee was headed by a Tennessee Democrat, Senator Estes Kefauver, known for his crusades against organized crime and antitrust violations. The Senate sessions would become known as the Kefauver hearings, and in the course of them, shocking new information came to light.
Civil defense had a two-pronged focus: on those who would stay in the city and seek shelter, and on those who would try to leave. In the event of a nuclear attack, which would likely target a big city, some people living in urban centers were advised to hurry to air-raid-type shelters that had been built underground. As for those who could leave, the Federal Civil Defense Administration said that they should evacuate the cities, promising that this was a better alternative. During the hearings, the senators had questions. In the mid-1950s, most land outside big cities was little more than open countryside. Where were citizens supposed to evacuate to? And what were they supposed to eat?
The director of the Federal Civil Defense Administration, Frederick “Val” Peterson, took the stand. The former Nebraska governor was under oath. He revealed that the plan of the administration was to dig roadside trenches along public highways leading out of all the big cities across the nation. The trenches were to be three feet deep and two feet wide. When the bombs hit the cities, Peterson said, people who had already made it out were to stop driving, abandon their automobiles, lie down in the trenches, and cover themselves with dirt. Senator Kefauver, learning this along with the public for the first time, was dumbfounded. The government could use science and technology to create power as great as that generated by the sun, but when it came to civil defense, this was the best they could come up with? What about “food, water [and] sanitation in [these] trenches?” the incredulous Kefauver asked. Peterson fumbled for an answer. “Obviously, in these trenches, if they are built on an emergency basis, there would be no provisions for sanitation,” he admitted. But there was an alternative plan. Instead of the dirt trenches, another idea being discussed involved using concrete pipes, four feet in diameter, to be laid down alongside the highways. When the bombs hit the cities, Peterson said, people who had already made it out would stop driving, abandon their automobiles, and crawl into the pipes. Sometime thereafter, Peterson explained, federal emergency crews would come along and bury the pipes with earth.
Senator Leverett Saltonstall, a Republican from Massachusetts, expressed astonishment. He told Peterson that he found it impossible to imagine millions of “shell-shocked evacuees waiting out a nuclear war inside concrete pipes,” without fresh air, water, sanitation, food, or medical care. And for who knew how long. Senator Saltonstall said he would rather lie down in a dirt ditch “than get into a concrete pipe a mile long, with no exit.” Saltonstall shared his vision of being crushed in the mayhem by fellow American citizens fighting to stay alive.
Next came the issue of food. Committee members wanted to know how the government was going to help feed evacuees after a nuclear exchange. Peterson replied that the United States would open food kitchens, but there would be little food to be served. “We can’t eat canned foods,” he explained, because radiation could penetrate tin cans. “We won’t eat refrigerated foods,” he conceded, because most electricity would be out. The truth was not pretty, he acknowledged, but was “stark, elemental, brutal, filthy and miserable,” he said under oath. “We will eat gruel made of wheat cooked as it comes out of the fields and corn parched and animals slaughtered as we catch them before radioactivity destroys them.” The committee told Peterson his agency’s plans for evacuation were inadequate. In a matter of hours, the notion of civil defense became the subject of national ridicule. And yet the nuclear tests continued unabated.
Over the next two years, the United States exploded eighteen nuclear weapons; the Soviet Union exploded twenty-five. Nuclear spending was at an all-time high, and design originality was key. The Pentagon ordered hundreds of high-yield hydrogen bomb warheads, like the one detonated during Castle Bravo, but also smaller, lighter-weight tactical atomic bombs. Herb York flew to Washington, D.C., with a full-scale mockup of Livermore’s newest design, the forty-eight-pound Davy Crockett nuclear weapon, in his carry-on bag. The Davy Crockett had the same yield as the atomic bomb dropped on Hiroshima, but advances in science meant that the powerful weapon was small enough to be handheld. Thanks to ambition and ingenuity, the Livermore laboratory had begun to pull ahead from behind. The computer designed by John von Neumann played an important role in allowing Livermore scientists to model new nuclear weapons designs before building them.
In the summer of 1955, John von Neumann was diagnosed with cancer. He had slipped and fallen, and when doctors examined him, they discovered that he had an advanced, metastasizing cancerous tumor in his collarbone. By November his spine was affected, and in January 1956 von Neumann was confined to a wheelchair. In March he entered a guarded room at Walter Reed Hospital, the U.S. Army’s flagship medical center, outside Washington, D.C. John von Neumann, at the age of fifty-four, racked with pain and riddled with terror, was dying of a cancer he most likely developed because of a speck of plutonium he inhaled at Los Alamos during the war. Two armed military guards never left his side.
For a while, von Neumann’s mind remained sharp, but as the end grew near, his mental faculties began to degrade. Beside him at his bed, von Neumann’s brother Michael read aloud from Goethe’s tragic play Faust. Michael would read a page and then pause. Lying on the hospital bed, eyes closed, faculties failing, for some time von Neumann could still pick up in the text precisely where his brother left off. But soon, even John von Neumann’s indomitable memory would fail. Friends said the mental decline was excruciating for him to endure. An atheist all his life, von Neumann used to joke about people who believed in God. In a limerick for his wife, Klara, he’d once written, “There was a young man who said, Run! / The end of the world had begun! / The one I fear most / Is that damn Holy Ghost. / I can handle the Father and Son.” Now von Neumann sought God and he called upon the services of a Roman Catholic priest.
But death grew near. In von Neumann’s final, frightened last days, even the priest could not offer a reprieve. Weeks before von Neumann died, Herb York went to Walter Reed hospital to pay his final respects. “Johnny was in a bed with high, criblike sides, intended to keep him from falling out or otherwise getting out on his own,” York recalled. “I tried to start a conversation about some technical topic I thought would interest and divert him, but he would say no more than a simple hello.” Von Neumann’s brain was failing him. Cancer was robbing him of the thing he valued most, his own mind. Soon he would not remember. In weeks there would be nothing left of him. John von Neumann died on February 8, 1957.
He left behind a single unfinished manuscript that he had been working on in his final months of life. It was called “The Computer and the Brain.” A copy was made for the Los Alamos Scientific Laboratory library, where it remains today. In this paper, von Neumann draws a comparison between the computer and the human nervous system. He theorizes that one day the computer will be able to outperform the human nervous system by infinite orders of magnitude. He calls this advanced computer an “artificial automaton that has been constructed for human use.” John von Neumann believed computers would one day be able to think.
CHAPTER THREE
Vast Weapons Systems of the Future
It was October 4, 1957, 6:00 p.m. Cocktail hour at the Officers Club at the Army Ballistic Missile Agency in Huntsville, Alabama, or “Rocket City, USA.” Neil H. McElroy, a corporate executive soon to be confirmed as secretary of defense, had just arrived in a military jet with an entourage of defense officials from the Pentagon. Inside the Officers Club, drinks flowed freely. Appetizers were passed among the men. McElroy stood chatting with Wernher von Braun, the famous German rocket scientist who now served as director of development operations at Huntsville, when a press officer named Gordon Harris rushed into the room and interrupted the party with an extraordinary announcement.
“The Russians have put up
a successful satellite!” Harris shouted.
The room fell silent. For several moments only the background music and the tinkling of ice cubes could be heard.
“It’s broadcasting signals on a common frequency,” Harris said. “At least one of our local ‘hams’ has been listening to it.” A barrage of questions followed.
It did not take long for news of Sputnik to become official. The Soviet news agency, TASS, released a statement providing technical information and specifics about Iskusstvennyy Sputnik Zemli, or “artificial satellite of the earth.” The Soviets had beaten the Americans into space. Not since Pearl Harbor had the Pentagon been caught by a surprise of such consequence.
The nation slipped into a panic over what was seen as superior Soviet scientific prowess. Eisenhower’s attempts to minimize the significance of Sputnik had a reverse effect, with many Americans accusing the president of trying to conceal U.S. military weakness. Sputnik weighed only 184 pounds, but it had been launched into space by a Soviet ICBM. Soon the Soviet ICBM would be able to carry a much heavier payload—such as a nuclear bomb—halfway across the world to any target in the United States.
The situation was made worse when, on December 20, 1957, someone leaked a top secret analysis of the Soviet threat, called the Gaither Report, to the Washington Post. The report “portrays a United States in the gravest danger in its history,” wrote the Post. “It shows an America exposed to an almost immediate threat from the missile-bristling Soviet Union.” If Sputnik had caused mild panic, the Gaither Report produced national hysteria.