Daniel Ellsberg will always be the patron saint of whistleblowers. He earned his place in history by leaking the Pentagon Papers, documenting that the American government knew the war in Vietnam was based on lies and going poorly long before it admitted anything of that sort to the public. Vietnam was the end of Ellsberg's official career. He real passion was nuclear war, and trying to make sure that one never occurred. In a twist of fate, the nuclear documents that Ellsberg also copied in the 1970s were lost in a landslide, but decades later much of that material has become available through FOIA and similar requests.

I thought I knew a fair bit about the Cold War and nuclear brinksmanship, and even so this book was astounding. The conventional wisdom is that nuclear war is MAD-Mutually Assured Destruction. Peace is preserved in a tense equilibrium where each side knows that any nuclear exchange will lead to annihilation of it's own population via a sure retaliation. The paradoxical credibility of peace by violence is restrained by the twin promises that nothing can stop the fire and that nukes will only be launched in response to a nuclear attack. The first point is true, the second point is a lie.

As Ellsberg points out, the American government has never disavowed the first use of nuclear weapons, or the potential of a nuclear first strike (First use is any unprovoked use. First strike is massive first use intended to prevent retaliation). Every American president since Truman has used the nuclear arsenal like a robber with a gun. That the gun has not been fired yet is secondary to the basic fact that it is loaded, aimed, and used to compel obedience.

This book is best when it hews closest to Ellsberg's work at RAND and in the White House. Coming out of the Marines and Harvard with a PhD in decision-making under uncertainty, he embarked on a survey of nuclear strategy in the Pacific in 1959 or so, and what he found was incredibly alarming. President Eisenhower had delegated the authority to launch a nuclear strike to CINCPAC in Honolulu, who had further devolved authority to theater commanders on Okinawa, Guam, Korea, and various ships. All of these commands were routinely out of contact with higher headquarters due to distance and poor radio communications. While there was in theory a 'two man rule' that prevented any single officer from broadcasting the order to launch a nuclear strike, in reality every ship and base had procedures for bypassing the two man rule.

Bases practiced alerts on a daily basis and were capable of launching aircraft on 10 minute notice, a stated objective of the attack plan. The attack plan was "fail-safe", in that if an aircraft had not received a go order by the time it reached bingo fuel and had to either commit to the attack or return to base, it would return to base. Strategic Air Command (SAC) practiced full alerts with armed bombers flying to their holding points. In the Pacific, Tactical Air Command merely taxied to the flightlines with bombs. This was both to save fuel and maintenance, and also because the bombs used were not one-point safe, and F-100s were difficult airplanes to flying, meaning there was a small but real chance a plane crash could lead to a nuclear detonation.

As Ellsberg pointed out, visiting a small airbase in the ass-end of Korea, a real alert would be the first time that these pilots had taken off with live bombs. There was also a non-zero chance that plane 8 of 12 would crash on take-off, and the remaining pilots would find themselves out of communications with command, their base enveloped in a mushroom cloud, and with the fate of their world in their hands.

Ellsberg asked the officer in command, a major, what would happen. Would the pilots returned to base as planned? "Yes they would. They're good boys. Well, probably... Hell, if one goes, they might as well all go!" The end of the world could be triggered by an honorable and dutiful officer at the very low rank of major, on his own orders, based on his own very partial understanding of the strategic situation. And there was nothing the entire chain of command, from the President on down, could do to stop it.

Worse than accidents was the actual proper plan. The effort involved in coordinating thousands of aircraft and bombs and avoiding mutual fratricide meant that there was only one plan, a massive all-out attack on the Soviet Union, Communist China, and the Warsaw Pact that would drop thousands of hydrogen bombs in a single spasm until nothing remained in the American arsenal. This plan was to be activated on the event of general war, a conflict with the Soviet Union larger than a skirmish. The plan itself, the Joint Strategic Capabilities Plan, was so secret that it was concealed from the President and the White House staff, the anodyne JSCP acronym also kept secret. It was a plan for genocide. The initial bombardment would kill hundreds of millions. Fire effects were too difficult to estimate, so they were assumed to cause zero casualties. Radioactive fallout would kill an estimated another five hundred million or so, wiping out allies and neutrals in Western Europe and South Asia. Plumes of fallout would drift around the globe, and ash lofted into the stratosphere would trigger a nuclear winter and years-long famine.

In one rather acid summary of his career, Ellsberg describes his life's mission as moving a piece of paper from one desk to a desk with higher authority. The truth about the Vietnam War shifted from the Pentagon to the public. JSCP from the Air Force to the President. Ellsberg joined the Kennedy administration on a part-time leave from RAND, and drafted a new nuclear war plan that proposed leaving cities untouched, hostages for a second round, and reducing the triggering events for nuclear war. There is a lot of canny bureaucratic knife fighting, and great descriptions of the proper deployment of informational memos around the Cuban missile crisis, for those who care about those sorts of things.

The latter half of the book weapons lags as Ellsberg discusses general nuclear strategy, rather than his own experience, but he makes an ironclad case that current nuclear policy in the United States is inherently unsafe and that the soft power gained by joining international arms controls norms would override the veiled, and not-so-veiled threats, made by American Presidents. We've been lucky that there have been no fatal technical glitches, and that at moments of maximum tension people who understood the consequences had the last word, but luck is not enough. Something has to change before the doomsday machine goes off.

As the motto of the Soviet Strategic Rocket Forces goes, "After us - silence".