On October 27, 1962, the 12th day of the Cuban missile crisis, a bellicose and rattled Fidel Castro asked Nikita Khrushchev, his patron, to destroy America.
“I believe that the imperialists’ aggressiveness makes them extremely dangerous,” Castro wrote in a cable to Moscow, “and that if they manage to carry out an invasion of Cuba—a brutal act in violation of universal and moral law—then that would be the moment to eliminate this danger forever, in an act of the most legitimate self-defense. However harsh and terrible the solution, there would be no other.”
We exist today because Khrushchev rejected Castro’s demand. It was Khrushchev, of course, who brought the planet to the threshold of extinction by placing missiles in Cuba, but he had underestimated the American response to the threat. Together with his adversary, John F. Kennedy, he lurched his way toward compromise. “In your cable of October 27 you proposed that we be the first to carry out a nuclear strike against the enemy’s territory,” Khrushchev responded. “Naturally you understand where that would lead us. It would not be a simple strike, but the start of a thermonuclear world war. Dear Comrade Fidel Castro, I find your proposal to be wrong, even though I understand your reasons.”
Castro was 36 years old during the missile crisis. He was 84 when I met him, in Havana, in late summer 2010. He was in semiretirement, though he was still Cuba’s indispensable man. I spent a week with him, discussing, among other things, the Nuclear Age and its diabolical complexities. He still embraced the cruel dogmas of Communist revolution, but he was also somewhat reflective about his mistakes. I was deeply curious about his October 27 cable, and I put this question to him: “At a certain point it seemed logical for you to recommend that the Soviets bomb the U.S. Does what you recommended still seem logical now?” His answer: “After I’ve seen what I’ve seen, and knowing what I know now, it wasn’t worth it.”
[Read: Jeffrey Goldberg discusses Israel and Iran with Fidel Castro]
The problem with wisdom is that it tends to come slowly, if it comes at all. As a species, we are not particularly skilled at making time-pressured, closely reasoned decisions about matters of life and death. The sociobiologist E. O. Wilson described the central problem of humanity this way: “We have Paleolithic emotions, medieval institutions, and godlike technology.” The main challenge of the 80 years since the Trinity atomic test has been that we do not possess the cognitive, spiritual, and emotional capabilities necessary to successfully manage nuclear weapons without the risk of catastrophic failure. Khrushchev and Castro both made terrifying mistakes of analysis and interpretation during the missile crisis. So, too, did several of Kennedy’s advisers, including General Curtis LeMay, the Air Force chief of staff, who argued that a naval blockade of Cuba, unaccompanied by the immediate bombing of missile sites, was “almost as bad as the appeasement at Munich.”
Today, the Global Operations Center of the U.S. Strategic Command, which oversees America’s nuclear forces, is housed in an Offutt Air Force Base building named for LeMay. This decision has always struck me as an indirect endorsement by America’s nuclear establishment of the bias toward action embodied by the sometimes-Strangelovian LeMay. Bias toward action is an all-purpose phrase, but I first heard it in the context of nuclear warfare many years ago from Bruce Blair, a scholar of nonproliferation and a former Air Force missile-launch officer. It means that the nuclear-decision-making scripts that presidents are meant to follow in a crisis assume that Russia (or other adversaries) will attempt to destroy American missiles while they are still in their silos. The goal of nuclear-war planners has traditionally been to send those missiles on their way before they can be neutralized—in the parlance of nuclear planning, to “launch on warning.”
Many of the men who served as president since 1945 have been shocked to learn about the impossibly telescoped time frame in which they have to decide whether to launch. The issue is not one of authority—presidents are absolute nuclear monarchs, and they can do what they wish with America’s nuclear weapons (please see Tom Nichols’s article “The President’s Weapon”). The challenge, as George W. Bush memorably put it, is that a president wouldn’t even have time to get off the “crapper” before having to make a launch decision, a decision that could be based on partial, contradictory, or even false information. Ronald Reagan, when he assumed the presidency, was said to have been shocked that he would have as little as six minutes to make a decision to launch. Barack Obama thought that it was madness to expect a president to make such a decision—the most important that would ever be made by a single person in all of human history—in a matter of minutes.
We are living through one of the more febrile periods of the nuclear era. The contours of World War III are visible in the Russian invasion of Ukraine. Russia has been aided by Iran and North Korea and opposed by Europe and, for the time being, the United States. Pakistan and India, two nuclear states, recently fought a near-war; Iran, which has for decades sought the destruction of Israel through terrorism and other means, has seen its nuclear sites come under attack by Israel and the United States, in what could be termed an act of nonproliferation by force; North Korea continues to expand its nuclear arsenal, and South Korea and Japan, as Ross Andersen details elsewhere in this issue, are considering going nuclear in response.
Humans will need luck to survive this period. We have been favored by fortune before, and not only during the Cuban missile crisis. Over the past 80 years, humanity has been saved repeatedly by individuals who possessed unusually good judgment in situations of appalling stress. Two in particular—Stanislav Petrov and John Kelly—spring to my mind regularly, for different reasons. Petrov is worth understanding because, under terrible pressure, he responded skeptically to an attack warning, quite possibly saving the planet. Kelly did something different, but no less difficult: He steered an unstable president away from escalation and toward negotiation.
In September 1983, Petrov was serving as the duty officer at a Soviet command center when its warning system reported that the United States had launched five missiles at Soviet targets. Relations between the U.S. and the U.S.S.R. were tense; just three weeks earlier, the Soviets had shot down a civilian South Korean airliner. Petrov defied established protocols governing such an alert and declared the launch warning to be false. He understood that the detection system was new and only partially tested. He also knew that Soviet doctrine held that an American attack, should it come, would be overwhelming, and not a mere five missiles. He reported to his superiors that he believed the attack warning to be a mistake, and he prevented a nuclear exchange between the two superpowers by doing so. (Later, it was determined that a Soviet satellite had mistakenly interpreted the interplay between clouds and the sun over Montana and North Dakota as missile launches.)
John Kelly, the retired four-star Marine general who served as White House chief of staff for part of Donald Trump’s first term, is known for his Sisyphean labors on behalf of order in an otherwise anarchic decision-making environment. Kelly, during his 17 months as chief of staff, understood that Trump was particularly dangerous on matters of national security. Trump was ignorant of world affairs, Kelly believed, and authoritarian by instinct. Kelly experienced these flaws directly in 2017, when Trump regularly insulted the North Korean leader, Kim Jong Un, who was widely regarded as inexperienced and unstable himself. After North Korea threatened “physical action” against its enemies, Trump said, “They will be met with fire and fury and frankly power, the likes of which this world has never seen before.”
[Read: John Kelly finally lets loose on Trump]
Kelly repeatedly warned Trump that such language could cause Kim, eager to prove his bona fides to the senior generals around him, to overreact by attacking South Korea. But Trump continued, tweeting: “Military solutions are now fully in place, locked and loaded, should North Korea act unwisely. Hopefully Kim Jong Un will find another path!” Kim later responded by firing missiles over Japan and calling Trump a “mentally deranged U.S. dotard.”
According to reporting in Michael S. Schmidt’s book, Donald Trump v. The United States: Inside the Struggle to Stop a President, Kelly told Trump, “You’re pushing him to prove he’s a man. If you push him into a corner, he may strike out. You don’t want to box him in.” Schmidt wrote, “The president of the United States had no appreciation for the fact that he could bring the country not just to the brink of a war at any moment—but a nuclear war that could easily escalate into the most dangerous one in world history.” Kelly realized that his warnings to Trump weren’t penetrating, so he played, instead, on Trump’s insecurities, and on his need to be a hero, or, at the very least, a salesman. “No president since North Korea became a communist dictatorship has ever tried to reach out,” Kelly told Trump, according to Schmidt. “No president has tried to reason with this guy—you’re a big dealmaker, why don’t you do that.”
Kelly’s diversion worked: Trump quickly became enamored of the idea that he would achieve a history-making rapprochement with North Korea. Kelly understood that such a deal was far-fetched, but the pursuit of a chimera would cause Trump to stop threatening nuclear war.
Trump remains an unstable leader in a world far more unstable than it was during his first term. No president has ever been anything close to a perfect steward of America’s national security and its nuclear arsenal, but Trump is less qualified than almost any previous leader to manage a nuclear crisis. (Only the late-stage, frequently inebriated Richard Nixon was arguably more dangerous.) Trump is highly reactive, sensitive to insult, and incurious. It is unfair to say that he is likely to wake up one morning and decide to use nuclear weapons—he has spoken intermittently about his loathing of such weapons, and of war more generally—but he could very easily mismanage his way, again, into an escalatory spiral.
[From the November 1947 issue: Albert Einstein on avoiding atomic war]
The successful end of the Cold War caused many people to believe that the threat of nuclear war had receded. It has historically been difficult to get people to think about the unthinkable. In an article for this magazine in 1947, Albert Einstein explained:
The public, having been warned of the horrible nature of atomic warfare, has done nothing about it, and to a large extent has dismissed the warning from its consciousness. A danger that cannot be averted had perhaps better be forgotten; or a danger against which every possible precaution has been taken also had probably better be forgotten.
We forget at our peril. We forget that 80 years after the world-changing summer of 1945, Russia and the United States alone possess enough nuclear firepower to destroy the world many times over; we forget that China is becoming a near-peer adversary of the U.S.; we forget that the history of the Nuclear Age is filled with near misses, accidents, and wild misinterpretations of reality; and we forget that most humans aren’t quite as creative, independent-minded, and perspicacious as Stanislav Petrov and John Kelly.
Most of all, we forget the rule articulated by the mathematician and cryptologist Martin Hellman: that the only way to survive Russian roulette is to stop playing.
This article appears in the August 2025 print edition with the headline “Nuclear Roulette.” When you buy a book using a link on this page, we receive a commission. Thank you for supporting The Atlantic.