The American public doesn’t talk enough about nuclear war.

Sure, sometimes it gets brought up. In the openings of disaster movies and in the plots of video games, we’re reminded of the horrors that would accompany any major exchange of nuclear weapons. But this makes ICBMs seem science-fictional. In politics and technology coverage—when we discus the realistic future of copious student-loan debt and the occasional self-driving car—nuclear weapons appear only sometimes, in incoherent debate answers or bizarre warmongering. It can feel like we lack even the tools to talk about nuclear war.

For 70 years, the Bulletin of Atomic Scientists has tried to get people to talk more about it. The centerpiece of its approach is the Doomsday Clock, a giant timepiece that metaphorically shows how close humanity is to global catastrophe. On Tuesday, the Bulletin announced that this year the clock would stand at three minutes to midnight, where it has been since last year when it was bumped up from five minutes to midnight.

As William Perry, the former U.S. secretary of defense, said at the Bulletin’s press conference: Three minutes to midnight is a “more dangerous, more ominous forecast than two thirds of the years during the Cold War.”

Yet it can be hard to figure out what exactly the Doomsday Clock is forecasting, because the Bulletin doesn’t countenance only nuclear conflict anymore. The scientists and security experts of the Bulletin now look to four different risk areas when they make their report. The biggest of these remain, first, the possibility of a major conflict between nuclear states, and, second, out-of-control climate change. But the Bulletin also follows the risks of a civilian nuclear-power disaster, especially around waste storage, and what it calls “emerging technologies,” like genetically modified pandemics or destructive artificial intelligence.

“What connects all of those four? For us, it’s about existential threat—the things that can end life on this planet as we know it,” says Rachel Bronson, the executive director of the Bulletin.

She explained why the clock now includes more than the nuclear threat:“For a moment in time each year, the bulletin is able to prompt this global conversation on these truly existential issues that are really hard to discuss. It gives us a benchmark for where a set of leading experts are,” she told me Wednesday. “We were trending in the U.S. yesterday, we were top ten, and that blew me away. That’s Jimmy Fallon territory.”

It’s true. More than 48 hours after its announcement ended, the Bulletin was still trending on Facebook and Twitter—and this wasn’t even a year when they moved the clock hand. But speaking to Bronson, and reading the Bulletin of Atomic Scientists’s annual briefing to accompany the Doomsday Clock announcement, I had the same thought about this being a rare opportunity to discuss these existential issues—though I’m not sure whether I shared Bronson’s positivity.

Very few organizations talk regularly about the destructiveness of nuclear weapons anymore, and those that do can can seem mysterious. It’s only them, and the deterrent power of the weapons themselves, that stand between us and fiery, anguished deaths. Yet when you look into the process of setting the clock, it seems unequal to the task: It’s 10 or 20 people, of exceptional but not supernatural ability, with well-rounded but not omniscient expertise, sitting in a room for a day and doing the best they can to communicate disaster.

* * *

Sitting in a room is more or less how these decisions get made. Despite science being in the Bulletin’s name, its methods are more Socratic than anything else. Every year, the ten scientists and security experts on its board gather for a one-day discussion where they review what worried them last year and anticipate new concerns. There are no minute-hand-divining devices and no instability-predicting supercomputers: Just a bunch of experienced adults, trying to come to a consensus about how bad things have gotten.

Bronson told me that the board members “know coming into it what the big events of the year were—positive and negative—and they know what they’re concerned about for years going forward. And it becomes a professional discussion. Do their colleagues see it the same way? Who sees it differently and why?”

“The mood in the room is very sober, and it’s serious. The board takes their job very seriously because they know that it gets a lot of attention. They know the public will respond to this,” she said.

The Bulletin’s staff is also always trying to adjust the balance of expertise in the room. The current membership of the Science and Security Board includes the former director of Argonne National Laboratory, the lead author of the UN Intergovernmental Panel on Climate Change, a former rear admiral in the U.S. Navy who is also an Arctic climate expert, and an expert on military planning and nuclear-associated firestorms. Members of the Bulletin’s board of sponsors—an advisory panel that includes more than a dozen Nobel laureates in physics and medicine—also sometimes sit in on the meeting, Bronson said.

One of these sponsors, Stephen Hawking, has said he is increasingly concerned about “bad” artificial intelligence. (For better or for worse, similar concerns among the Silicon Valley set have funded oodles of new research into existential risk, generally.) Bronson said evil A.I. was a risk that the Bulletin already took into consideration and that it was thinking about pursuing it further. But how many new and edgy threats to life-as-we-know-it should the Bulletin of Atomic Scientists attempt to consider, given that the old and traditional threat—and the only threat, by the way, mentioned in their name—still looms?

In 1991, as the Cold War drew to a close, the Doomsday Clock shifted to its most optimistic position ever: 17 minutes to midnight. The Soviet Union was splitting apart, and both American and the former Soviet militaries were engaged together in a concerted disarmament program. In the quarter-century since, the clock has edged closer and closer to doom. In 1998, it responded to successful nuclear tests in India and Pakistan; in 2002, it reached seven minutes to midnight, as the U.S. withdrew from the Anti-Ballistic Missile Treaty and as concerns about nuclear terrorism grew.

By 2007, North Korea had successfully tested its first nuclear weapon. Iran was also growing its nuclear-weapons program. But scanning the horizon, the Bulletin of Atomic Scientists saw new varieties of catastrophic possibility. The largest among these was climate change. So as the Board set the clock to five minutes to midnight that year, it adopted a new, hybrid Doomsday Clock approach. The clock would now account for more than just global war.

Was that the right choice? On Wednesday, Bronson said that the clock’s mandate has always been about species-wide catastrophe, no matter its origin. “What the clock is about is existential threat. In 1947, when the clock first appeared, the existential threat of the day was nuclear weapons. And they clearly still are. But in 2007, there was a recognition that something else threatens life on earth as we know it,” she said. “There’s no doubt that climate change can end life on earth as we know it.”

As a reporter who writes a weekly climate newsletter called Not Doomed Yet, I’m not quite as sure about that (although I wouldn’t want to doubt the Bulletin). Climate change endangers contemporary civilization and its myriad supply chains—there’s no question about that. Climate-associated cataclysm will kill millions of people this century. But the threat of climate change is only existential insofar as it increases the possibility of nuclear war.  

As Oliver Morton, an Economist editor who spent six years researching and writing about geoengineering, told The Atlantic last year: “What I really worry about with geoengineering is that conflict over its use will lead to a greater conflict that leads to a nuclear war. […] We don’t even know if anyone’s going to try geoengineering, but we know the wherewithal to have a nuclear war is out there in the world already.”

Which isn’t to say that climate shouldn’t be included in the clock. The clock is ultimately a political tool, and, by 2007, the state of international climate politics had become dire. The previous year, An Inconvenient Truth had made climate change real to millions, but the United States was taking no real steps toward mitigating it. The Kyoto Protocol had failed, but the annual UN climate conference, held in late 2006, barely discussed the possibility of limiting emissions.

Much of this has since improved. The Paris climate talks this year ended with an ambitious agreement to mitigate the worst effects of climate change—though the world’s path to actually following through on that ambition is still unclear. The U.S. energy system seems to be reforming, with halts and starts; and China may even be reining in its coal use. But the Board considers it too soon to tell whether the Paris Agreement—and the Iran Nuclear Deal, for that matter—will be successful, though it called them “bright spots.”

Would the Bulletin ever consider multiple clocks? In a small way, they already have. Bronson pointed me to the Bulletin’s pleasantly alliterative Doomsday Dashboard, an online attempt to pull apart some of the most concerning trends.

And the point, above all, is to force the news conversation to recognize existential risk, if only for a day. “I do think there is a power in bringing all these different threads together for a moment and saying, ‘How seriously are our policymakers taking this?,’” Bronson said. Multiple, issue-specific clocks would only water that down. Besides, interested parties can always read the Bulletin’s five-page synthesis report.

* * *

The clock has ticked to three minutes to midnight twice before: in 1949, after the Soviet Union first tested an atomic bomb and the nuclear-weapons race began; and in 1984, as American-Soviet relations deteriorated and both states deployed short-range nuclear missiles throughout Europe.

Only once has the Bulletin edged the clock to the two-minute mark: in 1952, when both the U.S. and Russia tested the first thermonuclear hydrogen bombs.

The world didn’t end during those moments of crisis. But looking back now, what’s striking is that the world came much closer to catastrophe than was even understood at the time. The United States and Russia nearly tumbled into World War III twice in the year 1983 alone.

One of those near-misses is now famous. On September 26, 1983, a bug in the U.S.S.R. early-warning system reported that five NATO nuclear missiles had been launched and were bound for Russian targets. The officer watching the system, Stanislav Petrov, had also designed the system, and he decided that any real NATO first-strike would involve hundreds of I.C.B.M.s. Therefore, he resolved the computers must be malfunctioning. He did not fire a response.

Only one month later, a NATO military exercise across Europe called Able Archer nearly started another war with Russia. As the U.S. practiced its nuclear routines in Western Europe, the U.S.S.R. was responding in kind and in earnest, transporting missiles across the country and assigning strategic targets. By 1990, the President’s Foreign Intelligence Advisory Board could report that: “In 1983, we may have inadvertently placed our relations with the Soviet Union on a hair trigger.”

Yet that last incident wasn’t publicly known until 2015, when the advisory board’s report was finally declassified. The New York Times published the first public account of the “war scare of 1983” less than three months ago.

In a post for the Bulletin in December, Perry, the former secretary of defense, laid out five “nuclear nightmares” for 2016. Chief among these is still a NATO war with Russia, followed by American conflict with China, an erratic launch by North Korea, an India-Pakistan war, and a nuclear-enabled ISIS. “These five nuclear nightmares add up to a danger to our people that is greater in some ways than the nuclear dangers we faced during the Cold War. But most Americans—especially our youth—are blissfully unaware of those dangers,” he writes.

Elsewhere, Perry has said he believes that the clock stands at five minutes to midnight for nuclear conflict, but only one minute to midnight for nuclear terrorism.

Though he never says it, that all these catastrophes “add up” to a grave but combined risk would seem to be exactly the problem. During the Cold War, the Bulletin of Atomic Scientists could follow a relatively discrete number of issues. Now, the catastrophic issues that it tries to watch are complex and interrelated and prone to agglomerate with each other. It’s left to discuss the many fail-points of contemporary society with a 20th-century metaphor—and, especially when it talks about nuclear security, it’s one of the very few organizations with a platform to do this.

Many writers and experts have noted of late that the world is, without ever quite deciding it, drifting back into a second nuclear arms race. The most worrisome trend to the Bulletin of Atomic Scientists, the one they keep bringing up, are the nuclear-weapons modernization programs ongoing in the U.S. and Russia. In years to come, the United States will spend between $100 billion and $1 trillion retrofitting its nukes for another four decades of service. Nuclear weapons are still a reality in the world, and a reality in our future—as much as climate change, as much as geoengineering, as much as evil artificial intelligence.

Since the dawn of the nuclear era, the Bulletin of Atomic Scientists has annually reminded us of the danger of the nuclear threat. It would be nice if they weren’t the only ones.