Welcome to Galaxy Brain—a newsletter from Charlie Warzel about technology, media, culture, and big ideas. You can read what this is all about here. If you like what you see, consider forwarding it to a friend or two. We’re still figuring things out in our new home so let me know what you think: [email protected].
This is a free edition of the newsletter, but you can subscribe to The Atlantic to get access to all posts. Past editions I’m proud of include: Guardians of the Internet, Don’t Alienate the Willing, and How to Spend 432,870 Minutes on Spotify in a Year.
Devoted Galaxy Brain readers might remember a post from my Substack days in which I interviewed Tom Neill, a Londoner who, while bored in lockdown, built a silly website tracking the container ship that was then currently (and rather gloriously) wedged in the Suez Canal. I am generally fascinated by what it’s like when something a person builds is at the center of a viral storm (as Neill’s site was). That interview is still one of my favorite newsletters. Today, I’m continuing the series—though under far more serious circumstances.
The first time I used Nukemap was sometime in 2013, when I was living in New York City. I’m not certain how I stumbled on the site, but I know I spent at least an hour toying with it. Nukemap is a nuclear-effects calculator, which is to say, a website that shows you the various radii of destruction, if a nuclear bomb went off in the given location. You can customize the area of detonation, the size of the bomb, and other details. The site is somewhat of a cult classic, with over 220 million “detonations” logged since it came online in 2012. The results it shows are, as you can imagine, sobering—especially now.
Last week, after Russia launched its invasion of Ukraine, President Vladimir Putin ordered his nuclear forces into a higher state of alert. It was the first time the Kremlin had done this since the Russian Federation was established in 1991. I’m not a nuclear expert at all, so I’m not going to speculate about what this means—you can read my colleague Tom Nichols on that—but the move has brought the notion of nuclear war back into the global conversation in a nontrivial way. As you might expect, Nukemap’s traffic has surged, and the site has, at some points, crashed.
I reached out to Nukemap’s creator, Alex Wellerstein, to talk about the site’s creation, how people have used it for the last decade, and what he’s seen change in the last week. Wellerstein is a historian of science and nuclear weapons and a professor at the Stevens Institute of Technology. Seriousness of the subject matter aside, it’s one of my favorite interviews in a while. The conversation has been lightly edited for clarity.
Charlie Warzel: Let’s start at the beginning—over a decade ago now—why’d you build Nukemap?
Alex Wellerstein: There’s the true story and the one that makes me look more rational and smart. Let’s start with the one that makes me look smart. I recognized that it’s difficult to conceptualize the size of nuclear weapons. I have a very hard time dealing with numbers and visualizing them and translating these equations into code that makes a visualization allows me to better understand these weapons for my job.
The real story, which is more haphazard, is that I first created a visualization tool for this a long time ago, back in 2004, before Google Maps. I used a terrible series of screenshots of MapQuest and it was a mess. I didn’t have the technology to do what I wanted.
That changed in 2012. I had started a blog, and as you, a content creator, know well, you need content! I was investigating a totally different nuke project, which required drawing circles to show nuclear radii. And in the process I realized the code had matured and what was hard to do in 2004 was trivial in 2012. I figured I’d bang it and some of my fellow academics and policy people would find it interesting. I didn’t imagine it’d get seen by the general public. I built it in the course of a weekend and I sent it to some colleagues and gave it a terrible name—something like Alex’s Amazing Nuclear Weapon Simulator. Everyone told me that name was awful, so I changed it to Nukemap.
Warzel: Did it take off right away?
Wellerstein: Not quite. It got a little traffic from my blog. My blog is about the history of nuclear weapons, so we’re not talking about tons of traffic here. But then, somehow, the U.K. tabloids picked it up and wrote a terrible story about this new viral nuke tool that was so scary. Nukemap hadn’t gone viral but, as I learned then, a story about something going viral can make it go viral. I got tons of traffic, and that made me take the project more seriously. [I figured] if this is a tool for other people, then I should probably spend more than two days working on it. So I upgraded the features, gave the code an overhaul, made fallout models and casualties models, all by 2013. And that’s where it is now, basically. For better or worse it’s the most engaged-with thing I’ve ever done. So far, about 40 million people have used it. It’s been a key part of my career and likely led to the job I have now.
Warzel: It’s always the things you don’t expect that have the biggest impact online, I find.
Wellerstein: I can come up with reasons why Nukemap works but ultimately some of this is idiosyncratic. If the U.K. tabloids don’t pick it up, does it get to this level of popularity? Now it is high in the search results for how big is a nuclear explosion? With algorithms and social media, the more attention you get the more you’re going to get. It becomes self-building.
Warzel: Has traffic been steady since the tabloid stories?
Wellerstein: Over time you get peaks. These range from crisis moments, like North Korea testing a missile or what is happening right now in Ukraine, to moments that are rather innocuous, like the 70th anniversary of the Hiroshima bombing, where people are looking at it in schools or wondering what it would be like if that bomb had been dropped on their hometown. I’ve found those moments where people are using Nukemap with existential dread are less spikes and more of a plateau that lasts for a while. Those events give you a user peak and a really long tail to it, and the baseline usage then steadily grows over time—the next peak is bigger than the last.
Warzel: Are there interesting insights you’ve been able to glean from it about how people use the tool?
Wellerstein: In terms of preliminary findings, I did some analysis a while back—it’s not up to date, but it hints at something really interesting. Something I’ve seen is that users in different countries use it differently. There are two categories of how people use Nukemap. The first is cathartic nuking, which is nuking somebody else. Say, Americans are mad at Russia, so they’re seeing what happens when you do it to somebody you don’t like. The second is experiential nuking, or nuking yourself to see what happens if it happens to me.
Warzel: How do Americans use it?
Wellerstein: Last time I ran this, Americans by far nuke themselves most of the time. They prefer experiential nuking. I’m not going to go so far as to say it’s narcissistic, but our main mode of using Nukemap is to look and see what will happen to us. When Americans nuke somebody else, the No. 1 country—last time I did it—was Japan. There’s a reason for this, and it’s that they’re re-creating the Hiroshima and Nagasaki events. After that, Americans tend to nuke all other countries equally. Japan is really the only one that stands out.
Warzel: How about other countries?
Wellerstein: Israel is one of the few countries that doesn’t nuke itself more than other people. Now, for lots of countries there aren’t huge sample sizes, so maybe this information isn’t that significant. Israel Nukemap users’ No. 1 target is Iran. If you look at Iranians, which is a small sample size, their No. 1 target besides themselves was Russia. Is this information useful? I don’t know. But you can tell little stories in your head about what’s going on. Or at least I do. The other data are not as interesting. People always choose the biggest bombs to drop first. It’s what you’d expect with risk calibration. They’re getting a sense of the worst-case scenario.
Warzel: I’ve used this website over the years on different occasions and find something almost weirdly soothing about using the tool. Do you get feedback as to how people are mostly using it? Is it in an academic capacity? Is it a kind of doomscrolling?
Wellerstein: If you look at what people post on social media—it’s all over the place. Some people are scared. Some do it for fun. Some people are doing deadly serious calculations about their families, based on the emails I’ve received from people. Some of them make me really uncomfortable—they’re describing their family situations and asking me if they’ll survive based on specific calculations. I try to give people the best general advice I can, and I’ll just say right now, I don’t think they need to get in the basement—not at this stage in the current conflict. But it’s always good to be aware and prepared. I also like to point out that in an actual nuclear war, you don’t know the targets and what the exact weapons will be. Nukemap is a tidy way of looking at this, but the reality wouldn’t be tidy. Plus, it’s a website that is created and run by one guy.
Warzel: But people seem to really rely on it! I wonder if the site design has anything to do with that?
Warzel: I think there’s something to that. There’s not really a timeless aesthetic on the internet, but I think that late-1990s vibe might be the closest thing to it. It is cool, though, that it’s a piece of old, reliable internet infrastructure.
Wellerstein: At least it doesn’t have a GIF of a guy doing construction on it!
That part leads to me feeling old. I’ll get a student saying, “Oh yeah, I used that in middle school!” And I’ll feel ancient. But I like that it’s become a standard reference. That’s the best kind of goal for this. And it’s interesting that it has been a standard reference across political lines. It’s been featured in Breitbart articles and mentioned by Newt Gingrich, but also, Jimmy Kimmel cited it once. To me, it’s hard to do something in the world of science and tech that has broad support, no matter your political orientation. You can tell big stories about using the tool. And the site doesn’t expressly say that nukes are good or bad or whatever. You, the user, experiment as you see fit. I once joked with my class and asked how different would Nukemap feel if there was, like, a big American flag plastered on it. I think it would have a huge difference there in how people perceive it, which is why I’m glad it is the way it is.
Warzel: On Nukemap’s fifth anniversary, you wrote, “My main frustration with Nukemap as a communication tool is that the top-down, concentric-circles approach is the view of the military planner. It’s the view of the nuclear targeteer, or as a friend and collaborator put it earlier this week, it’s the view of real estate.” Do you worry still that the tool could cause people not to engage as much with the “on the ground” reality of a nuclear attack?
Wellerstein: I’m a historian first and foremost. In my role as a nuclear historian I’ve looked at a lot of visualizations of nuclear bombs. It is tricky because the experience of a nuclear bomb in reality is not from 30,000 feet; it’s from the ground down. Now, there have been tremendous interpretations of this in culture—John Hersey’s 1946 book, Hiroshima, still hits very hard. There are graphic novels, fictional depictions of nuclear war, like The Day After. Those all deal with the reality. I would like to split the difference between what Nukemap is now and that. It’s why I added the casualty calculator. I saw people in 2012—around when North Korea was testing small nuclear weapons—that were testing the small nukes on the site and noting on social media that This blast barely fits across Central Park! It’s a baby nuke. And I found myself thinking, There’s a lot of people in an area the size of Central Park! It’s always hard to get people to see a city not as a collection of streets but as a place with humans. I built a 3-D version of Nukemap in 2013, but it isn’t supported by Google Earth anymore. But what I liked was that by tilting the map a bit you can see the volumetric cloud, and that triggers a different understanding in your brain.
Warzel: For those reading, how would you describe that understanding?
Wellerstein: My favorite trick is to use the new World Trade Center. Most people in New York have seen it from afar and many have been near its base. It towers over you. It’s dizzying if you’re underneath it. And to the top of its antenna is 1,776 feet. To get the Hiroshima mushroom cloud—which is, I should note, not even the same as today’s nukes—you’d have to stack 11 World Trade Centers on top of each other. That’s the kind of sublime awe of how big these weapons are. I’d like to show that.
I’m also interested in building a humanitarian-impact button—I got that suggestion at a conference. It would go over the blast zone of the nuke and say how many hospitals are nearby or how many schools, how many places of worship. It’s meant to be a punch in the gut to show the human costs.
Warzel: To come back to the present day for a moment … Since Russia invaded Ukraine last week, we’ve already seen some very unnerving nuclear saber-rattling from Putin. I’ve seen some reports that Nukemap has gone offline at moments in the last week. How big is the traffic surge you’ve experienced?
Wellerstein: It’s been really heavy usage of the servers to the point where I can’t say definitively how many people are trying to access it, because the server can’t handle it. I’ve done some fixes to try to help handle it and will have to do more. At some point we’re in levels of systems administration that I am not comfortable with. I do not understand DNS and Amazon CloudFront. I’ve been getting stuck at step B of a 100-step process. But it’s just me.
At any given second, we have at least 600 simultaneous users globally. The traffic is heavily American and heavily European—Central European and the United Kingdom. Lots of people are constantly on the website. The traffic graph looks less like a spike and more like a tsunami, like a massive cresting wave. The baseline traffic on a normal day without a crisis can range to about up to 20,000 people a day. On a very slow Sunday when the weather’s nice, it sees about 10,000 people. Right now, though, we’ve been at more like 150,000 people every day, and that’s probably just the level the server can take. Now that I just optimized it, maybe it can now handle a bit more.
Warzel: That is a significant increase!
Wellerstein: When I say tsunami, I don’t mean it metaphorically. The graph right now looks like a pic of a big wave. A steep rise that stays up for days and days.
Warzel: Is this unprecedented?
Wellerstein: During the scarier parts of the Trump administration there were times when it would go up and be craggy on the graph. A series of ups and downs. This is much more consistently high. Now every time somebody on site leaves, another person waiting in line replaces them. It’s taken up a lot of my time the last week to figure out how to fix the site to give access to those who want it. I’ve spent a couple of days rebuilding the server from the bottom. I’m sitting around looking up best practices for server maintenance and talking to people on forums to learn more. I’m pretty computer-savvy for a historian, but I’m still having trouble parsing things written for full-time engineers. If I was going to do this the right way, I would hire somebody to do it, but I don’t have the money for that. Nukemap doesn’t generate revenue—it’s being paid for generously by a sponsor.
Warzel: The circumstances are obviously grim, but it must be gratifying to see so many people using it.
Wellerstein: It’s been interesting to me to see the stuff online where people are complaining about not being able to access the site. They expect it to be, like, a fully operational service, but it’s just me, and I’m already doing a million things with my full-time job. But I’m deeply flattered by the idea that anyone out there would actually care that my site is down.
Warzel: You’re a nuclear-weapons historian, and there’s a bit of irony here that at the moment your expertise is in high demand, you also need to, like, do server maintenance!
I’m having trouble finding the words to ask this last question, but do you think there’s anything to learn from your experience over the past week? Is this just—nukes are in the news and people are nervous and Googling? How are you processing this, both as Nukemap’s creator but also as a professional who is closely watching this war?
Wellerstein: There’s this old XKCD cartoon that charts how scared a person should be, correlated with what kind of expert is on TV. The joke is: When you’ve gotten to the person who is an expert in meteors, you should be terrified.
The grim joke I have with colleagues is that the worse the world is, the more demand there is for me. When things are going swell, nobody wants to talk to me about nukes. When the world teeters off a cliff and everything is terrifying, I’m in demand. I accept that paradox. I accept the grim irony of it. It is why I pursued this job in the first place.
So when I look at the last week of Nukemap, I think I’m glad I can contribute something to this situation I’m proud of. I’m glad to play any kind of role. Even if that role is making people nervous, I believe they come to the site because they still want the knowledge. They wouldn’t be less worried if they couldn’t find out how big a nuclear explosion is. I think it’s fine however people use the site—even if people use it in ways I personally find distasteful. Because there’s some sort of education happening at the same time. I believe that you have to meet people wherever they are.
I wish we had a world where a website like this could be a purely academic curiosity, but we don’t. You can use this website to make whatever arguments you want, but the website is based on some actual data. And that feels like something.