The first urbanists were recorded in the pages of Genesis: “Come, let us build ourselves a city and a tower with its top in the heavens and let us make a name for ourselves; otherwise we shall be scattered abroad upon the face of the whole earth.” But God struck down the Tower of Babel and cursed his people to rely on Google Translate forever.

Despite this false start, the dream of building a great new city continues to this day, even in developed nations like the United States, where we already have a lot of them. We start new companies, new schools, new neighborhoods all the time. Why not a new San Francisco, Boston, or Miami? The yearning for a blank slate crosses the ideological spectrum, touching socialists, antidevelopment activists, curious policy makers, and, most recently, Silicon Valley investors attempting to build a city from scratch—among them Marc Andreessen, Patrick and John Collison, Michael Moritz, Nat Friedman, and Laurene Powell Jobs (who is also the founder of Emerson Collective, which is the majority owner of The Atlantic).

And they’re not just dreaming big or tweeting. As The New York Times reported in August, they’re backing California Forever, the parent company of Flannery Associates, which has acquired nearly 60,000 acres in Solano County, California, between San Francisco and Sacramento. That’s a lot of land—roughly twice the size of San Francisco or Boston, and slightly larger than Seattle. Housing developments crop up all the time, of course, and suburbs glom on to existing metropolitan areas. California Forever has something else in mind: a top-down community with brand-new infrastructure, where tens of thousands would live and, most important for the company’s vision, also work and play. It’s not your grandfather’s suburban development.  

[From the July/August 2023 Issue: Colorado’s ingenious idea for solving the housing crisis]

“We’ve gotten into a situation where it’s completely acceptable to talk about inventing general artificial intelligence, and that’s something we’ve accepted is going to happen, but it’s not possible to build a new town where people can buy homes,” Jan Sramek, the founder and CEO of California Forever, told me. (The comparison reveals more about his social environment than anything else; it is not commonly accepted that AGI is “going to happen.”)

But building a new city is hard, and this most recent push to do so—unlike with recent gains in AI—doesn’t reflect an exciting breakthrough in America’s technological, political, or financial capacity. Rather, it reflects an abiding frustration with the ridiculously sluggish process of building housing in America’s most productive cities and suburbs. The dream of a new San Francisco is, then, rooted in the nightmare that the old one may be past saving.

Details about the new proposed city in Solano County are hard to come by, but sketches on California Forever’s website portray an idyllic town, foregrounded by open space and densely built with multiple housing types. Windmills turn in the background. The website reads: “Our vision for walkable neighborhoods, clean energy, sustainable infrastructure, good jobs and a healthy environment is not about reinventing the wheel, but rather going back to the basics that were once the norm across America.”

This project has its advantages: The lack of urban or suburban development in the region means an absence of traditional groups that might fight against neighborhood change. Because California Forever has acquired so much land, local officials have a strong incentive to work with Sramek to prevent collapsing land values if his project fails. And Sramek is already considering ways to sweeten the deal for existing residents; he says one idea is “setting up a fund that would provide down-payment assistance for buying homes in the new community, which would only be accessible to current residents of Solano County.”

But financing urban infrastructure is exceedingly expensive. “Organic” cities, in which firms and workers agglomerate and then begin to demand that governments finance infrastructure, have a preassembled tax base. If you try to build the infrastructure first, paying for it becomes tricky.

Alain Bertaud, a former principal urban planner at the World Bank and an expert on urban development, told me: “A new city, especially a large one … has a problem of cash flow.” The city can’t raise taxes to build schools and hire teachers, for instance, but it needs to build schools and hire teachers before parents are willing to move—and be taxed—there. “If you look back to [recent] history … the only large new cities were new capitals like Brasilia, Chandigarh, Canberra, [where] the cash flow is not a problem [because] you have the taxpayers of the entire country paying for the cost.”

Thinking of cities as mere infrastructure is a categorical mistake. New York City is not the Empire State Building or the Brooklyn Bridge; London is not the tube; and Levittown, New York—America’s quintessential “first” suburb—is not its single-family homes. Infrastructure follows people, not the other way around. “You don’t go to a new city because the sewer system is fantastically efficient,” Bertaud said.

In general, the superstar cities we have today were not preselected from above; they were chosen by millions of workers in search of economic opportunity: Los Angeles (oil); San Francisco (gold); Boston (a port, academia); Seattle (lumber, aircraft, tech); New York City (a port, finance). Granted, workers tend to follow firms that follow transportation networks, which themselves are sometimes functions of state investments, but the principle is sound: Cities are people.

When people are choosing where to live, that decision is almost wholly dominated by job availability. What that means is people attract people. It’s a virtuous cycle in which people who move have kids and want teachers and day-care providers and taxi drivers and nurses, and those people want restaurant workers and iPhone-repair specialists, and so on. (Within a job market or when choosing between two equally promising job markets, people regularly consider the quality of life.)

But what if Sramek and his backers aren’t really building a new city after all, just a commuter suburb far away from the inner core? That’s what the pro-housing activist Jordan Grimes thinks is happening; he told the San Francisco Chronicle the project was “sprawl with a prettier face and prettier name.” Solano’s population has a lot of commuters already. Census data from 2016 to 2020 indicated that of the roughly 207,000 workers who lived in Solano, more than 40 percent commuted to another county. Compare that with San Francisco, where of its nearly 510,000 workers, a bit more than 20 percent commuted to another county.

I asked Sramek: Is he truly looking to build a city with its own job market, where residents will be responsible for policing, fire services, parks and recreation, wastewater, libraries? Or is he looking to develop housing, with some space for retail, restaurants, and other cultural amenities? “This is one of those issues that’s very open for community input,” he told me. “We do think that eventually this would become an incorporated city that does provide many of those services.”

Sramek isn’t a developer, and his investors are not the sort of people who hope that their hundreds of millions of dollars go into the construction of a few thousand single-family homes. Someone close to the project, who spoke on the condition of anonymity to discuss it freely, told me the aspiration is to prove to the rest of the world what’s possible in America: We can build an attractive, dense, and climate-friendly metropolis, and we can do it quickly. The source also suggested that a big Silicon Valley player might one day move its offices to the area. (I reached out to Andreessen, Patrick Collison, Friedman, and Powell Jobs. They declined to comment.)

Either way, new city or new sprawl, this project is going to run headfirst into the politics of development. Right now, the land is zoned largely for agricultural use. The county holds that changing the current designation to accommodate high-density urban infrastructure will require a ballot measure. Sramek told me he might try to put the question to voters as early as November 2024, but victory is far from assured.

According to some local officials, Flannery Associates alienated the local community by refusing to announce its intentions before it began acquiring land. (Sramek argues that doing so would have made land values skyrocket.) Congressional representatives alerted the Treasury Department, worried that foreign investors were buying up real estate for nefarious purposes. They noted that an Air Force base is nearby. “I will tell you they have poisoned the well,” John Garamendi, who represents a large part of Solano County, told me. “There’s no goodwill. Five years of total secrecy? Five years of not communicating with [local officials]?”

The process of building a city, difficult as it is, seems remotely rational only because trying to build within cities drives people mad.  

Sramek and his director of planning, Gabriel Metcalf, who once ran the influential San Francisco Bay Area Planning and Urban Research Association, say the idea for a new city came to them after deciding that working on incremental reforms would never yield the housing needed to make a dent in the overall housing shortage. As of now, the country needs more housing than almost anyone can imagine, a formidable challenge even if America’s political and legal systems were focused on meeting it—which, unfortunately, most of them are not. Instead of directing a building boom, states still devolve permitting decisions down to the hyperlocal level, where the default is to ban smaller, more affordable homes and where opposition from just a few people can quash desperately needed construction.

“It’s always hard to come to an existing place and try to change it very profoundly,” Sramek told me, when I asked him why he wasn’t focused on building in established cities.

“I spent my whole career on the infill side,” Metcalf told me. (Infill development is building on underutilized land within existing development patterns, such as turning a parking lot into a few townhomes.) “I believe in that completely, but we are only delivering a small fraction of what we need … Whether it’s trying to build a high-speed rail line or renewable-energy transmission line or high-density infill housing, there is a vetocracy in place that across America makes it incredibly difficult and slow to build the things we need to build.” (Funnily enough, that vetocracy includes one of the investors in the Solano County project: Andreessen. I reported last year that he co-signed a letter with his wife opposing new development in the wealthy town of Atherton.)

The socialist writer Nathan J. Robinson has also issued a call to build new cities, and he, too, seems to have given up on the idea of reforming existing places: “​​The exciting thing about building new cities from scratch is that it allows you to avoid the mistakes that are made in the ‘organic’ (i.e., market-built) city … A new city can avoid all of the disastrous errors that gave us the ugly suburban wastelands that constitute so much of contemporary ‘development.’”

American cities and suburbs have earned Sramek’s fatalism. And certainly, building a walkable, thriving new town in Solano County would be positive for anyone who found a home they loved there. But Sramek and his backers want to set an example, and good examples should be replicable. This one isn’t. Sites like Solano County—near bustling job centers that lack residential development—are few and far between.

Two types of places need a development boom: those that already have lots of people living in them, like Boston or Miami, and those that are growing quickly, like Georgetown, Texas, near Austin. To the extent they’re failing to build, it’s not because they lack inspiration. They’re failing because the politics are genuinely thorny. Many people oppose new development on ideological grounds, or because they think it’s a nuisance, or because they deny the existence of a housing shortage at all, or even because they believe it interferes with other priorities.

[Jerusalem Demsas: California isn’t special]

A new city, moreover, won’t necessarily escape these antidevelopment pressures in the future. It might expand for a while, but it will eventually face the same old problem: residents who don’t want change. Even in Manhattan, a place where residents are surrounded by high-density housing and cultural amenities that come from density, people regularly oppose new housing, new transit, and even new dumpsters.

Solving the housing crisis doesn’t require inventing new places for people to go; it requires big cities to embrace growth, as they did in the past, and smaller cities to accept change. Again, cities are people, and people are moving to Maricopa, Arizona, in the suburbs of Phoenix, and Santa Cruz, California, south of San Jose. These places may not feel ready to accommodate newcomers, but some will have to rise to the occasion.

What America needs isn’t proof that it can build new cities, but that it can fix its existing ones.

Gareth Edwards’s new blockbuster, The Creator, couldn’t have picked a better moment to arrive on the big screen. A sci-fi epic from a director who plays best in theaters (his previous films include Rogue One and Godzilla), The Creator is set in a world where artificial intelligence inhabits its own stratum of society, as a robotic underclass invented to serve humans. When he started co-writing the movie with his Rogue One collaborator Chris Weitz in 2018, “AI was up there with flying cars and living on the moon,” Edwards recently said. Now it’s a topic of constant global discourse, hailed and feared as an innovation that is reshaping societal norms on a daily basis.

The machine intelligence in The Creator is a far cry from the enigmatic chatbots dominating the news in 2023. In the film, AI has evolved into a species unto itself, a cadre of humanoid robots who initially function as part of American society but are eventually forced to leave the country after a cataclysmic incident sparks a global human-AI war. This is all explained in a brief, cable-newsy preamble, before the viewer is plunged into the middle of a conflict that practically never lets up. Here the topicality grows trickier, because The Creator is not really about AI as we currently understand it. Rather, it’s a broader metaphor for every insurgent foe America has fought since World War II.

This is not new for the sci-fi genre: Star Wars was designed by George Lucas, rather explicitly, as a Vietnam War allegory, with its Rebel Alliance imagined as a sort of Viet Cong resisting its larger imperial oppressors. The Creator functions along the same lines, but it boldly underlines how America has become the bad guy, seeking to wipe AI off the planet even though all of the world’s robots have already fled to Asia. Much as in James Cameron’s Avatar films, the U.S. military is presented as fearsome, ruthless, and largely immoral; it creates a terrifying orbital platform named NOMAD that circles the skies, dropping gigantic bombs on targets near and far.

All of this is fascinating—but it’s also largely background noise. The Creator is not too interested in world-building beyond these basics. Its protagonist, Joshua (played by John David Washington), is an undercover operative embedded with a group of robot guerrilla fighters, and the film follows him as he falls in love with a rebel fighter named Maya (Gemma Chan); he eventually deploys to Asia in search of a mysterious superweapon that takes the form of an adorable robot child named Alphie (Madeleine Yuna Voyles). The pace of the movie is relentless, chasing Joshua behind enemy lines; barely five minutes goes by without some massive explosion or exchange of gunfire.

[Read: The coming humanist renaissance]

This action takes place against many stunning backdrops: The Creator looks tremendous, partly because Edwards shot the film in gorgeous landscapes around the world (mostly in Thailand), taking every advantage of this natural beauty instead of depending on CGI sets. It’s an approach to blockbuster action that’s far more engaging than the empty, grayscale battle zones of several recent superhero films, and it looks just as good or better than Edwards’s previous two blockbusters, despite being made on a far smaller budget.

It’s just unfortunate that the story is overflowing with familiar tropes. Alphie, the well-meaning superweapon who functions as the film’s MacGuffin, descends from a long line of cute-kid characters designed to curry audience sympathy, but she lacks any real personality. Joshua’s love interest, Maya, is off-screen for almost the whole film, confined to a few ethereal flashbacks that leave her feeling like an underdeveloped plot device. Though Washington does his best to infuse Joshua with some real grit and spirit, the spectacle of The Creator is the film’s real star, which isn’t quite enough to sustain a running time of more than two hours.

Even so, The Creator is a high-level craft achievement that is undeniably cool on a big screen. I was intrigued by many of the ideas bubbling away within the movie’s larger world; I wanted to know more about the AI warriors and civilians Joshua encounters, and the society around them. American imperialism has been portrayed as villainous in other movies, but there’s some transgressive thrill to the film’s bluntly negative depiction. Still, the allegory stays thin throughout, light on details but easy for any audience member to recognize. Here, robots are just another Rebel Alliance to rally around. They’re easy to root for, yes, but they’re mere clones of underdogs past.

Welcome to Up for Debate. Each week, Conor Friedersdorf rounds up timely conversations and solicits reader responses to one thought-provoking question. Later, he publishes some thoughtful replies. Sign up for the newsletter here.

Last week, I asked for your thoughts on self-driving cars.

Replies have been edited for length and clarity.

Kathryn is bullish and looks forward to shedding the responsibility of driving:

Yes, driverless cars are the future, at least for people alive today. I’m sure there will be some later innovation in transportation we can’t even imagine yet. Cities should allow them to be tested on the street now, assuming the vehicle has passed something analogous to the driving test humans take to receive a driver’s license. I’d love to have these vehicles in my neighborhood. I live in an urban area and use our sidewalks and crosswalks as my primary mode of transportation. Given all the close calls I’ve had over the years with human drivers, I’d welcome anything that is safer. Driverless cars don’t need to be even close to perfect to be an improvement on the status quo.

I don’t like driving, although I have a license and access to a car and I drive a few times a month. What I like least about driving is that I could make one wrong move and possibly kill someone, including myself, or destroy my family’s financial stability. I appreciate that I can commute by bus and avoid facing that responsibility on a daily basis. I also enjoy relaxing on the bus. I can read and write for work or fun, listen to music and podcasts, and watch the world go by. I wonder if today’s drivers will start to value the decrease in legal liability and the increase in downtime that driverless cars can provide.

Mike is bearish:

Driverless vehicles of any type should be banned in any setting where they will have to interact with vehicles controlled by humans. They shouldn’t be allowed to be tested, and the technology should not be pursued. Where every vehicle is autonomous, there is in theory nothing wrong with them, but on the public roads, that environment will never exist, at least not in our lifetimes. I, for one, will never get in a vehicle that does not have a human driver.

Chris points out that innovation is rarely, if ever, stopped by its critics or regulators acting on their behalf:

Driverless cars will flood the streets regardless of whether they are ready for prime time. There will be accidents, injuries, and deaths. Victims and their lawyers will haggle with and sue multiple parties––and compensation will be slow in coming while everyone is still figuring out who is actually responsible for an accident (the vehicle maker, the software company, third-party technology, the passenger somehow, etc.).

Due to the intricacies of multiple-party liability, insurance companies covering driverless vehicles may be very slow to compensate victims, leaving the latter in medical and financial purgatory. The business of America is business, and driverless cars and trucks represent a business opportunity. The consumer and his or her safety is always an afterthought.

Leigh anticipates safer roads:

I live in New Orleans, where just in the past five days of driving my kids to school, I have experienced drivers blowing through stoplights; drivers who were clearly watching their phones and not the road; drivers doing 50 in a 20 mph school zone, and not worried about getting caught because they have no tag on their car. Driverless cars will end all this. They will be programmed to obey traffic laws. This will make life easier and more predictable for pedestrians, cyclists, and other drivers. Plus, Baby Boomers are getting older, and it would be great to have a driverless car take them to their doctor appointments instead of them continuing to drive past the point of safety. And how great would it be to sleep in a driverless car on your way to your travel destination? You could wake up refreshed, having traveled at night when the roads aren’t busy.

Cameron is skeptical of anything tech companies touch:

I think the discussion around autonomous vehicles—and their viability as a mode of transportation in the not-so-distant future—encapsulates bigger questions that The Atlantic has covered surrounding different elements of American social and political life. The oft-discussed degradation of civil society in the U.S., combined with a legacy of automobile-centric infrastructure development and noxious residential zoning regulations, has resulted in starved public transportation networks (where they even exist) and urban and suburban layouts that aren’t terribly traversable on foot to begin with.

Now, I realize that tackling these issues will require a significant amount of effort, investment, and political willpower, but I find myself increasingly disheartened whenever Silicon Valley—who are not accountable to the public outside of “market forces”—get the opportunity to treat the U.S. as a playground while our civic institutions convulse.

Maureen is betting on old-fashioned car culture:

Our century-long love affair with all things automotive dooms the driverless concept to a niche market: people whose physical condition precludes driving and those who prefer to be driven.

The vast majority of us regard driving as a birthright, obtaining a license as a rite of passage, and operating a vehicle as an expression of control––and as much fun as you can have while fully dressed. In short, we love control and speed. The less mature among us love road games like Cut You Off and Tailgate. An astonishing number love to work on cars, restore cars, race cars, and watch others race. Driverless cars will have their place, when they iron out the kinks, but it is not on the roads of this vast and beautiful nation.

Alan makes the case for human drivers:

A couple years ago, I was heading west on I-66 in the late afternoon. The road was under construction, recently repaved and with no lines yet painted. It had just finished raining, so the new asphalt was wet, but the sun was coming out, very low on the horizon and reflecting off the wet asphalt. I was effectively blinded, as were all the other drivers. Visor pulled low, I could focus on the car in front and make judgments on where to be. Stay in my unmarked “lane,” keep a reasonable distance, and just drive. But what would a computer do? Would it have the “intuition” to adopt a defensive driving mode and guesstimate where the lanes should be? Or would it just freak out and stop? And how would a computer drive a car on the snow-covered roadways of Buffalo or Bismarck in February? No lane markings to follow, just human understanding of how we navigate difficult conditions.

Richard offers additional examples of nonstandard road conditions:

Construction reduces a two-lane road to one lane. Some flag man in a high-visibility vest is holding a pole with a small sign that says “SLOW,” which means you can enter the opposite lane. Then he flips it to the other side, which says “STOP.” Will self-driving cars figure this out?

Stuck behind a postal delivery vehicle stopping at every house on a two-lane road with double yellow lines. Does the self-driving vehicle know it can pass the mailman? Ditto for the garbage truck, UPS, FedEx.

A car accident requires a police officer to take control of an intersection. Humans recognize the presence of a police officer and know to obey his hand signals and ignore the traffic light. Will a self-driving vehicle recognize the man as a police officer and understand his hand signals for “go” and “stop”?

A power outage causes nonworking traffic lights at intersections. Humans know to treat this intersection as a four-way stop. How does a self-driving vehicle interpret this situation? Does it even recognize that there is a nonworking traffic light?

A school bus is stopped on the other side of the road. In Ohio, if it is three lanes or fewer, traffic must stop on both sides. If it is four lanes or more, traffic on the opposite side can keep moving. Does a self-driving vehicle know this? Does a self-driving vehicle recognize a school bus?

I could go on in that vein, but you get the drift.

Steve contends with one of the Northeast’s most hazardous road conditions: Massachusetts drivers. He writes:

Living near Boston, I can tell you that, within 10 minutes of every drive, I run into a scenario that would be nearly impossible for a driverless car to navigate. Beyond the nearly unnavigable cow paths we call roads, there are so many times that eye contact with the driver, pedestrian, or pet is the only real way to avoid calamity. Not to mention the average Boston driver seems to find new ways every day to do something irrational. It will take decades to master that and even longer for Bostonians to trust any software to solve that. Instead, car companies should focus on two things: First, make driver-assist technologies amazing. Imagine a windshield that enhances everything (especially at night) and highlights potential risks, 360 cameras that help see issues and then accident-avoidance technologies that give 80-year-olds the reflexes of a teenager. Said simply, keep the person at the wheel, but make them awesome drivers.

Second, situational autopilot: Designated areas where driverless cars move on preprogrammed routes (think shuttles in airport lots or parts of Rome filled with driverless vehicles). Also, why not special lanes on highways (repurpose HOV lanes) that allow cars to link up, form a dance line of sorts, and speed down the highway?  Instead of focusing on an unreasonable goal not reachable for 20 years or more, why not take the remarkable technologies we’ve developed and get us to a much safer place than we are now?

Leo suggests that we shouldn’t count on driverless cars winning the day politically, even if they perform better than humans:

This debate may play out differently in other countries and cultures, but in America, freedom will trump safety in the end. There are, of course, any number of laws and regulations in our society, but the underlying ethos, the dominant paradigm, is that we live in the land of the free. Laws, regulations, and limitations are not prized as arbiters of a functional society so much as endured as necessary evils. And any person, community, or movement that pushes too hard for too many restrictions will pay a heavy price.

What politician or political party is going to sign their own death warrant by limiting or, god forbid, outlawing our right to drive our own cars? Even the limitations that already exist (such as speed limits) are flouted so regularly that in many cases it’s unclear why they exist at all (other than to raise money for local governments through ticket citations). The decades-long effort to stigmatize and heavily fine drunk drivers has indeed yielded some results, but there are still drunk drivers, and there always will be. For better and for worse, Americans will only tolerate so many infringements on their individual liberty.

So driverless cars will likely be deployed to some extent. They will penetrate our society at some level. But they are not the future. At best, they are one aspect of the future.

Karen won’t be buying a self-driving car:

I enjoy driving a manual transmission. I have found that it forces me to pay attention only to driving. I feel engaged with the car. Power steering, power brakes, fine; but I still don’t mind winding my car windows up and down. Power seats? Entirely unnecessary. I don’t even like the whole touchscreen thing. Some cars force you to use the touchscreen to open the glovebox! Why? What’s the deal with putting the heating and AC controls on the touchscreen? Or the radio, for that matter. And I will decide what music I want to hear. No music apps! You now have to buy a used car to get a CD player. Yes, I’m old. But I like to drive—not be driven, even if the systems get better and more accurate.

Tanner writes that “autonomous cars are still cars,” which he sees as a bad thing:

We would be better off investing in low-tech, less car-centric ideas; designing our communities where car trips are less necessary, supporting robust public transportation, making streets safer for all users, dedicating less space to cars and more space to more community-oriented things. Some will argue that driverless cars will solve the issues above by reducing the need to own a car, reducing crashes (with cars at least), reducing congestion, etc.

Even if true (I have my doubts), do we really want to be even more dependent on giant tech companies than we are? How do I, as a pedestrian or a bicyclist, communicate with a machine about my intentions at an intersection (no more gadgets, please)? Is the future one where everyone and everything requires sensors and gadgets to work safely?

There is a place for driverless cars. But the future, for me, is a 30-year-old bicycle that I can take anywhere.

Eugene Stoner was an unassuming family man in postwar America. He wore glasses and had a fondness for bow ties. His figure was slightly round; his colleagues called him a teddy bear. He refused to swear or spank his children. “Boy, that frosts me,” he’d say when he was upset. He liked to tweak self-important people with a dry sense of humor. He hated attention.

A lifelong tinkerer and a Marine veteran, he was also fascinated by the question of how to make guns shoot better. When an idea came to him, he scribbled it down on anything he could find—a pad of paper, a napkin, the tablecloth at a restaurant. He had no formal training in engineering or in firearms design. Yet it was inside Stoner’s detached garage in Los Angeles, during the 1950s, that the amateur gunsmith, surrounded by piles of sketches and prototypes, came up with the idea for a rifle that would change American history.

Today, this weapon is the most popular rifle in America—and the most hated. The AR-15 is a symbol of Second Amendment rights to millions of Americans and an emblem of a violent gun culture run amok to millions more. With a lightweight frame and an internal gas system, the military version can be fired as an automatic, unleashing a stream of bullets from a single pull of the trigger, or as a semiautomatic, allowing for one shot per trigger pull. The civilian semiautomatic version is now the best-selling rifle in the country; more than 20 million such guns are in civilian hands. And it is a weapon of choice for mass shooters—including the white supremacist who killed three Black people last month at a store in Jacksonville, Florida, armed with a handgun and an AR-15-style rifle emblazoned with a swastika.

[Juliette Kayyem: The Jacksonville killer wanted everyone to know his message of hate]

The consequences of the AR-15’s creation have coursed through our society and politics for generations in ways that Stoner never foresaw. He created the gun with a simple goal: to build a better rifle for the U.S. military and its allies during the Cold War. He wanted to protect the country he loved. Now his invention is fused in Americans’ minds with the horror of people going about their daily tasks—at school, the movies, the store, a concert—and suddenly finding themselves running for their lives. Few of the participants in America’s perpetual gun debate know the true, complicated history of this consequential creation—or of the man behind it. The saga of the AR-15 is a story of how quickly an invention can leave the control of the inventor, how it can be used in ways the creator never imagined.

We interviewed Stoner’s family members and close colleagues about his views of his gun. They gave us insight into what the inventor might have thought about the way the AR-15 is being used today, though we’ll never know for sure; Stoner died before mass shootings with AR-15s were common. Later in life, after years of working in the gun industry, he was asked about his career in an interview for the Smithsonian Institution. “It was kind of a hobby that got out of hand,” he said.

As a boy growing up in the Coachella Valley, in Southern California, in the 1920s and ’30s, Stoner was fascinated by explosions. Before the age of 10, he had designed rockets and rudimentary weapons. On one occasion, he begged a friend’s father for a metal pipe and the local drugstore owner for magnesium. Stoner built a primitive cannon and pointed it at a house across the street, but before he could open fire, his father ran to stop him. “I told you to do this at the city dump,” scolded Lloyd Stoner, a veteran of the Great War who had moved the family to California from the farmlands of Indiana in search of a better life.

Eugene Stoner never went to college. He joined the Marines during World War II and was tasked with repairing weapons on aircraft in the Philippines. When he came home, he brought his wife, Jean, an adventurous woman who idolized Amelia Earhart, a special present: gun parts from Asia that he assembled into a rifle. She loved it. The couple often went hunting and shooting together. “He was a very quiet person,” Jean said in an unpublished interview that the Stoner family shared with us. “But if you talked about guns, cars, or planes, he’d talk all night.”

After the war, Stoner got a job as a machinist making aircraft parts. Every day after he came home, he would eat the dinner that Jean had prepared (beef Stroganoff was his favorite), take a quick nap, and then walk to the garage to work on his gun designs. Like other hobbyist inventors of the era, he believed he could move the country forward by the power of his ingenuity. “We were like the 1950s family. It was California. It was booming after the war,” his daughter Susan told us. “I knew from my dad—I felt from him—the future was wide open.”

[Conor Friedersdorf: The California dream is dying]

Stoner had the ability, common among inventors, to imagine engineering solutions that others stuck in the dogmas of the field could not. For centuries, gunmakers had built their rifles out of wood and steel, which made them very heavy. At the time, the U.S. military was searching for a lighter rifle, and Stoner wondered if he could build one using modern materials. If humans were soaring into the atmosphere in airplanes made of aluminum, he figured, couldn’t the lightweight metal tolerate the pressures of a gun firing? By the early 1950s, he had figured out how to replace one of the heaviest steel components of a rifle with aluminum. Then he devised a way of using the force of the gas from the exploding gunpowder to move parts inside the gun so that they ejected spent casings and loaded new rounds. This allowed him to eliminate other, cumbersome metal parts that had been used in the past. The first time he tried firing a gun using this new system, it blew hot gas into his face. But he perfected the design and eventually received a patent for it.

In 1954, Stoner got the opportunity to bring his radical gun concepts to life. That year, as Stoner later recalled, he had a chance encounter at a local gun range with George Sullivan. A relentless pitchman, Sullivan was then the head of a Hollywood start-up called ArmaLite, a subsidiary of Fairchild Engine and Aircraft Corporation whose mission was to design futuristic weapons. Impressed with the homemade guns Stoner was shooting, Sullivan hired him as ArmaLite’s chief engineer.

The small yet brilliant ArmaLite team worked at a fevered pace, designing a series of lightweight guns made of aluminum and plastic. Most went nowhere. Nevertheless, the ambitious Sullivan set the firm’s sights on an improbable target: the U.S Army’s standard-issue rifle. The Eisenhower administration’s “New Look”—an effort to rein in Pentagon spending and shift it toward newer technologies—opened the door for private companies to get big military contracts. The outsiders from Hollywood decided to take on Springfield Armory, the military’s citadel of gun making in western Massachusetts that had equipped American soldiers since the Revolutionary War. Springfield’s own efforts to develop a new rifle had resulted in a heavy wood-and-steel model that wasn’t much more advanced than the M1 Garand used by GIs in World War II.

Eugene Stoner, wearing his trademark bow tie, holds his creation the AR-10. The AR-15 was a scaled-down version of this gun. (Photograph courtesy of Susan Kleinpell via Farrar, Straus and Giroux)

ArmaLite’s first serious attempt at a rapid-fire rifle made of plastic and aluminum was the AR-10—AR for ArmaLite or ArmaLite Research (accounts differ), and 10 because the weapon was the company’s tenth creation. The rifle combined the efficient internal gas system Stoner had devised in his garage and lightweight modern materials with a design that made the gun easy to shoot and keep on target. In December 1956, Time heralded the AR-10 as a potential savior for the bumbling U.S. military and listed Sullivan as the gun’s inventor, a claim that infuriated Stoner’s wife. Sullivan had also meddled with the design, insisting that more aluminum be used in making the gun’s barrel, a move Stoner resisted. During military trials, the AR-10 fared poorly. At one point, a bullet erupted from the side of the gun’s barrel, just missing the hand of the soldier firing the weapon—and seemingly dooming ArmaLite’s chances of landing a military contract.

But within the Pentagon, a cabal of high-ranking officers led by General Willard Wyman launched a back-channel effort to save Stoner’s gun. Wyman was a legendary military leader who, at age 46, had joined the D-Day invasion at Omaha Beach as an assistant commander of the First Infantry Division. He knew that the United States needed better firepower as the Cold War flashed hot. America’s enemies around the globe were being armed by the Soviet Union with millions of rugged AK-47s that could spray bullets in automatic mode and were highly effective in guerilla warfare. Wyman was certain that modern wars would be won not by long-range marksmen but by soldiers firing lots of bullets in close combat. They needed a rifle that used small-caliber bullets so they could carry more ammo. And he was worried that the tradition-bound gun designers at Springfield Armory weren’t innovative enough to meet the challenge. When Wyman’s superiors brushed him off, he secretly flew to Los Angeles and stunned Stoner and his team by striding into the ArmaLite office unannounced. Wyman told Stoner that he wanted ArmaLite to build a new version of the AR-10 that fired a smaller bullet.

[James Fallows: Why the AR-15 is so lethal]

Stoner and an ArmaLite draftsman named Jim Sullivan (no relation to George) set about designing the gun. It was simple, efficient, and easy to use. Early versions of the AR-15 weighed just more than five pounds unloaded, less than the hedge trimmers and handheld vacuums of the era. With all of Stoner’s innovations—lighter material, fewer parts, and the gas system, as well as an in-line stock and a pistol grip—Jim Sullivan found shooting the prototype AR-15 to be easy, even after he flipped the selector switch to automatic. “That made it so well handling,” he told us. “If you’re firing full auto, you don’t want a gun that lifts.” Sullivan found the rifle’s recoil to be minimal. As a result, follow-up shots were quick when he switched it to semiautomatic. “It looked a little far-out for that time in history,” Stoner later said in the Smithsonian interview.

As Stoner and his backers sought to persuade the military to adopt the AR-15 in place of Springfield’s rifle, they were often met with skepticism about the gun’s small bullets. During secret military hearings about the rifle in the winter of 1958, Stoner explained to a panel of generals that the AR-15 had “a better killing cartridge with a higher velocity” than the Soviet AK-47. The generals asked Stoner how a smaller bullet fired from his rifle could do so much damage. “The wound capability is extremely high,” Stoner answered. “It blows up on contact rather than drilling a nice neat hole.” A slower .30 caliber round, similar to the one used by Springfield’s wood-and-steel rifles, “will go right through flesh,” but the faster, smaller bullet from the AR-15 “will tumble and tear,” he said.

Those in the military who wanted Springfield’s rifle to prevail tried to sabotage Stoner’s gun, rigging tests and shading reports so that it would seem like it wasn’t ready for the battlefield. During official trials in Alaska, Stoner arrived to find that the aiming sights on his guns had been replaced with bits of metal that were badly misaligned, causing soldiers to miss their targets. The guileless inventor was caught up in the murky world of Pentagon intrigue.

[From June 1981: James Fallows’s ‘M-16: A Bureaucratic Horror Story’]

Eventually, through persistence and luck, and with the help of a cast of lobbyists, spies, and analytics-driven military leaders, Stoner’s rifle would be adopted. At a key moment when it seemed that the AR-15 would be killed off by military bureaucrats, the powerful, cigar-chomping Air Force General Curtis LeMay, the architect of the U.S. bombing campaign in Japan during World War II, was asked if he wanted to shoot the gun. On July 4, 1960, at a birthday party for Richard Boutelle, the onetime head of Fairchild, the gun’s backers set up ripe watermelons as targets at Boutelle’s estate in western Maryland. LeMay fired, causing a red-and-green explosion. The general marched into the Pentagon soon after and demanded that the military purchase the weapon. It would become the standard-issue rifle—renamed the M16, for the prosaic “Model 16”—just in time for the rise of U.S. involvement in Vietnam.   

A U.S. Marine holds his M16 rifle alert after being fired on by North Vietnamese soldiers in the jungle southwest of Da Nang on April 22, 1969. (Yvon Cornu / AP)

In Eugene Stoner’s and Jim Sullivan’s minds, their work was not just intellectually engaging but also noble, a way to help America defeat the Communists. At school, in the 1950s, the Stoner children learned what to do in the event of a Soviet nuclear attack. Sirens and bells went off regularly, and teachers ordered kids to hide under their desks and cover their heads, Stoner’s daughter Susan recalled. For her father, the task of making the best rifle for the U.S. military wasn’t burdened with moral quandaries. Many weapons inventors at the time thought about the technical challenges of their weapons first, and wrestled with the consequences of their creations only afterward. “When you see something that is technically sweet, you go ahead and do it and you argue about what to do about it only after you have had your technical success,” J. Robert Oppenheimer, the lead developer of the atomic bomb, said almost a decade after bombs were dropped on Hiroshima and Nagasaki.

[From February 1949: J. Robert Oppenheimer’s ‘The Open Mind’]

After Stoner created the AR-15, he continued designing guns and artillery for a variety of gunmakers. Through a company he co-founded, he worked on antiaircraft weapons for the Shah of Iran, before the 1979 revolution scuttled the deal. He helped design a handgun for the venerable gunmaker Colt that the company tried to sell on the civilian market, without much success. But none of his creations came close to the prominence of the AR-15. By the 1990s, he’d become a superstar in the gun world. Royalties from the M16 made him wealthy; Colt, which purchased the rights to the gun from ArmaLite, sold millions of the weapons to the military. Stoner was “a Second Amendment guy,” his daughter said, but he didn’t talk much about the messy world of politics, either privately or publicly. He preferred thinking about mechanisms.

Throughout his life, Stoner was troubled by losing control over the production of his most famous gun. In the 1960s, as the U.S. ramped up production of the rifle for the war in Vietnam, a Pentagon committee made changes to the gun and its ammunition without proper testing. The results on the battlefields in Vietnam were disastrous. Stories of GIs dying with jammed M16s in their hands horrified the public and led to congressional hearings. The shy inventor was called to testify and found himself thrust into an uncomfortable spotlight. Declassified military documents that we reviewed show that Stoner tried in vain to warn Pentagon officials against the changes.

Stoner paid far less attention to the semiautomatic version of his rifle that Colt began marketing to the public in the 1960s as “a superb hunting partner.” Even after Stoner’s patent expired, in 1977, the rifle was a niche product made by a handful of companies and was despised by many traditional hunters, who tended to prefer polished wood stocks and prided themselves on felling game with a single shot. But the rifle’s status shifted after 9/11. Many Americans wanted to own the gun that soldiers were carrying in the War on Terror. When the 1994 federal assault-weapons ban expired after a decade, the AR-15 became palatable for mainstream American gunmakers to sell. Soon, it was a symbol of Second Amendment rights and survivalist chic, and gun owners rushed to buy AR-15s, fearful that the government would ban them again. By the late 2000s, the gun was enjoying astounding commercial success.

AR-15 style weapons are displayed for sale at the 2022 Rod of Iron Freedom Festival, an open-carry event to celebrate the Second Amendment, in Greeley, Pennsylvania. (Jabin Botsford / The Washington Post / Getty)

When Stoner died from cancer, in 1997, obituaries hailed him as the inventor of the long-serving military rifle; they made no mention of the civilian version of the weapon. Stoner left clues about his thoughts about the gun in a long letter, sent to a Marine general, in which he outlined his wishes for his funeral and burial at Quantico National Cemetery, in Virginia. He saw the creation of a rifle for the U.S military as his greatest triumph. He didn’t mention the civilian version. The government had wanted a “small caliber/high velocity, lightweight, select fire rifle which engaged targets with salvos of rounds from one trigger pull,” Stoner wrote. “That is what I achieved for our servicemen.”

[Ryan Busse: The rifle that ruined America]

The inventor wouldn’t get to control how his proudest achievement would be used after his death, or the fraught, outsize role it would come to play in American society and politics. Since 2012, some of the deadliest mass shootings in the nation’s history—Sandy Hook, Las Vegas, Sutherland Springs, Uvalde—have been carried out by men armed with AR-15s. Now children practice drills to avoid being gunned down by attackers with AR-15s at their school.

The last surviving member of that ArmaLite team, the draftsman Jim Sullivan, was at times haunted by the invention’s later impact. When we visited him at his workshop in Arizona in 2019, Sullivan pulled out the original drawings for the AR-15 and smiled broadly as he described how he and Stoner had designed the gun. He picked up parts to demonstrate how it worked, explaining its functions like an excited professor. He was proud of the weapon and loved Stoner. He said that his years working at ArmaLite were the best of his life. After hours of talking about barrels, bolts, receivers, and Stoner’s gas system, he paused and looked down at the floor. He said he’d grown deeply disturbed by the violence being wrought with the invention he had helped create. He said that mass shooters wouldn’t be able to do what they do without weapons such as the AR-15.

“Every gun designer has a responsibility to …” he said, pausing before finishing his thought, “to think about what the hell they’re creating.”

This article has been adapted from Zusha Elinson and Cameron McWhirter’s book, American Gun: The True Story of the AR-15.