Silicon Valley has long had its own dialect. Venture capitalists are always talking about the value of being “high agency” and making “orthogonal bets.” But lately, tech vernacular has taken a peculiar turn—people have started describing themselves as if they were chatbots. If you misspeak, maybe you’re “hallucinating.” Don’t know the answer to a question? That’s because it’s not in your “training data.” Feeling forgetful? Perhaps you have a case of “context rot,” a phrase that refers to the degradation of a bot’s responses over a long conversation. “I’ve been telling my wife I have context rot for months,” Conor Bronsdon, the host of an AI-focused podcast, told me.  

Where I live, in San Francisco, such comparisons are inescapable. “I’ve described myself as high temperature,” a friend recently told me; in AI-speak, this means he is prone to randomness. AI experts now talk about updating their “weights” when they learn something new and training on “synthetic data” when going over internal thoughts. “Humans have a huge base model trained over billions of years of evolution,” one software engineer wrote on a popular tech forum. “It’s impressive how quickly we learn, but it’s arguably comparable to fine tuning.”

[Read: Machine Unlearning]

Such speech easily comes off as unsettling, if not aggressively bleak: Why describe beautiful, tender life in detached, algorithmic terms? At the same time, language is a fossil record of previous technological revolutions, and if this current transformation is anything like previous ones, some of this new slang may well stick around. AI experts have repeatedly cautioned that anthropomorphism, or the tendency to attribute human qualities to nonhuman objects such as chatbots, can be misleading. But now the reverse is occurring in everyday speech—a sort of modelmorphism, wherein people describe themselves as if they were large language models.

The comparisons really took off a few years ago, after a group of AI researchers argued that language models are “stochastic parrots” that link together language based on statistical patterns without possessing any understanding of meaning. Dissenters retorted that humans are also “stochastic parrots.” As one software engineer riffed, “Humans are basically a sophisticated Markov chain. They are very good at pattern matching, but have no understanding of anything.”

Language has historically evolved alongside technology. Expressions such as running out of steam and cog in the machine, for instance, were adopted into common speech after the Industrial Revolution as shorthand for the human condition. AI metaphors are simply figures of speech, Anna Ivanova, a cognitive scientist at Georgia Tech, told me: When someone says, “My gears are turning,” they don’t literally mean that there are gears or gearlike things in their brain that are actually moving; in much the same way, when people say that they have “context rot,” they don’t mean that they literally have a transformer-based AI model running in their brain.

The difference between earlier technologies and AI systems is that artificial neural networks are loosely inspired by the brain itself. Philosophers such as Raphaël Millière and Cameron Buckner have argued that LLMs could serve as potential models of certain aspects of human cognition. In the 2010s, scientists found parallels between how image-recognition neural networks and the human brain process visual information. There may also be “parallels” between how LLMs and humans represent language, Ivanova said. The MIT neuroscientist Ev Fedorenko and others have identified a type of “language network” inside the human brain, which Fedorenko believes is “very similar in many ways to early LLMs.” These comparisons are contested, but they are receiving research attention: “The recent success of LLMs is rooted in the same computational principles that govern biological brains,” scientists wrote in a position paper last week.  

“Because we do not understand the brain very well we are constantly tempted to use the latest technology as a model for trying to understand it,” the philosopher John Searle wrote in the 1980s. “I am told that some of the ancient Greeks thought the brain functions like a catapult.” In the 19th and 20th centuries, people turned to telegraphs and telephone switchboards as models for the mind: “The brain is no more than a kind of central telephonic exchange,” the philosopher Henri Bergson once wrote. By the ’60s, some scholars were arguing that the mind functions like a computer. “Cognitive scientists often say that the mind is the software of the brain,” the philosopher Ned Block explained in 1990. Today, LLMs make for an alluring comparison.

For now, these analogies are being made mostly by those working in the AI industry. But over time, some of the metaphors might start to seep into everyday language. If you look at the growth of railroads, “a whole range of language change began to happen” as people started to use the new transportation system, Naomi Baron, a linguist and professor emerita at American University, told me. Some of the AI slang isn’t introducing new phrases into English so much as transforming the meaning of existing speech. Take the word hallucinating, which is used to describe AI models confabulating. People are “taking a term created for what the human mind is capable of doing, projecting it onto AI, and then projecting it back onto humans with an AI flavoring,” Baron explained.

None of this is to say every AI-inflected term should get a free pass. I’ve started chastising my friends when I catch them unironically referring to themselves as chatbots. Using LLM metaphors to describe our minds could lead to a reductive and dehumanizing understanding of human cognition. Earlier this year, when OpenAI CEO Sam Altman was pressed on AI’s natural-resource demands, he said that “it also takes a lot of energy to train a human.”

Still, the people in tech I spoke with told me that using LLM-speak to describe their own mind helps them reflect on their cognitive habits. Language subsumes new idioms over time and masks their origins. People today say “My plans have been derailed” without much thought to the phrase’s railroad origins. Perhaps my great-grandchildren will grow up saying things such as “I have context rot” and “That’s not in my training data” without pausing to remember that those phrases first applied to machines.

President Trump gets what he wants, when he wants it, from the National Park Service. One and a half million dollars to replace brick pavers in Lafayette Park, across from the White House? Contracting officials at the NPS Washington headquarters found the money within five days last month. A $170,000 request to remove a fence around the Hains Point golf course, preparing the ground for a new “championship-quality” layout? Six days this summer. And the more–than–$5 million project to gild equine statues along the nearby Memorial Bridge? That took 12 days.

The rapid-approval system is specific to projects Trump wants, thanks to a reorganization at the Interior Department led by Secretary Doug Burgum. Trump issued an executive order to make the District of Columbia “safe and beautiful,” justifying the creation of an NPS docket of “White House priorities” that is handled separately from other park projects and is prioritized both for funding and contracting resources, according to documents I obtained and interviews with five current or former NPS officials. Meanwhile, many other projects throughout the nationwide system of parks are languishing.

The list I received includes Trump’s well-known projects on the National Mall and the refurbishment of the White House, along with several lesser-known efforts: The hiring and equipping of 450 new U.S. Park Police officers in the Washington region, and more than $12.8 million—some of which came from the Department of Justice—for police equipment, including new Flock license-plate-reading cameras, radios, tasers, laptops, body cameras, and body armor. Internal documentation for park projects typically includes “Consequences of Failure to Act,” which scores the urgency based on health and safety measures. When work began in January on fountains near the White House, officials described the need for the work in this way: “Such failure will disappoint visitors, as well as residents and employees of the White House.”

The new project-funding system was described by a contracting officer during an April meeting with NPS officials. “There’s three different lists: White House priorities, high priorities, and low priorities,” the official told colleagues on the virtual call, according to a written account that was provided to me. Those on the call, who do not work on White House priorities, were advised that projects on the low-priority list likely would not be completed, so they should attempt to get their needs moved to the high-priority list.  White House priorities, meanwhile, were nonnegotiable. As Washington is beautified at great expense, officials told me, vital needs at dozens of national parks are not being filled.

I obtained copies of the high- and low-priority lists that circulated this summer inside the NPS. As of mid-July, that low-priority list included 1,573 projects, including many that officials told me would have been funded under prior administrations. Projects that have not yet been funded from that list, according to a separate national-park database, include requests for new wildlife-resistant dumpsters in Bryce Canyon National Park, snow machines to provide fire remediation in Alaska’s remote Yukon–Charley Rivers National Preserve, and an archaeological survey of recently acquired land at Camp Nelson National Monument, a major Union supply base during the Civil War that trained Black soldiers.

The Interior Department dismissed the concerns over this approach when I reached out for comment and told me that the low-priority list I had obtained included items that had been “mis-prioritized and were corrected.” Interior leaders have also boasted about new funding streams, including revenue from more expensive national-park passes for foreign visitors. “In any year, the National Park Service must prioritize projects most in need,” the press office said in an unsigned statement. “The National Park Service has not only been focused on beautifying the district for the 250th celebrations in our nation’s capital but has also been working on many deferred maintenance projects throughout the country.”

[Read: ‘Rush project at request of POTUS’]

But those working in the parks say that it is not just money holding these projects back. It’s also a lack of staff, after rounds of workforce reductions, who are able to handle contracts, a problem that the contracting official on the April call referred to as “capacity.” When Elon Musk’s Department of Government Efficiency targeted the department last year, many contracting managers at the NPS took buyouts, and many of those who remained found themselves reassigned within the larger department. The NPS had 179 full-time administrative officers in June, a 21 percent decrease from the month Trump took office, according to Office of Management and Budget records. The result is a growing backlog of unfinished projects in far-flung parks, as government employees try to jury-rig temporary fixes, the officials said. “There is very little acquisition support,” one of the NPS employees told me. “This is going to get exponentially worse.”

Spencer Morgan, the site superintendent of the Theodore Roosevelt Inaugural National Historic Site, in Buffalo, New York, told me that his facility is still waiting on this fiscal year’s promised operating-support funding, which needs to arrive before the end of September, or he will have to institute major cuts. He has been told that an additional stream of routine-maintenance funding he was expecting will never arrive. Because the site is run in conjunction with a local nonprofit that pays Morgan’s salary, he said that he was comfortable speaking openly about the struggles his staff faces as funding and resources are instead directed to the White House list. “I am not beholden to hold my tongue at this point,” he said.

Inside the NPS, Morgan said, the staff has been decimated. “I started in this role in April of 2024, and the number of people who are not there since I started is gut-wrenching,” he told me. “There is a whole loss of institutional knowledge.”

Other NPS employees requested anonymity to speak with me because they fear retaliation from their supervisors. One employee told me that their park can’t hire employees to fill vacant slots but also can’t contract out work, because it’s difficult to get on the high-priority list. “We keep being told to be creative,” this employee said. “And we are running into dead ends every time we try to pivot.”

[Read: Let them eat gold]

This reduction in spending showed up in budgetary documents I reported on in June. The documents showed a decrease of $854 million, or 68 percent, in spending on projects in park regions outside the Washington area in the first eight and a half months of fiscal year 2026, compared with the full prior fiscal year. During that same period, spending around Washington increased by about $100 million, not counting about $310 million in donations that the NPS received from allies of the president, most of which is going to fund a new White House ballroom.

The frustration over these shifting priorities has been a frequent topic of conversation among park superintendents, who are struggling to maintain services this summer. At another NPS meeting in early July, an employee presented a list of 133 projects, totaling $35.8 million, that they said were ready to go but lacked contracting staff to make them happen, according to a person on the call.

It’s not clear whether all of the projects on the White House priority list will receive funding this fiscal year. The list also includes a $25 million plan to replace the Eagle One helicopter used by the U.S. Park Police (and at least once by Burgum when he was running late for an Air Force One flight), $475,000 to clean and repaint the White House exterior, drones “to support law enforcement” in the Washington region, a new park-police substation for Rock Creek Park, and streetlight replacement along the National Mall. The priority list also includes improvements for parts of the country where Trump is known to travel. The document shows a $77,300 expense to clear brush and plant new trees around Camp David, the presidential retreat in Maryland, as part of a much larger effort to plant 3,000 trees and provide a “security benefit” to the facility.

And Trump continues to add more projects to the list—or expand projects already on it. Take, for example, the brick pavers in Lafayette Park outside the White House gates. In May, Trump said that the park renovation would be funded by “a contribution that I made,” even though internal documentation shows public funds being used. After the Fourth of July celebrations, the NPS agreed to pay an additional $1.5 million to replace more of the brick pavers in that park. Then, in August, the president told reporters in the Oval Office that he was still deliberating on whether the brick pavers were good enough.

“Now we have a big decision to make,” he said. “Do we take the pavers that are red—you know, the red-brick pavers—and do we change them into a granite, to gray granite? They’re beautiful, but they’re not the original. So we’re gonna have to make that decision.” The National Park Service, no doubt, will be ready to follow the president’s lead if he pivots.

For several months, something discomfiting has been creeping onto San Francisco’s lampposts, bulletin boards, and restaurant windows: more and more signs, for run clubs and a July 4 parade and tacos, all of which are obviously entirely produced by generative AI.

This bothers me terribly. (Friends have called it a fixation.) But why does it bother me? A ChatGPT-generated poster for a garage sale seems legitimately dispiriting compared with a handwritten flyer, or even a haphazardly formatted Word document. What about a bodega using an AI-generated image of a hamburger instead of a random piece of clip art pulled from the internet? It’s not like most of these flyers were ever going to be high art anyway. Still, the issue nags at me constantly, whenever I pass one of these synthetic images plastered over the real world.

Last month, in part looking for a reprieve from the AI onslaught, I went to what promised to be an imaginative new exhibition of Henri Matisse and other fauvists’ work at the San Francisco Museum of Modern Art. The show was centered on Femme au chapeau (“Woman With a Hat”), a foundational painting in modern art. Matisse depicted his wife with colors that were totally detached from reality: teal patches of skin, an orange-and-green neck, a phantasmagoric hat. In 1905, when the painting debuted, this was scandalous: “A pot of paint has just been thrown in the public’s face,” one critic declared.

Walking through the galleries, I was taken aback by what struck me as a scandal of its own. The museum, in collaboration with Google Arts & Culture, had used generative AI to augment visitors’ experience—in ways I found, at least initially, rather off-putting.

Part of the exhibition’s conceit was to re-create Femme au chapeau’s historical context. The museum commissioned a local artist to develop wallpaper that could have been in the palace where the painting debuted and worked with an animator to create a visual projection of the walls of Gertrude and Leo Stein’s living room, where the canvas hung for many years. I loved all of that. But an AI-generated animation of the streets of Paris circa 1905, with shadowy figures walking to the gallery and a dirigible floating above, felt more like a gimmick than historical immersion. The museumgoers around me, however, seemed awed.

Then, near the gallery’s exit, came the second AI experiment: a row of screens, each showing an expanded Matisse painting made with Veo, Google’s video-generation model. A painting of a seashore became a 3-D rendering in fauvist style; a portrait of a girl with green eyes was reimagined to place the girl in Matisse’s studio.

The fauvists painted the world with flat slabs of highly saturated hues, delighting in color for the sake of color rather than realism. Using AI to move beyond the frame and give the images “realistic” depth seemed antithetical to their original vision. I recalled childhood memories of being entranced by various Matisse works at the Museum of Modern Art in New York City, marveling at his Dance and The Piano Lesson. The AI videos alienated me rather than inviting me in.

Yet maybe I’m just a grump? I am at least open to the possibility that my response is not so different from that of the pot-of-paint conservatives in the early 1900s. So I reached out to the museum’s chief curator, Janet Bishop, to learn more about these AI experiments—how they’d been conceived, what effect they had on the show, and what role AI broadly has to play in traditional art museums. We met at the galleries, where Bishop told me that curators wanted to “give a sense of what early-20th-century Parisians or visitors to the Grand Palais would have experienced in approaching this imposing exhibition hall and then going into the sprawling 1905 salon.” The AI installations resulted from extensive human research and oversight. I came away, to my surprise, not unconvinced of these experiments’ utility—though still a bit skeptical of their tastefulness.

This conversation has been edited for length and clarity.

Matteo Wong: How did your collaboration with Google Arts & Culture come about?

Janet Bishop: It was early in 2026. We were fairly far along in terms of planning for the exhibition. We had secured all the artwork by that point and were in the process of developing a couple of the ideas that we had for how to situate visitors, such as by re-creating the gallery where Matisse’s painting made its debut at the 1905 Salon d’Automne in Paris and the successive scandal.

Our original idea for the opening gallery, where we wanted to bring our visitors back in time to turn-of-the-century Paris, was to project a lantern slide that is an image of the Grand Palais. Then our museum director, Christopher Bedford, and head of education and community engagement, Gamynne Guillotte, got back from a really inspiring trip to the Google Arts & Culture offices in London and proposed the idea of setting still images in motion.

Through our curatorial research, we had postcards, still images, interior shots, exterior shots, and other material to work with. So in what became weekly meetings with Google Arts & Culture and project team members, we reviewed the images we had and developed this film.

Wong: What does having the animated film provide to a visitor that projecting the lantern slide and displaying postcards wouldn’t?

Bishop: We have some archival displays throughout the exhibition. But what we’re programming for in this entrance gallery is to give people a sense of what it felt like to approach the Salon d’Automne in a more immersive way.

Wong: Someone could look at this AI animation and its imperfections and perhaps find it kitsch. Were you worried about that at all?

Bishop: “AI-generated” makes it sound like AI did it by itself or that this was quick and easy. The project was very labor-intensive, and everything is historically accurate. One moment that was especially exciting to me was when we were looking at a draft, and there were all of these figures that were moving away from the building, and I said to the Google engineer, “Oh my gosh, it would be so amazing if we could have them walk toward the building,” and she said, “No problem.”

Wong: Is it harder now to get people to come here and engage with the art, especially younger folks? Or is this something that would have been just as helpful decades ago?

Bishop: You were talking about your experience as a child going to MoMA, and now we recognize that people have a lot of things that command their attention. One of the challenges in getting people to art museums is just pulling people away from their screens.

People are used to experiencing content in a really wide variety of ways. I know that it is important to our director that, given our proximity to Silicon Valley, we pay attention to developing technology and how it might enhance visitor experience. I loved the idea of the original lantern slide as a projection, but I can say that I think this film does something different in terms of giving our visitors a sense of the energy of Paris around the turn of the last century.

Wong: I’d also like to talk about the other AI collaboration with Google in the exhibition, near the exit. Can you tell me the origins of this one?

Bishop: The thought for what we called the “Experiment” was to enter into the world of a select number of Matisse paintings and imagine the context for those paintings beyond what is represented on the canvas. For instance, with The Girl With Green Eyes, we don’t know who the sitter was, but we do know that it was made in Matisse’s studio. So in the Experiment, you see the figure get up and walk across the studio. The details of the studio are consistent with his actual studio, which—we know what it looks like. And she looks out the window across the Seine.

Wong: Clearly a lot of visitors are enjoying this, but I wanted to ask why you thought it was important to help people imagine the world beyond the frame. My initial reaction, if I’m being honest, was that there’s a frame for a reason, and that the paintings aren’t naturalistic for a reason. What was so shocking 100 years ago was that these were, like, just flat planes of color, and that the colors were the focus. To turn that into a three-dimensional seascape seems almost wrong or off-putting.

Bishop: It’s part of our interest in engaging with different sorts of learners, people of different ages. They’ve been very popular with kids, and these images of paintings might encourage people to look more closely at the actual paintings. We’re experimenting, and I think experimentation is very definitely what we embraced with these features within the exhibition.

Wong: You mentioned earlier that you were excited by changing the historical AI video to have visitors walking toward the palace, rather than away from it. Was there a similar moment of excitement here?

Bishop: The big surprises came for me in the first piece, where in terms of the ability to conjure an experience, the use of AI as a tool made us more effective in achieving our goals than would have been possible without it. So that was really actually quite exciting. I think the Experiment is a little bit less integral to the story but an engaging feature that I’m happy that we pursued.

Donald Trump celebrated the defeat of one of his first impeachment nemeses—retired Army Lieutenant Colonel Alex Vindman, who lost the Democratic Senate primary in Florida on Tuesday—in typical form. “SleazeBag Vindman loses tonight to a Radical Left Lunatic. How cool is that???” Trump exulted on Truth Social.

Vindman served as a witness in Trump’s first impeachment. Whatever currency this once gave him with Democratic voters has long since expired. In this, he is not unique. Democrats such as Dan Goldman, George Conway, and Tom Steyer, all of whom pushed for or played prominent roles in Trump’s first impeachment, have discovered in primaries this year that this credential means little to their party’s voters. Diana DeGette, a House manager during Trump’s second impeachment trial, and Harry Dunn, a Capitol Police officer who battled insurrectionists, have also lost primaries.

The rout of the Democratic Party’s impeachment stars has been interpreted, generally, as a sign of the end of resistance politics. “The string of losses for resistance candidates from the first Trump term illustrates a new era for the restive Democratic base, where vocal opposition to the president through official channels and public remonstrations no longer guarantees success,” Politico reports.

[David Frum: The ‘resistance’ is over]

The “resistance,” once a dominant mode of Democratic politics during the first Trump administration, was never cool. But in the second term, the “resistance liberal” has fallen out of favor so completely that the phrase has become a derisive punch line across the political spectrum. Which is a shame, because the resistance was, broadly speaking, correct. It was right about Trump’s nature, and also right about how to defeat him.

Resistance politics depicted Trump as a crook and an aspiring authoritarian with a troubling fondness for dictators, especially Vladimir Putin. He was a con man, unfit for office, and had broken faith with the best traditions of American politics. The Robert Mueller investigation found, among other things, that Putin had dangled a lucrative no-risk deal to Trump during the 2016 campaign. The Senate Intelligence Committee, found, in a bipartisan report, that the campaign’s links to Russia posed a “grave counterintelligence threat.” The first impeachment charged Trump with having “placed his own personal and political interests above the national interests of the United States, sought to undermine the integrity of the U.S. presidential election process, and endangered U.S. national security.”

These much-derided processes turned out to supply an accurate, almost banal, description of how Trump has operated during his second term. His corruption and abuses of power have been let loose at a scale that was previously only imagined. Trump has all but severed America’s system of alliances, mocked and bullied Ukraine’s president, and showered praise upon Kim Jong Un.

What discredited resistance politics in liberal eyes is that it was closely associated with institutional procedures that were deemed to have ultimately failed to restrain Trump. Mueller’s investigation ended anticlimactically. Neither impeachment drew enough Republican support to remove Trump from office. Somewhat similarly, conservatives have sometimes claimed that the Russia investigation and first impeachment helped Trump politically by drawing to his side Republicans who resented the overreach.

One Republican who doesn’t seem to believe either of these arguments is Trump. The president has worked hard to ensure that no such humiliations recur during his second term, by stacking the Justice Department with loyalists, firing inspectors general en masse, purging the armed forces of “woke” officers, and generally eliminating any potential source of independence within the federal government.

Trump himself does not seem to think that the investigations that were avidly followed by resistance liberals did nothing to restrain his actions during his first term, or that they helped him shore up partisan support. And the absence of any such processes in his second term has not made Republicans in Congress feel freer to criticize the president. To the contrary, their fealty is more absolute than ever.

Even in the first term, resistance Democrats were dogged by their populist left-wing critics. This split first appeared in the spring of 2017, when Jon Ossoff ran to fill a special election for a House seat in Georgia, in a previously Republican district. Ossoff’s campaign, which attacked Trump’s corruption but declined to articulate a sharp progressive alternative, fell short.

Supporters of Bernie Sanders blamed Ossoff, and the party, for not taking up the Sanders message. “Ossoff’s campaign has failed to support or address any progressive issues. He embraces a centrist—not progressive—message, that both political parties waste taxpayer money, rather than politicians being corrupt by outside money,” the progressive writer Michael Sainato complained in the Observer. Winnie Wong, a co-founder of People for Bernie, told The Guardian of Ossoff, “He didn’t have a core progressive message and that ultimately is why he lost.” (Ossoff won a Senate race in 2020 and is leading the polls again this year, making slashing attacks on Trump’s personal corruption while steering clear of controversial left-wing policy stances.)

[Jonathan Chait: Liberals are now fighting a two-front war]

Like the Republican critics of resistance politics, progressive ones depicted it as substituting legal tactics for electoral victories. Yet the resistance was never limited to cheering on Mueller or hoping to persuade Republicans to convict the president of high crimes and misdemeanors. Resistance themes inspired the Democrats’ political message during Trump’s first term. And that message landed. Democrats retook the House in the 2018 midterms, despite a hot economy, and then defeated Trump’s reelection effort. Joe Biden, a link to a bygone era of civility and bipartisanship, ran in 2020 on a message of restoring the soul of America. He won.

Once in office, however, Biden ceded his domestic agenda to progressives who believed in Sanders’s vision of transformational change. They believed that only overturning the neoliberal order, handing control of domestic policy to neo-Brandeisians and climate hawks, and giving unions almost every policy they demanded would allow Democrats to regain the trust of working-class voters. This theory, not the resistance, failed. Biden’s support collapsed within his first year, and Kamala Harris’s half-hearted efforts to distance herself from her unpopular boss were too little, too late.

Trump’s return to office has left Democrats casting about for a stronger and more effective response. Perhaps the most obvious solution is the one they hit upon when Trump was first elected nearly a decade ago: Rather than promise to change everything about American politics, focus relentlessly on Trump’s unfitness. The crisis in the United States is not the general corruption of both parties; it is the specific corruption of Trump and his enablers. And the solution is simply to replace the corrupt, racist authoritarian with a governing party that respects the rule of law, equality, and basic decency.

This article was featured in the One Story to Read Today newsletter. Sign up for it here.

Anyone can get distracted now and then, and say things out loud he or she probably shouldn’t say. Perhaps an excited friend lets on about a surprise party, or a co-worker dishes up some unseemly office gossip. Or, if you’re the president of the United States, you might reveal some numbers about an adversary’s nuclear arsenal and then admit that those weapons, even a relatively small number of them, are pretty darn useful.

Donald Trump made this impromptu foray into nuclear-deterrence issues yesterday when he was inspecting the destruction he’s inflicted on the White House East Wing. A reporter asked whether he was again exchanging letters with North Korean leader Kim Jong Un, a reference to the president’s earlier epistolary bromance with Pyongyang’s dictator. “I can’t tell you that,” Trump responded. “But I get along with him. And you know what? The fact that I get along with him, that’s a good thing. Not a bad thing.”

It’s not always a “good thing” to “get along with” a murderous autocrat who runs a country that is functionally a prison camp, especially one with a nuclear arsenal. The size of that arsenal is a tightly held secret in Pyongyang, but it is not inconsequential: Experts working from open sources have made various guesses, usually in the range of roughly 60 thermonuclear weapons.

[Tom Nichols: Trump will make South Korea pay for his Iran humiliation]

Trump, however, blurted out an exact tally. “He has 57 very powerful nuclear weapons,” the president said. That’s a pretty specific number; the intelligence community almost never issues such precise figures, instead offering ranges of plausible guesses. Trump may be making it up out of nowhere, or perhaps talking about some subset of Pyongyang’s arsenal, as the arms-control expert Jeffrey Lewis suggested, or he’s garbling something he saw in a classified setting and didn’t understand. In the end, Lewis mused in a series of posts on social media, “Trump is trying to tell us a secret he heard, but for the life of me I can’t figure out what it means.”

Clearly, Trump has been getting briefed on developments about the North Korean nuclear inventory. And yet, he seems strangely unconcerned. The stockpile might be small, but it is in the hands of a dangerous regime. Even at its current estimated size, the Hermit Kingdom’s arsenal can kill tens of millions of people in Asia; soon enough, the North Koreans may be able to reach the United States with nuclear-armed long-range missiles.

The president, however, seemed more concerned with blaming other chief executives for the situation, and he assured everyone that all was well.

Should have never allowed it to happen. They should have never allowed it. If I were president, I wouldn’t have allowed it. But he’s got ’em. I get along with him very well. I can’t let Iran have a nuclear weapon. And I know Kim Jong Un very well, and he’s gonna be fine. As long as we have a smart president, he’s going to be fine. If we have a dumb president, then he’s probably not gonna be so fine.

“Smart” and “dumb” need a bit of unpacking here, of course. When Trump says “smart,” he means me, and when he says “dumb” he means not me. What he seems to be implying is that although North Korea’s nuclear arsenal is very dangerous, he has no intention of messing with either the bombs or the regime and that future presidents would be fools to do otherwise.

Trump never says how he would have stopped North Korea’s nuclear program, and despite his current equanimity about North Korea’s nukes, he has emphasized that Iran can’t have a nuclear weapon, because … well, his reasoning on this double standard isn’t clear. Is it because the North Koreans—fighting alongside Russia in the middle of Europe, sharing missile technology with U.S. enemies, and running giant crypto scams—are somehow less dangerous a regime than Iran? Or perhaps he thinks the Iranian ayatollahs aren’t as reasonable as the North Korean leader—a man whose regime routinely threatens to turn Western cities into a “sea of fire”? Or—more likely—Trump feels he can live with Kim’s nuclear ambitions because the two of them took nice walks during a meeting that sated Trump’s ego with lots of saluting and pageantry.

Whatever Trump’s feelings, his bumbled comments (and needless fumbling of likely classified information) nonetheless revealed how much Trump regards nuclear weapons as a sign of status, and sent a message of crystalline clarity to Iran: Nuclear deterrence, even with only a few bombs, works. Get a nuclear weapon as soon as you can if you want to be safe and treated with the kind of deference America shows to North Korea.

[Tom Nichols: Russia and America are rediscovering the limits of nuclear weapons]

Iran and North Korea are practically allies, and in all likelihood, the North Koreans have probably already said something like this to their friends in Tehran. Recall that when NATO began its 2011 intervention in Libya, officials in Pyongyang said that Libyan leader Muammar Qaddafi should have held on to his nuclear program, instead of bargaining it away in exchange for Western promises. Months later, Qaddafi was dragged into the street and torn to pieces by his own people. The North Koreans, according to one scholar, probably felt “alarmed, but also deeply satisfied with themselves” as they watched countries such as Iraq and Libya suffer a fate that the DPRK averted by having a nuclear shield.

And now the sitting president of the United States has affirmed North Korea’s strategic wisdom: Trump has shown the world that even a tiny nuclear arsenal deters America from taking military action, and in Trump’s case, enemies who acquire nuclear weapons will also receive fawning letters and precious visits from the most powerful man on the planet. Forgo those weapons, as the Iranians have done until now, and the result is destruction, attempts at regime change, and demands for unconditional surrender.

Build a nuclear bomb, and you’ll face sanctions and other hardships—but you won’t get stealth bombers pulverizing your country. Your regime will survive, and after everyone gets used to the idea of another nuclear-armed power in the world, you’ll get entreaties and offers for talks.

And if America has a “smart” president, you might even get a love letter or two.

Sign up for our newsletter about national security here.

Early this year, Ukrainian officials came to President Volodymyr Zelensky with a risky new plan to end the war with Russia. It would require expanding their bombing campaign to include a set of targets and a class of weapons they had long avoided. In addition to Ukraine’s ongoing strikes against Russian military bases, oil refineries, supply lines, and logistics hubs, the officials wanted to launch waves of AI-enabled drones at Moscow’s airports, hoping to stop international carriers from flying in and out of the Russian capital, according to two sources with direct knowledge of the plan, which has not been previously reported.

Both people spoke on the condition of anonymity to describe sensitive military operations. If executed, the plan would mark a terrifying milestone in the history of warfare: an attack in which swarms of autonomous drones equipped with AI guidance systems would seek out and hit their targets without a human pilot controlling their movements. Such systems have been tested by various countries in recent years, including the United States. But the legality of their use has been vigorously debated as a possible violation of the laws of war, and they have never been deployed at scale.

When officials presented the plan to Zelensky this winter, he expressed strong reservations, one of the sources said. He worried that it could be seen as “piracy” if the drones hit a civilian airliner in Moscow, especially if the strikes killed or injured passengers, airport staff, or members of the flight crew. But in his desperation to find a Russian pressure point that could restart the peace process, Zelensky allowed preparations for the strikes to continue for several months this spring and early summer, both sources said.

A spokesperson for Zelensky declined to comment. Another official in the president’s office denied the existence of a plan to launch swarms of AI-enabled drones at Moscow’s airports. “That’s total nonsense,” the official told me, asking not to be named in an article about this topic. Ukraine has long sought to force the closure of Russian airports, the official added, and its drone strikes against various targets in Russia have often achieved that goal for short periods of time. But Ukraine does not have long-range weapons that are precise and reliable enough to hit Russian airports without posing an undue risk to civilians, the official said.

[Read: Inside the feud over Ukraine’s drone success]

According to Russian aviation authorities, airports across the country were forced to close 114 times during a one-month period last summer, from June 28 to July 28, usually due to the threat of Ukrainian drones. Moscow was affected by the highest number of shutdowns—at least 17 in total at its four main airports. The Russian association of travel agents reported more than 290 of these incidents nationwide in the first quarter of this year. The airport closures, which Russians refer to as a kovyor, meaning “carpet,” usually last a few hours before authorities give the all-clear and allow flights to resume. “We’ve only ever had one operation against the airports,” the official said. “Force the Russians to declare ‘carpets’ all the time.”

By the middle of spring, the plan to force a more permanent shutdown of Moscow’s airports had advanced far enough to receive a code name: M&Ms, a reference to the many small, cheap, and nearly identical drones that the mission would require. To overwhelm Russian air defenses, the operation would need to send as many as 1,000 drones a night to airports in and around Moscow, said one of the sources familiar with it.

Attacks of that size would not be far outside the norm in this conflict. Earlier this week, Russian authorities reported shooting down 791 Ukrainian drones in a single night, one of the largest attacks of the war; more than 600 of them were flying toward the Moscow region, according to the city’s mayor.

“The idea is to shut down the Moscow airspace to civilian flights for an extended period of time,” one of the sources familiar with the plan told me. Since the start of the Russian invasion, in February 2022, civilian air travel over Ukraine has been suspended. But Russian airports have mostly remained open, and millions of passengers used them this year to travel abroad. Due to Western sanctions, only a few international carriers still fly in and out of Moscow, including Turkish Airlines and Emirates, which many well-to-do Russians use to reach hubs in Istanbul and Dubai.

Forcing those carriers to stop flying to Russia would make life less comfortable for officials, oligarchs, and other wealthy loyalists of President Vladimir Putin. “People in the elites might come to Putin and say: We have a problem here,” said one of the sources familiar with the Ukrainian plan. That kind of pressure from the elites, in combination with other wartime strains on Russia’s economy and society, could force Putin to negotiate a peace deal, the source said.

Both sources told me planning for the operation was halted in July, when Zelensky abruptly fired its mastermind, Defense Minister Mykhailo Fedorov. As the architect of Ukraine’s military drone program, Fedorov had worked on the plan for most of this year, keeping it a secret even from some of his closest aides. He helped assemble a team of Ukrainian computer programmers to develop the AI guidance system, and he secured financing from Ukraine’s allies in Europe to produce the autonomous drones at scale, the two sources said.

Having served throughout the war as a crucial member of Zelensky’s team, Fedorov changed this month into a potential rival for the presidency. In a video address posted on YouTube yesterday, the former defense minister called for Ukraine to hold elections and choose a new government. Ukraine has not held presidential or parliamentary elections since 2019, because Ukrainian martial law, which has been in effect since the start of the invasion, prohibits elections in a time of war. Opinion polls taken in recent weeks show that Fedorov has grown popular enough in Ukraine since his dismissal to pose a serious challenge to Zelensky in a presidential race. “Democracy cannot be held hostage by Russia,” Fedorov said in his speech.

The same day, the Ukrainian Parliament confirmed the nomination of a new defense minister, Yevhenii Khmara, a decorated military veteran. The vote in Parliament foreclosed Fedorov’s hopes of returning to his previous post anytime soon. It also threw into doubt the future of the operation known as M&Ms. One of the sources familiar with the plan expressed hope that the new defense minister would carry it forward. “We will try. Technically it’s possible,” the source said. “You can revive this project with a new team, but it would take, in my view, around three months.” By that time, another brutal winter of war will have arrived.

In the months before his dismissal, Fedorov pursued numerous high-tech plans to force the Russians to negotiate before the onset of winter. The Atlantic reported earlier this month that he had appealed for help from Elon Musk, and that Zelensky had asked President Trump to secure Musk’s assistance. The Ukrainians needed Musk to allow the use of Starlink, his satellite-communications network, as a navigation tool for drones to hit targets inside Russia, including Russian ballistic-missile launchers, with precision. Musk has refused to allow the use of Starlink over Russian territory. He has not replied to my messages seeking comment on the matter.

As an alternative to Starlink, the Ukrainians have rushed to develop reliable drones equipped with AI, which offer a way to get around Russia’s counter-drone defenses. Most attack drones require a pilot to guide them to their targets, through radio signals, mobile internet connections, or some other means of remote control. Russia, like most advanced military powers, uses electronic jamming to disrupt the signal between enemy drones and their pilots, forcing the weapons to crash or veer off course.

[Read: Ukraine’s plan to unnerve Putin]

Last summer, the Ukrainians found an ingenious way around this problem. They hid 117 drones inside five long-haul trucks and hired unwitting drivers to bring them close to Russian air bases. Once in position, the drones connected to the internet, reportedly using Russian SIM cards, and flew out of the trucks’ secret compartments. Drone pilots back in Ukraine used the internet connection to guide most of the weapons to their targets. The mission, known as Operation Spider’s Web, damaged or destroyed roughly two dozen Russian warplanes, including nuclear-capable bombers, that stood parked on the tarmac of their bases in Russia.

Ukraine’s latest plan would allow the drones to pilot themselves, using an AI-powered technique known as map matching, the two sources familiar with the plan told me. The technique involves uploading detailed maps of a drone’s intended flight path to its onboard computer, along with pictures of its targets. As the drone flies, its camera collects images of the terrain below. The onboard computer uses an AI model to compare those images with the ones stored in its memory, a bit like a person consulting a map on an unfamiliar highway. When the camera spots something that matches, or resembles, the image of the target, the drone flies into it and explodes. A human pilot does not need to confirm that the drone has found the right target before it strikes.

When this technology first appeared in Ukraine a couple of years ago, some drone manufacturers told me, they had declined to make it available to their clients in the military, because they found it too risky. One drone maker showed me an AI-targeting system with a kind of accuracy selector, which he could slide up and down on a screen. If he chose the lower setting, it would give the drone permission to attack even if its AI model is only 30 percent certain that it sees the right target. The highest setting tells the AI to reach 90 percent certainty before it decides to strike. Even that did not seem precise enough for the manufacturer, who wondered what would happen if a child on a bicycle were to enter the frame of the drone’s camera as it locked onto its target: “Who is responsible if that kid dies?”

Ukrainian officials, arms makers, and military officers have wrestled with such questions for years. Some have called for the Ukrainian Parliament or an international body such as the United Nations to develop guidelines for the use of AI targeting in warfare. No such rules exist in Ukrainian law. The U.S. and most of its allies have vague guidelines about a human needing to be “in the loop” before an AI model decides to attack. But most experts on the laws of war see such restrictions as outdated and insufficient.

Zelensky raised concerns about unintended targets being struck when it came to the plan to attack Moscow’s airports, one of the sources familiar with these conversations told me. The officials behind the plan offered work-arounds, the source said. They suggested, for example, that Ukraine would warn civilian airlines to avoid Russian airports and evacuate their planes before the swarm of Ukrainian drones arrived. “If they say to themselves, You know what, it’s getting too risky. Why even continue flying there?” the source said. “That would already have quite an impact on the Russian elites. Those are their main connections to the outside world.”

On a recent afternoon in Washington, I put on a pair of augmented-reality glasses for a tour of some very recent ruins: the White House East Wing. In recent months, black fencing has overtaken the area, where Donald Trump’s ballroom construction is speeding along as the Supreme Court decides its fate. Visitors can’t see much of anything from the street. As I wore the clunky eyewear, though, historic images of the structure flashed in front of me. The glasses showed me features wiped away by the president during his ongoing renovation spree, such as the Jacqueline Kennedy Garden and the White House Family Theater. Looking around, I could see a panorama of images visualizing Trump’s plans to transform Washington, D.C.: the Reflecting Pool, the Triumphal Arch, the Eisenhower Executive Office Building, and more.

The tour guide, a public historian named Lexi Cleveland, created this very 2026 experience as the executive vice president of ARtGlass, an AR company. (One of the co-founders is Greg Werkheiser, a cultural-heritage lawyer currently suing the Trump administration over plans to paint the Eisenhower Executive Office Building.) Cleveland prompted me with questions throughout. “What do we actually lose when we lose a building?” she asked. When the glasses showed me a Black Lives Matter mural that the city removed after Trump’s return to office, Cleveland juxtaposed it with the equestrian statue of President Andrew Jackson in front of the White House. The populist Jackson is one of Trump’s favorite presidents; expanding slavery and signing the Indian Removal Act are inextricable parts of his legacy. The construction of the statue, completed in 1852, involved an enslaved apprentice. “If you got to decide,” Cleveland asked, “does the statue stay, go, get contextualized with signage, or get moved to a museum?”

She plans to ask these questions to larger groups when the tours start in full this month, and she’s prepared to mediate debates. The approach feels fitting in a country engaged in an active struggle over how to remember its past.

History is on Trump’s mind. He wants to become part of it by reshaping the commemorative landscape of Washington. And he wants to reframe how history is told within Washington’s museums. For over a year, the White House has scolded and threatened the Smithsonian. But last month, the president took an unprecedented step toward pushing that feud into the real world. In an executive order, he called on the National Park Service to place signs outside the National Museum of American History, which the White House has said is subject to “ideological capture” at the hands of “a radical activist cohort.” The Park Service will direct visitors to “accurate information” about the nation’s past and mount its own temporary exhibit about the signatories of the Declaration of Independence, who the White House says aren’t showcased enough by the museum. The fight for public memory will soon spill into the streets.

[Read: The capital is a mess]

It will happen on the National Mall, where the National Guard continues to roam and not far from the transformed White House, with its abundant fencing—at this point, a core aesthetic of the second Trump administration. Across town, scaffolding outside the Kennedy Center blocks the part of the facade where Trump’s name once hung above the late president’s. Historic statues at Arlington Memorial Bridge have been swallowed up by blinding gold. Meanwhile, at two plazas in downtown Washington, statues that came down during the Black Lives Matter protests have been reinstalled: One of Confederate General Albert Pike returned to Judiciary Square last fall, and one of Caesar Rodney (which was toppled in Wilmington, Delaware), who signed the Declaration of Independence and enslaved hundreds of people, now overlooks Freedom Plaza.

Historians, preservationists, and activists are finding ways to work their own ideas and stories back into public space, in spite of the obstacles. The most visible example might be a series of viral art pieces by the guerilla art group the Secret Handshake that keep appearing on the Mall. (In March, the group plopped a gold toilet by the Lincoln Memorial to poke fun at Trump’s redevelopment projects.) But there are more sober ventures too. When Trump’s exhibit goes up outside the American History Museum, a group called Citizen Historians for the Smithsonian plans to send docents to “interpret” the signage for visitors—a nerdy form of counter-protest. Even the Smithsonian American Women’s History Museum, which has not yet been assigned a site because of partisan fighting in Congress, is oddly of a piece with this renegade spirit: It has resorted to an augmented-reality exhibit that dreams up some new monuments for D.C.

The projects get at the central disagreement fueling these feuds over history: whether, as Trump sees it, American history is best told through a simple, sweeping narrative of grandeur or, as these groups see it, through a compendium of diverse stories. If the city suddenly feels like it’s in rough-draft form, Trump isn’t the only one with a red pen in hand.

A few days after Trump issued his executive order about the National Museum of American History, I arrived at a very wholesome protest: about a dozen concerned museumgoers holding signs outside the staff entrance to the building and singing patriotic songs such as “My Country, ’Tis of Thee.” The makeshift choir is part of a larger group, Citizen Historians for the Smithsonian, that came together last year to archive material across the Smithsonian because it feared that the Trump administration might start altering exhibits. I found one of Citizen Historians’ leaders, Jessica Dickinson Goodman, explaining the recent drama to a group of tourists, a kind of test run for conversations to come.

Earlier this month, the group’s role expanded from DIY archivists to rogue docents when it put out a call for help interpreting the signage that Trump is adding to the Mall. In its newsletter, the group described the effort as nonconfrontational and fact-based. Dozens of volunteers came forward. Dickinson Goodman told me that they’re creating instructions for the types of topics that might arise: “We’ll start showing up as soon as we hear of construction starting.”

[Read: Trump’s trigger warnings]

One place where the grand, old-school version of history that the Trump administration seems to prefer prevails in the monuments and memorials on the National Mall. Here, the subject matter is clear, and the tone, though solemn, is usually celebratory.

The sole memorial dedicated to a named historical woman is for Eleanor Roosevelt, though the statue is incorporated into her husband’s larger commemorative site. For its 250th celebration, the Smithsonian American Women’s History Museum put out its augmented-reality exhibit, “Unhidden Heroines,” to highlight five women associated with five major monuments—and to underscore their absence from public art. It’s hard not to perceive an exhibition squished onto your phone screen as an indicator of where women’s history stands these days.

Still, it’s worth a “visit.” At the Jefferson Memorial, the geotagged exhibition focuses on Mary Katharine Goddard, who printed the first copy of the Declaration of Independence to list most of the signers. At the Martin Luther King Jr. Memorial, Dorothy Height talks about how she helped organize the March on Washington but did not speak at it. At the end of each anecdote, a “monument” of sorts pops up next to whatever your camera is pointing at. The figures appear so large on the screen, they sometimes dwarf the actual monuments. “That was intentional, of course, because they were towering women in their own right,” Karri Nussle, the museum’s associate director for advancement and external affairs, told me.

It was fun. Well done too. But sitting there with my phone and my headphones and the make-believe monuments, I also felt kind of sad. Eventually, the augmented-reality glasses come off. The phone screen goes black. The volunteer docents go home. What’s left? At the Jefferson Memorial, after I put my phone down, sending a cartoonish Mary Katharine Goddard back into the digital ether, it was just me and the man cast in bronze.

When some Taylor Swift fans listened to the lyrics of her 2025 song “Wi$h Li$t,” in which she sings about yearning to “settle down” with “a couple kids,” they heard a Republican dog whistle. On social media, they claimed that Swift’s desire for family life was “a conservative agenda on a silver platter.”

Swift, an avowed feminist who has endorsed Democratic political candidates, almost certainly did not set out to make “tradwife propaganda.” But her online critics are getting at something real: Having children—particularly lots of children—has become a bit right-coded, as the kids say. Americans and Europeans who identify as leftist or liberal tend to have fewer children than those who identify as conservative or right-leaning. In the United States, differences in desired fertility seem to start at a young age—one 2024 study found that Republican-identifying 12th graders said they want more children than Democrat-identifying 12th graders.

Of course, plenty of liberals have children, and plenty of conservatives have small families or remain childless. And many Americans identify as neither conservative nor liberal, but rather as something in between. Still, several academics told me that research conducted in the past few years has shown a fairly strong association between ideology and fertility. The latest evidence comes from a June report by the Institute for Family Studies, a think tank that promotes marriage and family life, which polled more than 7,000 Americans aged 18 to 54 and found that self-identified conservatives had 1.4 kids, on average, to liberals’ 1.09, a statistically significant difference. The study also found that liberals were more likely to not have children. The ideological fertility gap remained even after controlling for other variables that may influence family size, such as religion, race, income, age, marital status, and education.

[Read: A culture war in the bedroom]

Lyman Stone, the head of IFS’s pronatalism initiative and the researcher who ran the poll, told me that when he first saw the results, he thought, Duh. This trend of greater fertility on the right fits with a pattern that he and other demographers have noticed since the 1990s. There’s no clear explanation for why, precisely, liberals tend to have fewer children than conservatives do. Causation and correlation are difficult to untangle here. Liberals are more likely to pursue higher education and to marry later, both of which can lead to having fewer children, and some research suggests that having children may make a person become more socially conservative.

The researchers I spoke with put forth three general theories about why the partisan fertility gap may persist, beyond simple demographics. First, liberals seem to have more anxiety about both the state of the world and their own parenting abilities, which may make them more reluctant to start a family. Second, the values and aspirations of liberals and conservatives tend to differ in ways that shape how they choose to live their lives. Finally, for decades, liberals and conservatives have been sorting into disparate geographic locations and lifestyles, and their respective bubbles seem to reinforce the decision to have children, or not. The result is that a singer such as Swift expressing her desire to have children may, to some liberals, seem like she’s spinning propaganda for a traditional life they can’t imagine wanting.

Some of the most striking data points from the IFS survey suggest that liberals hold a bleaker view of parenthood than conservatives do. Among liberals, 18 percent of respondents were unsure whether they would be a good parent, compared with 9 percent of conservatives. More than a third of the liberals said they thought parenting is “very complicated, difficult, and stressful,” compared with less than a quarter of conservatives. And 19 percent of liberals thought their mental health was “not good enough to have children right now,” compared with 10 percent of conservatives. All three beliefs were associated with having fewer children. “These views of parenting shape your idea of how hard and how rewarding it’s going to be to be a parent,” Stone said.

Other researchers I interviewed suggested several factors that might lead liberals to have this grim perspective. One is that liberals are more likely to follow a broad array of mainstream news sources, which may make them more aware of societal problems such as climate change or rising authoritarianism, and thus more concerned about bringing children into a world with such seemingly intractable challenges. Keeping up with current events could also alert liberals to changes in state policies that make pregnancy riskier. “The media that you consume, is it telling you that you might not be able to get pregnancy care if you need it?” Karen Guzzo, a demographer at the University of North Carolina at Chapel Hill who studies family formation, asked me.

According to Guzzo, liberals may also be more likely to believe that daunting news events will have a measurable impact on their life. Conservatives, by contrast, tend to believe that they will persevere despite whatever happens in the world. Many studies have shown that conservatives tend to favor individual responsibility and free will as an explanation for people’s successes and failures. When faced with a hypothetical scenario of a person losing their job, for example, conservatives are more likely to assume that the person was a poor performer, rather than that the company had financial problems. Guzzo said that many liberals, by contrast, are more likely to believe that structural factors such as the economy or the housing market will shape their future.

Some researchers also wondered whether therapeutic ideas about trauma and family dynamics could play a role in discouraging liberals from having kids. Liberals are more likely than conservatives to seek out counseling, and one view that can emerge in therapy is that the client’s own parents damaged them in some way. If someone believes “my parents have harmed me,” Heather Rackin, a Louisiana State University sociologist who studies family formation, told me, “they might become really afraid of reproducing those harms on their children.”

Another factor possibly driving the fertility gap is that Democrats and Republicans, as a whole, have different values. Researchers told me that many Democrats believe that the pinnacle of achievement is having a good education and a fruitful career. For many Republicans, the primary goal in life is to raise children. Conservatives, Stone said, “tend to just be predisposed to think that family is what you’re supposed to do.”

Conservatives and liberals, Rackin said, seem to have different answers to the questions “What does it mean to be an adult?” and “What does it mean to live a good life?” According to various  studies and surveys, she said, “conservatives are more likely to see people as embedded within traditions and obligations and social roles, and parenthood is a role.” Liberals, Rackin added, seem more interested in self-definition, and in “constructing and expressing your own identity and aspirations. You don’t just accept roles because of tradition.” In one recent Pew Research Center poll, for instance, Republicans were more likely than Democrats to say that having children was “highly important” for living a fulfilling life, and in another, Democrats were more likely than Republicans to say that creative activities were “one of the most important things” to them. Democrats, then, may be less willing to choose parenthood when it competes as a priority with career, friendships, travel, and other pursuits.

[Read: The best parenting hack is being rich]

Religion explains much of this values gap. Some 80 percent of conservatives are Christian, for example, compared with 37 percent of liberals, and many denominations of Christianity encourage having children. But even in studies that control for religiosity, some of the difference in desired fertility between the parties remains. Something about the conservative worldview seems oriented toward family life. “I think even if you took nonreligious conservatives and nonreligious liberals and compared them,” Rackin said, they’d still have different definitions of the meaning of life, which would in turn give them different perspectives on whether to have children.

The final explanation for the partisan fertility gap stems from the fact that, over the past few decades, the two major political parties have become more and more geographically sorted. Democrats tend to live among Democrats; Republicans, among Republicans. Jacob Brown, a Boston University political scientist who studies this phenomenon, told me this doesn’t necessarily mean Republicans and Democrats physically moved away from each other. Instead, the two parties shifted over time such that they have come to represent different types of people—the Democratic Party appeals to people in cities; the Republican Party targets rural voters. Now people who belong to each party disproportionately socialize with other people of the same party. Less than a third of Democrats, according to one recent survey, have “at least some” Republican friends.

This sorting means that initially small differences in fertility between liberals and conservatives can compound over time. As liberals see many other liberal, child-free people around them, and conservatives see fellow conservatives with larger families, they may make decisions about their own family size based on the choices of their friends and neighbors. “You’re getting cues about what appropriate behavior is from those around you,” Rackin said.

Research from the Penn State sociologist Daniel DellaPosta shows how this kind of social reinforcement can work. In a 2015 study, DellaPosta and his colleagues examined a survey of 216 different activities and beliefs—such as watching a dance performance or reading horoscopes—and found that 58 percent of them were significantly correlated with political ideology. People who dismissed modern art as something “a child could do,” for example, tended to be conservative, whereas jazz aficionados were likely to be liberal. This isn’t because liberal psychology necessarily fosters a love for John Coltrane. Rather, liberals and conservatives tend to interact with people similar to themselves, and through those exchanges, they influence one another’s lifestyle, DellaPosta told me. A liberal jazz fan, in other words, will spread his enthusiasm for A Love Supreme to other liberals because his social network is likely to be made up mostly of liberals.

[From the March 2020 issue: The nuclear family was a mistake]

Something similar could be happening with decisions about whether to have children, or how many to have. “Political identity and all kinds of lifestyle preferences can become very closely aligned even when there’s no necessary or intrinsic connection between them,” DellaPosta told me. (He’s not the only one who’s found these associations: In a 2018 study, two researchers found that liberals are, in fact, more likely to drink lattes than conservatives are.)

Of course, some behaviors and beliefs do have an obvious connection to politics. Conservatives in DellaPosta’s study were more likely to endorse spanking kids, for instance, which aligns with conservatism’s more traditionalist worldview. Liberals were more likely to say that people who intend to get married should live together first, which corresponds with liberalism’s more sexually permissive mores. But DellaPosta’s work also shows that political preferences may influence decisions about having children not because there’s something Republican about having kids, but because political affiliation invisibly guides how people live their life.

Maybe no single one of these theories is right—but, rather, all of them are. Ideology may nudge people into parenthood or childlessness by subtly shaping what they value, what they fear, whom they surround themselves with, and what they want out of life. The end result is that, as everyone settles on one side of the parenthood fence or the other, many pick the side with people a lot like themselves.

Subscribe here: Apple Podcasts | Spotify | YouTube | Overcast | Pocket Casts

I remember each of my triumphs in a trivia game because they are so rare. (Highlight, 2023: Who is Barry Manilow?) I do not shine in bar-trivia night, or charades, or Sporcle. At home, my track record when shouting out Jeopardy answers was always worse than anyone else’s on the couch, including my own children, who are obviously much younger than I am. What is Walden? What is Mount Everest? What is a pansy? Whatever comes at the end of “What is,” I usually do not know. The children long ago moved on to the British quiz show University Challenge, which I watched only once and never again, to preserve my sense of authority as a parent and also my self-worth (“If the number of the year of the Battle of Hastings is divided by the number of wives of Henry VIII, which pope’s numerical designation do you get?”)

I’ve always comforted myself by saying that trivia was a pointless accumulation of junk in the brain, the mental equivalent of a cluttered closet. But this is obviously just envy. Knowing trivia is a skill like any other. It can be learned, practiced, and serve you well in many ways beyond shouting answers to the TV.

On this episode of Radio Atlantic: Drew Goins—an Atlantic staff writer, a Jeopardy champion, and the creator of The Atlantic’s new trivia game, Rabbithole—shares his secrets on winning at Jeopardy and cultivating a very particular brand of curiosity.

The following is a transcript of the episode:

Hanna Rosin: I have never had the desire to go on Jeopardy. Like, at all. And to be clear, I haven’t been invited either.

Alex Trebek: If you’re ready, then let us play Jeopardy.

Rosin: My track record when I’m sitting on the couch, shouting out answers, is not good.

[Music]

Rosin: I don’t have the patience or the talent to memorize trivia, and I don’t crave the spotlight or pressure that comes with it. But recently, I started to feel like even though I don’t want to be a Jeopardy champion, it might be worth it to start thinking like one.

I’m Hanna Rosin. This is Radio Atlantic.

Drew Goins is a senior editor at The Atlantic. He created The Atlantic’s new trivia game, Rabbithole. But more importantly for this conversation and his Instagram following, he’s also a Jeopardy champion.

Ken Jennings: Drew Goins, congratulations! $35,000 for you, and you’re—

Rosin: And his approach to the game and to championdom is not what I expected, which is: Just be a born genius.

Turns out, it helps to be curious. Be the kind of person who finds joy in knowledge for knowledge’s sake, which seems like a quality that could be good for life, not just Jeopardy.

Also, there are some tricks, secrets that only a champion would know.

Drew Goins: The key to winning at Jeopardy is not knowing everything; the key to winning at Jeopardy is knowing one thing about everything.

Rosin: Today on the show: how to win at Jeopardy.

Drew, welcome to the show.

Goins: Thank you. Happy to be here.

Rosin: I’m very honored to have a Jeopardy champion.

Goins: A Jeopardy champion and a Jeopardy loser—let’s not forget that.

Rosin: So what was the first moment when you saw Jeopardy and thought, That’s me. I wanna do that. That speaks to me?

Goins: I think this is gonna be a common answer for people of my age, which is: the original run of Ken Jennings.

He is obviously the host now. He won 74 games before that in the mid-aughts. And I remember watching and thinking, Oh my God, you know, this dweeby software engineer is the coolest person I have ever seen.

Ken will say he’s still kind of amazed that people find him cool or whatever. But I have this very vivid childhood memory of being on a family beach vacation and scurrying back in from the ocean because it was almost 7 o’clock, and Ken was still playing, and Jeopardy comes on at 7:00.

And that day at the, you know, boardwalk, I had won some inflatable raft to take into the ocean, and I spent the episode blowing it up with my mouth. We did not have a pump or anything. And so by the end of the episode, I was so lightheaded from blowing up this raft. And then Ken loses.

Trebek: —FedEx. His wager was $5,601. He winds up in second place with $8,799.

Goins: It’s his 75th game.

[Applause]

Trebek: All good things have to come to an end, don’t they? Well, too bad for Ken, but he’s going home with a lot of money.

Goins: And I’m, like, a middle schooler at the time. And I’m like, Oh, surely, I’m hallucinating. It must be the lack of oxygen in my brain. This can’t be happening. (Laughs.)

Rosin: (Laughs.)

Goins: And then I asked my dad, “Did that really happen?” And he said, “Son, I’m afraid so.” I went on to avenge the legacy of Ken Jennings.

Rosin: So how did you go about studying? Like, how does one actually study? I feel like I’ve seen documentaries about spelling-bee kids, so I do have a sense of what their life looks like. What does a Jeopardy kid’s life look like?

Goins: The spelling-bee comparison I think is gonna be elucidating because when these kids are prepping for the Scripps, they have a really discrete amount of information to learn. Obviously, the English language is very generative. There are tons and tons and tons of words, so it’s strange to say discrete, but they learn their roots, and they learn various suffixes and the way that morphemes combine to form words.

Jeopardy is different, at least at first blush, ’cause it feels like Jeopardy can ask about anything. And so it’s a really overwhelming thing to set out and actually prepare for Jeopardy. I think the first step of it, at least in my experience, happened just in my upbringing, before I was actively preparing for the show, which is just a lifetime of curiosity.

Because it does ask about so much—it just pays to be a really curious person who’s interested in a lot of things. But then once you get what’s called The Call (capital T, capital C)—that’s what Jeopardy people say when you actually get the invitation to come onto the show—that’s kind of when you kick into high gear for studying, studying.

And you realize pretty quickly once you set out to actually learn Jeopardy stuff, that the canon of what Jeopardy asks about and the depth at which Jeopardy asks things is pretty much the same as the spelling bee. Like, there are tricks to pick up, there are bounds to what you need to learn, and you just need to figure out what those bounds are.

Once you go on the show, if you happen to be invited back and do well, people say you kind of get stuck in “forever Jeopardy.” So I might go back again at some point. I’ve gone a couple of times to play Jeopardy. But now that I’m in that space, every day I have an app on my phone that has thousands of flash cards on it. And I am hitting new flash cards to remember what the currency of the United Arab Emirates is, or remember what Pascalization is—these things that you just accrete over time that will disappear if you don’t tend to them.

Rosin: Okay, so now I’m in the other world with you. We’ve moved through the portal, and as you’re in this portal and you’re accumulating this knowledge, what are some taxonomies that you have developed? Like, what is the world of Jeopardy? What’s the landscape? What does it look like? What are you learning?

Goins: I think what Jeopardy really celebrates and what the Jeopardy canon really celebrates is an old-fashioned idea of a well-rounded person. The example that I think of and that a lot of people point to is how much Jeopardy asks about opera.

[Music]

Jennings: It’s time for Final Jeopardy, where we’re getting a little high-brow today. “Opera Source Material” is the category.

Goins: The average Joe Schmo, myself included, before going on Jeopardy does not know a whole lot about opera. But once you hit the Double Jeopardy round, which is where things become a little more rarefied, a little more academic, none of this pop-culture smut that appears in the first round of Jeopardy, you will often get questions about opera.

If you want to know whether something is likely to appear in the Jeopardy canon, ask yourself at a dinner party of a dozen or so well-read guests, Would two or more people know the thing?

Jeopardy has pushed me to want to be a better, more knowledgeable person, including learning things that I didn’t know before. Sports, for instance: I am not much of a sports person, and Jeopardy asks a lot about it. And I think what I tell myself when learning about sports or doing flash cards, It’s like taking my vitamins, right? I tell myself, This is not just for Jeopardy. I think this makes me a more well-rounded person who can maybe have a conversation with somebody that I wouldn’t have had before.

Rosin: Did it ever lead to a genuine interest? Like, did you ever start listening to any of the operas that you learned about?

Goins: I was listening to Carmen yesterday, actually.

Rosin: Ah, okay.

Goins: And I had not been an opera person before.

Rosin: So you’ve talked about what you have to learn. How do you learn it? Memory, the memory palace, and memory tricks are a whole art form of their own. So now that you’ve gotten a sense of the landscape of things you have to know, how does one remember them?

Goins: Kids’ reference books.

Rosin: Kids’ reference books?

Goins: Kids’ reference books. So in addition to bounds of breadth that we’ve been discussing for the Jeopardy canon, it is also a very shallow canon. Jeopardy only asks one thing about most things. If Jeopardy asks about a European duchy, the European duchy is going to be Luxembourg. If it asks about an Iowa painter—I am sure there are great painters in Iowa who are not Grant Wood—

Rosin: (Laughs.)

Goins: —but if Jeopardy asks about it, the Iowa painter is gonna be Grant Wood.

You kind of just have to learn one thing about everything. And what’s great for that are kids’ reference books. I’m talking about kids’ encyclopedias, kids’ atlases, all of those, like, What Is guides. So I checked out from my public library at the time What Is the Super Bowl?

Rosin: (Laughs.)

Goins: And I promise I know what the Super Bowl is, but I wanted that one thing beyond just football.

Rosin: Yeah, yeah.

Goins: These kids’ reference books—this is a trick that super-champion James Holzhauer also points to as having helped him—are great because they are really well shaped for the ethos of Jeopardy, of a survey of information that goes only so deep. And they’re colorfully illustrated. They are a joy to read.

I grew up reading them, and when I went off to high school, college, whatever, I kind of reluctantly set them aside because I was like, Oh, well, I’m a serious adult now who needs to read books that don’t have pictures. And then when I got the call to go on Jeopardy, I was like, Ooh! For listeners, I’m rubbing my hands together gleefully right now. I was like, I get to read children’s reference books again.

Rosin: Yeah.

Goins: Because getting to learn a little bit about so much is just really, really delightful.

[Music]

Rosin: I was a little hung up on this kids’-reference-book thing, so I decided to put it to the test. I made Drew play a quick round of Jeopardy and had him explain how he knew or didn’t know the answers. And just like a real game of Jeopardy, Drew didn’t have access to the questions ahead of time. I swear.

Okay, here we go.

Rosin: Dennis Hopper won a Best Villain MTV Movie Award for menacing a bus in this film.

Goins: What is Speed?

Rosin: Yes. So that was kind of a guess.

Goins: That was a guess, but Jeopardy clues are written very economically. Very rarely is there a word in there that is not in some way pointing you to the answer.

So I’ve, like—I had no idea Dennis Hopper was in that movie, which I have never seen, but I know Speed to be possibly the only movie that takes place almost entirely on a bus.

Rosin: Interesting.

After 3,388 performances and an eight-year run, this musical was no longer “the word” when it closed April 13, 1980.

Goins: Another little clue in there. What is Grease?

Rosin: Yes! What was your clue?

Goins: “The word.”

Rosin: The word.

Goins and Rosin simultaneously: (Sings.) Grease is the word.

Rosin: Yeah, exactly.

Goins: See, I would not have known that just from the statistics of it, ’cause there are, you know, a lot of long-running musicals. But you have to know Jeopardy’s language too.

Rosin: Yeah. Yeah. Okay.

This colorful cake got a boost in popularity from Steel Magnolias, where it was an armadillo-shaped groom’s cake.

Goins: “This colorful cake.” So from there I’m thinking, Okay, what’s a cake with a color in the name? What is red-velvet cake?

Rosin: Yes. I see how this is working. This is essentially: The Jeopardy mind would know to discount the irrelevant and very obscure—not necessarily obscure, but the more detailed information in here.

Goins: Right. See, it’s great if you happen to have seen Steel Magnolias, which a lot of people have, but really well-written Jeopardy clues offer a backdoor into the answer as well.

Rosin: Yeah. Yeah, ’cause “armadillo-shaped groom’s cake” is not gonna get you there. But “colorful cake,” there aren’t that many well-known cakes with a color.

Goins: Yeah. So red velvet is a reasonable guess.

[Music]

Rosin: There’s no shortage of hacks and tricks you can learn to prepare for a Jeopardy game. But what can Jeopardy do for you? That’s after the break.

[Break]

Rosin: Okay, so I feel like we’ve covered the theory of Jeopardy. Now I wanna ask you about the reality.

Goins: Sure.

Rosin: What was your very first time on the show?

Goins: My first time on the show was in June 2024.

Jennings: Drew Goins is a journalist from Honolulu, Hawaii, and being from the Aloha State—

Goins: And when you go on Jeopardy for the first time, they don’t put you up in a hotel. They don’t pay for travel or anything. You are paying for your own way.

Even if you come in third, you get $2,000 as your prize. I think the thinking is: That covers your expenses, and you get to have this experience. But it is kind of isolating because you’re out in California, terra incognita, by yourself. And I showed up to the studio, and you go past the studio gates at Sony Picture Studios, and you’re ushered into this parking garage, and there are a circle of chairs and a metal detector, and that’s it.

But then you go through that security screening, and then you go on to Stage 10, which is where the show is actually recorded, and all of a sudden, you open up the doors, and you’re greeted by a model podium and this wall of Emmys, and a hockey jersey that had been given to Trebek with his last name on it, and all of a sudden it feels so real.

[Music]

Goins: And you get just a chill. Oh, I’m giving myself chills right now describing it, which is how much of a dork I am.

And I remember going into the greenroom for the first time. And in the months, years since I’ve played Jeopardy, I think of this greenroom as—this is gonna sound so hokey—but as one of my happy places.

Because the people who are in there are your fellow contestants, who are just as nervous as you are and just as big of fans as learning inconsequentia. As the nervous energy wears off a little bit, people start warming up and they start sharing fun facts about things, and showing off their little trivia, maybe out of nervousness, maybe out of friendliness. A mixture of the two, possibly.

And by an hour in, everybody’s just saying, “Oh, well, did you know this? Did you know this? Did you know that?” And everyone’s like, “Oh, yeah, of course, of course. Then what about this?” And it’s just this bacchanalian celebration of knowledge, and the knowledge is flowing left and right. It is so exciting to be in a room of people who aspire to what Jeopardy wants you to aspire to.

Rosin: Okay, so you’re there in your first game. This is 2024, and how does it go?

Goins: I lose. (Laughs.)

Rosin: And do you remember any of the—

Goins: I lose stylishly.

Rosin: What do you mean? You looked good losing, or?

Goins: I certainly didn’t look good. I looked [like a] deer in the headlights. No, I played a really, really good game, and I lost narrowly.

I was leading until the very end, when my competitor David Erb, he finds the final Daily Double and wagers really big. It’s a Civil War question. The answer is Lee’s Miserables, a play on Les Miserables, and he springs in front of me. And then he gets Final Jeopardy right.

Jennings: So David Erb’s lead of $22,000, looking pretty good right now. Did he come up with Macbeth? He did. Well done. What did you wager, David?

Goins: I get it wrong, adding insult to injury.

Jennings: Drew Goins had an impressive second-place total of $16,800 today. Did he think of Macbeth? “Who is Titus Andronicus?” That’s even a bloodier play but unfortunately not correct.

Goins: But I played well enough that they, a couple of months later, ring me up and say, “Would you like to come back for the Second Chance tournament?” So I feel surpassingly lucky to have had not just one opportunity to play this dream of mine, but two, when most people get zero bites at the apple.

Jennings: Drew, welcome back. I hope you enjoyed your first appearance on Jeopardy.

Goins: I ultimately win the Second Chance competition. So I go to the next one, and I finish second in that, but by some other odd twist of fate, one of the contestants who was supposed to go into the Tournament of Champions is not able to play. And so I take her spot, and then I make it to the semifinals of the Tournament of Champions, where I, sadly, once again, fall short, but was brought back again for this year’s Jeopardy Invitational Tournament, which, as I said earlier, “forever Jeopardy,” right? Who knows what’s next.

Rosin: By the way, how much did you win?

Goins: I won just shy of $100,000.

Rosin: Oh! What?

Goins: Yeah.

Rosin: That’s incredible.

Goins: Pretty good.

Rosin: And what do Jeopardy champions do with their money?

Goins: (Laughs.) Nothing! If you’re struggling to fall asleep any night, just call up a former Jeopardy champion and ask them what they did with their money, because I can guarantee you it is in a high-yield savings account, or it is in very safe index funds.

I have a friend, Adriana, who won a whole bunch of money on the show, multiple times more than I won, and when she went back for a return appearance, she was asked how her life has changed since winning the money, and she said, “Oh, well, I’ve become a little more liberal with my boba budget.”

Rosin: (Laughs.)

Goins: I was like, “I should hope so. I really should hope so.”

Rosin: So this is not like calling up lottery winners.

Goins: No. It is not. Yeah.

Rosin: So I’ve heard people defend Jeopardy by saying, for example: “In an era of disinformation, we should respect this accumulation of facts.” Do you buy that theory?

Goins: I think that some people who are real defenders of trivia and fans of it like to frame it as an act of resistance or defiance. But I think that asks trivia to do things that trivia is not well equipped to do. There’s this common response variable of being a good, educated person and enjoying Jeopardy, and that common response variable is loving knowledge.

But I think we need to set trivia apart from that. What is so beautiful about trivia is that it really doesn’t matter. It is just knowledge purely for the sake of knowledge. I pick up these facts, and I get this little tickle of excitement when I see something that’s just sparkly or delightful, not because I know that I’m gonna use it, let alone use it to stop the creeping tide of authoritarianism in the United States, but because it’s this beautiful little jewel.

It is good and delightful to know that the Tuskegee Airmen flew mostly P-51 Mustangs, but knowing that one little bit of trivia is not gonna stop anybody from removing the Tuskegee Airmen from a curriculum. That’s just not what trivia is for or capable of.

Rosin: So then how did you start to think about knowledge or you in relation to knowledge? It’s not necessarily knowledge as liberation or as a path to liberation. It’s not necessarily knowledge as protest. It’s not necessarily knowledge as a thing you have to conquer and amass, and conquer and amass.

So what is the value to you, or what has the value been to you over time?

Goins: Preparing for and playing Jeopardy has really led me to see knowledge like I’d never seen it before, as this ocean out there.

[Music]

Goins: I think of my beach trip to Myrtle Beach, South Carolina, where Ken lost, and the next day when I was out on my raft that I had blown up with my own two lips, and just floating on top of this enormous body, knowing that I would never come close to experiencing all or even a small part of it, but delighting and just splashing right below the surface.

And to me, that’s what trivia is like, where I can stand on the shore of this body and scoop out something that will bring me a little bit of joy.

Rosin: Whew, Drew, I’m with you out there. Thank you so much for joining us today.

Goins: Thank you. This was a blast.

Rosin: Listeners, if you want to try your hand at trivia, Drew has a new game on The Atlantic website. It’s called Rabbithole.

This episode of Radio Atlantic was produced by Rosie Hughes and Jacob Smollen. It was edited by Kevin Townsend. Rob Smierciak engineered and provided original music. We also had music from Breakmaster Cylinder. Sam Fentress fact-checked. Claudine Ebeid is the executive producer of Atlantic audio, and Andrea Valdez is our managing editor.

Listeners, if you enjoy the show, you can support our work and the work of all Atlantic journalists when you subscribe to The Atlantic at TheAtlantic.com/Listener.

I’m Hanna Rosin. Thanks for listening.

This is an edition of The Atlantic Daily, a newsletter that guides you through the biggest stories of the day, helps you discover new ideas, and recommends the best in culture. Sign up for it here.

For many observers of American politics, the past few years have been disillusioning, knocking down hopeful beliefs about what the people of the United States will support or tolerate. Even some of Donald Trump’s backers have become disaffected.

Yet somehow, a small caucus of non-MAGA Republican senators has managed to remain preciously naive about the president and the administration he runs. No matter how many times Trump or his aides show who they are and how they operate, these senators seem flabbergasted and saddened.

No case illustrates the ease with which they are fooled like the confirmation of Todd Blanche as attorney general. Blanche was confirmed last week on a 50–49 vote, but only after GOP Senators Thom Tillis and John Cornyn extracted a written assurance from Blanche that a $1.8 billion fund for payouts to Trump allies would not proceed. Senator Bill Cassidy also wavered, but said he was persuaded by the vow. (All three men will leave the Senate at the end of this term: Tillis is retiring, and Cornyn and Cassidy lost primaries to Trump-backed challengers.)

This weekend, Blanche was interviewed on Meet the Press. Kristen Welker asked whether he could “pledge that the Justice Department will always act independently of the White House.” Blanche refused—likely an honest answer, given his demonstrated willingness to follow Trump’s lead, but one that should worry anyone who believes in the rule of law; witness Trump’s repeated demands for political prosecutions, irrespective of evidence. Yet Blanche insisted to Welker that Trump “will never ask me to do something unethical or illegal. He never has. He never will.”

That might reassure someone whose political memory stretches back only as far as early August. But anyone old enough to remember the slush fund, or Blanche’s starring role in negotiating it (as laid out by Federal Judge Kathleen Williams in May), would know that Trump not only has asked Blanche to do something unethical, but that Blanche complied. In order to vote to confirm, Cornyn, Tillis, and Cassidy had to convince themselves first that Blanche’s conduct in that case is not indicative of his basic moral compass; and second that he would be more accountable once permanently confirmed by the Senate than he was when he still had to win over their votes. If Trump finds or attempts to find some way to resuscitate the slush fund—as some experts believe he might—many people will be dismayed, but only this trio of guileless senators will be shocked.

Cassidy’s willingness to vote for a Trump yes-man is all the more puzzling because he voted to convict Trump in his 2021 impeachment trial, and because the president led the charge to defeat him in a primary. (The impeachment vote was Trump’s main reason for targeting him.) But Cassidy has a long record of credulity. As my colleague Nicholas Florko has reported, Cassidy—a physician—cast a deciding vote to confirm Secretary of Health and Human Services Robert F. Kennedy Jr., but only after the nominee made promises about preserving the existing vaccination regime, a pledge he promptly broke. On ABC’s The Week this past Sunday, Cassidy said “it’s pretty evident” that Kennedy’s vow was a lie. And on CNN’s State of the Union, he called the White House’s latest vaccine move, attempting to unbundle the measles, mumps, and rubella injection, “a crazy, stupid thing.” Senator John Barrasso of Wyoming, another doctor, voted for Kennedy and has since been upset about the administration’s anti-vaccine steps. Who could have foreseen that the prominent vaccine skeptic would undermine federal vaccination policy?

Two other moderate Republicans, Susan Collins and Lisa Murkowski, voted against Blanche’s confirmation, issuing statements that noted the politicization of the Justice Department under his watch as acting attorney general. But they are not always so clear-eyed. Murkowski delivered one of 2025’s most unintentionally hilarious statements when she cast a crucial vote to pass Trump’s One Big Beautiful Bill Act through the Senate. “My sincere hope is that this is not the final product,” Murkowski said in a statement. “This bill needs more work across chambers and is not ready for the President’s desk. We need to work together to get this right.” Naturally, that didn’t happen; the mystery is why she thought the White House, or Mike Johnson’s House, would improve the bill once she’d surrendered her leverage.

Collins, however, is the dean of the credulous caucus. The Mainer, who faces a tough reelection campaign this fall, is perpetually concerned about Trump-administration actions, but seldom ready to actually use her power. Collins is a rare pro-abortion-rights Republican who previously said she wouldn’t support any Supreme Court nominee who opposed Roe v. Wade. In 2018, she voted to confirm Justice Brett Kavanaugh, who said Roe was “settled as precedent.” When Kavanaugh was part of the majority that voted to overturn Roe in 2022, Collins lamented the ruling but now says she didn’t regret her vote for Kavanaugh and would do it again. More recently, Collins helped pass a huge appropriation for ICE with few strings attached, then expressed dismay at the use of the funds to buy “shock gloves” for agents. Even her vote against Blanche was symbolic, because he had enough votes to be confirmed without her.

One could easily write these senators off as cynical, but if they’re acting for political gain, it’s not working well: Cassidy’s and Cornyn’s mild deviations from Trump led to the end of their careers; Tillis’s retirement helped him avoid what might have been a stiff primary challenge. Many of their concerns appear earnest. Perhaps they really just are that gullible, unable to adapt to a Trump age that is leaving them behind. In defending his Blanche vote, Cassidy shrugged off attacks: “I’ll be criticized for this vote. What’s new?” But voters might reasonably expect that their elected officials are savvier about the world than a babe in the woods.

Related:

• Why did Bill Cassidy do it?
• What Trump officials mean by “never”

Here are three new stories from The Atlantic:

• David Frum: The ‘resistance’ is over.
• Why the closing of Harvard’s writing center matters
• Idrees Kahloon on the case for chilling out about birth rates

Today’s News

1. The United Arab Emirates said it will halt all trade and financial transactions with Iran, after accusing Iran of launching missiles toward Emirati territory. The announcement came hours after President Trump spoke with the UAE’s leader. Iran has denied carrying out the missile attacks.
2. The democratic socialist Angie Nixon defeated Alex Vindman, a former National Security Council aide, a surprise win in Florida’s Democratic U.S. Senate primary last night. Nixon will face Republican Senator Ashley Moody in November.
3. Trump said today that he had urged Republican Representative Cory Mills to drop out of his Florida primary “but he wouldn’t listen.” Mills lost to the former television anchor Ryan Elijah last night, and Trump has since backed Elijah.

Dispatches

• Work in Progress: Democrats are failing the Waymo test, Rogé Karma argues.

Explore all of our newsletters here.

Evening Read

You Don’t Know How Bad the Pizza Box Is

By Saahil Desai

Happiness, people will have you think, does not come from possessing things. It comes from love. Self-acceptance. Career satisfaction. Whatever. But here’s what everyone has failed to consider: the Ooni Koda 12-inch gas-powered outdoor pizza oven.

Since I purchased mine a year ago, my at-home pizza game has hit levels that are inching toward pizzaiolo perfection. Like Da Vinci in front of a blank canvas, I now churn out perfectly burnished pies entirely from scratch—dough, sauce, caramelized onions, and all. By merely looking at a pie, I can tell you whether the cornicione is too puffy or just right, if the crust could use a bit more leoparding, and whether the dough should have spent another day in the fridge. I am now, in a word, pizza-pilled.

But enlightenment is not without its consequences.

Read the full article. (From 2023)

More From The Atlantic

• Charlie Warzel on the ‘Home Slopping Network’
• Photos: Drone attacks on Russia’s Wildberries warehouses
• Helen Lewis: Enes Kanter Freedom obviously doesn’t belong in the WNBA.
• A Democrat in name only

Culture Break

Explore. Bumble’s “women message first” rule is dead—a change that reflects the hazy norms of dating today, Faith Hill argues.

Go outside. Ensuring a daily dose of sunlight is good for your happiness, Arthur C. Brooks wrote last year.

Play our daily crossword.

Rafaela Jinich contributed to this newsletter.

When you buy a book using a link in this newsletter, we receive a commission. Thank you for supporting The Atlantic.

New drivers may never have to learn how to parallel park. At least not the old-fashioned way. Many new cars—a Ford Mustang Mach-E, a BMW sedan, a Tesla Cybertruck—will just maneuver into a spot for you with a click of a button. Even a decade-old Toyota Corolla has a backup camera and pedestrian detection. Cars that leave you to park all on your own are a dying breed: Any car manufactured after April 30, 2018, for sale in the United States is legally required to have a backup camera and, thus, a screen.

After all, cars are just computers now. The screen (or really, screens) in today’s vehicles does a lot more than display rearview footage. It is essentially a tablet wedged into the dashboard that allows you to mirror your phone through Apple CarPlay, make a hands-free call via voice commands, Zoom into a work meeting, or even play a video game (not recommended). Mercedes-Benz debuted a 56-inch display in several models that it calls the “hyperscreen.”

But the expectation—and in some cases, the requirement—that cars have an ever-growing set of electronic features comes with a cost. Just like laptops, cars depend on microprocessors and RAM, or random-access memory, to run all of their computations. “There’s just a baseline level of tech, and thus a baseline level of cost, required of every vehicle,” Karl Brauer, the executive analyst at iSeeCars, an automotive-research platform, told me. Technology is a major reason the average price of a new car in the U.S. has reached some $50,000, and that was before RAM became one of the most prized commodities in the world. AI companies are snatching up as much memory as possible for their data centers, causing a RAM shortage that has significantly raised the prices of phones, laptops, and just about any consumer-electronic device. Cars are next.

The problem runs much deeper than a screen. Cars started morphing into computers in the 1970s, when new emissions standards made electronic systems necessary to manage ignition and fuel injection. There’s been a gradual software creep ever since: power windows, adaptive cruise control, antilock brakes. Regulators now rate cars based not just on how they do during a crash but also on how well their electronic safety systems avoid a crash. “We’ve evolved every single component of the vehicle to involve a semiconductor,” Ivan Drury, the director of insights at Edmunds, told me.

More recently, Tesla in particular has played a big role in pushing the auto industry to stuff vehicles with even more digital technology. Whereas pistons and fuel lines make a gas-powered car move, electric vehicles are run by a computer that regulates their battery and motor. It’s no coincidence that Tesla helped popularize the giant in-car touch screen. Today’s new cars, electronic or otherwise, have driver-assistance features, receive over-the-air software updates, and can even load ChatGPT. BMW advertises the ability to have a distinct climate-control setting for every passenger and the option to install screens into seat backs, airplane-style.

In most vehicles on American roads, all of this software is controlled by hundreds of separate electronic units throughout the car. Features such as antilock brakes and lane-detection sensors don’t need all that much processing power, and they can be operated by chips that might account for 5 to 10 percent of a vehicle’s materials cost, Sam Abuelsamid, an analyst at Telemetry and a former automotive engineer, told me.

Again following the lead of EV start-ups such as Tesla and Rivian, legacy automakers are shifting from these ad hoc, tiny control units to one or two big computers that run all of the software in their vehicles—and, in turn, require much more RAM. These central computers allow for a streamlined supply chain, seamless software updates, and a faster internet connection, plus, as General Motors advertises, they will be ready for “future AI workloads.” You can’t really build a car that safely and reliably drives itself with a collection of puny computer chips. Such centralized computers need more advanced automotive microprocessors from Nvidia or Qualcomm and might eventually account for 20 percent or more of a vehicle’s cost, Abuelsamid said. That was before the AI-fueled memory shortage. In other words, cars are becoming more software-heavy and dependent on memory chips than ever, right when those chips are in high demand from tech companies willing to pay premiums that automakers cannot afford.

Over the next year, the memory shortage—sometimes known as RAMageddon—will likely raise vehicle prices by a few percentage points, on average, Abuelsamid said. The relative amount would be smaller than the shocking double-digit jumps for gaming consoles and MacBooks but in some ways more significant: A 4 percent price hike for cars amounts to some $2,000 on average. Cars with those centralized computers will be affected the most, but no vehicle will be spared. “Even if you don’t need the high-end chips, you’re going to pay more even for the low-end chips just because of the supply constraint,” Abuelsamid said. On a recent earnings call, Ford’s chief financial officer said that the company had paid $1 billion in higher materials costs due to the memory shortage and inflation. GM and Volkswagen, too, have noted rising chip costs to investors. (Ford and GM did not respond to a request for comment. A spokesperson for Volkswagen told me that the company has “recognized an increased demand for memory chips, primarily driven by growing requirements in other industries,” and that in recent years, Volkswagen has taken measures to “mitigate supply risks.”)

RAMageddon is poised to last for several years, but the consequences for car buyers may be permanent. Consider what happened during the pandemic, when supply-chain disruptions and rising demand for electronics produced a major chip shortage. Nearly every major car company had to slash production because they simply couldn’t procure enough chips, and shifted their focus to selling higher-end and higher-profit vehicles. Potential customers already willing to spend six figures on a car are much less likely to care about a 5 or 10 percent price hike, Drury said. Even now, car companies are continuing to focus on selling more profitable models. Since the pandemic, the average price of a new vehicle has jumped $11,000.

The AI-fueled chip crisis could play out more severely. The average price of a new car could, before long, jump to $60,000 and beyond. These rising costs are making cars even more similar to computers and all of the software they run: Perhaps in an effort to mitigate higher prices, some automakers are also introducing in-car advertisements and putting certain features, such as heated seats, behind a paywall. As cars have morphed into computers, the inevitable next step is for automakers to behave like modern tech companies.

Most Americans cannot afford a $60,000 new car, let alone a $50,000 one. When new-vehicle prices soar, the used-car market follows. That has been one durable consequence of the pandemic’s chip crisis: The average price of a three-year-old used car is up $9,000, or nearly 40 percent, since before the pandemic. To get the same prices, Americans already have to buy much older cars that have far more miles on them. If automakers continue to shift their focus to designing cars with more software, more screens, more self-driving features, “that increasingly prices people even out of the used-vehicle market,” Abuelsamid said. It’s one thing for the AI industry’s appetite for memory chips to raise the price of an Xbox, or even a new laptop. But the stakes are much more consequential when it becomes harder and more expensive to get groceries, pick the kids up from school, and go to work.

It is not every day that the laying off of a university bureaucrat generates a torrent of criticism from students, faculty, and alumni. But Jane Rosenzweig was not just any university bureaucrat, and she was not presiding over just any administrative fiefdom. The dismantling of Harvard University’s writing center, and the laying off of Rosenzweig, its director—both announced last month—has been met with outcry shading into outrage. It has produced a minor literature of student-newspaper coverage, disapproving social-media posts, and angry emails. An open letter from current and former writing-center tutors protesting the decision had more than 100 signatories.

The passionate response is partly a testament to Rosenzweig herself, who was a beloved figure on campus. Among the flood of comments from former writing-center employees were a number that described her as the best boss they’d ever had. The backlash is also about what Rosenzweig represents: She has been a champion of human writing in the age of AI, through both her work at the center—where undergraduate students received coaching on how to revise and refine papers for their classes—and her essays on Substack. “Our students are not products to be moved down a frictionless assembly line,” she wrote in 2024, “and the hard work of reading, writing, and thinking is not a problem to be solved.”

Rosenzweig, who worked as an editor at The Atlantic from 1994 to 1997, has become a kind of synecdoche for the very idea of taking writing and the liberal arts seriously at Harvard in the post-ChatGPT world. Brandon Kingdollar, a former tutor who graduated in 2024, wrote in a comment on the open letter that her layoff “sends a clear message to today’s students that nurturing their abilities to write and think for themselves is no longer a priority.” Harvard’s leadership, he added, seems to be saying that “it is perfectly acceptable” for students to “outsource these important skills to AI agents.” A Harvard faculty member who requested anonymity so that he could speak openly told me he worried that the elimination of the center could be seen as part of a broader trend of “real boosterism” of AI on campus—one that students feel acutely. “There is an impression among students that they’re being told by people in positions of authority that really getting good at using AI is essential to what college is going to be for them and for their future career success,” he said.

[Read: ChatGPT doesn’t have to ruin college]

Harvard’s undergraduate college has held AI training sessions and given students access to premium chatbots including Google’s Gemini, OpenAI’s ChatGPT Edu, and Anthropic’s Claude. The university also maintains a webpage directed at faculty members titled “Teach With Generative AI”—oddly phrased as a command—that lays out resources for “leveraging LLM features” and boasts about Harvard’s record of embracing “new technology across our classrooms, residential and virtual.” Students and faculty might have good reason to use AI products in certain educational contexts. But the university generally leaves the policing of these “resources” up to individual faculty members, even as AI use has become widespread among students. A survey conducted this year by The Harvard Crimson found that the average student used AI to complete 34.5 percent of their homework, and that more than one in 10 students used AI to complete 70 to 100 percent of their work.

Harvard’s dean of undergraduate education, Amanda Claybaugh, told me that “we must teach students what they need to know about AI, while also ensuring that AI doesn't interfere with everything else.” She also said that concerns about closing the writing center are “largely a problem of terminology.” She went on, “Consolidating the writing center sounds alarming until you understand that the ‘writing center’ was merely a sub-unit of the Harvard Writing Program, which will continue to oversee our first-year writing courses.” The school still employs many experts on writing pedagogy, she said, and Harvard’s writing program will make sure that students receive “the best, most up-to-date, guidance about writing in an age of AI.”

Some in the Harvard community see the demise of the center as simply a regrettable but necessary concession to financial conditions. The school has a reported $365 million budget shortfall and has been making a number of changes to address it. One tenured professor in the humanities, who requested anonymity because of how controversial the decision has become, told me that the school has had to do “a couple of rounds of very painful cuts.” The professor said that he believes that eliminating the writing center could also help address a well-documented problem of administrative bloat. Harvard’s administrative staff has expanded at several times the rate of faculty in recent years. “Everyone thinks, in the abstract, there should be belt tightening,” the professor said. “But whenever a concrete person is let go, everyone screams.”

Still, the fact that some administrators should be laid off and some academic centers should be closed does not make it wise to lay off this administrator and to close this center. When an institution begins subjecting something so fundamental to a college education as writing to the cold logic of spreadsheets and bureaucratic streamlining, it has conceded quite a lot already. This concession may ultimately open the door for administrators of the future to argue that writing instruction should be turned over to robots for the sake of savings and efficiency. “Harvard must make difficult budgetary decisions sometimes,” Mary D. Lewis, a French-history professor, wrote in an email sent to senior administrators. “But the writing center was not an expensive enterprise, and it will likely cost us more down the road to undo the damage done.”

Some of that potential damage will be contained to Harvard. Instructors may have less support in their efforts to defend traditional forms of learning from technological change. Students may have less incentive to do the challenging work of original writing instead of turning to ChatGPT. But the consequences of the decision to close the writing center may be felt far beyond the campus in Cambridge. Last year, I traveled to Montana State University for a week as a writer-in-residence. There, I met with students and faculty to talk about journalism and creative nonfiction. The highlight of the trip was a Q&A session with the university’s writing-center tutors. I left with an ultramarine MSU Writing Center T-shirt and an overwhelming admiration for the students and staff who were doing their best, with limited resources and watery office coffee, to make a case for going to the center rather than using the bots.

[From the September 2026 issue: Why I quit the tenure track]

The tutors I spoke with at MSU said that they try to model writing as a creative, collaborative activity, one that happens between a scholar and a community rather than between an individual and his or her screen. They meet with students one-on-one to brainstorm ideas, read essays out loud, and critique their personal styles. When I heard the news about Harvard’s writing center, my first thought was not about the students and instructors there, but about the ones back in Montana.

College and university bureaucracies tend to have a strong mimetic streak. Institutions with lower academic rankings often emulate the initiatives and priorities of their higher-ranked peers. Harvard is the apex predator in America’s higher-ed ecosystem, and it has substantial power to influence what counts as essential to a good education. The decisions that Harvard makes create a permission structure for other institutions to follow suit—after all, if one of the most venerable schools in the country doesn’t need a writing center in order to be “elite,” why would a public university, or a regional liberal-arts college, or a community college need one? Why does Montana State need one? Administrators, boards of trustees, and state legislatures will almost certainly be asking these questions. Harvard has made it easier for them to decide that original writing—and, by extension, original thought—isn’t worth the cost.

Is Donald Trump over? Can we all now relax and move on to the next thing?

That’s the message from a significant portion of Democratic primary voters. Just as Trump is bored with Iran and the economy and wants to play with monuments and gilding, so too are many Democratic primary voters in Florida and elsewhere bored with Trump and eager to play factional politics instead.

The political statistician Nate Silver summed up the mood on X today: “It’s kind of been a rough set of primaries for Heroes of the Resistance 1.0.” Silver then listed stalwart anti-Trumpers who have lost their contests in 2026 and concluded, “People want to move on.”

America is a fast-moving, dynamic country. It’s tough for Americans to stay focused on a single thing, even if that single thing is the threat to American democracy. As the celebrity spokesmodel Heidi Klum tells the audience on Project Runway: “In fashion, one day you’re in, and the next day, you’re out.” This law also apparently applies to Democratic candidates who took serious risks to challenge Trump. The resistance, it seems, is out.

The far left of the Democratic Party has never been as energized as the party’s center about Trump and democracy. A 2017 analysis in The Washington Post suggested that between 6 and 12 percent of people who voted for Bernie Sanders in the 2016 Democratic primaries voted for Trump in the general election. With many careful caveats, the analysis observed that these voters exceeded Trump’s margin of victory.

[James Kirchick: The DSA is a parasite]

In 2024, likewise, some leaders of the activist left noisily defected from the Harris-Walz ticket. For many of them, opposing Israel’s war against Hamas in Gaza mattered more than stopping a Trump presidency at home. The effect of this mobilization on the left is hard to measure, but The Nation was quick to insist that the Biden administration’s Gaza policy likely cost the party Michigan, because many Arabs and Muslims swung toward Trump. The article did not seriously consider whether culture politics, such as Kamala Harris’s support for transgender rights, influenced socially conservative Arab and Muslim voters, even though the city of Dearborn has long been a locus of anti-gay grassroots politics in the state. The claim that Harris lost to Trump because of Gaza may well be activist self-medication, but this kind of story about the election is telling.

This year, the dulling of anti-Trump impetus on the left is even more visible and consequential. Again and again, primary voters have made choices that subordinate general-election concerns to factional power-seeking. Democratic leaders—the dreaded “Establishment”—have promoted highly plausible candidates for statewide office in North Carolina, Texas, and other suddenly competitive states. But the insurgent left has again and again upset leadership plans with high-risk choices to serve its own factional agenda, putting at risk once-probable Senate victories in Maine, Michigan, and some House seats too.

That agenda seems to argue as follows:

Donald Trump is not that important. His attacks on the rule of law, on free and peaceful elections, on U.S. alliances, on trade and markets—those are all mere symptoms of a disease, not the disease itself. Political energy should be focused on treating the deeper problems of which Trump is a mere passing expression, even at the risk of losing the Senate in 2026 and forfeiting the presidency in 2028. “Winning” with Chuck Schumer–approved candidates is no win at all. Better to risk losing than to “win” on Schumer’s terms.

[Jonathan Chait: There’s nothing democratic about these socialists]

The Democratic left is making a complicated—and quite ironic—internal gamble here. The faction’s efforts to hijack the larger party to their own unpopular causes will work only if they can convince Democratic voters that Trump is not the most important and urgent problem facing Americans. Yet the only reason Democratic centrists would coalesce around the party’s leftist candidates in November is if they believe in the absolute urgency of stopping Trump.

Many of the far-left candidates are aggressively obnoxious from a centrist’s point of view. Centrist Democrats are not socialists, not revolutionaries, and not disposed to blame Zionist conspiracies for every trouble in American life, including police brutality and urban gentrification. Why on earth would centrist Democrats in Michigan vote for a radical leftist who calls most Democrats hypocrites when the other choice on the ballot is a Republican campaigning on reducing housing costs by simplifying the government-approval process?

In Florida last night, Democrats rejected Alex Vindman, the former national-security official who at immense personal and career risk blew the whistle on Trump’s 2019 plot to blackmail Ukrainian President Volodymyr Zelensky into fabricating campaign dirt against Joe Biden. Voters chose instead Angie Nixon, who campaigned as an explicit “socialist” in a state that is home to hundreds of thousands of Cuban American and Venezuelan American refugees from socialism, and who ran alongside anti-Israel figures such as Oliver Larkin and Rashida Tlaib in a state with the nation’s third-biggest Jewish population.  

From a Democratic-centrist point of view, if Trump’s attacks on the rule of law are not a supreme and imperative threat, then a Republican Senate majority led by John Thune is an endurable nuisance. Far-left candidates have a chance in the general election only if Democrats believe that Trump is the supreme and imperative threat.

The remains of the suddenly unfashionable Democratic resistance can only hope that Democrats in Texas, North Carolina, Michigan, and elsewhere will rise above the factional politics and remember in November what, and who, remains the most pressing danger to the nation’s law and democracy.  

It is 12:04 p.m., and I’ve been watching AI slop for the past three hours. Before me on the screen, a man is playing trumpet in fog. He’s been fired (bad market for trumpet players) but then, at last, he lands an interview. Trumpet Man sits at a large desk across from an obese fellow who asks how good he is with a mop. (Smash cut.) Trumpet Man is now in a janitor uniform with a name tag reading V. Helsing. Oh, he’s a musician-janitor turned vampire slayer. My stomach grumbles in anticipation of lunch as the video feed shifts to a new scene: An animated vampire bellows, “I am what I was born to be: Dracula!” A dull ache forms around my temples.

This is Fairground AI Creator TV, a new channel on Roku’s free, ad-supported streaming television (FAST) service. The channel curates and plays AI-generated short films from creators all over the world 24 hours a day, seven days a week, and has limited commercial interruption. (Some of the advertisements are also AI-generated.) Roku has hundreds of these FAST channels—live news, marathon programming of Murder, She Wrote, et cetera—and some are quite popular: The flagship Roku Channel, which rotates between a smattering of movies and live TV, reached an estimated 145 million people in 2024, according to the company. But Fairground is the first of its kind, which, depending on your point of view, may be a glimpse at the future of unconstrained imagination in media or a dystopian nightmare.

[Read: Is this what comes after AI slop?]

There is no easy way to start watching Fairground. Many of the films are grouped according to a theme—sci-fi, documentary, fantasy—but much of the broadcast is contextless. A three-minute comedic short might be followed by a 20-minute drama. Because generative AI is evolving so quickly, the quality varies aggressively from film to film, depending on which models a creator used to make the video and when the film was made. One of the Lord of the Rings rip-offs that I watched during an eight-hour binge had the gauzy, uncanny look of early generative AI, and the characters’ mouths were wildly out of sync with their voices.

Another of Fairground’s offerings, a talk show called Paws for Thought, hosted by a dog and cat, legitimately made me question my sanity. (Episodes titles include: “What Is the Job of a Bumblebee?,” “How Do You Tell the Difference Between a Boy Turkey and a Girl Turkey?,” “What Does a Person’s Size Have to Do With Anything?,” and, of course, “Animal Cruelty—Recognizing the Signs and Addressing the Causes.”) At one point, during an episode titled “Who Was Mr. Rogers?,” an animated cat that has a halting, robotic text-to-speech voice offers a short history of the famous children’s-show host before—why not?—pivoting to a Zoom interview with its mother. “Your dad and I are very proud of you,” the mother cat says.

Although Fairground’s Best of AI collections invite the viewer to “come see what happens when imagination has no ceiling,” each video feels cut from the same cloth. Even in the most well-constructed videos—the ones that seem to have genuine scripts and plots you can follow—the vibes are immediately off. Generative AI, for all of its utility, is a homogenizing technology. Despite being trained on the glorious quirks of humankind’s creative canon, the models push our ersatz artistic output into a bland, muddled middle ground. The majority of content that I consumed on Fairground was aimless and forgettable—offensive in how persistently inoffensive even the strangest entries ultimately were. Watching these videos feels like speaking with another human through a translator or reading CliffsNotes instead of the prose of a novel—you get the gist, but something ineffable and decidedly human is lost in the process.

The longer I watched Fairground’s channel, the more I felt like I was trapped in some kind of contextless dream state. A man appears on a billboard in Times Square to deliver a message that the world is going to end in 48 hours. “I know. I’ve been calculating this for two years. Trust me,” he says, before adding that “sovereignty needs a planet to exist on.” Seconds later, that same man is engaged in a slow-motion martial-arts fight scene. Out of nowhere, a title card pops up that says Last Fight. With no explanation, a tidal wave as tall as a skyscraper envelops a city. Credits roll. Then the next film starts: A man who bears an alarming resemblance to the political writer Matt Yglesias wanders through a field of wildflowers.

As the hours ticked by, I noticed that the edges of real life started to fray ever so slightly. When my attention wandered while I watched yet another futuristic car glide through yet another synthetic environment, I tapped over to my social-media feeds. I came across a cute picture of a sea turtle that looked like it was waving at the camera. I smiled and then was suddenly paralyzed, realizing that I was unable to tell whether this turtle was real or AI-generated. All of a sudden, the usual buffet of influencers popping up in my Instagram algorithm seemed immediately suspect, perhaps synthetic. In just a few hours, my compass had been scrambled.

[Read: AI slop is coming for your playlists]

I kept wondering: Who is this for? When I put this question to Colin Petrie-Norris, Fairground’s founder and CEO, he suggested that the project is a work in progress, that the company, the channel, and the technology are all “in the first inning.” He offered up all of the usual bromides about generative AI. “This is as bad as it’ll ever get. It’s going to get better,” he told me. “It’s this democratization of creativity.” He was quick to offer up stories about AI creators he works with, including a child in Senegal who makes movies above his mom and dad’s fish-and-chips shop. Ultimately, Petrie-Norris wants the channel to “earn the attention of a family” sitting down to watch something together.

Petrie-Norris said that Fairground doesn’t blindly license videos, but vets them and works with creators to rewrite scripts and prompts. He mentioned a “kid in his basement in Berlin” whom Fairground has coached on creating better action sequences; the executive also suggested that the “Steven Spielberg of tomorrow” may not be discovered without this new technology.

To his credit, Petrie-Norris has been right before. In 2013, he founded Xumo, one of the earliest FAST streamers, which he later sold to Comcast in a deal reportedly worth more than $100 million. He told me that, in his recent meetings with television companies, there has been a genuine interest in embracing generative AI.

Yet still, who is this for? The best answer seems to be: TV executives. It’s not hard to imagine these folks salivating over endless, potentially customizable, or even interactive programming, all of it passable for background viewing and made without the messy overhead of human staffing. “The Home Slopping Network,” one of my colleagues dubbed it.

You could guess where all of this leads: a professionalized industry of guilds and craftspeople replaced by user-generated content and revenue shares, in which the only way to make a living is to flood the zone, a kind of cinematic version of what’s already happened on social media. “Is it good?” is an irrelevant question, one replaced by more pressing queries such as “Is it passable, modestly engaging, and infinitely replicable?” The early audience numbers are, apparently, promising for Fairground. Petrie-Norris claimed that just a week after launch, Fairground was already “seeing in the millions of people.”

Near the end of my binge, just after an advertisement featuring (the real, I think) Pamela Anderson talking about a new Earth Day channel she’s launching, Fairground ran an interview with one of its featured creators. For the first time in many hours, somebody spoke words that made sense. The purpose of AI, she said, “is not to make less, but to make more.”

Hundreds of thousands of books are published in the United States each year, and this dramatic influx of titles largely runs the calendars of the publishing and media industries—usually to the detriment of any work that isn’t brand new. Even best sellers or novels by famous authors get lost in the deluge, and books that were beloved on release can fall off readers’ radar quickly. But many were popular or critically acclaimed for good reasons, and they’re worth revisiting.

Here is a list of 15 fiction titles from the past two decades that you may have forgotten about in the years since. Some are from familiar names such as Kazuo Ishiguro, Margaret Atwood, and Louise Erdrich; others are by authors you may not have heard of at all. These selections include plenty of drama, and there’s an undercurrent of gentle comedy, even in novels with dark themes or plots. Their characters define love in many different ways, and they seek fulfillment across geographies and time periods—contemporary London, Vichy France, Nigeria, North Korea. Ultimately, these stories are bound together by a compassion for their characters’ struggles and shortcomings—a quality that only our finest writers are able to cultivate.

The Blind Assassin, by Margaret Atwood

Atwood’s tenth novel, a science-fiction story wrapped in a whodunit, throws noir tropes into a thoroughly modern setting and dares readers to keep up. It sometimes loses readers to her more accessible titles, such as the justly famous The Handmaid’s Tale or the imaginative MaddAddam trilogy, but The Blind Assassin stands out for its complicated structure and themes. The dual narrative follows Iris Griffen, angry and elderly, and her sister, Laura, whose book, also called The Blind Assassin, has made Laura into a kind of martyr after her death by suicide. Iris describes the siblings’ strange post–World War I upbringing: Their mother died in childbirth, their father slowly pickled himself in whiskey, and tutors came and went. Thanks to the large number of people obsessed with the book, Iris decides to tell her side of the story—and the sisters’ very different lives converge. The literary pyrotechnics Atwood employs are great fun to read, as she asks: What do we owe our families, past and present?

Brick Lane, by Monica Ali

Ali’s debut novel feels Dickensian, thanks to a cast of various lively characters. It’s set in a Bangladeshi corner of London, but it begins in Bangladesh, when Nazneen, whose younger sister, Hasina, has fallen in love and eloped, bends to her parents’ will and marries an older man named Chanu. After he moves her to England and she has three children, Nazneen surprises everyone (herself included) by developing a new independence and starting to navigate the city on her own. She begins an affair with a younger Muslim man named Karim, which provides her with romance and sexual fulfillment, but offers no easy escape from her daily obligations. Hasina’s fate is no easier—her promising match deteriorates, and so do her job prospects; both sisters must forge strong bonds with the women in their lives to find hope for their futures.

[Read: Housegirl complicates the diaspora narrative]

The Known World, by Edward P. Jones

The enslaver in this dense, astonishing consideration of the rot at America’s core is a Black man, Henry Townsend, who was once enslaved himself. That’s meant to be shocking, but it’s also historically accurate; a number of Black slaveowners did exist in the pre–Civil War United States. Townsend lies on his deathbed at the novel’s beginning, his wife, Caldonia, by his side. The book’s agile, omniscient narration, structured as a series of interconnected stories, leaps back and forth in time as frequently as it jumps between characters. We learn about Henry’s father, Augustus, who disapproved of his son’s decision to keep people in bondage, and about Caldonia’s twin brother, Calvin, and his unrequited love for another man. Occasionally Jones inserts invented references to fictional documents, which unmoors readers and makes them question the history they’ve been taught. Jones’s story painfully illustrates how insidious the evil of slavery was for all in its path.

Never Let Me Go, by Kazuo Ishiguro

The English residential school Hailsham first appears simultaneously quaint and menacing. Its students, including the book’s narrator, Kathy, engage in wholesome activities but are slowly introduced to their destiny: They are clones who will eventually have their organs harvested. Ishiguro isn’t writing pure science fiction, though; instead, he’s pursuing more elegant, metaphysical ideas. Do any of us have control over our own life? What do we do with the time we’re allotted? As Kathy grows up with, then takes a job caring for, her friends Tommy and Ruth, she knows that their bodies will be used until they die. Ishiguro writes a delicate dance between life’s moments of sheer joy—making art, learning to cook, falling in love—and their inescapable end. Tommy’s and Ruth’s coming of age is complicated by the state-sanctioned nature of the organ donation and the disturbed ethics of their existence, which destabilize our notions of what it means to be human. Ishiguro’s The Remains of the Day is better known, but Never Let Me Go might be his most tender work.

Suite Française, by Irène Némirovsky

The noted Ukrainian-French Jewish author Irène Némirovsky was arrested and deported to Auschwitz in 1942, where she died later that year. Her ambitious work Suite Française was planned as five novels, but only two exist; they were both published for the first time in 2004. (A third part was outlined but never completed.) They’re a master class in comic writing about small-town prejudice in wartime, filled with characters who bumble their way through important decisions. Storm in June, the first book, follows the invasion of France and its surrender to German forces through the eyes of two families, one upper-class, one middle-class. The second novel, Dolce, examines life in a village in Vichy France, where a woman whose house is commandeered by German officers must make a difficult choice between love and honor. Composed in the middle of the events it describes, Suite Française offers a surprising and incisive insider’s take on how the Nazis convinced so many French citizens to accept German rule.

Special Topics in Calamity Physics, by Marisha Pessl

The especially precocious 16-year-old Blue van Meer enters her senior year at a new high school in North Carolina after moving there with her academic father, Gareth. The murder mystery she gets tangled in when she meets a group called the Bluebloods might remind readers of Donna Tartt’s work, but Pessl’s debut novel is compelling less because of its plot and more because of its unusual execution. Blue tells her story in a combination of arch references, allusions, and imperfect, sometimes clunky word collages, and she’s always telling readers that being a teen can get very old, very fast. Yet when the group starts investigating, the action gets very fast too, and we see that Pessl’s setup was a slow burn in service of a terrific denouement.

Cutting for Stone, by Abraham Verghese

The twins Marion and Shiva Stone were once conjoined, so the effects of a major surgery are present in their very bones. As the brothers grow into adulthood and their own careers as surgeons, Verghese closely examines the bodies they cut open. That level of detail is perhaps not for the squeamish. However, it’s used to clue us in to the relationship between a surgeon’s precision and a body’s mess, and also references political upheaval and regime change in Ethiopia, where the brothers grow up, over decades. Verghese tries to do it all in his initial work of fiction (he wrote two fine memoirs first): He dips into history, tragedy, comedy, drama, reference, and documentary. If he sometimes misses the mark, it’s hard to fault him, especially because the story is tied together by a romance made as neatly as the finest sutures.

Say You’re One of Them, by Uwem Akpan

Each of the five stories in this wrenching collection is told from the perspective of a different child in a different African country. Akpan, a Jesuit priest, employs his faith’s purest anger to remove the curtain between privileged readers and impoverished characters. A girl watches her Hutu father join a murderous mob seeking Tutsi people like her mother. Another set of children obey their mom, who urges them to sniff glue to help them forget their extreme hunger. Akpan invokes an unmistakable fairy-tale violence, most visible in the “Hansel and Gretel”–like “Fattening for Gabon,” in which siblings sent to an uncle for safety have no idea of his cruel plans. Though Say You’re One of Them won prizes and was an Oprah’s Book Club pick, it’s been overlooked since, likely because of the difficult and painful material—both of which are its author’s deliberate choices—and not because of the writing or storytelling, which are superb.

Major Pettigrew’s Last Stand, by Helen Simonson

Simonson’s book is worthy of a lead spot in any list of comedies of manners. When the staid, retired English Major Pettigrew starts a friendship with the local shopkeeper Mrs. Ali, who is Pakistani, he begins to realize for the first time how racist some of the local characters and conventions in his village are. As the Major and the Mrs. slowly fall in love, both contend with people who would prefer that the couple go their separate ways. Simonson uses well-worn archetypes (golf-club toadies, event-planning matrons) to great effect, showing how unnecessary those backward roles are in the modern world. A subplot about an heirloom shotgun feeds nicely into the book’s climax, and the prose has a controlled, even Austenian, sense of humor that illuminates everyone’s misplaced expectations.

[Read: Making peace with Jane Austen’s marriage plots]

The Orphan Master’s Son, by Adam Johnson

In a fictional but detailed version of North Korea, being a citizen of the state takes precedence over any personal identity. Johnson’s protagonist, Jun Do (think “John Doe,” as you’re meant to), starts life in great deprivation in an orphanage. He undergoes a series of strange adventures that includes a stint in the United States and capture, imprisonment, and torture by his government. Jun Do winds up impersonating an important military commander, whose actor wife Jun Do loves beyond reason. He grows into a complex character whom you root for, even when he suffers at the hands of one of the novel’s three narrators, an interrogator fast losing faith in his superiors. Although The Orphan Master’s Son won the Pulitzer Prize for Fiction, it deserves renewed attention 10 years on, at a time when agency, kindness, and love are more important to protect than ever.

Beautiful Ruins, by Jess Walter

Every Jess Walter novel is an immersive experience, but Beautiful Ruins literally opens with a watery submersion off the coast of Genoa, Italy, in 1962. A starlet named Dee Moray arrives there straight from the set of Cleopatra (yes, the one with Richard Burton and Elizabeth Taylor). Sparks fly between Dee and a local pensione proprietor named Pasquale, who works at the “Hotel Adequate View.” But just as you settle in for an Italian beachside romance, Walter lobs the story 50 years into the future, when a Hollywood production office is having its “Wild Pitch Friday.” For reasons that will eventually be made clear, Pasquale sits and waits to offer up his story, and Walter launches into a series of vignettes that demonstrate the many ways people rummage through their lives to make art, whether through movies, memoir, or novels. Beautiful Ruins is brash, sly, and delightful.

The Round House, by Louise Erdrich

Some critics prefer other Erdrich novels, including Love Medicine or The Night Watchman. But The Round House is among her best, and should be read for the voice of its narrator, Joe Coutts, an Ojibwe lawyer whose tale begins in 1988, when he’s 13. That summer, he and his father separately, and dangerously, pursue justice for his mother’s harrowing rape. Erdrich’s authorial genius is evident in the way she gives voice to a personal experience of deep pain and connects it to the ongoing violence of settler colonialism that crosses generations, while allowing the younger Joe an adolescent ease with his friends.

A Tale for the Time Being, by Ruth Ozeki

Ozeki, an award-winning author and a Zen Buddhist priest, considers how writing and mortality intertwine in A Tale for the Time Being. The story is full of wonderful specificity, such as how one of its narrators, the teenage Nao, keeps her diary and other papers in a Hello Kitty lunchbox. The other narrator, also called Ruth, has moved to Canada’s Vancouver Island in search of a simpler life, but when Nao’s lunchbox washes up on the beach, Ruth finds her existence changed by its contents. The novel explores depression and introduces, among other things, a subplot about a kamikaze pilot, but it never feels despairing, and Ozeki doesn’t burden readers with high-flying concepts. The “time being” of the title refers both to the fleeting nature of the present moment and to humans themselves, whose lives, Nao writes, are governed by the flow of time. The wordplay serves as a reminder that all stories and all people have their places, but also their endings.

Under the Udala Trees, by Chinelo Okparanta

Ijeoma, the narrator and protagonist of Okparanta’s debut novel, grows up during the Nigerian civil war in the 1960s. She is sent away from Biafra to family friends for safety, where she first falls in love with another girl, Amina. Even after their relationship is discovered and the two are separated, Ijeoma has a romance with another woman, Ndidi, until she bows to maternal pressure and marries a man. Although the backdrop to much of Ijeoma’s story is war and famine, the action can be rich and sweet, as she and her lovers discover their true selves. Okparanta’s no-nonsense prose and focused narration help make Ijeoma’s voice stand out in a book as gentle as it is real.

[Read: A novel that weighs the costs of love and motherhood]

Sea Monsters, by Chloe Aridjis

Luisa appears to us as a teenager in the late ’80s, roaming a beach town called Zipolite in Oaxaca, where she’s arrived with her brooding goth boyfriend, Tomás, in search of a group of runaway circus performers. Aridjis won the 2020 PEN/Faulkner Award for Fiction with this atmospheric and whimsical coming-of-age story, but it was published by a small press and should find more readers. Nothing much happens in Luisa’s tale—except experience, which may be one of the most challenging things to capture on the page. Luisa adores The Smiths and Joy Division and wants to carve a space for herself in a life that has so far been dictated by her scholarly parents and their values. If Luisa finds the feelings and places she’s longing for, what will happen to her? Will she be better off? Satisfied? Or will she not yet have the maturity she needs to appreciate what she wants? Aridjis’s novel asks these questions and more, but smartly, it doesn’t worry about the answers.

This article was featured in One Story to Read Today, a newsletter in which our editors recommend a single must-read from The Atlantic, Monday through Friday. Sign up for it here.    

Don’t try to tell this to a mother sitting in the bleachers during a four-hour swim meet; or enduring a birthday party involving toddlers and craft projects; or resting in an armchair on a peaceful evening, savoring the heft of a tiny body and the scent of an infant’s freshly washed hair. Interminable or sweetly languid though they may feel in the moment, the childbearing years are startlingly brief. Fertility, which typically ends in a woman’s mid-40s, occupies less than half of her adult life. And then, if she’s lucky, she has 30 or 40 years in which to do something else.

Most people don’t realize how unusual humans are, in the way that nonreproductive females (how shall I put this?) persist. Females of most other species can bear young until they die, and many do, or at best enjoy a brief respite from breeding before death. This is true not only of creatures you might expect, such as rabbits, but also of long-lived mammals such as Asian elephants, and of primates such as gorillas and chimps. The odd exceptions—the Japanese aphid, for example, enters a “glue bomb” stage after her reproductive phase, ready to immobilize a colony intruder—only prove the rule.

The mystery of why women go on and on and on after their procreative function has ceased has occupied some of the great minds of the ages. I am sorry to report that many of those minds have not been forward-thinking. “It is a well-known fact … that after women have lost their genital function their character often undergoes a peculiar alteration” and they become “quarrelsome, vexatious and overbearing,” Sigmund Freud pronounced. The male-dominated medical community of the mid-20th century was similarly dismissive. “The unpalatable truth must be faced that all postmenopausal women are castrates,” opined the gynecologist Robert Wilson, who elaborated on this theme in his 1966 best seller, Feminine Forever. The influential book, it later emerged, was backed by a pharmaceutical company eager to market hormone-replacement therapy.

Even the architects of the sexual revolution were fixated on fertility as a marker of femininity, an attitude that seems doubly unfair coming from the people who gave us the pill. “Once the ovaries stop, the very essence of being a woman stops,” wrote the psychiatrist David Reuben in 1969 in Everything You Always Wanted to Know About Sex but Were Afraid to Ask, adding that the postmenopausal woman comes “as close as she can to being a man.” Or rather, “not really a man but no longer a functional woman.”

Little wonder that women writers have felt the need to weigh in over the centuries. A few took an upbeat approach. At the age of 41, having just given birth to her sixth child, the suffragist Elizabeth Cady Stanton wrote her friend Susan B. Anthony in 1857 to say that their best activist years lay ahead. “We shall not be in our prime before fifty & after that we shall be good for twenty years at least.” Others were less sanguine. At 54, writing her memoir, Simone de Beauvoir gloomily prepared to say “goodbye to all those things I once enjoyed”; women, Freud had taught her, become miserable and sexless as they age. Betty Friedan, Gloria Steinem, Helen Gurley Brown, Germaine Greer—all warily chronicled their maturity, as of course did the writer who invented the concept of “passage,” Gail Sheehy. Nora Ephron felt bad about her neck, and her anxiety spawned a best seller.

Even now it’s hard for a woman not to dread the consequences of moving out of youth. One of the wryest recent meditations is an episode of Inside Amy Schumer, in which the eponymous comedian happens upon three of her comedic icons—Tina Fey, Patricia Arquette, and Julia Louis-Dreyfus—picnicking in a meadow. They are celebrating Louis-Dreyfus’s “last fuckable day,” as adjudicated by the media, Fey explains. Schumer, feigning astonishment, asks whether the media do this to men. The trio laughs and laughs.

Three new books about postmenopausal womanhood show that the conversation is changing. For the first time, The New York Times noted early this year, a sizable cohort of women is moving into the sixth and seventh decades of life with a surfeit of energy and workplace experience. Women are better educated than men. Many spend early middle age constrained by work-life challenges, like athletes training with ankle weights. Once the weights come off, they have the muscle to run. Literally: The 2020 slate of female presidential candidates is Exhibit A.

The landscape looks different due to the #MeToo movement as well. In some ways, it has divided women by generation, yet even older women who may regret a return to the idea of feminine fragility are overjoyed to see workplace predators toppled. The unseating of men like Charlie Rose and Matt Lauer has opened the way for women like Christiane Amanpour and Gayle King to occupy top spots, where they exemplify what 60‑something really looks like: pretty freaking great.

The current conversation is also informed by evolutionary biology, which evaluates traits based on their reproductive purpose. Given that menopause is nonreproductive by definition, biologists consider it a “big evolutionary puzzle,” the novelist Darcey Steinke writes in her memoir, Flash Count Diary: Menopause and the Vindication of Natural Life. According to the prevailing view, a human female possesses all the eggs she will have while still in the womb; the number promptly begins diminishing, and by her mid-40s, the remaining ova have deteriorated. To an evolutionary biologist, this is interesting and weird. To Steinke, it was miserable and hard. Her book is lyrical but a bit depressing, because she herself was depressed.

Some women experience few symptoms during menopause, but Steinke suffered nearly two awful years of hot flashes, acute episodes that were like “four-minute surprise anxiety attacks.” She sensed mortality stalking her: “For the first time, I feel I have a time stamp, an expiration date.” She writes vividly and a little wistfully about sex, mourning her lost desirability, as she sees it, and the waning of her own desire. She feels angry; she yells at her husband. “Early times of sexual frenzy seem almost impossible now.”

Every woman is of course entitled to—can’t escape—her own response to menopause. But Steinke’s melancholy reflections sound a bit retrograde, as if she can’t escape those insufferable doctors, the Wilsons and the Reubens, with their pompous pronouncements about the wreckage that remains when estrogen, like a tide, drains away. “Without hormones my femininity is fraying,” she writes. In a transitional state herself, she identifies with people who are transitioning out of their birth gender—not that the empathy brings much relief.

Steinke also identifies with one of the few other species that enjoy a long postmenopausal life: killer whales. In the ocean, nonreproductive females play an important role. With the wisdom of years, they guide their pod to the best salmon. Steinke kayaks in waters off the coast of Seattle, hoping to commune, and is rewarded with a magnificent breaching. “The wild matriarchs have given me hope,” she writes. “They are neither frail nor apprehensive, but in every way leaders of their communities.”

That menopause may enable a new role and stature for women is the central argument of The Slow Moon Climbs: The Science, History, and Meaning of Menopause, by Susan Mattern. A historian at the University of Georgia, she steps away from the personal to consider “humanity’s massive primeval past.” Once upon a time, scientists assumed that women (and men) were designed to live to about 50, and that menopause was an accident, a by-product of medical progress. Yet even in primitive societies, it turns out, a portion of women lived well past middle age, which suggests that menopause is a feature, not a bug, of human evolution.

Mattern has her own audacious theory as to why: Menopause is a key to our success as a species. In humanity’s hunter-gatherer days, tribes needed a balance of producers and consumers—people who brought in food, and people who ate it. Most adults did both. Not so children, who remain dependent during the long period of brain development. Members who could bring in food for more than one person without adding to the population were crucial.

Enter the postmenopausal female. The anthropologist Kristen Hawkes studied a modern foraging tribe, the Hadza, and found that an energetic group of older women brought “more food into camp than any other age and sex category.” This paved the way for the Grandmother Hypothesis: Not only do older women serve as food producers, but they are providers of “allocare,” communal child care. In the Hadza and other tribes, Mattern writes, women “reach peak foraging productivity in their 50s and continue to produce a caloric surplus through old age.” She points out that tribes have been known to kill members who can’t contribute. If grandmothers aren’t murdered, she reasons, that is because they are useful.

Mattern makes the case that menopause probably emerged in humans when we diverged from chimpanzees millions of years ago. It gave Homo sapiens an advantage over other species of hominids such as Neanderthals and Denisovans, she proposes. Limiting childbearing to younger women, whose offspring could be cared for by older women, enabled the species to bounce back from an epidemic or a crisis: Those fertile women could reproduce quickly, but no woman could do so forever, sparing the tribe the risk of overpopulation. With the advent of farming, menopause still served an important purpose. The most prosperous time for a peasant family was postmenopausal, Mattern argues, when older children could help and the family no longer had new members to support. Nowadays, with fewer children and more resources, she brightly adds, “women past menopause, who historically used their energy surplus to help their families survive, can now use it in other ways.” While Steinke experienced menopause as a shutting-down, Mattern sees it as an opening-up.

Both views are true, as the New York Times columnist Gail Collins shows in No Stopping Us Now: The Adventures of Older Women in American History, which spans the colonial era to the present. Her takeaway is that older women fare well when circumstances permit them to be productive: “If you’re important economically, you’re important.” Providing allocare is well and good, but “eras in which older women were able to earn money or increase their family assets were eras in which they were … popular,” Collins writes drily.

Tracing their shifting status, she observes that during the colonial period, women continued spinning, weaving, and the like into old age. Physically, of course, the stage could be hell—pelvic disorders, childbirth damage, rotting teeth. Older enslaved women were exiled and horribly neglected. Yet often, the only thing worse than being a woman was being a man: Male mortality was higher, and widowhood could be a blessing—widows, at least white ones, could own property (unlike married women). Then, as the country became more populated, men monopolized the jobs. Women had less of an economic purpose, and they lost status.

Collins highlights the great age of social reform in the mid-19th century as another period when older women enjoyed prestige, though less because they had economic power than because they wielded moral authority. Notably, they used that authority to make the case for, among other things, female clout well into later life. A standout among women’s-rights advocates, Elizabeth Cady Stanton argued that women could occupy different spheres at different life stages, moving from narrow domestic concerns to community-minded platforms. Stanton, Collins writes, “believed that menopause had redirected all her ‘vital forces’ from her reproductive organs to her brain.” Vital she and Susan B. Anthony certainly were as they barnstormed the country, made speeches on tabletops, played cards with soldiers. Their age allowed them to “have adventures.” At the same time, one observer wrote, “stately Mrs. Stanton has secured much immunity by a comfortable look of motherliness.”

As the white-collar workplace expanded in the industrial era, women were shouldered out of it, eroding female social power across the age range. Older black women were a mainstay of early civil-rights struggles, but the contributions of activists such as Mary Church Terrell and Mary McLeod Bethune were sidelined, the credit given to younger men. And women’s return to the workforce during World War II gave way to postwar pressure to depart it, which delivered mixed results for older women. In the 1960s and ’70s, important strides were made thanks to female workers such as flight attendants, who filed class-action lawsuits protesting rules that obliged them to retire if they married or reached age 35. Yet as women’s earning potential grew along with their work span and sexual freedom, the more senior among them faced newly corrosive pressures.

It may not be a coincidence that Wilson, Reuben, and their ilk pushed the perils of “the menopause” during this era. The centrality of sexual liberty to the women’s movement arguably left second-wave feminists more vulnerable to insecurity about their bodies and looks. Susan Mattern proposes that the very concept of a menopausal syndrome was the invention of a culture that aimed to psychologically weaken women in a strong period of life—at a historical moment when female power was rising. “Dominant groups,” she observes, “can be very creative in inventing new ways of oppressing people.”

Yet I’m struck, reading these accounts, that Stanton intuited what remains true today: Women have a different life trajectory than men, and the place of menopause in it is liberating in a way that’s worth considering. To describe a passage of life, even a painful one, can itself be a form of empowerment. Men, too, feel loss and insecurity as they age, and perhaps could use a map of sorts themselves. Crossing the midlife point, many struggle to recalibrate professional ambition (as Arthur Brooks’s article in The Atlantic’s July issue revealed) and to build stronger social and intimate ties. Online-dating sites betray men’s own anxiety about physical decline (“My friends say I look much younger than 60!”) and suggest that many men are perfectly willing to date women across the age range. The sex books are right: Men aren’t as picky about women’s bodies as women fear. Men crave sex, but they also crave conversation, a partner with confidence and achievements. Even Simone de Beauvoir changed her mind. In The Coming of Age, a book about the experience of getting older, she wrote that she had crossed a “frontier” and found peace. She had also taken a younger lover. “It has been far less sombre,” she reported, “than I had foreseen.”

Sign up for Work in Progress, a newsletter where Rogé Karma investigates the mysteries of a complicated economy.

The driverless future is here—but, to borrow from William Gibson, it’s not evenly distributed. Waymo’s self-driving robotaxis are available in cities in California, Arizona, Florida, Georgia, Tennessee, and Texas. Aside from California, where the technology originated, those are all red or purple states. In every other blue state, meanwhile, driverless taxis are still not allowed to operate. If you live in Miami or Dallas, you can summon a Waymo to get where you’re going. If you live in Boston or New York, you can’t.

This is not an accident. In Democratic jurisdictions, a coalition of labor unions, anti-tech activists, and politicians have kept driverless vehicles off the streets. (Colorado, which passed a bill authorizing driverless vehicles in 2017, before the current backlash, is on track to become the exception later this year.) They argue that the new technology would make the roads more dangerous and throw millions of human drivers out of work, all to enrich a handful of tech oligarchs.

A version of this fight is now playing out in Washington, D.C., where Kamala Harris won more than 90 percent of the vote in 2024 and where local officials recently introduced a bill that would provide a pathway for robotaxis to begin operating. Some residents have welcomed that prospect. But others are digging in their heels. Bill Davis, the president of the local Teamsters union, declared in a statement that the legislation “doesn’t benefit anyone outside of Silicon Valley billionaires” and “comes at the expense of middle-class professional drivers and anyone else who wants safer streets in our city.”

Driverless cars will indeed cost some workers their jobs. But unlike so many other recent Silicon Valley innovations, they have already demonstrated the potential to bring immense societal benefits. About 40,000 Americans die in car accidents every year. The number of crash-related injuries is much larger. As many as 10,000 crash victims enter emergency rooms around the country every day. Waymos are so much safer than human drivers that deploying them widely could avert countless injuries and deaths every year.

[Saahil Desai: Move fast and break nothing]

They will also make transportation more affordable: A taxi without a driver is much cheaper. Waymo is currently more expensive than Uber and Lyft, but that’s largely because supply is artificially constrained and demand is high. The price differential is dropping fast and will go away soon. Waymos also show promise in reducing carbon emissions and air pollution (the cars are electric), and providing reliable access to people who live in poor neighborhoods (who are routinely turned down by rideshare drivers). Saying no to Waymo is saying no to a technology that will change the world for the better.

There is a name for a political movement that taps into people’s resistance to change and defense of the status quo. William F. Buckley famously described a conservative as “someone who stands athwart history, yelling Stop.” Progressivism was founded on the opposite ethic: that scientific and technological progress can bring about a better, more equitable future, provided that it’s steered in the right direction. In this sense, self-driving cars are a test for American progressives—a test of whether they have become so captured by interest groups, so hostile to corporations, and so fearful of technological change that they’ve transformed into enemies of progress itself.

In much of the country, the popular conception of self-driving cars is something between a bumbling idiot and a mechanical death trap. Videos of Waymo vehicles making ridiculous errors go viral online. Earlier this year, a Waymo hit a child who ran into the road near an elementary school in Santa Monica, California, causing minor injuries. When a jurisdiction considers whether to allow Waymos, stories like these come up again and again. “I’m also deeply concerned about the safety issues autonomous vehicles have already had in other cities,” Helen Raynor-Hughes, a longtime D.C.-area resident, testified during a public hearing in July, citing “reports of crashes, traffic disruptions, and situations where these vehicles failed to respond properly to pedestrians, cyclists, and other drivers.”

But anecdotes can mislead. The question is how robotaxis’ error rate stacks up against that of human drivers. In a recent study, researchers at the Insurance Institute for Highway Safety compared the crash and accident data on a per-mile basis for Waymo vehicles with those of human drivers in the cities where Waymos operated from 2021 to 2024. They found that Waymos were involved in 68 percent fewer crashes serious enough to be reported to police and 81 percent fewer crashes that caused an injury. These findings are “likely conservative” because human drivers regularly fail to report accidents, Eric Teoh, the report’s lead author, told me.

Several other studies have produced even more favorable results. These studies are co-authored by Waymo researchers but have been peer-reviewed by independent academics. Every expert I spoke with for this story, including Waymo skeptics, told me that they consider the research to be high-quality. “Reductions this big are basically unheard of when it comes to traffic safety,” Noah Goodall, a senior research scientist at the Virginia Transportation Research Council, told me. “Not even seat belts or airbags achieved the kinds of numbers we’re seeing here.” (Waymo, which is owned by Alphabet, is the only company that has demonstrated such a strong safety record. The others, including Tesla, have not disclosed the data needed to make safety claims.)

Even these figures probably underestimate how safe Waymos are. The journalists Timothy B. Lee and Kai Williams recently reviewed the details of the 78 accidents that Waymo vehicles had been involved in over a seven-month period and that were serious enough to cause an injury or trigger an airbag deployment. They found just one example where Waymo “could plausibly bear some responsibility” for a taxi crashing into another vehicle or a pedestrian—the Santa Monica–school incident—and another three where culpability was “mixed or unclear.” “We went into this looking for any evidence that Waymo could even remotely be at fault for a given crash,” Lee told me. “But those cases were just really hard to find.”

The big caveat to Waymo’s safety record is its difficulty with so-called edge cases—unpredictable situations for which its algorithm is unprepared and in which human judgment really comes in handy. When a power outage knocked out a third of San Francisco’s traffic lights last December, at least 60 Waymos abruptly idled mid-road, causing severe gridlock and delays. Something similar happened to a few dozen Waymos this year when the driverless vehicles encountered unexpected road closures during San Francisco’s Fourth of July celebrations. Waymos have obstructed emergency vehicles in several cities, typically because of their tendency to freeze when they encounter alarms, sirens, or temporary road closures. The company has not done nearly enough to fix the problem. “Competent driving is about a lot more than crash statistics,” Philip Koopman, an emeritus professor of engineering at Carnegie Mellon, told me. “If Waymo isn’t able to address some of these problematic patterns, it is going to be very hard to convince the public they are safe.”

Even so, the fact that Waymos are considerably safer than human drivers is undeniable. A robot driver can’t drink and drive, fall asleep at the wheel, or check its phone. It won’t flout the rules of the road when it’s running late or trying to impress its friends. And it is equipped with more than a dozen different cameras, plus radar and laser sensors, meaning it doesn’t have blind spots. Bryant Walker Smith, an associate law professor at the University of South Carolina, has advised governments on how to regulate self-driving cars. “What I always tell policy makers is that we absolutely should be concerned about automated driving,” he told me. “But we should be terrified of human driving.”

Safety might be the average American’s top concern about driverless cars, but the average American isn’t spearheading the political resistance to Waymo. Labor unions are. The campaign against driverless vehicles has been led by a coalition of unions representing taxi, rideshare, and truck drivers who argue that self-driving cars will destroy millions of working-class jobs. (Although experts believe that self-driving trucks and delivery vehicles are at least several years away from being a regular presence on roads, labor unions see robotaxi legalization as a stepping stone to a fully automated driving future.) Indeed, Helen Raynor-Hughes, the area resident who testified about Waymo’s safety record, began by identifying herself as a Service Employees International Union member who drives for Uber and Lyft.

In 2026 alone, this Labor United Against Waymo coalition has helped defeat or delay legislation to authorize driverless vehicles in Illinois, Minnesota, Maryland, Virginia, Oregon, and Washington State. In January, New York Governor Kathy Hochul declared that she would pursue a bill to allow robotaxis to operate in the state, only to backtrack a few weeks later following an outcry from labor leaders.

When I arrived at D.C.’s city hall in July for a public hearing on the proposed bill to authorize Waymo, I passed dozens of protesters from various labor groups chanting and holding signs with slogans such as WAYMO? HELL NO and AI DOESN’T VOTE. We Do. Inside, about a dozen labor leaders and rideshare drivers demanded that city officials vote against the bill. Crystal Middleton, a D.C. resident who drives Uber and Lyft part-time to support her family, testified that if Waymo were adopted in the city, “many D.C. families, including mine, will struggle to make ends meet.”

I’m sympathetic to that argument. When I was growing up, my dad spent years driving Uber and Lyft on the weekends to help pay the bills. Like many drivers, he resented the meager pay and lack of benefits. (There’s an irony that unions and Democratic lawmakers are rallying to defend the type of precarious gig work that they have spent more than a decade criticizing.) But it’s hard to imagine what our family would have done without the extra income. The impulse to protect workers like him—not to mention the delivery and truck drivers who fear that they’ll be next—from being displaced by a new technology is understandable.  

That approach, however, ignores the larger number of people who would benefit from a change to the status quo. At the public hearing, local nonprofit leaders and politicians, including Shane Simenstad, the director of public policy for Safe Kids Worldwide, pointed out that rideshare drivers often refuse to enter less desirable parts of the city, leaving residents of poor neighborhoods without a reliable way to get to work or school. Claire Stanley, the director of advocacy and governmental affairs for the American Council of the Blind, spoke about being denied Uber and Lyft rides because of her service dog. Multiple women testified about their fear of being harassed or assaulted by rideshare drivers. (According to a 2025 New York Times investigation, Uber received a report of sexual assault or misconduct nearly every eight minutes on average in the U.S. from 2017 to 2022.)  

Some of the most moving testimonies came from the victims of car crashes and their loved ones. Rameez Tase, a local resident and avid cyclist, told the story of a friend who had been nearly killed after being hit by a car. “Anyone who knows me well knows that my biggest fear of death is being struck and killed by a distracted driver—and leaving my three young children without a father for the rest of their lives,” he said. Cesar Toledo, a member of the D.C. Commission on Poverty, recounted a childhood “marked by the terrifying reality of riding in vehicles driven by family members under the influence.”

Car accidents kill roughly as many Americans each year as breast cancer does, and twice the number of people who die from homicides. It is a leading cause of death among children and young adults. “The number of traffic deaths in America is the equivalent of a Boeing 737 crashing every single day,” Jonathan Slotkin, a neurosurgeon who regularly treats crash victims, told me. “That’s what we’ve chosen to tolerate as the price of mobility.”

Waymos, to be clear, wouldn’t solve this problem right away. Although rideshare and taxi drivers get into accidents at higher rates than the average driver, they make up a small fraction of cars on the road. But operating robotaxis is a key step for Waymo and other companies to perfect their self-driving technology, which could then be sold to vehicle manufacturers like Toyota to include as an optional feature in every car. “If we had a drug that showed these kinds of results, we’d be doing everything we could to get it out to as many people as fast as possible,” Slotkin said.

[Read: The huge problem Waymo didn’t see coming]

Widespread adoption of driverless vehicles would further a litany of goals that anyone should celebrate, especially those who claim to be progressive: slashing traffic injuries and deaths; helping the poor and disabled secure stable employment; making cities safer places to bike and walk; reducing the prevelance of sexual assault and misconduct; putting more electric vehicles on the roads. And yet progressives are the ones standing in the way. “Honestly, it’s just embarrassing,” Colin Mortimer, the director of the Center for New Liberalism, a center-left advocacy group, told me. “Liberals are always talking about how we’re the party that believes in science, that believes in data, that believes in helping the vulnerable. But on this issue we’re doing the exact opposite.”

Some Washington, D.C., legislators believe that the nation’s capital can pioneer a different approach to Waymo. In May, the D.C. Council member Charles Allen introduced a bill that would authorize driverless vehicles to begin operating in the district, but only if their makers meet certain key conditions designed in part to address the edge-case problem and reduce the impact on the city’s labor force. These include completing stringent testing and safety requirements and submitting a plan for interacting with first responders. Even then, the total number of driverless vehicles would be capped at a few hundred during the initial rollout. Each vehicle would also be subject to a per-mile tax used to fund public transit, as well as workforce training and compensation for displaced rideshare and taxi drivers. “We have to be honest with ourselves: This technology is already here and is only going to become a bigger part of our lives going forward,” Allen told me a few weeks after the hearing. “And so we can either bury our heads in the sand and have this change happen to us. Or we can put guardrails in place to make sure the technology works for us and our communities.”

In Allen’s view, too many restrictions would risk throttling a technology that could otherwise bring huge benefits; too few would invite a high-profile incident that draws a public backlash. Getting the balance right will be difficult. My own view is that the bill should have more restrictive testing requirements around edge cases—for instance, simulating power outages or major emergencies—before driverless vehicles can be deployed. But it should also create a less bureaucratic process for increasing the number of vehicles once those requirements have been met. The bill also faces several political obstacles, including an incoming mayor who expressed skepticism about Waymo expansion on the campaign trail. But Allen is trying to do something that no city has done before: chart a middle course between letting the new technology rip and blocking it altogether.

One-hundred and fifty years ago, the United States was experiencing the advent of a brand-new transportation technology. Railroads reduced cross-country travel from months to days, turned the U.S. into an industrial powerhouse, and enabled the modern consumer economy. But they also displaced workers, such as canal operators and stagecoach drivers; led to skyrocketing inequality; and produced corrupt monopolies presided over by ruthless oligarchs. A new political movement formed in response. Its adherents passed legislation that protected workers, created novel regulatory agencies, broke up powerful monopolies, and developed safety standards. Neither techno-utopians nor luddites, they believed that technology could be a force for progress—but only if it was actively shaped and regulated by the public. They thus called themselves progressives. Their political descendants would do well to remember their example.

This article was featured in the One Story to Read Today newsletter. Sign up for it here.

Was there ever a time when people were not panicking about population? In 1798, the English economist Thomas Malthus wrote that “the power of population is so superior to the power in the earth to produce subsistence for man, that premature death must in some shape or other visit the human race.” Humanity, then just 1 billion strong, did not fall into the Malthusian trap. Thanks to the Industrial Revolution, it instead underwent a huge expansion in both size and living standards.

More than a century later, when the world population stood at more than 2 billion, the new concern was impending shrinkage. In 1938, Joseph Spengler, an American economist, confidently predicted that “within the next quarter-century true depopulation—a persistent long-run excess of deaths over births—will manifest itself in nearly all the countries of Europe.” By the time Spengler’s deadline had passed, the postwar Baby Boom was in full swing and population concerns had reverted to Malthusianism. In 1968, Paul Ehrlich, a population biologist at Stanford, argued in The Population Bomb that “the battle to feed all of humanity is over” and, regrettably, humanity had lost. In the next two decades, he warned, “hundreds of millions of people will starve to death in spite of any crash programs embarked upon now.” Instead, the world population went from 3.5 billion to more than 5 billion, living standards increased, and child mortality fell.

Today, as humanity numbers just over 8 billion, the concerns have reversed polarity once more. Right around the time of Ehrlich’s doomsday prediction, global fertility rates began a decline that has not stopped. In the United States, Europe, and even India, births are now below the replacement rate of 2.1 per woman. In 2024, South Korea reached a “total fertility rate”—a measure of how many children the average woman will have over her life—of just 0.75, which if held constant would mean that 100 Koreans should expect to have 14 grandchildren and only five great-grandchildren. (The country’s total fertility rate appears to have ticked up slightly since then.) Very few countries outside Africa are above replacement levels. On current trends, the global population may peak in the 2050s at 9 billion, at which point it would start shrinking—for the first time since the bubonic plague of the 14th century.

[Marc Novicoff: Are men to blame for the birth-rate collapse?]

Now the French worry about dénatalité; Hungarians talk about nemzethalál, national death; the Serbs call the baby bust bela kuga, “the white plague.” For prominent figures in the global New Right, depopulation is seen as a threat to civilization. Giorgia Meloni, Italy’s prime minister, has warned that without bold action, her country could be “destined to disappear.” Elon Musk regularly calls low birth rates the greatest threat to civilization. But this population panic may age just as poorly as the prior ones. Even if the fertility decline doesn’t reverse itself anytime soon, the consequences are unlikely to be apocalyptic.

All of these dire projections have relied on a simple mathematical construct: the exponential curve. When you extrapolate an exponential curve with a positive growth rate far out to the future, it ascends toward infinity. One with a negative growth rate descends toward zero. Ehrlich, for example, wrote that if the growth rates of the 1960s continued “for about 900 years, there would be some 60,000,000,000,000,000 people on the face of the earth.” (That’s 60 quadrillion.) Today, long-run predictions look equally dire in the other direction: Newspapers in South Korea warn about “extinction” by 2750; Japan has a running clock, a bit like America’s national-debt counter, ticking down the seconds until the country has only one child left—expected around August 2644.

The problem with these terrifying models is not the mathematics—a population that halves in size every generation will indeed eventually go extinct, just as one that doubles will eventually overrun the planet—but the assumptions. For such projections to be accurate, people would have to resolutely continue to have a certain number of children for centuries, up to the point of planetary exhaustion or human disappearance. They would have to stay the course even as this demographic decline transformed family structures, housing prices, labor markets, and the workings of the welfare state. Yet the experience of the past century, with its baby booms and busts, demonstrates just how dubious that assumption is.

Family sizes seem to change as the world does, in ways that are hard to predict. For the past half a century, in almost every country on Earth, people have been choosing to have fewer children, despite experiencing the best average living standards that our species has enjoyed so far. This synchronized decline confounds many of the baby bust’s potted explanations in American discourse, from both the left and the right. Liberals argue that people are having fewer children because, among other inequities, child care is insufficiently subsidized, parental leave is insufficiently generous, and men do too little housework. But that can’t explain the serious decline in fertility in progressive utopias such as Finland, Norway, and Sweden. Social conservatives blame decaying norms around marriage and family values. Yet even in countries with strong taboos against out-of-wedlock birth and with strong heterosexual-marriage norms, such as India and Turkey, birth rates have also shrunk to below replacement levels.

One of the best predictors of a country’s fertility rate is its human-development index—a composite measure of life expectancy, education, and average income. As societies get richer and better educated, their citizens desire fewer children. And the past century of globalized capitalism has drastically improved living standards everywhere. In 1950, a bare majority of the world’s population lived in extreme poverty, long defined as having an income of less than $1.90 a day. Today, the share who live in extreme poverty is 10 percent, despite the new threshold of $3 a day. Sociologists trying to make sense of demographic trends have theorized that people assured about their material needs start to value their own individuality and self-actualization more than they do having kids. If this is the cause of the fertility collapse, it defies easy correction. It also would explain why all of the policy attempts that advanced countries have made at encouraging more births have generally proved unsuccessful.

Two claims may be true at once: Humanity is not doomed to wither away, but a sharp uptick in fertility does not look imminent. This latter reality raises the question of whether a sustained bout of population shrinkage—even one that eventually reverses—is itself harmful.

Plenty of economists would argue that it is. The essence of “endogenous growth theory,” which earned Paul Romer a Nobel Prize in 2018, is that economies ultimately grow because of ideas, each generated by individuals but then usable by others. In the case of a stagnating population, fewer people are around to have ideas. This is a scenario that the Stanford economist Charles I. Jones has described as the Empty Planet result: Rather than rising exponentially, economic growth stalls out as innovation plateaus. Note that this model predicts not collapsing living standards but rather stagnating living standards sometime in the distant future. Last year, the economists Dean Spears and Michael Geruso, both at the University of Texas, published a book called After the Spike, making an emphatic case for stabilizing the population even if living standards do not decrease, on the grounds that “more good is better.” “I think that life in 2150 is going to be a lot better than life today,” Spears told me recently. But, he argues, it would be better with more people around, because everyone would benefit from the accumulated work in medicine, technology, and even entertainment.

Even so, there will be ways to soften the blow. Economies facing a dearth of new ideas could compensate by devoting greater shares of workers toward such tasks. “If you told me the population would flatline or even fall in half, you could still have as many engineers and people coming up with new ideas as you had before just by drawing more people into that work,” Dietrich Vollrath, an economist at the University of Houston, told me. AI also seems set to increase the pace of discovery, including at the highest echelons of research and development—witness the recent string of pathbreaking mathematical proofs in just the past few months. A recent working paper by economists at MIT and the London Business School argues that “lower birth rates are associated with higher growth in GDP per working-age adult across countries … with no negative impact on aggregate GDP or earnings.” (Emphasis mine.) The paper’s methodology is being actively debated, but perhaps more important than the observed correlation is the authors’ proposed explanation: that when prime-age workers become scarce, companies adapt by developing productivity-boosting technology.

[Read: Americans are in denial about elder care]

This is not to say that the birth-rate decline will be painless. In a paper published this year, David Weil, an economist at Brown University, modeled what life would look like in two possible futures: one in which the U.S. has roughly replacement-level fertility, and another in which fertility is half that. Living standards in the extremely low-birth world, measured by consumption per capita, would be 8.7 percent lower, he calculated. As Detroit and other Rust Belt cities have shown, the problems of shrinkage—the unused school buildings and the underutilized infrastructure—are harder to manage than the problems of growth. When a population has fewer people, businesses have fewer customers available to them and may not even form in the first place—in the same way that the natural audience limit on Hungarian-language podcasts is much smaller than that of English-language ones.

Societies will be dominated more and more by the elderly and remade as a result. Government projections from South Korea anticipate that senior citizens will make up a majority of the country’s population starting in 2045. There are several reasons to think that this graying would be an economic drag. Most successful start-ups are led by people in their 40s, not people in their 60s. There will be intense demand for care, for both physical and cognitive ailments, which is labor-intensive and difficult to do well at scale. Because people are having fewer babies all around the world, the strategy of importing care workers from abroad will become untenable. Governments already spend considerably more on the elderly than on the young, because unwinding entitlement promises is so politically unpalatable. When gerontocracy becomes genuinely majoritarian, this problem will worsen too. Already, there is no available demographic escape route from a crack-up over entitlements. Within the next decade, America’s Social Security trust fund is expected to run out, which would trigger an immediate 22 percent cut in benefits. Even an immediate baby boom the size of Musk’s dreams would do little to alleviate the financial strain, because children generally take time to turn into taxpayers.

Vollrath, the University of Houston economist, views the ongoing baby bust as analogous to such past changes as urbanization and the decline of agriculture. “This is a clear change in preferences,” he said. “It has unintended consequences that are going to drag out for a couple of decades. That may change the growth rate. But ultimately—it’s always interesting to say this as a growth economist—who gives a shit about the growth rate? It’s not actually the thing you’re trying to optimize. You’re trying to optimize welfare and living standards.” From that perspective, he argues, declining fertility is unlikely to be catastrophic. As theories of the future go, this one lacks the apocalyptic oomph of more dire population projections. But it has a more important virtue as a prediction: It may be what actually happens.