The AtlanticModify

2026-06-22 00:00:00

for my father

My father’s hands flapped in a spiral of smoke—a weak light.
What did I dream then, a child drenched in image? Sleek light,

falling honeyed rivers, purpled fruit. What did I need
to imagine my body, calm in migration? I wanted to seek light.

Dawn sank into my hands like rain. I wanted to evaporate
& ask God to reveal my face. I wanted to speak light

& watch the earth settle into being. Each splash of wilderness
unraveled into clean, solid lines. From there I would leak light.

From there I would take flight, my body sloped & pliant
in this arena of disorder. But in the dark beak of night

that light still shivered. The world with its oblique
tilt. Every day I arrived & arrive. My physique light,

my mouth blazing verse. With prayer I swill inward
those weeks I lie rooted. Flood my cheek, light

traveling into all skin: I am learning to find pleasure
in uncertainty. Teach me your technique, light.

Wait for it to come to you, I heard once in a car. O radiant
risk, I am ready. Give me your mystique, light.

Untouched by flame, my father now shakes his hair
that suddenly grows to its full, shiny length—an antique light.

2026-06-21 22:00:00

2026-06-21 20:00:00

On a fall afternoon 15 years ago, I met an idealistic researcher outside a Stanford coffee shop to discuss our shared dream: using AI to detect cancer. He had wiry hair, a penchant for talking with his hands, and a reputation for brilliance. He worked at a research lab that developed early screens for cancer; I, at 20, had just learned that I carried a mutation that conferred a very high risk of breast, ovarian, and other cancers. Over the following years, he offered guidance on how to enter his field, prepared me to apply for the scholarship that would fund my Ph.D., and warned me away from cancer-screening companies that made exaggerated claims.

But from there our paths diverged. I became an AI professor. He co-founded Anthropic. My mentor was Dario Amodei, the man who leads one of the most powerful AI companies in the world. In a utopian 2024 essay titled “Machines of Loving Grace,” he predicted that superhuman AI—smarter than Nobel Prize winners, freely using computers, and collaborating with millions of copies of itself—could soon compress a century of scientific progress into a single decade, and potentially reduce cancer mortality by 95 percent.

Which should sound pretty good to me. At 35, my cancer risks are catching up with me. A few weeks ago, surgeons removed my ovaries, instantly inducing menopause and destroying my ability to naturally bear children. By 40, the risk of breast cancer for carriers of my mutation rises to one in four, double the lifetime risk for the average woman. My mother, who also carries the mutation, was diagnosed with breast cancer at 45. Now would be a fabulous time in my life for a superintelligent AI to cure cancer.

Why, then, do I find myself rooting for delays in the creation of this AI—hoping, in my heart of hearts, that GPT-6 will be a disappointment?

Part of the answer is that, despite the extraordinary speed of AI development, I do not believe that AI is likely to cure cancer anytime soon—certainly not enough to bet my life on it. This skepticism is shared by most of the AI experts in a survey I recently advised, who generally expect slower progress than the leaders of AI labs. AI systems are strongest in settings such as chess, where they can generate infinite data (by playing over and over again), experiment freely, and observe exactly what happens. Many important settings, including math and coding, share these properties, and AI has yielded remarkable progress there. But cancer is different. Cancer data are finite and come from biological experiments and clinical trials that cannot run at silicon speeds. Experimenting freely on cancer patients would be unethical. And cancer data only imperfectly illuminate the complex processes by which our own cells betray us. There are, in short, many barriers to curing cancer beyond a lack of intelligence.

The intelligence our existing AI systems provide is also already formidable and underused. We have yet to take full advantage of systems such as the Nobel Prize–winning AlphaFold, which predicts protein structures with stunning accuracy but has not yet yielded revolutions in drug development; or the AI algorithms that match or beat radiologists at many types of image analysis; or the chatbots that now aid scientists with research. My Ph.D. students used to write code to analyze medical data; now they express their ideas in plain English and let AI do the rest. They operate essentially as professors, constrained only by their own imagination. My student recently came to me giddy with excitement over an AI-aided medical discovery.

So as daunting as a cure for cancer remains, I am certain that AI will contribute to it. And if curing cancer were the only result of building ever more powerful AI systems, I would cheer for their arrival. But the problem is that their impacts are much broader, and we are moving too quickly to ensure that these impacts are positive.

The recent chaotic release of Anthropic’s latest model, Fable 5, illustrates how unprepared we are to handle the broader repercussions of these models. Anthropic, fearing that the model might be misused to develop bioweapons, initially kneecapped its ability to answer most basic biological questions, which the company said was a temporary measure. This made the model, ironically, far less useful for cancer research than its less powerful predecessors. A couple of days later, the U.S. government issued a national-security directive prohibiting foreign nationals from using the model, likely due to concerns that it could be used for cyberattacks. In response, Anthropic shut the model down entirely. Reasonable people disagree about how risky this model is and whether Anthropic or the government is overreacting. But clearly, our institutions aren’t remotely ready to respond to these rapid deployments. (Anthropic did not respond to a request for comment about Fable 5’s rollout, nor to other questions.)

Many developers of these models, including Dario Amodei, agree that AI is progressing more quickly than society is adapting. The solution they propose is for society to speed up, not for AI to slow down, which they view as unrealistic; the very title of Amodei’s latest essay, “Policy on the AI Exponential,” frames AI progress as an iron arc to which society must bend. But speeding ahead will inevitably mean more of the type of chaos that surrounded Fable 5’s release. More fundamentally, it will shorten our time to respond to the many societal challenges that powerful AI may raise, including mass unemployment, skyrocketing inequality, repressive surveillance, and autonomous warfare. Each of these—and many others that match their scope—is an enormous problem, no less obviously important than curing cancer, for which we lack good solutions. It is not at all clear that crafting an international response to all of these issues at breakneck speed is easier than slowing AI down.

I myself am ferociously impatient; since the day I learned I carried my mutation, I have lived with the constant awareness that life is finite. But I will wait a little longer for a cure—even if it means losing my fertility and living under the shadow of risk—if it lets us approach this new world more carefully, and ensure that, in curing cancer, we do not lose the things that make cancer worth curing.

Of all the things we stand to lose, I worry perhaps most about how we will find meaning if we obviate our own minds. Amodei struggles repeatedly with this question in his essays, calling it “more difficult than the others.” I admire his attempt to confront the question but find his answer unconvincing. “I spend plenty of time playing video games, swimming, walking around outside, and talking to friends,” he writes in “Machines of Loving Grace.” But I doubt that he would want to spend the rest of his life doing only those activities—certainly I would not. He suggests that humans will still find meaning in deep intellectual pursuits, such as doing research, even if AI can do them much better. For my own part, I would neither spend months struggling with a research problem I knew AI could solve instantly nor find as much pleasure in the answers it provided. I do not want to be merely a spectator to the universe, whatever wonders AI may reveal.  

Or take this essay. I will be heartbroken when a chatbot can extract my innermost feelings and, having gorged itself on the words of a million artists, regurgitate Fitzgerald-worthy prose I cannot match. For me, writing is a process bound up in self-discovery and human connection. My sister suggested the idea for this essay; my wife, seeing me suddenly and deeply sad as I reflected on it, touched my cheek, offering a comfort that no AI therapist could. Afterwards, I wrote late into the night at the handmade dining-room table I inherited from my grandparents. I thought of how my family would gather for long dinners around this table—the adults loosened with wine, the children excited to be part of it all, everyone laughing and talking over one another and debating physics and philosophy—trying, in our slow, suboptimal, human way, to figure things out.

2026-06-21 19:00:00

For several months last year, a Ukrainian housewife, 35 and lonely in a marriage that had gone cold, traded WhatsApp messages with a Chechen commander, Achmad, stationed somewhere in Ukraine’s occupied south. They wrote about their days, their disappointments, what they hoped to do when the war ended. She asked about the front. He told her.

“Send me a picture,” she said. “I want to see your life.”

One afternoon, he obliged—a photograph taken inside the barracks, of himself and another soldier grinning for the camera. Behind them, pinned to the wall, was a map of the compound showing the unit’s position.

The housewife did not exist. “She” was a middle-aged officer named Serhiy working for Ukraine’s military-intelligence directorate, part of a concerted effort to draw secrets from the men sent to occupy his country.

“Serhiy was great at flirting,” his commander told me. “Guys in our team started asking him for dating advice.” Shortly after Achmad sent that photograph, the coordinates it revealed were struck by a Ukrainian drone.

[Anne Applebaum: Ukraine is not losing. Russia is not winning.]

Ukraine’s resistance is alive and more lethal than ever. But it has changed dramatically since its early days. A man I will call Dmytro (he requested anonymity for reasons of safety) has served with a resistance team inside occupied Kherson from the first days of the full-scale invasion. “We took insane risks then,” he told me. “Nobody thought the Russians would be here long.” Partisan cells sprang up organically—people who knew one another, sometimes ex-military, improvising as they went. Symbolic acts of resistance happened daily. Ukrainians flew their flag and blared patriotic songs in public. The image of a grandmother pressing sunflower seeds into a Russian paratrooper’s hand—“so that sunflowers grow here when you die”—traveled around the world.

As it became clear the Russians intended to stay, such open defiance faded. Today, expressing support for Ukraine in Russian-occupied areas is likely to earn a trip to “the basement,” a euphemism for Russian torture chambers. Dmytro described Russia’s repression as a kind of machine: “It takes time to get spinning, then it has its own momentum.” High-resolution surveillance cameras now blanket city centers, and interrogations are a feature of daily life. Resistance leaders from Mariupol estimate that nearly half of the adult population there has been polygraphed by the Russian security service.

Even peaceful acts can meet with extreme repression. In 2022 and 2023, occupation forces in Mariupol effectively banned the colors blue and yellow. Residents describe receiving aid packages, including school supplies, with yellow and blue markers ripped out of their boxes. Today, Russia’s camera net is sophisticated enough to track individuals block by block, masked or not. Petro Andriushchenko escaped from Mariupol and now coordinates a cell still operating there. “Pro-Ukrainian graffiti can get you killed,” he told me. “Even disguised, your movements can be traced backwards to find where you live.”

The kind of symbolic resistance once waged by the general public has now given way to intelligence work, carried out by serious operatives. Managed by handlers in unoccupied Ukraine, these agents help identify targets, verify coordinates, and pass them to the Ukrainian military. The location of Achmad’s barracks, although traced through online subterfuge, was almost certainly confirmed by an agent on the ground. The result is a movement that has grown both quieter and deadlier. It now feeds the “middle-strike” campaign—a sustained drone offensive against targets deep inside the occupied territories, including air defenses, logistics hubs, command posts, and personnel. A crucial link in that kill chain is information from loyal Ukrainians behind enemy lines.

The partisans I spoke with included coordinators directing operations from free Ukraine, operatives working inside occupied territory, and volunteers scattered across Ukraine and abroad. Many have family members still living under Russian occupation and must closely guard their resistance activities. Inside the occupied territories, most agents work alone, their only connection via encrypted communication with their handlers. One operative, code-named Sestra, has no idea how far the network around her extends, except that it kills Russians almost every day. “What you do not know,” she explained to me, “you cannot betray under interrogation.”

The Russian military is a meat grinder. Commanders send infantry forward in waves that Ukrainian officers refer to as “human radar”: The piles of bodies reveal Ukrainian strong points. As crude as Russia’s infantry operations may be, its electronic countermeasures are very sophisticated and continually reshape how the resistance communicates and survives.

Until June 2022, Ukraine’s mobile carriers Kyivstar and Vodafone kept operating in the occupied zones, because the Russians had not yet stood up their own infrastructure. Then the Ukrainian networks went dark, and the resistance had to improvise. Early fixes were crude: VPNs, Wi-Fi nodes. Methods have since been refined, and the details are closely guarded. What is clear is that any phone purchased inside the occupied territories is useless for resistance work. Devices sold there come preloaded with monitoring software developed by Russian intelligence. That app is called Druge—Друг—which means “friend” in Russian.

Druge monitors communications, photographs, and location data, relaying all of it back to Russian intelligence. One woman who had recently escaped from an occupied zone told me that her mother, who still lived there, tried to delete Druge. The icon disappeared, but the app kept running in the background. At checkpoints, Russian soldiers examine every phone. Not having Druge installed is a red flag to them; having an encrypted app, such as Signal, guarantees a phone’s owner a trip to the basement.

Phones smuggled in from free Ukraine are the linchpin of resistance communications. For a time, one reliable route ran through Deutsche Post: parcels mailed from Germany, routed through Russia, delivered to innocuous addresses in occupied territory—a government office, a shop—and wrapped so that any tampering would be evident. That route is now closed. Others remain, though the specifics are closely held. In emergencies, a phone can be delivered by drone, which demands real-time coordination and carries its own risks. These “clean” devices do not have Russian spyware installed, and they lack SIM cards that would connect to the local mobile network. Because cell towers can detect when a new phone enters their coverage area, resistance members compose encrypted messages on a clean phone then send them via internet, using the hotspot of a second device already recognized by the network.

[Phillips Payson O’Brien: Ukraine has finally given up on Trump]

Few resistance agents have professional training. Most learn on the job. Partisans pass around hard-copy tradecraft manuals to avoid using vulnerable digital channels. Within Kherson’s partisan brigade, one of the most sought-after is a Soviet-era handbook describing CIA catfishing tactics in Africa during the Cold War. No online version exists, but a well-worn original circulates among the resistance.  

“Your CIA was good at this,” Dmytro said. “You bastards knew how to use sex.”

Several Ukrainian print shops have developed methods for hiding instruction manuals inside best-selling books. A guard at a Russian checkpoint, thumbing through an artificially tattered paperback, will likely have no idea that some of the pages explain how to kill him.

Inside the occupied territories, women form the backbone of the resistance. Many hold positions in Russia’s civil administration—at clinics, schools, and government offices—and report to Ukrainian intelligence. They exploit the occupiers’ assumptions: Russian soldiers often fail to imagine that women can be combatants. Few suspect that a grandmother passing their barracks every morning, shopping bags in hand, is the first link in a kill chain.

Some women volunteer with Russia-based charities supplying aid to military units. Every interaction is a collection opportunity. In late May, a government-linked Russian aid organization circulated a warning about this over Telegram: It had identified individuals, the agency noted, “who offer to deliver humanitarian aid to the Special Military Operation zone” as “a tactic employed by hostile forces to gather intelligence regarding the deployment of Russian troops.”

The most valuable sources are the occupiers themselves. In free Ukraine, agents build online relationships with occupation soldiers. Most of the operatives are women—though some men, such as Serhiy, have a gift for it. Native Russian soldiers tend to be difficult marks; they are transactional, Serhiy’s commander told me. “They always ask in the first five minutes: ‘Are we going to fuck or not?’” Chechens, by contrast, are “much more likely to seek a real relationship” and are easier to manipulate.

Since its creation in the early 1990s, the National Academy of the Security Service of Ukraine has trained operatives to cultivate intelligence assets. There are rumors of a new course that requires students to develop online connections with real occupation soldiers. An instructor at the academy, who requested anonymity for security reasons, would not confirm the course’s existence but insisted that “every intelligence agency does this—even yours.” When pressed about a specific aspect of the rumor, that the highest grades go to students who deliver target coordinates by the end of the course, he smiled. “That would indeed earn high marks.”

Most instruction for resistance agents, however, remains unofficial. Olena Biletska runs the Ukrainian Women’s Guard, a volunteer organization launched after Russia’s initial assault in 2014 to train women to survive under attack or occupation. By 2022, training had been delivered to more than 60,000 participants, some of whom remained in occupied territory. Most courses cover basic self-defense and survival, but others apply to resistance work. Pipelines smuggle training materials behind enemy lines. One course focuses on defeating polygraphs; others cover urban surveillance and intelligence-gathering.

Outside Ukraine, a diaspora helps vet target coordinates obtained through these resistance networks. Refugees from the occupied territories, who have detailed local knowledge, provide insight for the middle-strike drone campaign. A woman I will call Roksana, who asked for her name to be withheld to protect her network inside Ukraine, served in a clinic near Kherson on the occupied south bank of the Dnipro River. She barely escaped with her life after refusing to work for the Russian military. Now, living abroad, she helps verify targets for Ukrainian military intelligence. “I know my village—every street, every farm, every warehouse,” she told me.  

For Roksana and some of the other women operatives I spoke with, the determination to destroy the Russian occupation was forged in dark experience. “Almost every day, for the first few weeks after the invasion, we would hear about another body in the street,” Roksana said. “If it was a woman, they were often abused.” The doctor Tetyana Kostyantynivna runs the women’s center at one of Kyiv’s largest hospitals. In 2022 and 2023, her facility treated a steady stream of sexual-assault survivors from the occupied territories, ranging in age from 4 to 75. “Over the past four years,” she told me, “we have become a world leader in new methods for gynecological reconstructive surgery.”

During her own escape, Roksana passed through 33 Russian checkpoints, several of which were surrounded by dead bodies. Some of those corpses were of women and showed what she understood to be clear signs of sexual violence. At one checkpoint, a Russian soldier fired into the back of Roksana’s car while it sat parked, hitting a passenger in the legs. The soldiers did it for sport. But Roksana’s group made it through, and today, she has no reservations about guiding drone strikes against her own village, if doing so helps drive the Russians out. “We can rebuild warehouses,” she said, “but the Russians can’t rebuild Russians.”

[Read: Something is wrong with Russia’s children]

Women are crucial to the Ukrainian resistance. “They can go places, do things, that men cannot,” Andriushchenko, who runs agents inside Mariupol, told me. “Also, they are ruthless.” Several resistance leaders call their female agents vidma, a term that appears often in Ukrainian folklore. Its closest translation is “witch,” but it has a very different connotation here. The word derives from vidate, which means “to know.” Lesia Orobets, a former member of the Ukrainian Parliament, explained: “Vidmas were wise. They understood the secrets of the surrounding environment. Here in Ukraine, our vidmas were respected for their knowledge, not burned for it.”

These warrior-witches have become Ukraine’s most feared intelligence assets, moving through occupied territory like shadows. Orobets travels abroad often, where she is sometimes asked, “What happens if Ukraine runs out of men?”

“Be careful what you wish for,” she says. “If Ukraine’s women are in charge, there won’t be a Russian left alive.”

In the early months of the occupation, children played a role in the resistance. They slipped through checkpoints easily, took instantly to encrypted apps, and were extraordinarily brave. But the risks they took were no less grave than those faced by adults. An errant social-media post—or simply “liking” content supportive of Ukraine—was enough to get a child hauled in for interrogation. Those sessions could involve unspeakable violence, especially for girls. I interviewed one who was only 11 when her village near Kherson had been occupied. Implicated in “resistance” activities, she was dragged from her home. As we began our conversation, she apologized for her stutter. “I did not used to have this problem,” she told me, “until the Russians took me to the basement.”

The resistance says that it now enforces an absolute ban on children taking part. Dmytro, from the Kherson brigade, explained that “it’s not just about the risk to the kids. It’s about the risk to the whole unit.” The death of a child at Russian hands can devastate morale. Andriushchenko described the case of two teenage boys from Melitopol who had been interrogated by Russia’s security service. Their bodies were never returned, almost certainly because of how badly the boys had been tortured. “Their deaths hit us hard,” he said. “What they did to them.”

In practice, the ban on children helping the resistance has its limits. In one occupied city, teenagers have learned to move around the Russian camera network. For a while, they spray-painted Ukrainian colors on the sides of abandoned buildings. Now, given the risks of carrying blue or yellow paint, they chalk or scratch the Ukrainian letter Ї, which does not exist in the Russian alphabet, wherever they can get away with it. Even this, resistance leaders discourage. Sestra, the agent operating inside Mariupol, describes how “a single piece of graffiti can mean torture, a cellar, or deportation to Russia for the child, and arrest or the stripping of parental rights for the family.”

“It’s not worth it,” Andriushchenko said. “We need intel, not art. When they turn 18, they’ll get their turn.”

If resistance fighters are the first link in the kill chain, drone operators are the last. Iegor Kravchenko, whose call sign is “Ram,” commands a company in Ukraine’s 426th Unmanned Systems Regiment. Every night, his unit launches attack drones into the occupied territories. “A significant percentage of those missions,” he told me, “rely on intel provided by the resistance.”

Today, the middle-strike drone campaign is the main engine of partisan activity. It has become extraordinarily efficient. For a high-priority target—an air-defense system, a command post, a munitions dump—anywhere from 15 minutes to a few hours can elapse between the transmission of coordinates and the strike. There have been moments when an operative was still chatting online with a soldier as a drone hit his position. “Our goal,” Orobets said, “is to make sure Russian soldiers never reach the front line.”

[Read: Building tanks while the Ukrainians master drones]

I asked the partisans why they would talk with me at all, sharing intimate details of the war’s most dangerous operations. In part, they are sending a message to the occupier: You are hated here. Sestra put a finer point on this: “I want every Russian soldier who has set foot on our land to carry that paranoia with him—suffocating, relentless, every second of every day. I want him to look at the grandmother at the market, at the bus driver, at the doctor in the clinic, at the ordinary passerby on the street—and to see in each of them his own potential destruction.”

The Ukrainian operatives also want Americans to know that Ukraine is fighting for every inch of its land. Asked whether Ukraine would tolerate a peace deal ceding occupied territory, Biletska answered, “Have you seen Bucha? Kherson? Mariupol? That’s not peace.”

As for Achmad, the Chechen commander who revealed the location of his own barracks, Andriushchenko was blunt: “He’s gone dark online, but we suspect he still doesn’t realize he was flirting with a middle-aged chuvak,” meaning a dude.

I asked whether he worried about exposing Achmad as a source. “No. I hope his unit learns what he did,” Andriushchenko said. “And then I hope they cut off his balls.”

2026-06-21 19:00:00

In early 2025, J. D. Vance paid a visit to Les Invalides, in Paris, where he was invited to clutch the sword of the Marquis de Lafayette, a hero of the American Revolution. In a speech the next day, Vance drew a parallel between that sword and artificial intelligence, calling them both “weapons that are dangerous in the wrong hands but are incredible tools for liberty and prosperity in the right hands.”

Whether the rollout of AI in the U.S. ends up in the right hands will depend to some degree on the vice president himself. Since President Trump returned to office, Vance has taken a prominent role in articulating how the administration should approach the AI revolution. Many Republicans, including Trump, broadly favor a hands-off approach. Vance, in a series of speeches and interviews, has offered a more substantive framework for the interplay among government, AI companies, and workers—with the occasional political barb thrown in.

He, too, has called for avoiding regulations that slow innovation. But he also believes that some forms of power are too important to leave to Big Tech to self-regulate. And he has sought to tackle one question on the minds of many Americans: What will happen to workers?

Vance’s approach reflects the tension between the two main forces that fueled his political rise. Before he ran for Senate, Vance worked in venture capital, and his ascent was backed by Silicon Valley figures such as Peter Thiel and David Sacks—many of the same people who are now in the aggressively anti-regulation, pro-market camp of the GOP. Yet Vance also built his political reputation on his self-described hillbilly upbringing and on giving voice to the frustrations of working-class voters—who feel forgotten by Washington, view both government and big companies with suspicion, and fear that AI is coming for their jobs.

Now Vance is attempting to build a distinctly MAGA vision of AI that rewards innovation and is global in ambition, protective of American workers, skeptical of regulation, and wary of concentrated corporate power. Depending on what you think about AI (and about Vance), that could be viewed as a reasonable middle ground designed to keep some guardrails on AI’s development and to help protect American livelihoods—or viewed as the tap dance of a politician aiming to appease everyone and potentially satisfying no one.

Either way, Vance is in no doubt about the stakes. The day after touching Lafayette’s sword, he told the audience of executives and policy makers gathered for the Artificial Intelligence Summit: “If we choose the wrong approach on other things that could be conceived of as dangerous—things like AI—and choose to hold ourselves back, it will alter not only our GDP or the stock market but the very future of the project that Lafayette and the American founders set off to create.”

Vance often compares workers’ fears about the impact of AI to concerns that followed the introduction of ATMs in the 1970s, when many people predicted that bank tellers would become obsolete. “What actually happens is we have more bank tellers today than we did when the ATM was created, but they’re doing slightly different work,” Vance told the Interesting Times podcast in May 2025. “More productive. They have pretty good wages relative to other folks in the economy. I tend to think that is how this innovation happens.”

[Read: What happens if Trump seizes AI companies]

I asked the economist James Bessen, who has written extensively about the impact of ATMs, whether that was an apt analogy. “It’s not even the right account of what happened with bank tellers,” he told me. When ATMs were first introduced, the number of tellers in any given bank branch likely fell. That dynamic also made it cheaper to open new branches, which banks raced to do after deregulatory measures in the 1990s allowed them to expand their footprint. More branches meant more tellers, though they were performing different roles. That picture changed again after 2010, as online banking spread. This time, the new technology led to a dramatic reduction in the number of teller jobs.

A more fitting comparison, Bessen told me, is the 19th-century automation of the textiles industry. Automation made cloth cheaper, demand surged, and employment and productivity both grew. But by the mid-20th century, those gains had already been reaped, and as productivity continued to grow and demand plateaued, jobs disappeared. The lesson is that technological progress doesn’t guarantee permanent job creation, as Vance suggests. What is clear, Bessen added, is that people across the workforce will need to acquire new skills if they want to keep working.

Vance disagrees. “I don’t buy the premise,” he told me in an interview last week. “I have not yet seen the evidence that you’re going to see widespread job destruction because of AI.”

Yet Vance said he worries that the benefits of AI will fall disproportionately on the wealthy, as has happened in previous industrial revolutions. “You certainly saw a massive concentration of wealth and income upwards,” he said. “I do worry about that. I think it’s a serious concern, and I think it’s something, certainly, policy makers should not sit idly by and let happen.”

Vance acknowledges that he is an optimist. And he has a venture capitalist’s focus on productivity. Americans, he says, should be less concerned about AI replacing workers and more worried about falling behind in developing technologies that make workers more productive. “If the robots were coming to take all of our jobs, you would see labor productivity skyrocketing in this country—but you actually see labor productivity flatlining,” he said at the Winning the AI Race Summit last July.

The bigger problem, he believes, is the tendency by some companies to prioritize foreign talent. Vance has criticized Microsoft and other tech firms for laying off workers while simultaneously applying for H-1B worker visas to hire foreign labor. Corporate claims of a domestic talent shortage are a “bullshit story,” he has said. “When we try to grow our economy, frankly, through importing cheap labor, that, I think, is a dead end,” he said at the summit. Microsoft declined to comment.

Will Rinehart, a tech-policy expert at the right-leaning American Enterprise Institute, told me that he has discussed with the vice president’s staff ways to promote labor development for an AI world. “What’s interesting about Vance is that he does cut his own line, where, in part, he is with the administration on no excessive regulation of AI, but at the same time, he sees this as an opportunity for pro-worker development and pro-worker growth,” Rinehart said.

How that might translate into policy, however, isn’t clear—and on this, the United States isn’t alone. “It’s not surprising to me that they don’t have a plan for that,” Nathaniel Persily, a co-director of the Stanford Law School AI Initiative, told me. “Frankly, no country in the world right now seems to have a thought-out plan for how to deal with AI-related labor-force impacts.”

The Republican Party itself is split over what to do about AI, a fight reflected in a meeting at Vance’s office last year. Seated across from each other were two influential figures in Trump’s orbit: David Sacks, who stepped down in March as the administration’s AI czar (but continues to advise the White House), and the conservative attorney Mike Davis, who was trying to stop Sacks from inserting language into legislation that would have limited congressional oversight over the technology and eliminated state jurisdiction.

Vance, who is close to both men, tried to talk through their differences. The exchange grew heated at times, and Vance urged Sacks and Davis to find common ground, according to officials familiar with the exchange.

Weeks later, in December, Trump signed an executive order that sought to curb states’ ability to pass their own AI regulation, and directed the administration to work with Congress on a national framework for governance. Republican governors from 17 states responded by calling on Congress to strip any language that banned state-level AI regulations from a congressional budget reconciliation. California and other Democratic-run states have pushed their own AI rules.

Months later, those tensions still simmer. Florida’s Republican governor, Ron DeSantis, has argued that the White House can’t legally strip states of their right to protect citizens using an executive order alone, and has pushed ahead with proposals for Florida to put its own safeguards around AI. Earlier this month, Florida became the first state to sue OpenAI and CEO Sam Altman, claiming that ChatGPT puts corporate profits ahead of safety and is unsafe for minors. Kate Waters, a spokesperson for OpenAI, told me in a statement that the company is working to enhance AI safety. “We know there is more work to do,” she said. “We are committed to getting this right.”

In February, Fox News’s Martha MacCallum asked Vance in an interview whether AI policies should be left to the states or be federalized. He said the country needed “one standard” because the tech companies are, in his words, “so complicated.”

“What’s happening in California affects what happens in Ohio, and vice versa. I think that, eventually, you’re going to have some standard applied, whether it’s a federal standard or whether it’s one state standard dominating,” he added. “Frankly, the worst possible outcome would be to have far-left California dominate the entire AI regulatory map. That is, unfortunately, what the Californians would like to happen.”

Vance doesn’t want to see the Europeans take the lead, either. In his Paris speech, he argued that perhaps the greatest threat to AI was that governments would burden it with rules before the benefits could be realized.

Vance argued that overly restrictive rules—in other words, what he sees happening in Europe—strengthen the position of established firms, making it harder for new companies to compete. In Paris, Vance also described how he viewed the technology itself. He talked about AI in the way that previous generations talked about oil, steel, and nuclear reactors—as a source of national power that could determine which countries build stronger economies, field more capable militaries, and exert greater influence in the decades ahead. “AI, we believe, is going to make us more productive, more prosperous, and more free,” he said.

[Read: The conversions of J. D. Vance]

Given those stakes, Vance has repeatedly warned against allowing a small number of companies to control the industry. That view sets him apart from many traditional Republicans, who have long been skeptical of antitrust enforcement. Vance sees danger not only in government overregulation but in a handful of dominant firms gaining too much power. The new industrial revolution will be thwarted, he said, “if we allow AI to become dominated by massive players looking to use the tech to censor or control users’ thoughts.”

His interest in corporate restraint goes only so far. The construction of AI data centers is controversial in both parties. Supporters argue that the sprawling facilities are essential to winning the global AI race; detractors warn that they strain power grids, consume enormous amounts of energy and water, and hand even more influence to Big Tech. Vance has mostly sided with the builders.

For all of Vance’s skepticism of regulation, he makes exceptions where national security, human judgment, or democracy are on the line. Addressing the graduating class at the U.S. Air Force Academy in May, Vance warned that artificial intelligence was transforming warfare faster than military institutions have historically adapted to new technologies. He  cautioned them to “use technology to make you better, but never submit to it.”

In making his case, Vance invoked Pope Leo XIV, whose recent writings on AI have called for stricter ethical constraints on autonomous weapons and warned that some military systems are already moving beyond meaningful human control. Vance has publicly disagreed with the pontiff on other matters, but he embraced those concerns.

AI is rapidly becoming embedded in national-security systems. OpenAI and other major tech firms have scaled up their partnerships with the Department of Defense. Anthropic, which has been put on a national-security blacklist, is an exception. (Trump last week  said that talks with Anthropic to resolve the issue were “going fine.”) The company’s relationship with the federal government fractured earlier this year following a fierce dispute over military guardrails, including whether AI systems should be allowed to make autonomous decisions.

Defense Secretary Pete Hegseth has argued that the Pentagon’s AI initiatives should operate “without ideological constraints that limit lawful military applications” and that the military is “building war ready weapons and systems, not chatbots for an Ivy League faculty lounge.” He issued a directive requiring a handful of AI defense contractors to permit the military to use their technologies for all “lawful operational use,” without exceptions.

Vance, who, like Hegseth, is a military veteran, has focused on a different concern: ensuring that even as machines become more capable, human beings retain responsibility for choices that carry moral weight. “Decisions over life and death must be made by humans and not machines,” he said at the Air Force Academy.

The dilemma facing policy makers now is whether the lines that Vance favors will hold—or whether the AI revolution will overwhelm those seeking to guide its development. On that, Les Invalides and Lafayette may provide different history lessons than the ones Vance cited in Paris.

On July 14, 1789, Parisian insurgents seized weaponry from Les Invalides that they used to storm the Bastille later the same day. In the resulting chaos, Lafayette, as head of the French National Guard, sought to steer a middle path between the monarchy and the insurrectionists. He failed, and was eventually forced to flee France.

2026-06-20 23:31:25

Editor’s Note: Washington Week With The Atlantic is a partnership between NewsHour Productions, WETA, and The Atlantic airing every Friday on PBS stations nationwide. Check your local listings, watch full episodes here, or listen to the weekly podcast here.

Last night on Washington Week With The Atlantic, panelists joined to discuss the signing of an agreement between the United States and Iran, and what Donald Trump’s deal with the regime may mean for other countries.

“The international community is looking at what happened not only in Iran but in Ukraine, and seeing that this idea of large powers coming in and definitively defeating other weaker nations is not necessarily the case anymore,” Nancy Youssef, a staff writer at The Atlantic argued last night.

What this may mean going forward is that militaries across the world, including the U.S., will look at their technological, drone, and AI capabilities to “figure out what advances they need to make given this rapid moving and changing battlefield dynamic,” she said.

Joining the editor in chief of The Atlantic, Jeffrey Goldberg, to discuss this and more: Jonathan Karl, a chief Washington correspondent at ABC News; Karim Sadjadpour, a contributing writer at The Atlantic and a senior fellow at the Carnegie Endowment for International Peace; David Sanger, a White House and national-security correspondent at The New York Times; and Youssef.

Watch the full episode here.