Itemoids

Marc Andreessen

The DeepSeek Wake-Up Call

The Atlantic

www.theatlantic.com › newsletters › archive › 2025 › 01 › the-deepseek-wake-up-call › 681512

This is Atlantic Intelligence, a newsletter in which our writers help you wrap your mind around artificial intelligence and a new machine age. Sign up here.

Earlier this week, almost overnight, the American tech industry entered a full-on panic. The latest version of DeepSeek, an AI model from a Chinese start-up of the same name, appeared to equal OpenAI’s most advanced program, o1. On Monday, DeepSeek overtook ChatGPT as the No. 1 free app on Apple’s mobile-app store in the United States.

So far, China has lagged the U.S. in the AI race. DeepSeek suggests that the country has gained significant ground: The chatbot was built more quickly and with less money than analogous models in the U.S., and also appears to use less computing power. Software developers using DeepSeek pay roughly 95 percent less per word than they do with OpenAI’s top model. One prominent AI executive wrote that DeepSeek was a “wake up call for America.” Because DeepSeek appears to be cheaper and more efficient than similarly capable American AI models, the tech industry’s enormous investments in computer chips and data centers have been thrown into doubt—so much that the top AI chipmaker, Nvidia, lost $600 billion in market value on Monday, the largest single-day drop in U.S. history. Sam Altman, the CEO of OpenAI, said that it was “invigorating to have a new competitor” and that, in response, the company would move up some new software announcements. (Yesterday morning, OpenAI said that it is investigating whether DeepSeek used ChatGPT outputs to train its own model.)

But many prominent American researchers and tech executives celebrated DeepSeek, as well. That’s because “the most notable feature of DeepSeek may be not that it is Chinese, but that it is relatively open,” I wrote on Monday. Whereas the top American AI labs at OpenAI, Google, and Anthropic have kept their technology top-secret, DeepSeek published an in-depth technical report and is allowing anybody to download and modify the program’s code. “Being democratic—in the sense of vesting power in software developers and users—is precisely what has made DeepSeek a success,” I wrote. Start-ups and researchers love this relative transparency. In theory, competitors can use DeepSeek’s code and research to rapidly catch up to OpenAI with far fewer resources—you might not need colossal data centers to get to the front of the AI race. (The Atlantic recently entered into a corporate partnership with OpenAI.) However, there’s substantial uncertainty about just how much cheaper DeepSeek was to build, based on reports about the start-up’s hardware acquisitions and uncertainty about how the model was trained.

Meanwhile, for national-security hawks, the fear is that an open-source program that won’t answer questions about the Tiananmen Square protests could become a global technological touchpoint. DeepSeek could face similar privacy concerns as TikTok: Already, the U.S. Navy has banned its use, citing security concerns.

Any predictions, for now, are highly speculative. The global AI race is far from over, and forthcoming products from Silicon Valley could leap ahead once again. At the very least, U.S. tech companies may have to reconsider whether the best way to build AI is by keeping their models a secret.

Illustration by The Atlantic. Source: Getty.

China’s DeepSeek Surprise

By Matteo Wong

One week ago, a new and formidable challenger for OpenAI’s throne emerged. A Chinese AI start-up, DeepSeek, launched a model that appeared to match the most powerful version of ChatGPT but, at least according to its creator, was a fraction of the cost to build. The program, called DeepSeek-R1, has incited plenty of concern: Ultrapowerful Chinese AI models are exactly what many leaders of American AI companies feared when they, and more recently President Donald Trump, have sounded alarms about a technological race between the United States and the People’s Republic of China. This is a “wake up call for America,” Alexandr Wang, the CEO of Scale AI, commented on social media.

But at the same time, many Americans—including much of the tech industry—appear to be lauding this Chinese AI. As of this morning, DeepSeek had overtaken ChatGPT as the top free application on Apple’s mobile-app store in the United States. Researchers, executives, and investors have been heaping on praise. The new DeepSeek model “is one of the most amazing and impressive breakthroughs I’ve ever seen,” the venture capitalist Marc Andreessen, an outspoken supporter of Trump, wrote on X. The program shows “the power of open research,” Yann LeCun, Meta’s chief AI scientist, wrote online.

Read the full article.

What to Read Next

The GPT era is already ending: “The release of o1, in particular, has provided the clearest glimpse yet at what sort of synthetic ‘intelligence’ the start-up and companies following its lead believe they are building,” I wrote in December. The new AI panic: “The obsession with frontier models has now collided with mounting panic about China, fully intertwining ideas for the models’ regulation with national-security concerns,” Karen Hao wrote in 2023.

P.S.

After several major tech executives announced their support for Donald Trump, many liberal internet users are now alleging that they are being censored on certain social-media platforms. “To some, this pattern was as unmistakable as it was malicious,” my colleague Kaitlyn Tiffany writes. “Social media was turning against Democrats.” And they are panicking.

— Matteo

Donald Trump Is Just Watching This Crisis Unfold

The Atlantic

www.theatlantic.com › politics › archive › 2025 › 01 › trump-airplane-crash › 681511

You might be forgiven for forgetting—ever so briefly—that Donald Trump is president of the United States. Sometimes it seems like he does, too.

In the middle of the night, as news about the plane crash at Ronald Reagan Washington National Airport was breaking, Trump posted on Truth Social:

The airplane was on a perfect and routine line of approach to the airport. The helicopter was going straight at the airplane for an extended period of time. It is a CLEAR NIGHT, the lights on the plane were blazing, why didn’t the helicopter go up or down, or turn. Why didn’t the control tower tell the helicopter what to do instead of asking if they saw the plane. This is a bad situation that looks like it should have been prevented. NOT GOOD!!!

He raises some valid points—ones that many people might be wondering about themselves. The difference between them and him is that he is the leader of the federal government, able to marshal unparalleled resources to get answers about a horror that happened just two and a half miles from his home. He’s the commander in chief of the U.S. armed forces, and the crash involved an Army helicopter. But Trump isn’t really interested in doing things. Like Chauncey Gardiner, the simple-minded protagonist of Being There, he likes to watch.

This morning, Trump held an astonishing briefing at the White House where he and his aides unspooled racist speculation, suggesting (without any evidence) that underqualified workers hired under DEI programs had caused the accident. “We do not know what led to this crash, but we have some very strong ideas and opinions, and I think we’ll state those opinions now,” Trump said, and he did. Vice President J. D. Vance and Defense Secretary Pete Hegseth criticized diversity efforts from the lectern as well. (Trump also misrepresented Federal Aviation Administration programs.) Trump insisted that he wasn’t getting ahead of the investigation by speculating, and that he could tell diversity was to blame because of “common sense.”

Trump also paused to accuse former Transportation Secretary Pete Buttigieg of “bullshit,” and narrated videos and information he’d seen in the news, interspersing his personal observations as a helicopter owner and passenger. “The people in the helicopter should have seen where they were going,” Trump said. At times, he appeared to blame both the helicopter pilots and air-traffic control. Perhaps it would be better to actually gather some information, but Trump is more interested in pontificating.

The pilots, DEI, air-traffic controllers, Buttigieg—the only common thread appeared to be that everyone was to blame, except for Trump himself.

No one could reasonably hold Trump responsible for the crash, just 10 days into his term—though that is the bar he has often tried to set. “I alone can fix it,” he has assured Americans, telling them that he personally can master and control the government in a way no one else can. He promised to be a dictator, though only on day one. Yet even while discounting his bluster, it would be nice to see the president doing something more than watching cable news and posting about it.

If he’s not going to do that, he could offer some consolation. Almost exactly 39 years ago, after the loss of the Space Shuttle Challenger, President Ronald Reagan memorably described how the astronauts aboard had “‘slipped the surly bonds of earth’ to ‘touch the face of God.’” Trump is giving us “NOT GOOD!!!”

Though exasperating, this passivity is no surprise. It was a running theme of Trump’s first administration and is already back in the second. In May 2016, Trump reportedly offered fellow Republican John Kasich a chance to be vice president, in charge of domestic and foreign policy; Trump would be in charge of “making America great again.” During Hurricane Harvey, in 2017, he struggled to show empathy for victims or do more than gawk at (and tweet about) the destruction. A few months later, he tried half-heartedly to do more after Hurricane Maria, producing the indelible visual of the president tossing paper towels to victims, like a giveaway at a minor-league baseball game.

[Read: That time Trump threw paper towels at Puerto Ricans]

Marc Andreessen, the venture capitalist and Trump ally, has claimed that Trump wasn’t even running the government during his first term. During the worst of the coronavirus pandemic, Matt Yglesias notes, Trump was more interested in offering punditry on how the government was doing than acting like the head of the executive branch. And on January 6, 2021, according to federal prosecutors, Trump sat at the White House watching the violent sacking of the Capitol and doing nothing to stop it.

This approach to governance—or refusal to approach it, rather—is inextricably tied to Trump’s Gardiner-like obsession with television. The president watches hours of news every day, and if reports from inside the White House didn’t bear witness to this, his all-hours social-media posts would. Because he has little grounding in the issues facing the government and little interest in reading, television frequently seems to set his agenda. Political allies learned that the best way to get a message to Trump was to appear on Fox News. (Trolls, similarly, learned that a good way to rankle him was to take out ads on the channel.) Trump has used the Fox roster as a hiring pool for his administration.

One vignette from the first Trump administration illustrates the dynamic. In April 2019, as the White House was juggling half a dozen serious controversies, Trump called into Fox & Friends and yakked at length about whatever happened to be on his mind until even the hosts couldn’t take it any longer. Finally, Brian Kilmeade cut in and brought things to a close. “We could talk all day, but looks like you have a million things to do,” he said. Trump didn’t appear concerned about it.

[Read: Donald Trump calls in to Fox & Friends]

What’s odd is that even as Trump acts so passively, his administration is moving quickly to seize unprecedented powers for the presidency. In part, that’s because of the ideological commitments of his aides, but Trump also has a curious view of presidential power as an à la carte thing. He’s very interested in acquiring and flexing power to control the justice system, punish his enemies, and crack down on immigration, but he’d just as soon get the federal government out of the emergency-management business.

The presidency is not a spectator sport, though. At the end of Being There (spoiler alert), a group of political advisers conspires to put Chauncey Gardiner forward as the next president. The movie’s central joke is that the childlike, TV-obsessed protagonist has inadvertently fooled the nation’s most powerful circles into believing that he is profound, simply by stating directly what little he sees and understands. Joke’s on us.

China’s DeepSeek Surprise

The Atlantic

www.theatlantic.com › technology › archive › 2025 › 01 › deepseek-china-ai › 681481

One week ago, a new and formidable challenger for OpenAI’s throne emerged. A Chinese AI start-up, DeepSeek, launched a model that appeared to match the most powerful version of ChatGPT—but, at least according to its creator, was a fraction of the cost to build. The program, called DeepSeek-R1, has incited plenty of concern: Ultrapowerful Chinese AI models are exactly what many leaders of American AI companies feared when they, and more recently President Donald Trump, have sounded alarms about a technological race between the United States and the People’s Republic of China. This is a “wake up call for America,” Alexandr Wang, the CEO of Scale AI, commented on social media.

But at the same time, many Americans—including much of the tech industry—appear to be lauding this Chinese AI. As of this morning, DeepSeek had overtaken ChatGPT as the top free application on Apple’s mobile-app store in the U.S. Researchers, executives, and investors have been heaping on praise. The new DeepSeek model “is one of the most amazing and impressive breakthroughs I’ve ever seen,” the venture capitalist Marc Andreessen, an outspoken supporter of Trump, wrote on X. The program shows “the power of open research,” Yann LeCun, Meta’s chief AI scientist, wrote online.

Indeed, the most notable feature of DeepSeek may be not that it is Chinese, but that it is relatively open. Unlike top American AI labs—OpenAI, Anthropic, and Google DeepMind—which keep their research almost entirely under wraps, DeepSeek has made the program’s final code, as well as an in-depth technical explanation of the program, free to view, download, and modify. In other words, anybody from any country, including the U.S., can use, adapt, and even improve upon the program. That openness makes DeepSeek a boon for American start-ups and researchers—and an even bigger threat to the top U.S. companies, as well as the government’s national-security interests.

To understand what’s so impressive about DeepSeek, one has to look back to December, when OpenAI launched its own technical breakthrough: the full release of o1, a new kind of AI model that, unlike all the “GPT”-style programs before it, appears able to “reason” through challenging problems. o1 displayed leaps in performance on some of the most challenging math, coding, and other tests available, and sent the rest of the AI industry scrambling to replicate the new reasoning model—which OpenAI disclosed very few technical details about. The start-up, and thus the American AI industry, were on top. (The Atlantic recently entered into a corporate partnership with OpenAI.)

DeepSeek, less than two months later, not only exhibits those same “reasoning” capabilities apparently at much lower costs, but has spilled at least one way to match OpenAI’s more covert methods to the rest of the world. The program is not entirely open-source—its training data, for instance, and the fine details of its creation are not public—but, unlike with ChatGPT, Claude, or Gemini, researchers and start-ups can still study the DeepSearch research paper and directly work with its code. OpenAI has enormous amounts of capital, computer chips, and other resources, and has been working on AI for a decade. In comparison, DeepSeek is a smaller team formed two years ago with far less access to essential AI hardware, because of U.S. export controls on advanced AI chips, but it has relied on various software and efficiency improvements to catch up. DeepSeek has reported that the final training run of a previous iteration of the model that R1 is built from, released in December, cost less than $6 million. Meanwhile, Dario Amodei, the CEO of Anthropic, has said that U.S. companies are already spending on the order of $1 billion to train future models. Exactly how much the latest DeepSeek cost to build is uncertain—some researchers and executives, including Wang, have cast doubt on just how cheap it could have been—but the price for software developers to incorporate DeepSeek-R1 into their own products is roughly 95 percent cheaper than incorporating OpenAI’s o1, as measured by the price of every “token”—basically, every word—the model generates.

DeepSeek’s success has abruptly forced a wedge between Americans most directly invested in outcompeting China and those who benefit from any access to the best, most reliable AI models. (It’s a divide that echoes Americans’ attitudes about TikTok—China hawks versus content creators—and China’s other apps and platforms.) For the start-up and research community, DeepSeek is an enormous win. “A non-US company is keeping the original mission of OpenAI alive,” Jim Fan, a top AI researcher at the chipmaker Nvidia and former OpenAI employee, wrote on X. “Truly open, frontier research that empowers all.”

But for America’s top AI companies, and the nation’s government, what DeepSeek represents is unclear. The stocks of many major tech firms—including Nvidia, Alphabet, and Microsoft—dropped this morning amid the excitement around the Chinese model. And Meta, which has branded itself as a champion of open-source models in contrast to OpenAI, now seems a step behind. (The company is reportedly panicking.) To some investors, all those massive data centers, billions of dollars of investment, or even the half-a-trillion-dollar AI-infrastructure joint venture from OpenAI, Oracle, and SoftBank, which Trump recently announced from the White House, could seem far less essential. Maybe bigger AI isn’t better. For those who fear that AI will strengthen “the Chinese Communist Party’s global influence,” as OpenAI wrote in a recent lobbying document, this is legitimately concerning: The DeepSeek app refuses to answer questions about, for instance, the Tiananmen Square protests and massacre of 1989 (although the censorship may be relatively easy to circumvent).

None of that is to say the AI boom is over, or will take a radically different form going forward. The next iteration of OpenAI’s reasoning models, o3, appears far more powerful than o1 and will soon be available to the public. There are some signs that DeepSeek trained on ChatGPT outputs (outputting “I’m ChatGPT” when asked what model it is), although perhaps not intentionally—if that’s the case, it’s possible that DeepSeek could only get a head start thanks to other high-quality chatbots. America’s AI innovation is accelerating, and its major forms are beginning to take on a technical research focus other than reasoning: “agents,” or AI systems that can use computers on behalf of humans. American tech giants could, in the end, even benefit. Satya Nadella, the CEO of Microsoft, framed DeepSeek as a win: More efficient AI means that use of AI across the board will “skyrocket, turning it into a commodity we just can’t get enough of,” he wrote on X today—which, if true, would help Microsoft’s profits as well.

Still, the pressure is on OpenAI, Google, and their competitors to maintain their edge. With the release of DeepSeek, the nature of any U.S.-China AI “arms race” has shifted. Preventing AI computer chips and code from spreading to China evidently has not tamped the ability of researchers and companies located there to innovate. And the relatively transparent, publicly available version of DeepSeek, rather than leading American programs, could mean Chinese programs and approaches become global technological standards for AI—akin to how the open-source Linux operating system is now standard for major web servers and supercomputers. Being democratic—in the sense of vesting power in software developers and users—is precisely what has made DeepSeek a success. If Chinese AI maintains its transparency and accessibility, despite emerging from an authoritarian regime whose citizens can’t even freely use the web, it is moving in exactly the opposite direction of where America’s tech industry is heading.

MAGA Is Starting to Crack

The Atlantic

www.theatlantic.com › technology › archive › 2025 › 01 › maga-trump-tech-nationalist-conflict › 681422

On Sunday night, in the basement ballroom of the Salamander Hotel in Washington, D.C., Charlie Kirk was happier than I’d ever seen him. “I truly believe that this is God’s grace on our country, giving us another chance to fight and to flourish,” Kirk, the head of Turning Point USA, a conservative youth-outreach organization, said to cheers from the hundreds of MAGA loyalists who had come out for his pre-inaugural ball. “What we are about to experience is a new golden era, an American renaissance.”

The celebrations have continued now that Donald Trump is back in the White House, as he has signed a flurry of executive orders to make good on his campaign promises. But this might be the best mood that MAGA world will be in for a while. The president’s coalition is split between two distinct but overlapping factions that are destined for infighting. On one side are the far-right nationalists and reactionaries who have stood by Trump since he descended down his golden escalator. Among them are Stephen Miller, who is seen as a chief architect of Trump’s anti-immigration agenda, and Steve Bannon, Trump’s former chief strategist and the former executive chair of Breitbart News. On the other side is the tech right: Elon Musk and other Silicon Valley elites, including Peter Thiel and Marc Andreessen, who have become ardent supporters of the president. Already, these groups are butting heads on key aspects of Trump’s immigration crackdown. In Trump’s second term, not everyone can win.

During the campaign, it was easy for these two groups to be aligned in the goal of electing Trump. Members of the nationalist wing took glee in how Musk boosted their ideology on X, the social platform he owns. With his more than 200 million followers, Musk has helped spread far-right conspiracy theories, such as the false claim that Haitian immigrants in Ohio are eating people’s pets. Meanwhile, the tech right has relished attacks on DEI efforts in the workplace—attacks that have allowed them to more easily walk back hiring practices, against the wishes of their more liberal employees.

But the two groups also want different things. The nationalist right wants an economy that prioritizes and assists American-born families (specifically, traditional nuclear ones), sometimes at the expense of business interests; the tech right wants a deregulated economy that bolsters its bottom line. The nationalist right wants to stop almost all immigration; the tech right wants to bring in immigrant workers as it pleases. The nationalist right wants to return America to a pre-internet era that it perceives as stable and prosperous; the tech right wants to usher in a bold, globally focused new economy.   

Already, the cracks have started to show. Last month, Trump’s pick of the Silicon Valley venture capitalist Sriram Krishnan as an AI adviser led to a bitter and very public spat between the two camps over visas for highly skilled immigrants. (“FUCK YOURSELF in the face,” Musk at one point told his critics on the right.) At the time, I argued that the MAGA honeymoon is over. The disagreements have only intensified. Last week, after former President Joe Biden used his farewell speech to warn about the influence of Silicon Valley oligarchs and the “tech industrial complex,” the white-nationalist influencer Nick Fuentes posted on X that “Biden is right.” Bannon in particular has not relented: Earlier this month, he told an Italian newspaper that Musk is a “truly evil person” and that would get the billionaire “kicked out” of Trump’s orbit by Inauguration Day. (Considering that Musk is reportedly getting an office in the West Wing, Bannon does not seem to have been successful in that quest.) In an interview with my colleagues Ashley Parker and Michael Scherer, Bannon described the tech titans as “nerds” whom Trump was humiliating. Seeing them on Inauguration Day was “like walking into Teddy Roosevelt’s lodge and seeing the mounted heads of all the big game he shot,” Bannon said.

In a sense, he is right. During the inauguration ceremony, tech billionaires—including Musk, Meta CEO Mark Zuckerberg, the Amazon founder Jeff Bezos, Google CEO Sundar Pichai, and Apple CEO Tim Cook—sat directly behind Trump’s family on the dais. They are not all as forcefully pro-Trump as Musk, but they have cozied up to the president by dining with him at Mar-a-Lago and making million-dollar donations to his inaugural fund (in some cases from their personal bank accounts, and in others from the corporations they head).

In doing so, they’ve gotten his ear and can now influence the president in ways that might not line up with the priorities of the nationalist right. On Monday, during his first press conference from the White House this term, Trump defended the H-1B visa program: “We want competent people coming into our country,” he said. Later, Bannon responded on his podcast, lamenting the “techno-feudalists” to whom Trump is apparently listening.

Both factions still have overlapping interests. They are both fed up with a country that they see as having grown weak and overly considerate to the needs of the vulnerable, at the expense of the most productive. America lacks a “masculine energy,” as Zuckerberg recently put it. Some members in both camps seem interested in trying to reconcile their differences, or at least in not driving the wedge further. On the eve of the inauguration, just before Turning Point USA’s ball, the right-wing publishing house Passage Publishing held its own ball in D.C.—an event intended to be a night when “MAGA meets the Tech Right.” The head of Passage Publishing, Jonathan Keeperman, has been keen on playing peacemaker. Last month, he went on Kirk’s podcast and tried to frame the debate over visas as one where his reactionary, nativist wing of the right could find common cause with the tech right. By limiting immigration and “developing our own native-born” STEM talent, he said, Silicon Valley can “win the AI arms race.”

Kirk couldn’t keep his frustration toward the tech elite from seeping out. “Big Tech has censored us and smeared us and treated us terribly,” he said. “Why would we then accommodate their policy wishes?” It’s easy to imagine Musk asking the same question. He and his peers run some of the most powerful companies in the world. They’re not going to give that up because a few  people, on the very platforms that they own, told them to. Each side is steadfast in what it wants, and won’t easily give in.

We already can guess how this will end. During his first administration, despite making populist promises on the campaign trail, Trump eventually sided with the wealthy. Bannon, Trump’s chief strategist during the start of his first term, pushed for tax hikes on the wealthy. Seven months into his presidency, Trump fired him, and then proceeded to pass tax cuts. In his new administration, the nationalist right will certainly make gains—it is thrilled with Trump’s moves around birthright citizenship and his pledge to push forward with mass deportations. But if it’s ever in conflict with what Trump’s rich advisers in the tech world want, good luck.

Remember, it was Zuckerberg, Bezos, and Musk who sat on the dais at Trump’s inauguration. Bannon, Keeperman, and Kirk were nowhere in sight.

Trump’s Second Term Might Have Already Peaked

The Atlantic

www.theatlantic.com › politics › archive › 2025 › 01 › trump-inauguration-executive-orders › 681403

Ever since Donald Trump emerged as a credible threat to return to the White House, the guardrails that seemed to restrain him in his first term—political, legal, psychic—have collapsed with astonishing speed. His nominees are sailing through their confirmation hearings, including some who are underqualified and ideologically extreme. Titans of business and media are throwing themselves at his feet as supplicants. He has obliterated long-standing norms, unashamedly soliciting payoffs from corporations with business before the government. (The Wall Street Journal reports that Paramount, whose parent company needs Trump’s approval for a merger, is mulling a settlement of one of his groundless lawsuits.) Steps that even his allies once dismissed as unthinkable, such as freeing the most violent, cop-beating January 6 insurrectionists, have again reset the bar of normalcy.

These displays of dominance have convinced many of Trump’s critics and supporters alike that his second term will operate in a categorically different fashion from the first. Where once he was constrained by the “deep state”—or, depending on your political priors, by the efforts of conscientious public servants—Trump will now have a fully subdued government at his disposal, along with a newly compliant business and media elite. He will therefore be able to carry out the sorts of wild policy objectives that failed to materialize during his first term.

The earliest indications, however, suggest that this might prove only half true. Trump has clearly claimed some territory in the culture wars: He is now dancing with Village People in the flesh, not merely to a recording of the group’s most famous track. And when it comes to getting away with self-dealing and abuses of power, he has mastered the system. But a politician and a party that are built for propaganda and quashing dissent generally lack the tools for effective governance. As far as policy accomplishments are concerned, the second Trump term could very well turn out to be as underwhelming as the first.

Trump has promised a grand revolution. At a pre-inaugural rally, he announced, “The American people have given us their trust, and in return, we’re going to give them the best first day, the biggest first week, and the most extraordinary first 100 days of any presidency in American history.” He branded his inauguration “Liberation Day,” labeled his incoming agenda a “revolution of common sense,” and boasted, “Nothing will stand in our way.” After being sworn in on Monday, he signed a slew of executive orders in a move that has been termed “shock and awe.”

[David A. Graham: The Gilded Age of Trump begins now]

Those orders fall into a few different categories. Some are genuinely dangerous—above all, the mass pardon of about 1,500 January 6 defendants, which unambiguously signals that lawbreaking in the service of subverting elections in Trump’s favor will be tolerated. Others, including withdrawing from the World Health Organization and freezing offshore wind energy, will be consequential but perhaps not enduring—that which can be done by executive order can be undone by it.

What’s really striking is how many fall into the category of symbolic culture-war measures or vague declarations of intent. Trump declared a series of “emergencies” concerning his favorite issues, just as Joe Biden had. His order declaring an end to birthright citizenship seems likely to be struck down on constitutional grounds, although the Supreme Court can always interpret the Fourteenth Amendment’s apparently plain text as it desires. He is re-renaming a mountain in Alaska—which, in four years’ time, could be renamed yet again, perhaps after one of the police officers who fought off Trump’s insurrection attempt. He has ordered the federal government to officially recognize only two genders, male and female. “You are no longer going to have robust and long drop-down menus when asking about sex,” an incoming White House official said. Ooooh, the federal intake forms will be shorter!

Meanwhile, Trump has already scaled back many of his most grandiose day-one promises from the campaign. Broker an end to the Ukraine war before taking office? He has “made no known serious effort to resolve the war since his election,” The New York Times reports. Ask again in a few months. Bring down grocery prices? Never mind.

Trump’s supporters probably realized that some of his campaign pledges were hyperbolic. Even by realistic standards, however, Trump seems unprepared to deliver on some of his biggest stated goals. Take his signature domestic policy. Trump loudly promised throughout the presidential campaign to impose massive global tariffs once he took office. And yet, even that proposal remains theoretical. Trump’s executive order on trade instructs, “The Secretary of Commerce, in consultation with the Secretary of the Treasury and the United States Trade Representative, shall investigate the causes of our country’s large and persistent annual trade deficits in goods, as well as the economic and national security implications and risks resulting from such deficits, and recommend appropriate measures,” and then proceeds to issue more solemn calls for study of the matter.

Presidents don’t always come into office with fully formed plans, but Trump doesn’t even have concepts of a plan, or any way to resolve fundamental tension between his belief that foreign countries should pay tariffs and the reality that tariffs raise prices for Americans. Another White House document announces, “All agencies will take emergency measures to reduce the cost of living.” What measures? We can be fairly sure that there is no secret plan waiting to be unveiled.

None of this is to say that Trump will accomplish nothing. At a minimum, he will restrict immigration and sign a regressive tax cut. But even his policy successes will likely sow the seeds of a thermostatic backlash in public opinion. Americans favor mass deportation in the abstract, but their support dwindles when they contemplate specifics. An Axios poll found that strong majorities oppose separating families, employing active-duty military to locate undocumented immigrants, and using military funds to carry out immigration policy. Even some high-level Trump allies have warned that mass deportations will cause immediate economic disruption.

Trump’s fiscal agenda is where the desires of his wealthy benefactors, the preferences of his voters, and economic conditions will clash most violently. The previous two Republican presidents to take office—George W. Bush in 2001, and Trump in 2017—inherited low inflation and low or falling interest rates. Both were able to cut taxes and raise spending without facing any near-term economic costs. In his second term, Trump faces an economy that, while growing smartly, is still plagued with high interest rates relative to the pre-COVID norm. If Trump revises the old playbook of cutting taxes now and worrying about the cost later, he may discover that “later” happens right away.

One answer to the dilemma would be to pay for tax cuts with deep cuts to social spending on the poor, a staple of past Republican budgets. Yet Trump’s strength with low-income voters turns that maneuver into another potential source of backlash. Last month, The Washington Post’s Tim Craig interviewed low-income Trump voters in a poor town in Pennsylvania who earnestly believe that he will not touch their benefits.

[Russell Berman: What Trump can (and probably can’t) do with his trifecta]

Meanwhile, some of Trump’s most prominent backers refuse to acknowledge that any tough choices await. In a recent interview, the New York Times columnist Ross Douthat presented Marc Andreessen, one of the Silicon Valley billionaires hoping to influence Trump’s domestic agenda, with concerns that Elon Musk’s plans to cut the budget would alienate voters. In response, Andreessen insisted that the very suggestion reflected “absolute contempt for the taxpayer,” repeating versions of the line rather than engaging with the problem. Musk himself recently reduced his goal of cutting $2 trillion from the budget to a mere $1 trillion. When the brains of the operation are picking arbitrary round numbers and then revising them arbitrarily, one begins to question their grasp on the challenge they face.

Whether Trump pays any political price for failing to deliver on unrealistic promises—or for succeeding at delivering on unpopular ones—is an open question. Political difficulties won’t generate themselves. They will require an energetic and shrewd opposition. And a major purpose of Trump’s maneuvers to intimidate corporate and media elites is to head off a backlash by gaining control over the information environment.

One of Trump’s greatest strengths as a politician is to constantly redefine his policy goals so that whatever he does constitutes “winning.” The success of this tactic reflects the degraded intellectual state of the Republican Party’s internal culture. The conservative movement rejected institutions such as academia and the mainstream media decades ago, building up its own network of loyal counterinstitutions that would construct an alternate reality. This has helped Republicans hold together in the face of corruption and misconduct that, in a bygone era, would have shattered a governing coalition. (Today, Watergate would just be another witch hunt.) But the impulse to disregard expertise and criticism has also disabled Republicans’ ability to engage in objective analysis. The past two Republican administrations accordingly both ended in catastrophe, because the president had built an administration of courtiers who flattered his preexisting beliefs, whether about weapons of mass destruction and Iraq or COVID and the economy.

[George Packer: The end of democratic delusions]

None of those pathologies has disappeared. To the contrary, the MAGA-era GOP has grown more cultlike than ever. The rare, feeble attempt to steer Trump away from bad decisions is usually buried in obsequious flattery. The Trump presidency will be, by definition, a golden age, because Trump will be president during all of it. But it is a measure of his allies’ decrepitude that, whatever positions he ultimately lands on, they are prepared to salute.

Trump has struck fear into his party and America’s corporate bosses. His inauguration was a display of mastery, a sign that none will dare defy his wishes. But a leader surrounded by sycophants cannot receive the advice he needs to avoid catastrophic error, and to signal that his allies can enrich themselves from his administration is to invite scandal. In his inaugural spectacle of dominance and intimidation, Trump was planting the seeds of his own failure.

Trump’s Anti-Immigrant Coalition Starts to Fracture

The Atlantic

www.theatlantic.com › podcasts › archive › 2025 › 01 › trumps-anti-immigrant-coalition-starts-to-fracture › 681257

This story seems to be about:

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

Last month, Donald Trump appointed the venture capitalist Sriram Krishnan as his senior AI-policy adviser. Krishnan, an Indian immigrant and U.S. citizen, was seen by some as being friendly to H-1B visas, which are often used in Silicon Valley to allow skilled laborers to work in the tech industry. This sent part of the MAGA faction into a frenzy, spurred by troll in chief Laura Loomer, who declared the appointment a betrayal of the “America First” movement.

The argument over H-1Bs exposes an important fissure in the MAGA alliance that worked together to help elect Trump. How Trump navigates this rift will give us clues about what his real priorities will be as president.

In this episode of Radio Atlantic, we talk with Ali Breland, who writes about the internet, technology, and politics, about this new rift in Trump’s camp and other places it might show up. And we’ll go beyond the politics, with staff writer Rogé Karma, to discuss what a solid body of research shows about the relationship between immigrant labor and the American worker—because even though some prominent Democrats, such as Bernie Sanders, agree with Loomer that there is a negative effect from H-1B visas on American workers, research doesn’t back them up.

The following is a transcript of the episode:

Hanna Rosin: There are already cracks starting to show in the MAGA alliance, and those cracks happen to show up in the issue that Trump has declared one of his top priorities, which is drastically reshaping U.S. immigration policy. Trump appointed to a senior position someone seen as being friendly to H-1B visas, the visas that allow people with specialized skills to work in the U.S. People in Silicon Valley love these visas. They depend on them. And maybe more importantly, the H-1B visa lovers include Elon Musk.

But the “America First” wing of Trump supporters—sometimes known as the nativist right—they do not love these visas. “America First,” to them, means, literally, Americans first. No exceptions.

I’m Hanna Rosin. This is Radio Atlantic. On today’s show, we’ll talk about this MAGA infighting. In the second half of the show, we’ll get into what’s actually true about the relationship between immigration and the American worker, because it turns out that even a lot of Democrats don’t get that one right. But first, let’s dive into the recent news and what it means. To help me with that is Ali Breland, an Atlantic staff writer who writes about the internet, politics, and technology.

Hey, Ali.

Ali Breland: Hey. Thank you for having me.

Rosin: So, Ali, this fracture in the MAGA alliance seemed to start around Christmas, when Trump announced a senior AI-policy adviser. Who is he, and how did people respond?

Breland: Yeah, his name is Sriram Krishnan. He’s this Silicon Valley figure who has a long history. He works in tech, and he was being appointed to be an adviser on Trump’s AI team, which is being headed up by another big guy in tech: David Sacks, who’s a part of the infamous “PayPal Mafia” that includes Peter Thiel, Elon Musk, etcetera.

Rosin: So these are, like—this is a faction. Like, these guys are becoming more and more powerful, sort of Trump’s tech allies.

Breland: Yeah, there’s some different ideological things happening, but for the most part, they’re largely on the same page. And a lot of people right now are kind of calling them the “new tech right,” or just, like, the “tech right.”

Rosin: So they’re on one side, and then how did the discussion around H-1B visas get going?

Breland: Yeah, so there’s this provocateur troll in Trump World called Laura Loomer. She’s been kind of this weird thing on the right for a long time. She’s chained herself to the headquarters of Twitter in protest of her account being banned at one point. But she sees this appointment, and she decides to make hay of it.

She pulls out a tweet that Krishnan made about country caps for green cards, rather, and high-skilled immigration. And she points to these things and says, This is not what we want. This is not “America First.” These things are not good for our constituency. And so that’s, like, the sort of obvious bit of it.

The other bit, too, is you can kind of see how race is this animating issue in this fight. David Sacks had already been appointed by Trump to be his chief adviser on issues of AI and crypto. David Sacks has talked about H-1B visas. He’s pushed Trump on this. He’s successfully gotten Trump to say that he would support the continued use of H-1B visas.

But Loomer didn’t attack him on that and didn’t turn this into a huge issue. Instead, she went after Sriram Krishnan, who is South Asian. And I think, you know, her targeting him, specifically, on this issue and associating him with that kind of speaks to the sort of nativist sentiment undergirding all of this.

Kind of right after the election, I sort of thought that maybe there was a chance that there was going to be some sort of fractious element at some point in the future, because these are two sides that kind of believe sort of different things.

The tech right is reactionary, like the nativist right that includes people like Laura Loomer, people like Steve Bannon. They sort of all have this streak of being frustrated with the progress that’s taken place in America. They are frustrated with what they see as, like, American weakness. But the distinction is that the tech right also loves business. They love being rich. They love making a lot of money and having their industry be benefitted.

The sort of nativist right cares much more about the American constituency and, specifically, the white American constituency—and benefitting what they see as, like, the natural order of whiteness and the average American, and things that some people in the tech right kind of care about but prioritize less than their own companies and less than their own industry.

Rosin: It’s really complicated because they both have ideas like, There’s an optimum society; there’s a right way that things should be. And then they’re slightly different. So what is each side’s ideal “America made great again” look like?

Breland: Yeah, I think it on the sort of nativist right, the ideal America is this place that prioritizes—with some exceptions, more so now—but fundamentally, it’s this white, sort of very classic, conventional, conservative vision for what the United States is. It’s this, like, return fantasy to a version of the 1950s America that prioritizes white American interests above other people—again, with exceptions. There’s—you know, these people would all say that they’re not racist, that they’re just meritocratic, or things like that.

The tech right is more agnostic to those kinds of things. People like Marc Andreessen and Peter Thiel kind of, to some degree, see value in that. But they only see value as far as that doesn’t get in the way of their vision for creating this sort of all-star team of Americans that can sort of dominate the global stage in technology and dominate economically.

And so they’re willing to go to look to other countries to bring people in; to try to, like, get the best talent, according to them; to try to solve the toughest engineering problems; and to do things like beat China, which is something that they’re all very obsessed with.

Rosin: So they’re less concerned about where people come from. I mean, what makes it especially complicated and charged that this came up so soon is that it came up in immigration. Trump has made controlling immigration one of his top priorities. How did Trump himself end up weighing in on this?

Breland: After a few days of silence—perhaps because this was happening literally over Christmas and the days after—Trump did say that he does support H-1B visas. And he seemed to kind of take Elon’s side on this.

I wasn’t super surprised, because on an episode of the All-In Podcast—which is a sort of who’s who of the tech right; it includes David Sacks—Trump was pressed on the H-1B visa issue, and he did say, Yeah, I support it; I’m down for this. This was in the summer. And so it was consistent for him to come back up with this. And the other thing it’s sort of consistent with, in a sort of more general, patterny kind of way, is that in the past, when there is sort of tension between his sort of more nationalist, nativist base versus the wealthier interests that are in his coalition—not always, but—he often tends to go with the sort of interests of the wealthy, the people who have given him the most amount of money, people who he probably respects because he has a great deal of respect for people who have built wealth.

And so it wasn’t super surprising to see him break that way, especially because it seems like his larger immigration priority is not regarding H-1Bs, and he seems more flexible on that. His larger immigration priority is people who, as he would say, came here illegally and are not quote-unquote “high-skilled workers.”

And so on the sort of issue of mass deportation, this doesn’t signal that he’s, like, going to break from that at all. He’s talked a lot, very aggressively, about conducting mass deportations and quote-unquote “securing the southern border.” And they talk about the southern border, specifically, because they’re talking about a different kind of immigrant, and they have a different set of priorities when it comes to people coming across the southern border.

Rosin: Interesting. So then, maybe, the thing to explore is the nativist right, not just Laura Loomer. Laura Loomer is, you know, a little more on the fringes. But what about someone like Stephen Miller, who will be Trump’s deputy chief of staff for policy and who is credited with shaping a lot of the more draconian immigration policies in the last administration. He has solid power in this administration. Have we heard from him or someone closer to power about what they think about H-1B visas?

Breland: Miller hasn’t weighed in directly on this specific moment and this specific issue. He sort of gave a cryptic tweet that signaled that he is still anti-H-1B.

But he’s been very consistent on this in the past, and there’s no reason to believe that he would change, as someone who is, like, motivated primarily by this sort of nativist perspective that is, again, sort of galvanized by racial animus and, in many cases, just outright racism. I don’t think he’ll change his perspective, and he’s going to fight on this, and so there’s going to be weird tension moving forward.

Elon seemed to—I don’t want to say he walked back from this position, but, like, after a few days of fighting, he did seem to try to want to soften the blows and sort of extend an olive branch. People in sort of fairly influential but niche figures in this sort of nationalist, reactionary wing of the party also tried to sort of smooth over the tension and make it seem like there was common cause being found. And so they have an interest among themselves in trying to come together and paint themselves as a united front and sort of reach a consensus on this.

Rosin: Yeah, I mean, it’s still early. He hasn’t even taken office yet. But could you imagine a universe where, then, it just moves forward, and we quietly make an exception for elite workers and do mass deportations for everyone else? Like, is that where immigration policy could land?

Breland: Yeah, exactly. I mean, I think that—from my perspective and the things I pay attention to—that seems exactly the direction it’s going to go in.

The tech right is aware of the mass deportations [but] has not really talked out against them. Elon Musk has tweeted acknowledging them and sees them as an inevitability that he doesn’t seem to have a clear problem with. That could change when we sort of get, like, harrowing images of ICE conducting raids and things like that, but right now, that’s the track that we’re on.

Rosin: So if what you said is true, and if the past history holds, he is going to make an exception for elite immigrant workers. What does that imply about how he might handle other economic issues?

Breland: Yeah, if we extrapolate this out, which we can—both from this example but then, also, from how 2016 through 2020 went—Trump is probably going to side, I guess, with more of the wealthier faction, which includes the tech right, which includes people in his coalition, who are people like the hedge-fund manager Scott Bessent, who also sort of have this prioritization of more, like, economically laissez-faire issues. They have this sort of more traditional, conservative perspective on economics. And that’s something that’s going to run into tension with what the nationalists want. They want this sort of economic nationalist perspective that is a departure from this hyper-free-market sort of way of viewing the world that’s been the dominant conservative perspective for the past several decades.

Rosin: So essentially, this rift that you pointed out in the MAGA world—between, you know, Is he going to take the side of the elites, or is he going to take the side of all the workers? even if that means the nativist right—that’s a rift you can track kind of up and down various issues for the next many years, just to see, Okay, whose side does he take on a lot of these issues?

Breland: Exactly. Yeah. AI and automation is going to be a really big one in this area, too, because the tech right obviously cares a lot about AI and automation. They’re very pro-AI and automation. They see this as, like, an existential issue in the United States versus China, and that the U.S. must—to continue its being, like, the most important country in the world—that must beat China on this.

But a lot of the sort of more nationalist right doesn’t agree with this. They see this as a different kind of issue. Tucker Carlson, who I think kind of squarely falls in this nativist camp and is one of its most influential members, has outright said that he opposes—not necessarily the development of AI and automation but—its implementation and use.

He’s talked directly about never using AI for, like, things like driverless trucks. But Elon at Tesla is directly making self-driving trucks. And so yeah, there’s a lot of weird places where these sort of fractures are going to play out.

Rosin: And Tucker Carlson takes that issue because it’s a betrayal of the American worker.

Breland: Precisely.

Rosin: Interesting. So this is, actually, the central fissure of the Trump administration, basically?

Breland: Yeah. Yeah, it seems like that. I do want to say that this is kind of a unique issue, in that it draws in race, which is a very big thing, and it draws in immigration. And so it might get a uniquely high amount of attention. But there’s still going to be versions of this fight that might not play out as aggressively that are going to happen over the next four years.

Rosin: Well, Ali, thank you for pointing out this line to us. We’ll be watching it for the next four years, and thank you for joining me.

Breland: Yeah, thank you so much for having me. I appreciate it.

Rosin: After the break, we explore what’s behind the politics. Trump and his allies made the argument often in the campaign that immigrants take away jobs from Americans. It’s an argument that, on the surface, has some intuitive logic. But it actually doesn’t work like that. More soon.

[Break]

Rosin: Joining me is Atlantic staff writer Rogé Karma, who mainly covers economics. Rogé, welcome to the show.

Rogé Karma: It’s great to be here. Thanks for having me.

Rosin: Sure. So an early rift broke out in the Trump administration over H-1B visas, which we’ve been discussing on this show, with the nativist right saying what people say about all kinds of immigration: These immigrants take jobs away from American workers. So what do we know about the relationship between H-1B visa holders and the American worker?

Karma: Well, luckily, the H-1B program allocates workers randomly to companies based on a lottery. And that allows researchers to study what actually happens to the companies that did get workers, as opposed to the companies that didn’t.

And I agree with you. I think there’s a real sort of “man on the street” argument. There’s a sort of view that there’s a fixed pool of jobs, and so any immigrant that we bring in is going to take away a job that would otherwise go to an American. But when researchers have looked at this, the overwhelming majority of the studies have actually found no negative impact on either employment or wages, which I think at first sounds a little bit counterintuitive.

But the reason is a few fold. One: Companies who get H-1B workers actually end up growing and scaling up faster than the companies who don’t. And then because of that, they have to then hire a bunch of more native-born workers around that immigrant. The second reason is innovation.

One of my favorite statistics comes from Jeremy Neufeld, who’s a fellow at the Institute for Progress. And he pointed out that 30 percent of U.S. patents, almost 40 percent of U.S. Nobel Prizes in science, and more than 50 percent of billion-dollar U.S. startups belong to immigrants. Now, not all of those are H-1B holders, but there’s a lot of evidence that the companies who are awarded H-1B visas—they produce more patents, more new products, get more VC funding, and all of that actually creates jobs. So on the whole, I actually don’t think there’s a lot of evidence for this broader nativist claim about this program.

Rosin: Let’s make this a little more concrete. So let’s just play out a theoretical company. Here’s a theoretical company, hires H-1B visa holders. How does it work? Like, innovation is a vague word. How does it actually play out?

Karma: I think what’s important to remember here is that getting one of these H-1B visas is actually pretty difficult. And so the idea that a company is going to be able to systematically bring in foreign workers to replace their native ones using this program—it’s just really hard to do because there’s such a low chance they’re even going to get those workers in the first place. And so a lot of times when companies use this program, what they’re doing is they’re looking for a very important skill set.

So let’s use semiconductors as an example. This is an industry, when it comes to the manufacturing of semiconductors, that U.S. companies haven’t really done for a while. A lot of the most advanced chips are made in places like Taiwan, and so a lot of the best talent is abroad. And so if you’re a U.S. semiconductor manufacturer, the industry in the U.S. estimates that even if we had the best job-training programs possible, that would only fill about 50 percent of the high-skilled demand for the labor force in this field.

And so you need to bring in folks who have this highly specialized knowledge, probably because they’ve worked in other countries. But then, what that allows you to do, once you have a subset of foreign-born workers who can do this sort of specialized manufacturing—what you then have is people to come in and support around them. And then because a company has that need met, they’re able to then hire a bunch of other workers to fill other needs that they have but that don’t require that same kind of specialized knowledge.

And on the other flip side is that we actually have some studies that look at: What happens to the companies that don’t get H-1B visas? What happens to those companies? Do they hire more native workers? Do they invest in more job training? And it turns out that they don’t. In fact, they end up often just either (A) producing less or growing less quickly, or (B)—and this is a finding of a lot of the recent literature—they end up outsourcing the jobs instead. And so instead of bringing in this new worker and then hiring more native workers around them, they just say, Well, look, we have an office in China, or we have an office in Singapore, or we have an office in Hong Kong or India. Let’s just hire more there because we’re not going to be able to get the talent that we need here.

There are a handful of outlier studies, but I think, right now, the broad consensus in the field is that the H-1B program, even for all its flaws, doesn’t seem to have these negative employment or wage effects.

Rosin: So that’s what the research shows. It’s fairly definitive until now, and yet even some Democrats have repeated the line, The H-1B visas take away American jobs—for example, Bernie Sanders. What do you make of that?

Karma: Well, I think where Bernie’s coming from—and I think where a lot of Democrats are coming from and, quite frankly, some Republicans—is that there are two things that are true here at once. The first thing that’s true is that we don’t find these huge negative effects from the H-1B program. And the second thing that’s also true is that, despite that, the H-1B program has a lot of flaws, a lot of loopholes that companies have learned how to game.

So one of these is that a significant portion of H-1B visas are used by so-called outsourcing firms, which are these companies that basically bring in foreign workers. They train them here, and then, when their H-1B visa expires, they employ them in their home countries for a fraction of the cost. And so they’re functionally using the H-1B visa to train workers here and then employ them at lower labor costs elsewhere.

That’s just bad, on the face of it. The fact that we still don’t see negative effects, overall, is really telling, but we should fix that loophole by, among other things, raising the minimum wage for H-1B visa holders, making the program merit-based instead of random—like, you can more closely regulate how companies use those workers.

So I think part of what Bernie Sanders is getting at, part of what some of these critiques are getting at, is that this program does have a lot of flaws that allow corporations to game it. And it’s actually kind of shocking that, despite all these flaws, it still hasn’t produced these horribly negative results.

But imagine how much better it could be if we fix them. So I really think that this might be a place where you see the sort of messy realities of immigration politics running up against what, really, people all across the political spectrum agree is a pretty common-sense set of reforms. But that doesn’t always mean it makes good politics.

Rosin: Right. Right. Okay. So we’ve been talking exclusively about the H-1B visas because they came up in the news, but the whole of Trump’s promise is not specifically about H-1B visas at all; it’s a promise of mass deportation and immigrant labor, in general. I know that you’ve been looking into the research about the relationship between immigrant labor and the American worker. What did you find?

Karma: Well, I went into this because I kept hearing Donald Trump, J. D. Vance, Stephen Miller make these kind of claims that sound kind of intuitive—that when immigrants come in, they take jobs from natives, right? There’s a sort of Econ 101 logic, which says that when the supply of any good goes up, including labor, the price of that good, like wages, goes down.

And so I kept hearing these arguments and thinking, Well, maybe there’s something to this, and so let’s actually look at what is happening. And it turns out that the sort of Trump-Vance view was pretty much the conventional wisdom for most of the 20th century, both among policymakers and economists, until a study came along that sort of shattered the consensus.

And so to tell you about the study, I’m gonna go back a little bit. So in 1980, Fidel Castro, the president of Cuba, opened up emigration from his country. He lifted the ban on emigration. And what that allowed is for 125,000 Cubans to leave from Mariel Harbor to Miami, Florida, an event that ended up becoming known as the Mariel Boatlift. And in just a few short months, Miami’s workforce expands by about 25 times as much as the U.S. workforce expands every year because of immigration. And this created the perfect conditions for what economists call a “natural experiment.” It was like this big, massive shock that only happened to Miami.

And so what the economist David Card later realized is that you could compare what happened to workers in Miami to workers in other cities that had not experienced the boatlift, track how wages did in both, and then see what actually happened. And his view was, Look—if there is a negative effect of immigration on wages, Miami in the 1980s is exactly where it should show up. It’s this big, unprecedented shock. That makes what he ended up finding so shocking, because he ends up finding that this huge influx of immigrants has virtually no effect on both employment or wages of native-born workers in Miami, including those without a college degree.

Rosin: And why? I mean, it seems counterintuitive.

Karma: It seems completely counterintuitive. There are a few reasons, but I think the big one—and the big thing that the common-sense view of immigration misses—is that immigrants aren’t just workers. They’re also consumers. You know, they’re people who buy things, like healthcare and housing and groceries. And so at the same time that they’re, you know, competing with Americans for jobs, they’re also buying lots of things that then increase the need for more jobs.

And I think this sounds counterintuitive, but we think about it in other contexts all the time, right? When’s the last time you heard a Republican politician railing against the upcoming group of high-school graduates because they were about to come in and compete with, you know, people currently in the workforce?

You probably haven’t, because we understand that population growth has these two sides to it: that people are consumers who create demand for jobs and workers who take jobs. And so I think that’s the gist of the problem with the conventional view.

Rosin: So that was a singular study. Has that held up over time?

Karma: It has. And so after that study, it got a lot of researchers interested, and this has now been studied in countries all over the world, from Israel to Denmark to Portugal to France, and almost all of the high-quality studies come back with very similar results.

I think the one complication in all of this—the one challenge—has been, Well, what about the least-skilled workers? What about: Okay, maybe on average, immigrants don’t hurt the employment prospects or the wages of native-born workers, but what about the least-skilled workers? What about high-school dropouts, folks without a high-school diploma? And a lot of the more recent literature has shown that even that group doesn’t suffer when immigrants come in.

And so I think the broad consensus in the literature now is that immigration does have costs. It can exacerbate inequality. Tellingly, the wages of other immigrants often get hurt by new immigration. You could see some negative effects in certain sectors, even if it’s balanced out by other sectors, but on the whole, it appears to be really beneficial for basically all classes of native workers.

Rosin: So at this point, there’s a large body of research saying the arrival of immigrants—even sudden arrival of immigrants—doesn’t have a great effect on the American worker, may even have a positive effect. Now, what about the disappearance of immigrant labor? Because Trump’s promise is mass deportations. I’m not sure if you can just flip, you know, the findings of this research. Like, is there a similar natural experiment or study that shows how that might affect workers or the economy?

Karma: There is, actually. And I think the claim from Trump and his advisers is that the ultimate pro-worker policy is mass deportation, right? Because what happens when you get rid of a bunch of immigrant laborers is now those employers have to hire natives at higher wages, because there’s a sort of artificially created labor shortage.

Rosin: Right.

Karma: And again, very intuitive. But when we actually look at what happens in the real world, we see something very different. So the best study on this, I think—although there’s a few—is from the Secure Communities program, which is a Department of Homeland Security program that between 2008 and 2014 deported about 500,000 immigrants. And because the program was rolled out community by community, it created this really nice natural experiment where you could see what happened to the communities that had experienced it and the ones [that] hadn’t.

You could compare them and see what the overall effect [was]. And what researchers found, actually, shocked me—it shocked many of them—was that for every hundred immigrants that were deported, you actually ended up with nine fewer jobs for natives. That’s not just temporary work. That’s, like, nine jobs permanently gone in this community.

And there are many studies that reinforce this finding from all across history, from the Bracero program, studies on the H-2B program—which is like H-1B, but for lower-skilled immigrants—studies going all the way back to the Great Depression that all find similar things.

And the reason is that immigrants are deeply interwoven into their local economies. And so take the restaurant industry. If you’re a restaurant owner, and suddenly you lose a big chunk of your workforce, to the point where you either have to have higher labor costs and at the same time you have less demand, there’s a good chance you have to go out of business altogether. And when you go out of business, that doesn’t just hurt the immigrants who are working for you. That also hurts the native-born workers.

And so there are all these sort of synchronicities, all of these interconnections, that allow immigration to have this positive sum effect. But then as soon as you—if you rip out the immigrants, then native workers often get caught in the crossfire.

Rosin: Yeah. So if the research is so consistent—so strong—and makes a lot of sense, if you think about it a tiny bit more deeply, why do you think this sentiment persists? Is it just a feeling, you know? Because it persists on both the right and the left. It’s not as if the left is fighting back. They don’t necessarily advocate mass deportations, but they are also not fighting back against this idea that immigrants take away American jobs.

Karma: I think part of the fixation on the economics of immigration is a way for many people like us—elites, people in the media—to try to find a more materialist explanation for a set of instincts that I think many of us are uncomfortable with. And I think that is actually kind of a tragedy.

I think if people oppose immigration or feel strongly about immigration because of certain cultural beliefs or concerns about national identity, it’s important to take those concerns seriously. And I think it’s actually a problem, and even a bit patronizing, that we tend to project these sort of more wonky economic concerns onto that.

Rosin: Yeah. I had a conversation with Representative Ritchie Torres of New York right after the election, who talked about how a lot of the immigrants in his neighborhood had a surprising amount of anti-undocumented immigrant sentiment.

And it made me wonder about—I don’t even know how to define this, but sense of chaos, just a feeling of things not being in control. It’s sort of the way people feel about crime. There just seems to be a sense that things have run away, and you can’t get ahead. It’s a vague thing, but it is related to—There’s just so much out of control, and I need someone to stop it.

Karma: I actually think that’s a really important point. One of the greatest shifts in public opinion on immigration has happened in the last few years, where in 2020, according to Gallup, only 28 percent of Americans said they wanted immigration decreased.

Four years later, that number was 55 percent. So it had almost doubled. And that is much larger and much faster than even the public-opinion shift on something like gay marriage. So this is a huge, almost unprecedented shift. And as I dug into why, what came up over and over again is this feeling of chaos, this feeling that we are not in control of our own border. And when you actually look at questions about how people feel towards immigrants themselves, they hadn’t changed nearly as much.

People weren’t necessarily anti-immigrant, as much as they felt like the immigration process had gotten out of control and the immigration process was no longer serving the country. And so I think it is really important to distinguish [between] those two things. And I think a lot of the public-opinion shift we’ve seen over the last few years—it isn’t about economics. It’s really about this sense of control and chaos.

Rosin: Yeah. So maybe the place to end is this: Have you talked to anyone or done any thinking about how, in a situation like this, you close the gap? Because we, as journalists—it’s frustrating to us to know that there is an answer. You know, there’s an answer that research has provided. There are truths and facts. And separate from that, there is a perception. So have you thought of or seen anybody talk interestingly about how you bridge a gap like that, where people feel one way that is discordant with what the reality is?

Karma: Unfortunately, like any good journalist, I’m not quite as good at the solutions as I am about identifying the problems. But I will say, I think at the root of a lot of this is the fact that there’s an underlying scarcity. Right?

So I think an example of this is housing. Recently—you know, we haven’t talked about this, in particular—but J. D. Vance and Donald Trump made a big deal in their campaign about how immigrants were responsible for driving up housing costs. That argument has never held weight in American politics before, because it is only over the last decade that housing costs and a housing shortage has become a big problem. When there is material scarcity, people look for a villain; people look for someone to blame. And so I think one answer to, for example, the blaming [of] immigrants for housing costs is to say, Well, if we fix the housing shortage such that people don’t feel that scarcity, maybe we can avoid some of that.

I think the other sort of way I’d look at this is: In some senses, one of the most pro-immigrant things you could do is reduce the amount of chaos, right? So I think there’s actually a sort of middle ground here where you could reduce a lot of the chaos at the border while expanding legal immigration in a way that keeps immigrants coming in but creates a more orderly process that people feel comfortable with. And you can actually get more positive sentiment as a result.

I just think what makes it difficult is the politics are almost perfectly aligned to make that difficult from happening. And it’s been, you know—immigration reform is something that politicians have been talking about for more than 20 years now, and it hasn’t happened.

Rosin: Well, that was really helpful. Rogé, thank you so much for joining me today and talking about this.

Karma: Thank you so much for having me. It was a pleasure.

Rosin: This episode was produced by Kevin Townsend and edited by Claudine Ebeid. Rob Smierciak engineered, and Sara Krolewski fact-checked. Claudine Ebeid is the executive producer of Atlantic audio, and Andrea Valdez is our managing editor.

My thanks to Ali Breland and Rogé Karma for joining me. If you’d like to hear Rogé go even deeper on the research into immigration’s economic impact, you can hear him on another Atlantic podcast called Good on Paper. It’s hosted by staff writer Jerusalem Demsas, and that episode is linked in the show notes.

I’m Hanna Rosin. Thank you for listening.