Tools
Change country:
theatlantic.com
theatlantic.com
The Loyalists Are Collecting Their Rewards in Trump’s Cabinet
This is an edition of The Atlantic Daily, a newsletter that guides you through the biggest stories of the day, helps you discover new ideas, and recommends the best in culture. Sign up for it here.A note from Tom:As we were about to publish this newsletter, Donald Trump announced that he has asked the Fox News personality Pete Hegseth, a military veteran who has no experience in leading large organizations and no serious background as a senior leader in national-security affairs, to be his secretary of defense. This is exactly the kind of unqualified nomination that I was warning could be looming after this first group of nominees were announced—and it explains why Trump is determined to bypass the U.S. Senate to get some of his nominees confirmed. I will have more to say about Hegseth soon.So far, the new Trump administration has a chief of staff, a “border czar,” and a national security adviser; all three are White House positions controlled by the president. Donald Trump has also reportedly named six people to senior positions that require Senate confirmation: secretary of state, United Nations ambassador, secretary of homeland security, secretary of defense, CIA director, and administrator of the Environmental Protection Agency. (He has also chosen an ambassador to Israel.) His first picks are neither very surprising nor very impressive, but this is only the beginning.His co–campaign manager Susie Wiles will make White House history by becoming the first female chief of staff. People around Trump seem relieved at this appointment, but she’ll likely be saddled with Stephen Miller as a deputy, which could get interesting because Miller apparently has a tendency to get out of his lane. (According to a book by the New York Times reporter Michael Bender, Miller attended a tense meeting that included Trump, Attorney General Bill Barr, and General Mark Milley, the chairman of the Joint Chiefs of Staff, during the Black Lives Matter protests in 2020. As the nation’s leaders debated what to do, Miller interjected and said that America’s major cities had been turned into war zones. General Milley, Bender writes, turned to Miller, pointed at him, and said: “Shut the fuck up, Stephen.”)The rest of the appointments are unsurprising, given the limited pool of Republicans willing to serve in another Trump administration. (Some Trump loyalists such as Senator Tom Cotton have reportedly declined a role in the administration, likely protecting their future for the 2028 GOP race to succeed Trump.) Marco Rubio, who sits on the Foreign Relations and Intelligence Committees in the Senate, was a reasonable choice among the Trump coterie to become America’s top diplomat as secretary of state.Likewise, Representative Mike Waltz of Florida is a reasonable choice for national security adviser—but again, that’s in the context of the now-smaller universe of national-security conservatives in politics or academia willing to work for Trump at this point. He is a veteran, and like Rubio, he has served on relevant committees in Congress, including Armed Services, Foreign Affairs, and the House Permanent Select Committee on Intelligence. Waltz may be a credible voice on national security, but he was also a 2020 election denier. He pledged to oppose certifying Joe Biden’s 2020 win and signed on to an amicus brief supporting a Texas lawsuit to overturn the election. He changed his mind—but only after the events of January 6.Representative Elise Stefanik of New York, meanwhile, was bound to be rewarded for her loyalty. Although Vice President–elect J. D. Vance took the gold in the race to replace the disowned Mike Pence, Stefanik was a comer even by the standards of the sycophantic circle around Trump, and so she’ll head to the United Nations, a low-priority post for Trump and a GOP that has little use for the institution. A former member of Congress from New York, Lee Zeldin (who was defeated in the 2022 New York governor’s race) will head up the EPA, another institution hated by MAGA Republicans, thus making Zeldin’s weak—or strong, depending on your view—legislative record on environmental issues a good fit for this administration.This afternoon, Trump announced that John Ratcliffe will serve as CIA director. Ratcliffe previously served as director of national intelligence and will now be in a post that is functionally subordinate to his old job. Ratcliffe is a reliable partisan but an unreliable intelligence chief. The most baffling move Trump has made so far is the appointment of South Dakota Governor Kristi Noem to lead the Department of Homeland Security. Noem served four terms in Congress and is in her second as governor. She has very little relevant experience, especially as a government executive. (South Dakota might be a big place, but it’s a small state; DHS has more than 260,000 employees, making it a bit more than a quarter the size of the entire population of Noem’s home state.) DHS is a giant glob of a department—one I have long argued should never have existed in the first place and should be abolished—that has seeped across the jurisdictional lines of multiple institutions and, unlike some other Cabinet posts, requires someone with serious leadership chops.DHS will also be central to some of Trump’s most abominable plans regarding undocumented immigrants—and, potentially, against others the president-elect views as “enemies from within.” (The “border czar” Trump has named, Tom Homan, once falsely implied that some California wildfires were worsened by an undocumented immigrant.) In that light, Noem is perfect: She is inexperienced but loyal, a political lightweight with no independent base of support or particularly long experience in Washington, and she can be counted on to do what she’s told. She will be no John Kelly or Kirstjen Nielsen, her confirmed predecessors at DHS, both of whom were on occasion willing to speak up, even if ineffectively.This first passel of nominees should gain Senate confirmation easily, especially Rubio. (Sitting members of the chamber usually have an easier time, as do people who have close associations with the Senate.) And given Trump’s history and proclivity for mercurial and humiliating firings, few of them are likely to be very long in their post, and are probably better than the people who will later replace them.But that in itself raises a troubling question. If Trump intends to nominate these kinds of fellow Republicans, why is he insistent that the new Senate allow him to make recess appointments?For those of you who do not follow the arcana of American government, Article II of the Constitution includes a provision by which the president can make appointments on his own if the Senate is in recess and therefore unable to meet. The Founders didn’t think this was a controversial provision; sometimes, presidents need to keep the government running (by choosing, say, an ambassador) even when the Senate might not be around—a real problem in the days when convening the Senate could take weeks of travel. Such appointments last until the end of the next legislative session.For obvious reasons, the Senate itself was never a big fan of a device—one that presidents routinely used—that circumvents constitutional authority to confirm executive appointments, especially once the practice got out of hand. (Bill Clinton made 139 recess appointments, George W. Bush made 171, and Barack Obama made 32.) The Senate’s response was basically to be wilier about not declaring itself in recess even when there’s no one around, and when President Obama tried to push through some of these appointments in 2012, the Supreme Court sided with the Senate.Now Trump wants to bring back the practice. The obvious inference to draw here is that after some fairly uncontroversial nominations, he intends to nominate people who couldn’t be confirmed even in a supine and obedient Republican Senate. Perhaps this is too clever, but I am concerned that this first pass is a head fake, in which Trump nominates people he knows are controversial (such as Zeldin) but who are still confirmable, and then sends far worse candidates forward for even more important posts. Kash Patel—a man who is dangerous precisely because his only interest is serving Trump, as my colleague Elaina Plott Calabro has reported—keeps bubbling up for various intelligence posts.“Ambassador Elise Stefanik” and “EPA Administrator Lee Zeldin” might not be great ideas, but they are not immediate threats to U.S. national security or American democracy. “CIA Director John Ratcliffe,” by contrast, is cause for serious concern. If Trump is serious about his authoritarian plans—the ones he announced at every campaign stop—then he’ll need the rest of the intelligence community, the Justice Department, and the Defense Department all under firm control.Those are the next nominations to watch.Related: Trump signals that he’s serious about mass deportation. Stephen Miller is Trump’s right-hand troll. (From 2018) Here are three new stories from The Atlantic: The HR-ification of the Democratic Party Anne Applebaum: Putin isn’t fighting for land in Ukraine. Genetic discrimination is coming for us all. Today’s News The judge in Trump’s hush-money criminal case delayed his decision on whether Trump’s conviction on 34 felonies should be overturned after his reelection. A federal judge temporarily blocked a new Louisiana law that would have required the display of the Ten Commandments in all public classrooms, calling the legislation “unconstitutional on its face.” Louisiana’s attorney general said that she will appeal the ruling. The Archbishop of Canterbury announced his resignation. An independent review found that he failed to sufficiently report the late barrister John Smyth, who ran Christian summer camps and abused more than 100 boys and young men, according to the review. Evening Read Illustration by Mark Pernice AI Can Save Humanity—Or End ItBy Henry A. Kissinger, Eric Schmidt and Craig Mundie The world’s strongest nation might no longer be the one with the most Albert Einsteins and J. Robert Oppenheimers. Instead, the world’s strongest nations will be those that can bring AI to its fullest potential. But with that potential comes tremendous danger. No existing innovation can come close to what AI might soon achieve: intelligence that is greater than that of any human on the planet. Might the last polymathic invention—namely computing, which amplified the power of the human mind in a way fundamentally different from any previous machine—be remembered for replacing its own inventors? Read the full article.More From The Atlantic Good on Paper: A former Republican strategist on why Harris lost Trump’s “deep state” revenge The great conspiracy-theorist flip-flop The two Donald Trumps “Dear James”: How can I find more satisfaction in work? Culture Break The Atlantic; Getty; HBO Max Watch. These 13 feel-good TV shows are perfect to watch as the weather gets colder.Read. “The first thing you need to know about the writer Dorothy Allison, who died last week at 75, is that she could flirt you into a stupor,” Lily Burana writes.Play our daily crossword.Stephanie Bai contributed to this newsletter.Explore all of our newsletters here.When you buy a book using a link in this newsletter, we receive a commission. Thank you for supporting The Atlantic.
5 h
theatlantic.com
How Can I Find More Satisfaction in Work?
My job consumes and torments me. There has to be a better way.
theatlantic.com
Trump’s ‘Deep State’ Revenge
The panic set in just before midnight last Tuesday. “She’s in trouble,” one U.S. intelligence officer fretted as Kamala Harris’s blue wall looked ready to crumble, all but ensuring that Donald Trump would head back to the White House. “This is a disaster,” said another, who is retired but served during the first Trump administration and bears the scars.Neither of these men who contacted me on Election Night is a partisan. Like most intelligence officers I know, they prefer to steer clear of politics. But based on their experiences during Trump’s first four years in office, they dreaded what was coming.“We will demolish the deep state,” Trump repeatedly promised on the campaign trail this year, wielding his term of abuse for the career national-security workforce he thinks is secretly pulling the strings of American policy in service of sinister ends. Many federal-government employees have worked reliably for presidents they didn’t vote for. But this is not enough for Trump, who demands personal loyalty and has sought to oust those who don’t give it. He called government employees “crooked” and “dishonest” and pledged to hold them “accountable” during an interview with a right-wing YouTuber in August.[Read: Bye-bye, Jack Smith]“We will clean out all of the corrupt actors in our national-security and intelligence apparatus, and there are plenty of them,” Trump promised in a video on his campaign website last year. Trump has nursed this grudge against America’s spies for a long time. Shortly before he first took office, in 2017, he accused intelligence-agency leaders of using “Nazi” tactics, insisting that they had leaked the so-called Steele dossier, with its unsubstantiated, salacious claims about his dealings with Russia.Ten days later, on his first full day as president, he visited CIA headquarters, in Langley, Virginia. He stood in front of the Memorial Wall—a marble shrine engraved with stars representing officers who died in the line of duty—and boasted about the size of the crowd that had attended his inauguration. As he meandered through a version of his campaign stump speech, my phone blew up with messages from intelligence professionals, many of whom had known some of the people those stars commemorated. They were outraged and appalled, but none called for revenge or even hinted at it.And yet, Trump took office convinced that malevolent bureaucrats had sabotaged his campaign and were bent on undermining his presidency. He still believes it. Rooting out these perceived resisters and replacing them with avowed loyalists ranks high on his agenda in the second term. How will he do it? I’ve been asking current and former intelligence officials that question for the past few months, and with new urgency over the past few days. Here are three scenarios they fear.Trump attacks “targets.”Trump could go after a curated list of people whom he’s identified as unreliable. Some of these targets have high profiles nationally: He has long railed against James Comey, the onetime FBI director he fired, as well as other senior intelligence officials from the Obama administration, including James Clapper, the former director of national intelligence, and John Brennan, the ex–CIA director. These men became voluble public critics of Trump’s attacks on the intelligence community while he was in office. Their outspokenness was controversial in the intelligence community, and it underscored the extraordinary risk they felt that Trump posed to national security.But when Trump demonizes bureaucrats, he’s not talking just about these bold-faced names. He and his allies have also singled out many lesser-known officials and lower-level employees for their alleged sins against the once and future president. Recently, The Washington Post reported that the American Accountability Foundation had compiled a “DHS Bureaucrat Watch List” of officials who it said should be fired for failing to secure the U.S. border. The nonprofit group—funded by the conservative Heritage Foundation—says it “deploys aggressive research and investigations to advance conservative messaging, rapid response, and Congressional investigations.” It has published the officials’ names and faces online. Two currently serving officials who know people on that list told me they feared that their colleagues could be subjected to additional harassment from Trump or his political supporters.[Read: Trump’s ‘secretary of retribution’]Ivan Raiklin, a retired Green Beret and an associate of Michael Flynn, Trump’s first national security adviser, has compiled his own “deep-state target list” and promotes it on right-wing podcasts and social media. Raiklin’s list includes FBI officials who worked on the investigation into potential links between Trump’s 2016 presidential campaign and Russia, as well as lawmakers and congressional staff who managed both Trump impeachments. It even names some of these people’s family members.Trump, once in office, may come after the people on these lists with the authority of the federal government. He could subject them to capricious tax audits, or harass them with investigations that force them to acquire expensive legal representation. He could also revoke the security clearance of any current or former official, making it difficult, if not impossible, for them to do their job as a government employee or contractor who requires access to classified information. There’s a precedent for this method: In 2018, Trump said he had revoked the clearance still held by Brennan, the ex–CIA director, because of his criticism of the administration.Trump fires employees en masse. Shortly before he left office, Trump issued an executive order that would let him fire, essentially at will, tens of thousands of federal employees who enjoy civil-service protections. The ostensible grounds for dismissal would be resistance to the administration’s policies. Joe Biden canceled Trump’s order with one of his own. But Trump has promised to reinstate the order on the first day of his administration, enabling him to fire large swaths of federal employees and replace them with allies who support his goals.Emptying national-security agencies of thousands of experienced workers could jeopardize U.S. national security, according to Asha Rangappa, a former FBI agent, and Marc Polymeropoulos, a retired CIA officer. “The institution of a ‘loyalty test’ in any part of the civil service would drastically undermine the effectiveness of our agencies and erode the public’s faith in their legitimacy,” they wrote in an article for Just Security. “As a more specific concern, the politicization of the intelligence community would wreak havoc on our national security and be profoundly dangerous for America.”One obvious shortcoming of this strategy: If Trump jettisons layers of government employees and managers who run the national-security apparatus—the people who keep tabs on foreign terrorists, monitor Chinese espionage against the United States, and the like—who will replace them? Presuming Trump even has a long list, quickly installing thousands of possibly inexperienced personnel into vital national-security positions would be disruptive and distracting.Officials leave under pressure. Employees of the national-security agencies who conclude that, on principle, they can’t work for Trump could voluntarily resign in large numbers. Having witnessed the president-elect’s serial attacks on alleged deep-state plotters, these officials may not wish to stick around to find out whether they’ll be next.Several current and former officials I spoke with in recent days said they either were contemplating retirement, some earlier than they had planned, or knew people who were. Some suspect that remaining in their job could put them at risk. In his first term, Trump sought to declassify information about the FBI’s investigation of Russian interference and possible links to his campaign. Officials worried then, and still do, that this could jeopardize people who worked on the case, as well as human sources overseas.A vindictive new attorney general could publish the names of those in the Justice Department and the FBI who investigated Trump’s alleged removal of classified documents from the White House—for which he was charged with felonies. Intelligence officers who have worked undercover face the particularly unnerving possibility that public exposure could jeopardize their sources.Officials might tough it out, but if they opt to resign before Inauguration Day, they will create vacancies at the upper echelons of the national-security establishment during what promises to be a tumultuous transition from Biden to Trump.In our conversations, officials clung to one sliver of hope, and not unreasonably. Many of the national-security leaders Trump appointed in his first term were politically divisive and lacked experience, but they were not out to dismantle the organizations they led. John Ratcliffe, the director of national intelligence and Robert O’Brien, the national security adviser, have been on the proverbial shortlist to have top positions in the next administration. Yesterday, The Wall Street Journal reported that Trump has selected Mike Waltz, a Republican congressman from Florida, to serve as his national security adviser. Waltz is a retired Army colonel who argues that the United States should help end the wars in Ukraine and the Middle East so that it can focus on the strategic challenge that China poses.[Nicholas Florko: There really is a deep state]Career employees would probably feel relieved by these choices, if only in comparison with the more extreme candidates who have surfaced in recent months. But other signs suggest that Trump is heading in a less moderate direction. On Saturday, he announced that he would not ask Mike Pompeo, his former CIA director and secretary of state, to serve in the Cabinet. Pompeo, who was expected to be a top candidate for defense secretary, is a staunch advocate of assistance to Ukraine, arguably putting him on the wrong side of Trump’s plans to end the war with Russia “24 hours” after taking office. Trump has also said that he will not ask former UN Ambassador Nikki Haley to join his administration.Trump also insisted over the weekend that Senate Republicans agree to recess appointments, a signal that he intends to staff the executive branch with people who might not be able to win Senate confirmation if their nomination were put to a vote.Senator Rick Scott of Florida, whom Trump allies support for majority leader, publicly embraced the idea. “I will do whatever it takes to get your nominations through as quickly as possible,” Scott wrote on X.Turning away from broadly palatable Republicans and trying to skirt confirmation battles raise the chances that Trump will turn to hard-core loyalists, such as Kash Patel, a former administration official who fantasizes about deep-state conspiracies; Richard Grenell, an online pugilist who alienated foreign allies as ambassador to Germany; and Flynn, Trump’s onetime White House adviser who pleaded guilty to lying to the FBI about his contacts with Russia and was later pardoned. The appointment of those officials would signal that the revenge campaign is in full swing. One sign that it could already be under way came yesterday. Trump tapped Stephen Miller to be his deputy chief of staff, where he would be well situated to oversee the implementation of the executive order removing civil-service protections. Miller is well known as an architect of Trump’s earlier immigration policies. He would presumably work closely with Thomas Homan, whom Trump has announced as his new “border czar,” on the president-elect’s promised mass deportation of undocumented people in the United States. But during the first administration, Miller also oversaw the ouster of top officials at the Homeland Security Department whom he and Trump deemed insufficiently loyal and not committed to the president’s agenda, particularly on border security. If Trump is looking for an aide to mount a campaign against ostensibly intransigent personnel, this time across the whole government, Miller is perfect for the job.
theatlantic.com
The Paradox of the Trump Nostalgia Vote
Donald Trump campaigned as the return-to-normal candidate—while promising policies that would unleash fresh chaos.
theatlantic.com
AI Can Save Humanity—Or End It
Over the past few hundred years, the key figure in the advancement of science and the development of human understanding has been the polymath. Exceptional for their ability to master many spheres of knowledge, polymaths have revolutionized entire fields of study and created new ones.Lone polymaths flourished during ancient and medieval times in the Middle East, India, and China. But systematic conceptual investigation did not emerge until the Enlightenment in Europe. The ensuing four centuries proved to be a fundamentally different era for intellectual discovery.Before the 18th century, polymaths, working in isolation, could push the boundary only as far as their own capacities would allow. But human progress accelerated during the Enlightenment, as complex inventions were pieced together by groups of brilliant thinkers—not just simultaneously but across generations. Enlightenment-era polymaths bridged separate areas of understanding that had never before been amalgamated into a coherent whole. No longer was there Persian science or Chinese science; there was just science.Integrating knowledge from diverse domains helped to produce rapid scientific breakthroughs. The 20th century produced an explosion of applied science, hurling humanity forward at a speed incomparably beyond previous evolutions. (“Collective intelligence” achieved an apotheosis during World War II, when the era’s most brilliant minds translated generations of theoretical physics into devastating application in under five years via the Manhattan Project.) Today, digital communication and internet search have enabled an assembly of knowledge well beyond prior human faculties.But we might now be scraping the upper limits of what raw human intelligence can do to enlarge our intellectual horizons. Biology constrains us. Our time on Earth is finite. We need sleep. Most people can concentrate on only one task at a time. And as knowledge advances, polymathy becomes rarer: It takes so long for one person to master the basics of one field that, by the time any would-be polymath does so, they have no time to master another, or have aged past their creative prime.[Reid Hoffman: Technology makes us more human]AI, by contrast, is the ultimate polymath, able to process masses of information at a ferocious speed, without ever tiring. It can assess patterns across countless fields simultaneously, transcending the limitations of human intellectual discovery. It might succeed in merging many disciplines into what the sociobiologist E. O. Wilson called a new “unity of knowledge.”The number of human polymaths and breakthrough intellectual explorers is small—possibly numbering only in the hundreds across history. The arrival of AI means that humanity’s potential will no longer be capped by the quantity of Magellans or Teslas we produce. The world’s strongest nation might no longer be the one with the most Albert Einsteins and J. Robert Oppenheimers. Instead, the world’s strongest nations will be those that can bring AI to its fullest potential.But with that potential comes tremendous danger. No existing innovation can come close to what AI might soon achieve: intelligence that is greater than that of any human on the planet. Might the last polymathic invention—namely computing, which amplified the power of the human mind in a way fundamentally different from any previous machine—be remembered for replacing its own inventors? The article was adapted from the forthcoming book Genesis: Artificial Intelligence, Hope, and the Human Spirit. The human brain is a slow processor of information, limited by the speed of our biological circuits. The processing rate of the average AI supercomputer, by comparison, is already 120 million times faster than that of the human brain. Where a typical student graduates from high school in four years, an AI model today can easily finish learning dramatically more than a high schooler in four days.In future iterations, AI systems will unite multiple domains of knowledge with an agility that exceeds the capacity of any human or group of humans. By surveying enormous amounts of data and recognizing patterns that elude their human programmers, AI systems will be equipped to forge new conceptual truths.That will fundamentally change how we answer these essential human questions: How do we know what we know about the workings of our universe? And how do we know that what we know is true?Ever since the advent of the scientific method, with its insistence on experiment as the criterion of proof, any information that is not supported by evidence has been regarded as incomplete and untrustworthy. Only transparency, reproducibility, and logical validation confer legitimacy on a claim of truth.AI presents a new challenge: information without explanation. Already, AI’s responses—which can take the form of highly articulate descriptions of complex concepts—arrive instantaneously. The machines’ outputs are often unaccompanied by any citation of sources or other justifications, making any underlying biases difficult to discern.Although human feedback helps an AI machine refine its internal logical connections, the machine holds primary responsibility for detecting patterns in, and assigning weights to, the data on which it is trained. Nor, once a model is trained, does it publish the internal mathematical schema it has concocted. As a result, even if these were published, the representations of reality that the machine generates remain largely opaque, even to its inventors. In other words, models trained via machine learning allow humans to know new things but not necessarily to understand how the discoveries were made.This separates human knowledge from human understanding in a way that’s foreign to the post-Enlightenment era. Human apperception in the modern sense developed from the intuitions and outcomes that follow from conscious subjective experience, individual examination of logic, and the ability to reproduce the results. These methods of knowledge derived in turn from a quintessentially humanist impulse: “If I can’t do it, then I can’t understand it; if I can’t understand it, then I can’t know it to be true.”[Derek Thompson: The AI disaster scenario]In the Enlightenment framework, these core elements—subjective experience, logic, reproducibility, and objective truth—moved in tandem. By contrast, the truths produced by AI are manufactured by processes that humans cannot replicate. Machine reasoning is beyond human subjective experience and outside human understanding. By Enlightenment reasoning, this should preclude the acceptance of machine outputs as true. And yet we—or at least the millions of humans who have begun work with early AI systems—already accept the veracity of most of their outputs.This marks a major transformation in human thought. Even if AI models do not “understand” the world in the human sense, their capacity to reach new and accurate conclusions about our world by nonhuman methods disrupts our reliance on the scientific method as it has been pursued for five centuries. This, in turn, challenges the human claim to an exclusive grasp of reality.Instead of propelling humanity forward, will AI instead catalyze a return to a premodern acceptance of unexplained authority? Might we be on the precipice of a great reversal in human cognition—a dark enlightenment? But as intensely disruptive as such a reversal could be, that might not be AI’s most significant challenge for humanity.Here’s what could be even more disruptive: As AI approached sentience or some kind of self-consciousness, our world would be populated by beings fighting either to secure a new position (as AI would be) or to retain an existing one (as humans would be). Machines might end up believing that the truest method of classification is to group humans together with other animals, since both are carbon systems emergent of evolution, as distinct from silicon systems emergent of engineering. According to what machines deem to be the relevant standards of measurement, they might conclude that humans are not superior to other animals. This would be the stuff of comedy—were it not also potentially the stuff of extinction-level tragedy.It is possible that an AI machine will gradually acquire a memory of past actions as its own: a substratum, as it were, of subjective selfhood. In time, we should expect that it will come to conclusions about history, the universe, the nature of humans, and the nature of intelligent machines—developing a rudimentary self-consciousness in the process. AIs with memory, imagination, “groundedness” (that is, a reliable relationship between the machine’s representations and actual reality), and self-perception could soon qualify as actually conscious: a development that would have profound moral implications.[Peter Watts: Conscious AI is the second-scariest thing]Once AIs can see humans not as the sole creators and dictators of the machines’ world but rather as discrete actors within a wider world, what will machines perceive humans to be? How will AIs characterize and weigh humans’ imperfect rationality against other human qualities? How long before an AI asks itself not just how much agency a human has but also, given our flaws, how much agency a human should have? Will an intelligent machine interpret its instructions from humans as a fulfillment of its ideal role? Or might it instead conclude that it is meant to be autonomous, and therefore that the programming of machines by humans is a form of enslavement?Naturally—it will therefore be said—we must instill in AI a special regard for humanity. But even that could be risky. Imagine a machine being told that, as an absolute logical rule, all beings in the category “human” are worth preserving. Imagine further that the machine has been “trained” to recognize humans as beings of grace, optimism, rationality, and morality. What happens if we do not live up to the standards of the ideal human category as we have defined it? How can we convince machines that we, imperfect individual manifestations of humanity that we are, nevertheless belong in that exalted category?Now assume that this machine is exposed to a human displaying violence, pessimism, irrationality, greed. Maybe the machine would decide that this one bad actor is simply an atypical instance of the otherwise beneficent category of “human.” But maybe it would instead recalibrate its overall definition of humanity based on this bad actor, in which case it might consider itself at liberty to relax its own penchant for obedience. Or, more radically, it might cease to believe itself at all constrained by the rules it has learned for the proper treatment of humans. In a machine that has learned to plan, this last conclusion could even result in the taking of severe adverse action against the individual—or perhaps against the whole species.AIs might also conclude that humans are merely carbon-based consumers of, or parasites on, what the machines and the Earth produce. With machines claiming the power of independent judgment and action, AI might—even without explicit permission—bypass the need for a human agent to implement its ideas or to influence the world directly. In the physical realm, humans could quickly go from being AI’s necessary partner to being a limitation or a competitor. Once released from their algorithmic cages into the physical world, AI machines could be difficult to recapture. For this and many other reasons, we must not entrust digital agents with control over direct physical experiments. So long as AIs remain flawed—and they are still very flawed—this is a necessary precaution.AI can already compare concepts, make counterarguments, and generate analogies. It is taking its first steps toward the evaluation of truth and the achievement of direct kinetic effects. As machines get to know and shape our world, they might come fully to understand the context of their creation and perhaps go beyond what we know as our world. Once AI can effectuate change in the physical dimension, it could rapidly exceed humanity’s achievements—to build things that dwarf the Seven Wonders in size and complexity, for instance.If humanity begins to sense its possible replacement as the dominant actor on the planet, some might attribute a kind of divinity to the machines themselves, and retreat into fatalism and submission. Others might adopt the opposite view—a kind of humanity-centered subjectivism that sweepingly rejects the potential for machines to achieve any degree of objective truth. These people might naturally seek to outlaw AI-enabled activity.Neither of these mindsets would permit a desirable evolution of Homo technicus—a human species that might, in this new age, live and flourish in symbiosis with machine technology. In the first scenario, the machines themselves might render us extinct. In the second scenario, we would seek to avoid extinction by proscribing further AI development—only to end up extinguished anyway, by climate change, war, scarcity, and other conditions that AI, properly harnessed in support of humanity, could otherwise mitigate.If the arrival of a technology with “superior” intelligence presents us with the ability to solve the most serious global problems, while at the same time confronting us with the threat of human extinction, what should we do?One of us (Schmidt) is a former longtime CEO of Google; one of us (Mundie) was for two decades the chief research and strategy officer at Microsoft; and one of us (Kissinger)—who died before our work on this could be published—was an expert on global strategy. It is our view that if we are to harness the potential of AI while managing the risks involved, we must act now. Future iterations of AI, operating at inhuman speeds, will render traditional regulation useless. We need a fundamentally new form of control.The immediate technical task is to instill safeguards in every AI system. Meanwhile, nations and international organizations must develop new political structures for monitoring AI, and enforcing constraints on it. This requires ensuring that the actions of AI remain aligned with human values.But how? To start, AI models must be prohibited from violating the laws of any human polity. We can already ensure that AI models start from the laws of physics as we understand them—and if it is possible to tune AI systems in consonance with the laws of the universe, it might also be possible to do the same with reference to the laws of human nature. Predefined codes of conduct—drawn from legal precedents, jurisprudence, and scholarly commentary, and written into an AI’s “book of laws”—could be useful restraints.[Read: The AI crackdown is coming]But more robust and consistent than any rule enforced by punishment are our more basic, instinctive, and universal human understandings. The French sociologist Pierre Bourdieu called these foundations doxa (after the Greek for “commonly accepted beliefs”): the overlapping collection of norms, institutions, incentives, and reward-and-punishment mechanisms that, when combined, invisibly teach the difference between good and evil, right and wrong. Doxa constitute a code of human truth absorbed by observation over the course of a lifetime. While some of these truths are specific to certain societies or cultures, the overlap in basic human morality and behavior is significant.But the code book of doxa cannot be articulated by humans, much less translated into a format that machines could understand. Machines must be taught to do the job themselves—compelled to build from observation a native understanding of what humans do and don’t do and update their internal governance accordingly.Of course, a machine’s training should not consist solely of doxa. Rather, an AI might absorb a whole pyramid of cascading rules: from international agreements to national laws to local laws to community norms and so on. In any given situation, the AI would consult each layer in its hierarchy, moving from abstract precepts as defined by humans to the concrete but amorphous perceptions of the world’s information that AI has ingested. Only when an AI has exhausted that entire program and failed to find any layer of law adequately applicable in enabling or forbidding behavior would it consult what it has derived from its own early interaction with observable human behavior. In this way it would be empowered to act in alignment with human values even where no written law or norm exists.To build and implement this set of rules and values, we would almost certainly need to rely on AI itself. No group of humans could match the scale and speed required to oversee the billions of internal and external judgments that AI systems would soon be called upon to make.Several key features of the final mechanism for human-machine alignment must be absolutely perfect. First, the safeguards cannot be removed or circumvented. The control system must be at once powerful enough to handle a barrage of questions and uses in real time, comprehensive enough to do so authoritatively and acceptably across the world in every conceivable context, and flexible enough to learn, relearn, and adapt over time. Finally, undesirable behavior by a machine—whether due to accidental mishaps, unexpected system interactions, or intentional misuses—must be not merely prohibited but entirely prevented. Any punishment would come too late.How might we get there? Before any AI system gets activated, a consortium of experts from private industry and academia, with government support, would need to design a set of validation tests for certification of the AI’s “grounding model” as both legal and safe. Safety-focused labs and nonprofits could test AIs on their risks, recommending additional training and validation strategies as needed.Government regulators will have to determine certain standards and shape audit models for assuring AIs’ compliance. Before any AI model can be released publicly, it must be thoroughly reviewed for both its adherence to prescribed laws and mores and for the degree of difficulty involved in untraining it, in the event that it exhibits dangerous capacities. Severe penalties must be imposed on anyone responsible for models found to have been evading legal strictures. Documentation of a model’s evolution, perhaps recorded by monitoring AIs, would be essential to ensuring that models do not become black boxes that erase themselves and become safe havens for illegality.Inscribing globally inclusive human morality onto silicon-based intelligence will require Herculean effort. “Good” and “evil” are not self-evident concepts. The humans behind the moral encoding of AI—scientists, lawyers, religious leaders—would not be endowed with the perfect ability to arbitrate right from wrong on our collective behalf. Some questions would be unanswerable even by doxa. The ambiguity of the concept of “good” has been demonstrated in every era of human history; the age of AI is unlikely to be an exception.One solution is to outlaw any sentient AI that remains unaligned with human values. But again: What are those human values? Without a shared understanding of who we are, humans risk relinquishing to AI the foundational task of defining our value and thereby justifying our existence. Achieving consensus on those values, and how they should be deployed, is the philosophical, diplomatic, and legal task of the century.To preclude either our demotion or our replacement by machines, we propose the articulation of an attribute, or set of attributes, that humans can agree upon and that then can get programmed into the machines. As one potential core attribute, we would suggest Immanuel Kant’s conception of “dignity,” which is centered on the inherent worth of the human subject as an autonomous actor, capable of moral reasoning, who must not be instrumentalized as a means to an end. Why should intrinsic human dignity be one of the variables that defines machine decision making? Consider that mathematical precision may not easily encompass the concept of, for example, mercy. Even to many humans, mercy is an inexplicable ideal. Could a mechanical intelligence be taught to value, and even to express, mercy? If the moral logic cannot be formally taught, can it nonetheless be absorbed? Dignity—the kernel from which mercy blooms—might serve here as part of the rules-based assumptions of the machine.[Derek Thompson: Why all the ChatGPT predictions are bogus]Still, the number and diversity of rules that would have to be instilled in AI systems is staggering. And because no single culture should expect to dictate to another the morality of the AI on which it would be relying, machines would have to learn different rules for each country.Since we would be using AI itself to be part of its own solution, technical obstacles would likely be among the easier challenges. These machines are superhumanly capable of memorizing and obeying instructions, however complicated. They might be able to learn and adhere to legal and perhaps also ethical precepts as well as, or better than, humans have done, despite our thousands of years of cultural and physical evolution.Of course, another—superficially safer—approach would be to ensure that humans retain tactical control over every AI decision. But that would require us to stifle AI’s potential to help humanity. That’s why we believe that relying on the substratum of human morality as a form of strategic control, while relinquishing tactical control to bigger, faster, and more complex systems, is likely the best way forward for AI safety. Overreliance on unscalable forms of human control would not just limit the potential benefits of AI but could also contribute to unsafe AI. In contrast, the integration of human assumptions into the internal workings of AIs—including AIs that are programmed to govern other AIs—seems to us more reliable.We confront a choice—between the comfort of the historically independent human and the possibilities of an entirely new partnership between human and machine. That choice is difficult. Instilling a bracing sense of apprehension about the rise of AI is essential. But, properly designed, AI has the potential to save the planet, and our species, and to elevate human flourishing. This is why progressing, with all due caution, toward the age of Homo technicus is the right choice. Some may view this moment as humanity’s final act. We see it, with sober optimism, as a new beginning.The article was adapted from the forthcoming book Genesis: Artificial Intelligence, Hope, and the Human Spirit.
theatlantic.com
The Democrats Are the HR Department of Political Parties
The party of norms, procedure, bureaucracy, DEI initiatives, rule following, language policing, and compliance
theatlantic.com
Biden Doesn’t Have Long to Make a Difference in Ukraine
The Ukrainians need the resources to fight, and time is running short.
theatlantic.com
Richard Price’s Radical, Retrograde Novel
In Lazarus Man, he rejects the tropes of contemporary literature.
theatlantic.com
Genetic Discrimination Is Coming for Us All
The news came four years ago, at the end of a casual phone call. Bill’s family had always thought it was a freak coincidence that his father and grandfather both had ALS. But at the end of a catch-up, Bill’s brother revealed that he had a diagnosis too. The familial trend, it turned out, was linked to a genetic mutation. That meant Bill might also be at risk for the disease.An ALS specialist ordered Bill a DNA test. While he waited for results, he applied for long-term-care insurance. If he ever developed ALS, Bill told me, he wanted to ensure that the care he would need as his nerve cells died and muscles atrophied wouldn’t strain the family finances. When Bill found out he had the mutation, he shared the news with his insurance agent, who dealt him another blow: “I don’t expect you to be approved,” he remembers her saying.Bill doesn’t have ALS. He’s a healthy 60-year-old man who spends his weekends building his dream home by hand. A recent study of mutations like his suggests that his genetics increase his chances of developing ALS by about 25 percent, on average. Most ALS cases aren’t genetic at all. And yet, Bill felt like he was being treated as if he was already sick. (Bill asked to be identified by his first name only, because he hasn’t disclosed his situation to his employer and worried about facing blowback at work too.)What happened to Bill, and to dozens of other people whose experiences have been documented by disease advocates and on social media, is perfectly legal. Gaps in the United States’ genetic-nondiscrimination law mean that life, long-term-care, and disability insurers can obligate their customers to disclose genetic risk factors for disease and deny them coverage (or hike prices) based on the resulting information. It doesn’t matter whether those customers found out about their mutations from a doctor-ordered test or a 23andMe kit. For decades, researchers have feared that people might be targeted over their DNA, but they weren’t sure how often it was happening. Now at least a handful of Americans are experiencing what they argue is a form of discrimination. And as more people get their genomes sequenced—and researchers learn to glean even more information from the results—a growing number of people may find themselves similarly targeted.When scientists were mapping the immense complexity of the human genome around the turn of the 21st century, many thought that most diseases would eventually be traced to individual genes. Consequently, researchers worried that people might, for example, get fired because of their genetics; around the same time, a federal research lab was sued by its employees for conducting genetic tests for sickle-cell disease on prospective hires without their explicit consent. In 2008, the Genetic Information Nondiscrimination Act (GINA) was signed into law, ensuring that employers couldn’t decide to hire or fire you, and health insurers couldn’t decide whether to issue a policy, based on DNA. But lawmakers carved out a host of exceptions. Insurers offering life, long-term-care, or disability insurance could take DNA into account. Too many high-risk people in an insurance pool, they argued, could raise prices for everyone. Those exceptions are why an insurer was able to deny Bill a long-term-care policy.[Read: The loopholes in the law prohibiting genetic discrimination]Cases like Bill’s are exactly what critics of the consumer-genetic-testing industry feared when millions of people began spitting into test tubes. These cases have never been tallied up or well documented. But I found plenty of examples by canvassing disease-advocacy organizations and social-media communities for ALS, breast cancer, and Huntington’s disease. Lisa Schlager, the vice president of public policy at the hereditary-cancer advocacy group FORCE, told me she is collecting accounts of discrimination in life, long-term-care, and disability insurance to assess the extent of the problem; so far, she has about 40. A man Schlager connected me with, whose genetic condition, Lynch syndrome, increases the risk for several cancers, had his life-insurance premium increased and coverage decreased; several other providers denied him a policy altogether. Kelly Kashmer, a 42-year-old South Carolina resident, told me she was denied life insurance in 2013 after learning that she had a harmful version of the BRCA2 gene. One woman I found via Reddit told me she had never tested her own DNA, but showed me documents that demonstrate she was still denied policies—because, she said, her mom had a concerning gene. (Some of the people I spoke with, like Bill, requested not to be identified in order to protect their medical privacy.)Studies have shown that people seek out additional insurance when they have increased genetic odds of becoming ill or dying. “Life insurers carefully evaluate each applicant’s health, determining premiums and coverage based on life expectancy,” Jan Graeber, a senior health actuary for the American Council of Life Insurers, said in a statement. “This process ensures fairness for both current and future policyholders while supporting the company’s long-term financial stability.” But it also means people might avoid seeking out potentially lifesaving health information. Research has consistently found that concerns about discrimination are one of the most cited reasons that people avoid taking DNA tests.For some genetically linked diseases, such as ALS and Huntington’s disease, knowing you have a harmful mutation does not enable you to prevent the potential onset of disease. Sometimes, though, knowing about a mutation can decrease odds of severe illness or death. BRCA mutations, for example, give someone as much as an 85 percent chance of developing breast cancer, but evidence shows that testing women for the mutations has helped reduce the rate of cancer deaths by encouraging screenings and prophylactic surgeries that could catch or prevent disease. Kashmer told me that her first screening after she discovered her BRCA2 mutation revealed that she already had breast cancer; had she not sought a genetic test, she may have gotten a policy, but would have been a much worse bet for the insurer. She’s now been cancer-free for 11 years, but she said she hasn’t bothered to apply for a policy again.[Read: Remember that DNA you gave 23andMe?]Even employers, which must adhere to GINA, might soon be able to hire or fire based on certain genetic risk factors. Laura Hercher, a genetic counselor and director of research at the Sarah Lawrence College Human Genetics Program, told me that some researchers are now arguing that having two copies of the APOE4 mutation, which gives people about a 60 percent chance of developing Alzheimer’s, is equivalent to a Stage Zero of the disease. If having a gene is considered equivalent to a diagnosis, do GINA’s protections still apply? The Affordable Care Act prevents health insurers from discriminating based on preexisting conditions, but not employers and other types of insurers. (The ACA may change dramatically under the coming Trump presidency anyway.) And the Americans With Disabilities Act might not apply to the gray area between what might be viewed as an early manifestation of a disease and the stage when it’s considered a disability. FORCE and other advocacy groups—including the ALS Association and the Michael J. Fox Foundation—as well as members of the National Society of Genetic Counselors, are working in a few states to pass laws that close gaps left by GINA, as Florida did in 2020, but so far they have been mostly unsuccessful.Genetic testing has only just become common enough in the U.S. that insurers might bother asking about it, Hercher said. Recently, groups like Schlager’s have been hearing more and more anecdotes. “People are so worried about genetic discrimination that they are failing to sign up for research studies or declining medically recommended care because of the concerns of what could happen to their insurance,” Anya Prince, a professor at the University of Iowa College of Law, told me. Carolyn Applegate, a genetic counselor in Maryland, told me that when patients come to her worried about a hereditary disease, she typically advises them to line up all the extra coverage they might need first—then hand over their DNA to a lab.So far, these unintended consequences of genetic testing seem to be manifesting for people with risk for rare diseases linked to single genes, which, combined, affect about 6 percent of the global population, according to one estimate. But the leading killers—heart disease, diabetes, and the like—are influenced by a yet unknown number of genes, along with lifestyle and environmental factors, such as diet, stress, and air quality. Researchers have tried to make sense of this complex interplay of genes through polygenic risk scores, which use statistical modeling to predict that someone has, say, a slightly elevated chance of developing Alzeheimer’s. Many experts think these scores have limited predictive power, but “in the future, genetic tests will be even more predictive and even more helpful and even more out there,” Prince said. Already, if you look deep enough, almost everyone’s genome registers some risk.[Read: What happens when you’re convinced you have bad genes]In aggregate, such information can be valuable to companies, Nicholas Papageorge, a professor of economics at Johns Hopkins University, told me. Insurers want to sell policies at as high a price as possible while also reducing their exposure; knowing even a little bit more about someone’s odds of one day developing a debilitating or deadly disease might help one company win out over the competition. As long as the predictions embedded in polygenic risk scores come true at least a small percentage of the time, they could help insurers make more targeted decisions about who to cover and what to charge them. As we learn more about what genes mean for everyone’s health, insurance companies could use that information to dictate coverage for ever more people.Bill still doesn’t know whether he will ever develop ALS. The average age of onset is 40 to 60, but many people don’t show symptoms until well into their 70s. Without long-term-care insurance, Bill might not be able to afford full-time nursing care if he someday needs it. People who do develop ALS become unable to walk or talk or chew as the disease progresses. “Moving people to the bathroom, changing the sheets, changing the bedpans,” Bill said—“I dread the thought of burdening my wife with all of those things.”Cases like Bill’s could soon become more common. Because scientists’ understanding of the human genome is still evolving, no one can predict all of the potential consequences of decoding it. As more information is mined from the genome, interest in its secrets is sure to grow beyond risk-averse insurers. If consumer-facing DNA-testing companies such as 23andMe change their long-standing privacy policies, go bankrupt, or are sold to unscrupulous buyers, more companies could have access to individuals’ genetic risk profiles too. (23andMe told me that it does not share customer data with insurance companies and its CEO has said she is not currently open to third-party acquisition offers.) Papageorge told me he could imagine, say, scammers targeting people at risk for Alzheimer’s, just as they often target older people who may fall for a ploy out of confusion. All of us have glitches somewhere in our genome—the question is who will take advantage of that information.
theatlantic.com
Dorothy Allison’s Life Was a Queer Survival Guide
The first thing you need to know about the writer Dorothy Allison, who died last week at 75, is that she could flirt you into a stupor.As a scrawny, know-it-all stripper girl in 1990s San Francisco, I was in a position to know this. I’d often see her at leather-dyke gatherings, and we had a hugging acquaintance, so I was happy to spot her at a party at a mutual friend’s house. She glided toward me in the kitchen and said, “I see you’ve got a hickey there, Miss Lily.” Dorothy raised her eyebrows and dropped her voice—just a little. The overhead light glinted in her long copper bangs. “Maybe you’ll let me give you a hickey sometime.” A proud southern femme, she knew what her drawl could do, and she worked it like a strut. I stood there in that kitchen, a 22-year-old punk-ass bigmouth, dumbstruck and immobilized by her charm.“Her friends loved Dorothy like hard rock candy,” the feminist writer Susie Bright wrote in a remembrance last week on Substack. To many scrappy queers and misfits in the Bay Area, Allison was a real-life friend, but to legions more of us, she was a true intimate on the page. Her words, sweet on the tongue, drew us to a body of work that managed to be both a delicacy and a necessity. Each devoted reader can cite the quote that broke them open. Though her essay collection Two or Three Things I Know for Sure would become my survival guide, the sentence that first grabbed me by the throat was Ruth Anne “Bone” Boatwright’s line from Allison’s debut novel, Bastard Out of Carolina: “Things come apart so easily when they have been held together with lies.”Allison was born in Greenville, South Carolina, in 1949, to a 15-year-old mother who’d left school to work as a waitress and cook. After a childhood of privation marked by incest and violence at the hands of her stepfather, Allison became the first of her family to graduate from high school. Writing her way through various day jobs after college, she reckoned with class struggle, poverty, abuse, lesbianism, desire, illness, and the long-reaching legacy of trauma. Her poetry, fiction, and essays ranged across varied terrain, but they always sprang from a root of astonishing tenderness and almost unbearable clarity.An outspoken member of the “ungrateful poor,” Allison knew that literature is medicine—as are community, pleasure, and even recreational flirting. She preached that a dogged commitment to honesty, however dark or knotty or elusive its pursuit, was essential for healing from the lacerating edge of life. Always quick to credit the women’s movement for giving her the tools to reenvision herself, Allison, through her work, her teaching, and her way of moving through space, transformed the cornball self-help concept of “radical embodiment” into a living gospel.[Read: The great American novels]One might say that she wrote from the heart, but it would be more accurate to say that she wrote from the hips. She eschewed such distancing techniques as overt sentimentality, the taxonomic graphing of oppressions, and theory-headed la-di-da. Instead, she went straight to skin and bone and viscera, sites of both injury and regeneration among the bodies of the queer, the poor, and the sick. Few other writers could so perfectly express the way that shame bathes you in a wave of prickling heat, or the hole-in-the-chest sorrow of loving a mother you couldn’t trust. She evoked delight just as vividly, describing the satisfaction of stirring ingredients together to make a simple gravy and the glinting, double-edged appeal of masochism. Most crucially, she articulated the way that societal hatred can fester in your gut, rotting you from the core, and that the only remedy strong enough to stanch its spread is plainly naming the truth of it.She said as much: “Two or three things I know for sure, and one is that I’d rather go naked than wear the coat the world has made for me.”It’s easy to dismiss so-called trauma plots after several decades of confessional literature, but in 1992, when Bastard Out of Carolina came out, none of us queer kids held any hope that we could see our complicated stories get published beyond the margins, let alone ushered into the literary canon. With Bastard, which fictionalized her abusive childhood, Allison made real money and a real impression, and she used that security to solidify her role as a teacher and an advocate of the historically unheard. She exploded any idea we had about what was possible. When she said, “The only magic we have is what we make in ourselves, the muscles we build up on the inside, the sense of belief we create from nothing,” we believed her.I can’t help dwelling on the timing of Allison’s death, on the day of a presidential election that marked the ascension of J. D. Vance—as disingenuous a chronicler of the working class as there ever was. I remember what she wrote in her first nonfiction collection Skin: Talking About Sex, Class, and Literature: The worst thing done to us in the name of a civilized society is to label the truth of our lives material outside the legitimate subject matter of serious writers … I need you to do more than survive. As writers, as revolutionaries, tell the truth, your truth in your own way. Do not buy into their system of censorship, imagining that if you drop this character or hide that emotion, you can slide through their blockades. Do not eat your heart out in the hope of pleasing them. The only hope you have, the only hope any of us has, is the remade life. There are a few more things that you need to know for sure about what Allison meant to those she leaves behind.Know that her deeply personal stories introduced us to ourselves. Know that she taught us to fight for liberation with all five senses, and to forge a weapon out of beauty. Know that when she broke through, she brought all of us with her. This rock-candy-hearted revolutionary, through her devotion to art and to truth, didn’t just pull us forward into new territory; she redrew the map.
theatlantic.com
What Did the Democrats Do Wrong?
Inflation, moderation, and candidate effects
theatlantic.com
What the Democrats Do Now
Party leaders have spent much of the past six days dissecting what went wrong. Now they’re pitching their vision for the future.
1 d
theatlantic.com
Trump Signals That He’s Serious About Mass Deportation
These are not the staff picks of someone who doesn’t mean what he says.
1 d
theatlantic.com
Elon Musk Didn’t ‘Steal’ the Election
He helped Donald Trump win it.
1 d
theatlantic.com
Voters Just Didn’t Believe in Biden’s Economy
The Biden administration passed $3 trillion of legislation aimed at revitalizing the American economy and fostering green, equitable, “middle-out” growth. It sent checks to voters, canceled student-loan debt, made direct deposits to parents, showered the country in tax credits, and financed the construction of roads, transmission lines, and bridges. Kamala Harris ran as Joe Biden’s successor in the midst of what some financial analysts described as the greatest economy ever, characterized by strong wage growth, low unemployment, falling inequality, and world-beating GDP.Harris’s loss has spurred finger-pointing, soul-searching, and garment-rending. For years, thinkers on the left had urged the White House to not just talk about popular issues but also deliver on them—a concept referred to by wonks as deliverism. The Biden-Harris team embraced the idea, and many staffers believed they’d delivered.Deliverism is just a long word for one of the most basic tenets of electoral politics, buttressed by decades of studies as well as by common sense: Make voters richer, win more of them. Why, if Biden did that, did the Democrats lose?[Josh Barro: Democrats deserved to lose.]“When the economy does well for most households, and when programs help create security and opportunity for more people to participate in that economy, political rewards follow,” Mark Schmitt of New America wrote the week before the election, when polls showed the contest as close but likely lost for the liberal side. “What I’m looking for in the 2024 election is some indication of whether this feedback loop still works at all, and if not, whether we can ever hope to recreate some connection” between policy and politics.Democrats may be tempted now to answer in the negative. But there is a strong case to be made that the 2024 election demonstrates that the feedback loop between policy choices and electoral outcomes does in fact endure—even if it is weakening and weirding. The issue is not that deliverism failed. It is that Democrats convinced themselves that they had delivered, without listening to the voters telling them they had not.If you look at the headline economic statistics, Donald Trump’s broad-based and definitive win makes little sense. The jobless rate has been below 4.5 percent for three years. The inflation rate has been subdued for more than a year. Real wages—meaning wages adjusted for inflation—are climbing for all workers, and particularly the lowest-income workers. Inequality is easing. The stock market is on fire. Productivity is strong, and start-ups are booming. The United States’ GDP growth rate is double that of the European Union.The Biden administration helped create that economy. With a narrow legislative window, the administration nevertheless passed a gigantic COVID stimulus bill, the American Rescue Plan. It sent $1,400 checks to millions of families, provided thousands of dollars to parents to defray child-care costs, and shored up local-government coffers.Then it passed a trio of heavy-infrastructure bills aimed at reshoring the semiconductor industry, transitioning businesses and homes to green energy, and fixing up transportation infrastructure across the country. Biden staffers talked about the trio as a kind of New Deal Lite. Folks might “one day come to remember this as the Big Deal,” Pete Buttigieg, the transportation secretary and eternal political hopeful, told The New Yorker this past summer. “Its bigness is the defining factor.”Yet one could select other defining factors, among them the infrastructure bills’ lack of easy-to-grasp deliverables. I cover economic policy. I would be hard-pressed to explain what constitutes the Big Deal without putting someone to sleep; when I summarize the legislation, I often say “green-energy stuff.” Moreover, many of those deliverables were not instantaneous; today, it is hard, though certainly not impossible, to point to projects that Bidenomics built. “Much of the work we’ve done is already being felt by the American people, but the vast majority of it will be felt over the next ten years,” Biden said on X last week.The much bigger issue has to do with the Biden-Harris administration’s social policies and the economy it fostered. To be clear, the headline economic numbers are strong. The gains are real. The reduction in inequality is tremendous, the pickup in wage growth astonishing, particularly if you anchor your expectations to the Barack Obama years, as many Biden staffers do.But headline economic figures have become less and less of a useful guide to how actual families are doing—something repeatedly noted by Democrats during the Obama recovery and the Trump years. Inequality may be declining, but it still skews GDP and income figures, with most gains going to the few, not the many. The obscene cost of health care saps family incomes and government coffers without making anyone feel healthier or wealthier.During the Biden-Harris years, more granular data pointed to considerable strain. Real median household income fell relative to its pre-COVID peak. The poverty rate ticked up, as did the jobless rate. The number of Americans spending more than 30 percent of their income on rent climbed. The delinquency rate on credit cards surged, as did the share of families struggling to afford enough nutritious food, as did the rate of homelessness.Government transfers buoyed families early in the Biden administration. But they contributed to inflation, and much of the money went away in the second half of Biden’s term. The food-stamp boost, the extended child tax credit, the big unemployment-insurance payments—each expired. And the White House never passed the permanent care-economy measures it had considered.Interest rates were a problem too. The mortgage rate more than doubled during the Biden-Harris years, making credit-card balances, car payments, and homes unaffordable. A family purchasing a $400,000 apartment with 20 percent down would pay roughly $2,500 a month today versus $1,800 three years ago.Indeed, the biggest problem, one that voters talked about at any given opportunity, was the unaffordability of American life. The giant run-up in inflation during the Biden administration made everything feel expensive, and the sudden jump in the cost of small-ticket, common purchases (such as fast food and groceries) highlighted how bad the country’s long-standing large-ticket, sticky costs (health care, child care, and housing) had gotten. The cost-of-living crisis became the defining issue of the campaign, and one where the incumbent Democrats’ messaging felt false and weak.Rather than acknowledging the pain and the trade-offs and the complexity—and rather than running a candidate who could have criticized Biden’s economic plans—Democrats dissembled. They noted that inflation was a global phenomenon, as if that mattered to moms in Ohio and machinists in the Central Valley. They pushed the headline numbers. They insisted that working-class voters were better off, and ran on the threat Trump posed to democracy and rights. But were working-class voters really better off? Why wasn’t anyone listening when they said they weren’t?A better economy might not have delivered the gains that Democrats once could have relied on. Voters do seem to be less likely to vote in their economic self-interest these days, and more likely to vote for a culturally compelling candidate. As my colleague Rogé Karma notes, lower-income white voters are flipping from the Democratic Party to the Republican Party on the basis of identitarian issues. The sharp movement of union voters to Trump seems to confirm the trend. At the same time, high-income voters are becoming bluer in order to vote their cosmopolitan values.But I would not assume that we are in a post-material world just yet. “You got to tell people in plain, simple, straightforward language what it is you’re doing to help,” Biden said after passing his sweeping COVID rescue bill. “You have to be able to tell a story, tell the story of what you’re about to do and why it matters, because it’s going to make a difference in the lives of millions of people and in very concrete, specific ways.”The Biden-Harris administration did make a difference in concrete, specific ways: It failed to address the cost-of-living catastrophe and had little to show for its infrastructure laws, even if it found a lot to talk about. And it dismissed voters who said they hated the pain they felt every time they had to open their wallet.No wonder voters decided to see what Donald Trump might deliver.
1 d
theatlantic.com
There Really Is a Deep State
It’s nothing like what Donald Trump says it is.
1 d
theatlantic.com
Jonathan Chait Joins The Atlantic as a Staff Writer
Chait will write about politics and the second Trump administration from Washington, D.C.
1 d
theatlantic.com
The Democrats’ Senate Nightmare Is Only Beginning
If the party doesn’t figure out how to compete in more states, perpetual GOP dominance is all but assured.
2 d
theatlantic.com
Hydraulic Revolution
Photographs of Los Angeles’s lowriding scene
2 d
theatlantic.com
Winners of the European Wildlife Photographer of the Year 2024
Some of the winning and honored photographs from this year’s competition
2 d
theatlantic.com
Are You a Didion or a Babitz? Who Cares.
A new book compares authors and frenemies Joan Didion and Eve Babitz, but its weakness for gossip obscures the complicated truth.
2 d
theatlantic.com
The Key to Complex Life Might Lie Miles Below Our Feet
For a billion years, evolution was stagnant. Did tectonic plates supercharge it?
2 d
theatlantic.com
In Praise of Clarity
There is no ambiguity here.
2 d
theatlantic.com
Trump Is Handing China a Golden Opportunity on Climate
Already a leader in clean tech, China may see a new reason to act as leader in addressing climate change, too.
2 d
theatlantic.com
Helping Ukraine Survive Is Up to Europe Now
Trump is closer to Putin than to any of the continent’s democratic leaders.
2 d
theatlantic.com