Photographs by Eva O’Leary

Landon was supposed to be at her high-school graduation on Saturday. Instead, she was preparing to travel to Washington, D.C. By Monday, the 17-year-old from Gulfport, Mississippi, was wearing a long, sleek blue dress and dancing in front of the U.S. Capitol with about 100 of her transgender peers at the Trans Youth Prom.

Landon’s plans changed earlier this month after Harrison County School District Superintendent Mitchell King informed her that she needed to dress as a boy during the graduation ceremony, and not in a dress and heels as she’d planned. (Landon’s parents requested that she be identified by her first name only to protect her privacy.) After a federal judge on Friday denied a motion filed by the American Civil Liberties Union on her behalf demanding that she be allowed to wear her dress, she decided not to attend.

Landon had been out as a trans girl at her high school for four years and told me her classmates were in her corner. But ultimately she decided not to walk, because she did not want to attend an event that “does not support me,” she said. “There was no care; there was no compassion. It felt disrespectful,” she said.

But at the Capitol, there was no shortage of love and support. This prom was created by four kids—Daniel Trujillo, 15; Libby Gonzales, 13; Grayson McFerrin, 12; and Hobbes Chukumba, 16—with the help of some adult organizers as a way of celebrating being trans.

Left: Quetzal Gonzalez, 16, escapes the heat by eating an ice pop under a tree. Right: An attendee holds up a sign outside the U.S. Supreme Court after the Trans Youth Prom. Two hundred trans youths and their parents, as well as trans adults, protested outside the Supreme Court, where the words Equal Justice Under Law are etched above the entrance.

After exiting a bus in front of the Capitol Reflecting Pool, the kids and teens marched in their formal wear down a grassy runway flanked by cheering organizers, parents, and supporters who held up trans and gender-nonbinary flags and signs that read Trans kids have always existed, Trans youth are powerful, and Celebrate trans joy. The youth entered the party, dance music already blasting, through a large arch made of color-changing glass topped with the words You are loved.

Unlike a typical high-school prom, the event also included teens that had already completed high school. Alongside the drag queen Stormie Daie—the event’s emcee—they took turns showing off their moves inside a dance circle to hits by Lizzo, Lil Nas X, Madonna, and others. Occasionally, they would duck into the shade of nearby trees or chat and take pictures with family and friends.

Harleigh Walker, 16, and her father, Jeff Walker, in front of the Capitol Reflecting Pool. The Walkers traveled to Washington, D.C., from Auburn, Alabama, to attend the event. Left: Willow Soto holds a trans flag and flowers at the steps of the Supreme Court. Right: Landon holds the trans and gender-nonbinary flags next to her mother, Samantha. Landon decided not to attend her high-school graduation after she was informed that she would not be able to wear a dress to the ceremony. Many of the youths who attended traveled with their parents from hometowns where their gender identity is affirmed and embraced. Some came from states where politicians have been debating and legislating the details of their life.

Many of the youths who attended had traveled with their parents from hometowns where their gender identity is affirmed and embraced. Some came from states where politicians have been debating and legislating the details of their life—whether they can receive transition-related medical care, for example, or compete on sports teams that match their gender identity. In 2023 alone, more than 500 bills that would restrict the rights of transgender people have been introduced. (More than 70 of them have passed so far.)

Attendees of the Trans Youth Prom marched from the Capitol Reflecting Pool to the Supreme Court. In 2023 alone, more than 500 bills that would restrict the rights of transgender people have been introduced.

Left: Hildie Edwards walks through a decorative gate at the Trans Youth Prom. Right: The drag queen Stormie Daie, who served as the Trans Youth Prom’s emcee, walks through a large arch made of color-changing glass.

Michelle Callahan-DuMont and her husband, Andy DuMont, flew with their 9-year-old daughter, Violet, from Tucson, Arizona, to D.C. for the prom. Violet wore a sequined gold-and-blue dress that she told me she’d bought at Dillard’s. Her favorite music to dance to is techno, she said. As Michelle told me, “We’re always talking about the sad things; we’re always talking about the scary things, talking about maybe having to move out of state. And so for this weekend, we’re just having fun.” She said that during a dinner for the event the night before, Violet had told her parents: “This is the best day of my life.”

Zoé Anspach at the Trans Youth Prom on Monday To attendees, the happiness on display at the Trans Youth Prom was, in itself, an act of protest.

The choice of prom, a quintessential coming-of-age milestone for American kids, was deliberate: a message that lawmakers can’t take away people’s childhoods, Chase Strangio, one of the event’s adult organizers and the deputy director for transgender justice at the ACLU, told me. “This is trans people defining their joy on their own terms, coming together to say, ‘Enough. We don’t need you to debate us any longer. We know exactly who we are.’”

After the speeches and dancing, about 200 people in attendance marched down Constitution Avenue, chanting and cheering to music, before stopping at the steps of the Supreme Court. Underneath the words Equal Justice Under Law etched on the front of the high court, they chanted: “Trans rights are human rights!”

Left: Violet DuMont, 9, traveled with her parents to D.C. from Tucson, Arizona. Her mother, Michelle Callahan-DuMont, said, “We’re always talking about the sad things; we’re always talking about the scary things, talking about maybe having to move out of state. And so for this weekend, we’re just having fun.” Right: Prom attendees march to the Capitol. The group protested anti-trans legislation with banners and signs outside the Supreme Court.

Three and a half years since the start of a pandemic that has killed millions of people and debilitated countless more, the world is still stuck at the start of the COVID-19 crisis in one maddening way: No one can say with any certainty how, exactly, the outbreak began. Many scientists think the new virus spilled over directly from a wild animal, perhaps at a Chinese wet market; some posit that the pathogen leaked accidentally from a local laboratory in Wuhan, China, the pandemic’s likely epicenter. All of them lack the slam-dunk evidence to prove one hypothesis and rule out the rest.

That’s not to say nothing has changed. Those embroiled in the origins fracas now have much more data to scrutinize, debate, and re-debate. In March, I reported that the case for a zoonotic origin had acquired a consequential new piece of support: An international team of scientists had uncovered genetic data, collected from a wet market in Wuhan in the weeks after the venue was closed on January 1, 2020, that linked the coronavirus to wild animals. This evidence, they said, indicated that one of those creatures could have been shedding SARS-CoV-2, the virus that causes COVID-19; one of the most intriguing bits of data pointed to raccoon dogs, a foxlike creature that was already known to be vulnerable to the virus. The finding wasn’t direct evidence of an animal infection, but, stacked alongside other clues, ​​“this really strengthens the case for a natural origin,” Seema Lakdawala, a virologist at Emory University who wasn’t involved in the research, told me at the time.

Not everyone agreed that the finding counted as a substantial new insight. When the researchers who originally collected the samples, many of them from the Chinese Center for Disease Control and Prevention, published their own analysis of the data in April—a revision of an earlier report—they emphasized that there was no clear evidence that the virus had been introduced to the market by a wild animal. Then, this month, Jesse Bloom, a computational biologist at Fred Hutchinson Cancer Center, in Seattle, posted a third analysis of the market data, inspired in part, he told me, by his concern that the public discussion of the initial findings, and their connection to raccoon dogs, had overinflated their worth. The international team’s report, he argued, hardly moved the needle on the origins debate at all—certainly not “much beyond where it was before,” he told me.

Bloom’s analysis, too, set off a wave of fervor—including a fresh spate of claims that he told me were “exaggerated,” or even outright wrong; some even asserted, for instance, that his preprint proves that raccoon dogs “weren’t infected, which is not an accurate summary,” he said. All the while, researchers have been squabbling on social media over the minutiae of statistical methodology, and what constitutes a meaningful amount of viral RNA; some have even come to loggerheads publicly at conferences.

At the crux of this particular fight is a difference of interpretation, with one camp of researchers contending that the recent data matter a lot, and another asserting that they matter much less, or perhaps not even a little bit at all. Under most other circumstances, a scientific scuffle this deep in the weeds might hold the attention of a few dozen people for a few months at best. Here, though, the central topic is one of the most consequential in recent memory—a virus that’s left its mark on the world’s entire population, and will continue to do so. Which has made it easy for pitched battles over differences in scientific opinion to become a public spectacle—and difficult, maybe even impossible, for the debate to ever end, no matter what evidence might emerge next.

The genetic sequences analyzed in the March report contained evidence of a zoonotic origin that is more circumstantial than direct. Researchers extracted them from swabs taken from surfaces in and around Wuhan’s Huanan Seafood Wholesale Market from January to March of 2020, weeks after the first known COVID cases were documented in Wuhan. That makes these environmental samples “a useful part of the story,” Alice Hughes, a conservation biologist at the University of Hong Kong, told me. Though, by themselves, “they are limited in what they are able to tell us.”

By the time the swabs were collected by China CDC researchers, Chinese officials had hastily closed Huanan; many vendors had likely disappeared with their animals, or culled them en masse. The swabs could show only where the virus had once been, or which animals the venue had sold—more akin to dusting a crime scene for fingerprints than catching a vagrant in the act. And although they could show where animal and viral genetic material had mixed, they couldn’t guarantee that those two types of genetic material had been deposited at the same time. Nor could they distinguish between, say, a sick creature sneezing on the bars of its cage and an infected human coughing on an enclosure housing healthy wildlife. Those answers could have come from swabs taken directly from the noses or mouths of live animals for sale at Huanan in late November or early December of 2019. But as far as researchers know, those swabs don’t exist—or at least, the public has no record of them.

The sequences from these environmental samples, then, are “what we have,” says Katherine Xue, a computational virologist at Stanford who previously worked with Jesse Bloom, the author of the May preprint, but was not involved in any of the new reports. And “we want to do what we can with what we have.” When the international team behind the March analysis found that several market samples contained genetic material from both the virus and a wild animal known to be susceptible to it—including the common raccoon dog—they said that the best explanation for this commingling was an infection.

As I reported at the time, the data don’t constitute direct evidence of an infected raccoon dog at the market. “But this is exactly what we would observe if infected raccoon dogs were in fact present in this location,” says Kristian Andersen, a computational biologist and virologist at the Scripps Research Institute and one of the authors of the March analysis. Which, they wrote in their analysis, “identifies these species, particularly the common raccoon dog, as the most likely conduits for the emergence of SARS-CoV-2 in late 2019.”

Other researchers, though, think that calling the evidence even supportive of an animal origin for the outbreak is a stretch. The samples were taken too long after the outbreak’s start to be meaningful, some said; the data were too shaky to even hint at the idea of an infected raccoon dog, others insisted, much less one that might have passed the virus to us.

Bloom, too, was unswayed. The swabs contained genetic material from many creatures at the market—some of them alive, some dead; some that we now know can host the virus, others that almost certainly do not. In Bloom’s analysis, he explains that the species repeatedly highlighted as potential hosts weren’t the animals that were most frequently and notably commingled with the virus in the market swabs. “If you’re trying to figure out if there is a meaningful association between raccoon dog and viral genetic material,” he told me, there should be a lot of raccoon-dog genetic material in the places where the virus was found, and far less where the virus was not.

But that wasn’t the case for raccoon dogs—or “any of the animals that could conceivably have been infected,” Bloom told me. Instead, in his analysis he saw the virus most closely linked to several kinds of fish, which aren’t known to be viable hosts for it. People, Bloom told me, were the probable source of SARS-CoV-2 in those spots. All of that “probably just suggests that it had been spread around the market by humans by the time” the swabs were taken, diminishing the samples’ usefulness.

Several other scientists not involved in Bloom’s preprint were quick to point out the limits and flaws in his approach. To draw meaningful conclusions from this type of analysis, researchers would need samples amassed at about the same time, with the same collection goals in mind. That wasn’t the case for these samples, Zach Hensel, a biophysicist who has been publicly critical of Bloom’s report, told me. Researchers took them over the course of many weeks after Huanan’s closure, altering their tactics as more intel came to light. A first foray into the market, for instance, targeted the parts of the venue where COVID cases had been identified, a strategy that would, by design, turn up more virus-positive samples; another, conducted days later, focused on stalls that had been discovered to have housed wildlife, regardless of their proximity to sick people. Many samples in the latter set, then, would be expected to be virus-negative—and were. Sloshing them together with the first set of swabs and trying to pull patterns out could end up masking actual associations between the virus and any wild animal hosts.

Bloom also points out that many of the swabs that turned up mammalian DNA, including one containing raccoon-dog genetic sequences that some members of the international team initially emphasized, had relatively little material from the virus on them. But genetic material, especially RNA—the basis of SARS-CoV-2’s genome—degrades fast; a difference of even a few days could artificially deflate how important a particular swab looked. Alice Hughes also pointed out that certain market locales highlighted in Bloom’s preprint, including surfaces around duck or fish tanks, might have better preserved viral RNA simply because they were cold or damp. When I brought up these concerns with Bloom, he admitted “there are certainly a lot of confounders” that could have skewed his results. His main goal, he said, was just to show that “the samples are not sufficient to answer whether or not there were infected animals.”

Bloom’s re-analysis doesn’t mark a major shift in thinking for Hughes, who told me she thinks “there is reasonable support for a zoonotic origin.” Felicia Goodrum, a virologist and an immunologist at the University of Arizona who has written repeatedly on the origins debate but was not involved in the team’s analysis, agrees. The Huanan market is “most likely where the spillover occurred,” she told me. “I really, truly believe that, based on the accumulation of the evidence.”

Data never sit alone in a vacuum: They’re amassed, interpreted, and reinterpreted alongside the totality of evidence that precedes them. By themselves, the sequences from the Huanan market couldn’t say much. But they fit a broader, more detailed scenario that researchers on the team behind the March analysis had been exploring for years.

History has always supported a zoonotic scenario: A wet-market spillover is what researchers are fairly certain started the SARS outbreak in China in 2002, potentially via infected masked palm civets. In this latest outbreak, the Huanan Market was one of only four wet markets in all of Wuhan that has consistently been documented selling an array of live, coronavirus-susceptible wildlife; the earliest known COVID cases were detected near the venue, centering “on it like a bull’s-eye,” says Michael Worobey, an evolutionary biologist at the University of Arizona and one of the authors of the March report. Scientists analyzing genetic sequences collected from the venue have also detected two distinct coronavirus lineages from the outbreak’s earliest days—a likely indication, some researchers have argued, that the pathogen spilled over from animals into humans twice.

The missing clincher for them is which creature might have initially carried the virus into the market. The raccoon-dog swab was particularly compelling to the team not only because it contained gobs of animal genetic sequences, and very few human ones—but also because it had been plucked from a stall where Eddie Holmes, one of the report’s authors, had snapped a photo of a raccoon dog in a cage years before. The clues to a possible animal host, Worobey told me, were “right in the very stall we said they would be.”

But data are also amassed, interpreted, and reinterpreted by humans, who have their own biases. The experts now quarreling over the importance of the recent data approached the new evidence having already drawn tentative conclusions and made their opinions known. Kristian Andersen was an early proponent of a zoonotic origin, and has repeatedly denounced the notion of a lab leak; Worobey was later to voice his support for the zoonotic hypothesis, but is now no less enthusiastic. And long before they and their colleagues stumbled across the data that yielded their March analysis, which didn’t become publicly available until recently, the researchers had been hoping that such sequences would appear—noting in a 2022 paper that this sort of intel could constitute an essential and still missing puzzle piece. Now that the evidence has emerged, and fits with their established thinking, it feels validating, Worobey told me.

Bloom, by contrast, has long positioned himself as an agnostic moderate, and isn’t yet budging from his neutral territory. Others who have come out vocally in favor of a lab-leak scenario have cast their own doubts on the international team’s analysis. In a landscape so sparsely populated by data, it gets all too easy for people to fill in the gaps with speculation; “what starts off as a weak preference,” Hughes told me, “becomes almost like a religion.” I’ve been reporting now for three years on many controversial COVID stories, along the way interviewing hundreds of opinionated scientists about dozens of thorny questions. Through it all, this debate has stood out for being so ignitable. Individual data points have become catalysts; single statements have been endlessly scrutinized. And experts have staked out territory and stuck to it almost dogmatically—many of them to the point of avoiding admitting past mistakes. COVID’s origins are now shrouded in combustible gas, with matches scattered everywhere: Lighting up a single point, normally harmless enough, inevitably sets off a conflagration.

All of this leaves the world trying to peer through the smoke. “All hypotheses are on the table,” Maria Van Kerkhove, the World Health Organization’s technical lead on COVID-19, told me. “We can’t take any off.” To her mind, though, “there’s much more evidence to support a zoonotic origin.”

More evidence could still emerge. The international team isn’t yet done analyzing the Chinese researchers’ original data set, which was recently released in fuller form. They’re eager to mine the sequences to tease out the subspecies of some of the market’s potential SARS-CoV-2 hosts, which could inform searches for the virus out in nature or on animal farms; other experiments, analyzing how degraded certain genetic samples are, could hint at how much time passed between the moment the biological material was dropped and the moment it was picked up. Van Kerkhove has also separately been pressing the Chinese researchers for more information on how these and other samples might have been collected, and any intel on where the market’s animals might have been sourced from—which could guide searches for evidence of the virus or its relatives on farms or in the wild. These bits of data, too, would all be incremental,with no single shred of evidence acting as total proof or disproof. But each could constitute a clue, Van Kerkhove told me, to continue nudging the conversation along.

In the grand scheme of things, though, the world probably won’t ever get data that will conclusively end the debate. Even if scientists were to turn up virus-positive samples from a live creature from the market—direct evidence of an infected animal—it would remain technically possible that a human caught the virus first, then passed it on to the venue’s wildlife. But data that aren’t debate-ending can still be notable. And the recent sequences from the market swabs could easily, and frustratingly, end up being one of the best clues to the pandemic’s roots that the world is likely to get.

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.

Senator Tim Scott today joined the ranks of GOP candidates hoping to displace Donald Trump as the party’s nominee. America would be better off if one of them could win, but the GOP is no longer a normal political party.

First, here are four new stories from The Atlantic:

Beware of the food that isn’t food. Harlan Crow wants to stop talking about Clarence Thomas. Where living with friends is still technically illegal A firearm-owning Republican’s solutions for gun violence

Thanos From Queens

Tim Scott of South Carolina joined the field of Republican contenders for the GOP presidential nomination today. He’s polling in single digits among primary voters, as are all of the other (so far) declared candidates. Only Governor Ron DeSantis of Florida is managing to get out of the basement—rumors are that he will announce his candidacy this week—and even he is getting walloped by Donald Trump in polls of the Republican faithful.

Scott seems like a classic no-hoper presidential prospect but a strong choice for vice president, which of course is why some weaker candidates run and then bow out (see “Harris, Kamala”). The current GOP field, however, includes at least some politicians who should be credible alternatives to Trump: In any other year, people such as DeSantis, Nikki Haley, and Asa Hutchinson, all current or former governors from the South, would be obvious contenders. Instead, their campaigns are flailing about in limbo while the rest of the field is populated by the likes of the wealthy gadfly Vivek Ramaswamy and the radio-talk-show host Larry Elder.

Of course, in a normal year, a twice-impeached president who has been held liable for sexual abuse would do the decent thing and vanish from public life.

The United States desperately needs a normal presidential election, the kind of election that is not shadowed by gloom and violence and weirdos in freaky costumes pushing conspiracy theories. Americans surely remember a time when two candidates (sometimes with an independent crashing the gates) had debates, argued about national policy, and made the case for having the vision and talent and experience to serve as the chief executive of a superpower. Sure, those elections were full of nasty smears and dirty tricks, but they were always recognizable as part of a grand tradition stretching all the way back to Thomas Jefferson and John Adams—rivals and patriots who traded ugly blows—of contenders fighting hard to secure the public’s blessing to hold power for four years.

Such an election, however, requires two functional political parties. The Republicans are in the grip of a cult of personality, so there’s little hope for a normal GOP primary and almost none for a traditional presidential election. Meanwhile, Republican candidates refuse to take a direct run at Donald Trump and speak the truth—loudly—to his voters; instead, they talk about all of the good that Trump has done but then plead with voters to understand that Trump is unelectable. (Hutchinson, who is unequivocal in his view of Trump, has been an honorable exception here and has called for Trump to drop out.)

The electability argument about Trump is not only amoral, but it also might not even be true: Trump might be able to win again. In normal times, there’s nothing wrong with “electability” arguments. It is hardly the low road, if presented with two reasonable candidates in a primary, to choose the one who can prevail in a general election. But such a choice assumes the existence of  “reasonable” candidates. Instead, some of the Republicans who are running or leaning toward running against Trump are saying, in effect, that Trump really should be the candidate, but he can’t win—instead of saying, unequivocally, that no decent party should ever nominate this man again, whether he can win or not.

Republican contenders are caught in a bind. If they run against Trump, they will likely lose. But if they don’t run against Trump, they will certainly lose—to Trump, and then everyone in America loses. GOP primary candidates want to pick up Trump’s voters without overtly selling them Trump’s lies and conspiracy theories, which is why the “electability” dodge is nothing but pandering and cowardice. Not that any of these hopefuls have tried to lay a punch on Trump: Haley is AWOL—is she even still running?—and DeSantis is busy clomping around with flaming wastebaskets on his feet as he tries to stomp out fires he’s already set.

Tim Scott is an especially vexing case, because he has a life story that should have made him the natural anti-Trump candidate in every way. A religious man who triumphed over poverty, got an education, and became a successful businessman, his life and character are a photo-negative image of Trump’s. And yet, Scott can’t help himself: He’s “thankful” for Trump’s years in office.

None of these Republicans are going to overcome the Thanos from Queens, who, with a snap of his fingers, will soon make half of the GOP field disappear.

These Republicans are likely waiting for a miracle, an act of God that takes Trump out of contention. And by “act of God,” of course, they mean “an act of Fani Willis or Jack Smith.” This is a vain hope: Without a compelling argument from within the Republican Party that Fani Willis and Jack Smith or for that matter, Alvin Bragg, are right to indict Trump—as Bragg has done and Willis and Smith could do soon—and that the former president is a menace to the country, Trump will simply brush away his legal troubles and hope he can sprint to the White House before he’s arrested.

No one is going to displace Trump by running gently. A candidate who takes Trump on, with moral force and directness, might well lose the nomination, but he or she could at least inject some sanity into the Republican-primary process and set the stage for the eventual recovery—a healing that will take years—of the GOP or some reformed successor as a center-right party. DeSantis would rather be elected as Trump’s Mini-Me. (It might work.) Hutchinson has tried to speak up, but too quietly. Haley, like so many other former Trump officials, is too compromised by service to Trump to be credible as his nemesis. Tim Scott is perfectly positioned to make the case, but he won’t.

A Republican who thinks Trump can be beaten in a primary by gargling warm words such as electability is a Republican in denial. Trump is already creating a reality-distortion field around the primary, as he will again in the general election. Is it possible that the GOP base would respond to some fire and brimstone about Trump, instead of from him? We cannot know, because it hasn’t been tried—yet.

Related:

Why outspoken women scare Trump America’s lowest standard

Today’s News

The head of Russia’s Wagner mercenary group has vowed to transfer the Ukrainian city of Bakhmut to the Russian army by June 1. Ukraine insists that the city has not been entirely captured. Arizona, California, and Nevada agreed on a plan to reduce water usage from the drought-stricken Colorado River.   Speaker Kevin McCarthy said U.S.-debt-ceiling talks were on the “right path” ahead of a meeting with President Joe Biden this evening.  

Dispatches

Famous People: Lizzie and Kaitlyn down “Pumptinis” at a live screening of the scariest show on TV.

Explore all of our newsletters here.

Evening Read

Apple TV+

Martin Scorsese’s Killers of the Flower Moon Is a Triumph

By David Sims

David Grann’s nonfiction book Killers of the Flower Moon: The Osage Murders and the Birth of the FBI is the sprawling story of a criminal investigation undoing a systemic evil. It lays out in riveting detail the mystery of the Osage murders of the 1920s, when dozens of Native Americans were killed in a grand conspiracy to exploit their oil-rich land. Grann digs into the societal phenomenon surrounding the Osage, many of whom became ultra-wealthy after generations of displacement and persecution. But the book’s through line is the federal investigator Tom White, who helped solve the murders on the orders of a young J. Edgar Hoover.

Martin Scorsese’s adaptation, which premiered at this year’s Cannes Film Festival and will be released in theaters this October, takes a very different narrative approach.

Read the full article.

More From The Atlantic

The Roys stumble into the real world. A world without Martin Amis My friend, Tim Keller

Culture Break

Illustration by The Atlantic. Source: Getty.

Read. The Princess Casamassima, a novel written more than 100 years ago (and originally serialized by The Atlantic!), is a political novel that could’ve been written today.

Listen. The first podcast episode of our new podcast series How to Talk to People, which explores the barriers to good small talk.

Play our daily crossword.

P.S.

I’m concerned about events at the Zaporizhzhia nuclear plant, where the Russians have apparently dug in for a fight. I’m especially concerned that the Kremlin, facing a Ukrainian counteroffensive, might be planning a nuclear disaster in retaliation for losing more ground. That hasn’t happened yet, and I promise I’ll come back to this if events change.

In the meantime, however, the danger at the Ukrainian nuclear installation has jogged loose a memory of a lost bit of music from the 1980s. After the 1986 disaster at Chernobyl—also in Ukraine—the New Zealand musician Shona Laing released a song in 1987 titled “Soviet Snow.” (You can see the video here.) Given my, ah, heterodox musical tastes, you might be surprised that I would like something with such obvious environmental advocacy. (Don’t tell the other young Ronald Reagan voters, but I also bought Bruce Cockburn’s Stealing Fire album in 1984, and I still like it.) There is an urgency and panic in the song, a strong New Wave feel over Laing’s plea:

Are we wide awake? Is the world aware?

Radiation over Red Square

Creeping on to cross Roman roads

I remember feeling a great unease hearing that song the first time. Thirty-six years later, I am feeling that same unease once more.

— Tom

Katherine Hu contributed to this newsletter.

Arizona, California and Nevada agree to cut water usage from the Colorado River by 13 percent over three years.

Several centuries ago, as scientists began to embrace the startling idea that Earth was not the center of the universe, they also began to ponder its startling implication: that the stars in the night sky might be suns in their own right, orbited by their own worlds. Until the 1990s, that idea was no more than a hypothesis. By then, telescopes had become sufficiently advanced to reveal the hard evidence: A star about 50 light-years away was wobbling, a sign that a small world was tugging on it. This world was called an exoplanet.

Astronomers have since discovered more than 5,300 exoplanets and counting, and they’re studying the atmosphere of these worlds to determine if the molecules suspended in their clouds could sustain life. They’ve even recently captured evidence of an exoplanet getting swallowed up by its dying star. (Truly, telescopes have gotten really good.) Although most people probably can’t name an exoplanet—something like “HD 108236 b” doesn’t exactly roll off the tongue—the fact that the cosmos is full of them is now well known.

The exo- prefix extends beyond the realm of planets. Astronomers have found exoauroras, exoasteroid belts, and even exorings, such as the ones that surround Saturn. Exomoons haven’t been proved to exist, but astronomers believe that they are more numerous than exoplanets themselves. The search for such celestial objects and structures has intensified in recent years. Just this week, astronomers announced that they have found an exoradiation belt, an invisible cocoon of charged particles, held in place by a planet’s magnetic field. We are firmly in the era of exothings.

[Read: Faraway planets don’t seem so distant anymore]

There is plenty of scientific possibility in the discovery of exothings, and also comfort. Not to be dramatic, but Earth is a small island in an endless sea. Ours is a cosmically lonely existence. So it’s nice to look out across that sea, dark and unknowable, and occasionally spot other islands, other familiar shores. The study of exothings is a way of conversing with the rest of the cosmos and saying, “Oh, you’ve got one of those too?” And because any hypothetical life must live somewhere, exothings could someday answer one of our most existential questions: Are we alone?

When Melodie Kao, an astronomer at UC Santa Cruz, went looking for exothings, she set her sights 20 light-years away, on a brown dwarf—a rather unusual object that is neither a star nor a planet, with a mass somewhere between the two. Kao had previously studied this particular brown dwarf and found that its intense radio emissions produced auroras similar to our own northern lights. She couldn’t account for what was causing the aurora, so she turned to our own solar system for inspiration, and realized that planets with auroras also have radiation belts. No one had ever found one outside the solar system before. But when Kao went looking for one around her brown dwarf, voilà, there it was: a radiation belt, invisible to the human eye but billowing in radio wavelengths, in all its exoglory.

Exoasteroid belts are downright stunning, particularly in infrared wavelengths. Astronomers recently used the most powerful observatory in operation, the James Webb Space Telescope, to observe a star about 25 light-years away with a known asteroid belt. They were surprised to find not one single ring of material, but three. The dusty region was likely shaped by the same forces that produced the versions in our own solar system: the movements of planets, hidden somewhere in those concentric rings. “It’s sort of comforting that the same processes likely play out, to some extent, very similarly,” Andras Gaspar, an astronomer at the University of Arizona who led the observations, told me. Perhaps, he added, we’re “not the only ones who are deciphering how our universe works.”

An exoasteroid belt (NASA / ESA / CSA / A. Gaspar (University of Arizona) / A. Pagan (STScI))

Although the universe is full of repeats, each iteration brings its own little quirks. The exoradiation belt that Kao discovered, for example, is “almost 10 million times more intense” than Earth’s, Kao said. The exoasteroid belt is far more complex than our main asteroid belt, between Mars and Jupiter, as well as the Kuiper Belt farther out, a region beyond Neptune of icy objects and dust. And the only known set of exorings is 200 times larger than Saturn’s—possibly because it sits around a young sunlike star, not a planet.

Some exothings have proved elusive, even though astronomers would bet all the money in the universe that they exist. Exomoons are one of them: Our solar system is full of moons, so others must be too. But no one’s been able to prove they exist. In 2018, a pair of astronomers said they had found evidence of what could be the first-known exomoon, orbiting a planet around a star 8,000 light-years from here. But when other teams took up the same data and conducted their own analyses, they got mixed results. The search continues.

[Read: We’ve found 5,000 exoplanets, and we’re still alone]

It continues as well for that most elusive exothing: a truly Earthlike planet, nice and rocky, with a chemically rich atmosphere and temperatures that would allow water to lap on the surface. Mansi Kasliwal, an astronomy professor at Caltech who was involved in the discovery of the unfortunate exoplanet consumed by its star, told me she thinks a lot about the idea of an exohome. Not another place in the universe where humans might relocate (though she hopes future generations take a crack at that before our sun swallows our own planet in a few billion years), but some other speck out there where life managed to emerge and thrive. “It doesn’t even have to be identical to Earth,” she said. “It just needs to be habitable and hospitable, a planet that is home to some species.”

Our collection of exothings large and small is growing every year. “Until recently, I had a lot of exoplanet envy,” says Jackie Villadsen, a physics and astronomy professor at Bucknell University who worked with Kao. “Now we can do exoradiation belts.” Compared with exoplanets—which can have oceans of lava and raindrops made of glass—radiation belts might seem a little boring. But exothings, whatever they are, lend themselves to daydreaming. Consider the exoasteroid belt that Gaspar studies. The star at its center is about 440 million years old, much younger than our approximately 4.6-billion-year-old sun. When our sun was that age, Earth was still lifeless. “It’s very doubtful that any life could have emerged, let alone sophisticated, intelligent life forms,” Gaspar said.

But maybe sometime in the cosmic future, a spark will flash. Someday, perhaps, a group of beings on a planet inside that system will dispatch little robots into the depths, to the other planets and moons nearby. They will sprinkle signs of their existence beyond their cozy atmosphere, all the way out to their asteroid belt and beyond. After all, it happened here.

One-hundred percent AI. That’s what the software concluded about a student’s paper. One of the professors in the academic program I direct had come across this finding and asked me what to do with it. Then another one saw the same result—100 percent AI—for a different paper by that student, and also wondered: What does this mean? I did not know. I still don’t.

The problem breaks down into more problems: whether it’s possible to know for certain that a student used AI, what it even means to “use” AI for writing papers, and when that use amounts to cheating. The software that had flagged our student’s papers was also multilayered: Canvas, our courseware system, was running Turnitin, a popular plagiarism-detection service, which had recently installed a new AI-detection algorithm. The alleged evidence of cheating had emerged from a nesting doll of ed-tech black boxes.

This is college life at the close of ChatGPT’s first academic year: a moil of incrimination and confusion. In the past few weeks, I’ve talked with dozens of educators and students who are now confronting, for the very first time, a spate of AI “cheating.” Their stories left me reeling. Reports from on campus hint that legitimate uses of AI in education may be indistinguishable from unscrupulous ones, and that identifying cheaters—let alone holding them to account—is more or less impossible.

Once upon a time, students shared exams or handed down papers to classmates. Then they started outsourcing their homework, aided by the internet. Online businesses such as EssayShark (which asserts that it sells term papers for “research and reference purposes only”) have professionalized that process. Now it’s possible for students to purchase answers for assignments from a “tutoring” service such as Chegg—a practice that the kids call “chegging.” But when the AI chatbots were unleashed last fall, all these cheating methods of the past seemed obsolete. “We now believe [ChatGPT is] having an impact on our new-customer growth rate,” Chegg’s CEO admitted on an earnings call this month. The company has since lost roughly $1 billion in market value.

Other companies could benefit from the same upheaval. By 2018, Turnitin was already taking more than $100 million in yearly revenue to help professors sniff out impropriety. Its software, embedded in the courseware that students use to turn in work, compares their submissions with a database of existing material (including other student papers that Turnitin has previously consumed), and flags material that might have been copied. The company, which has claimed to serve 15,000 educational institutions across the world, was acquired for $1.75 billion in 2019. Last month, it rolled out an AI-detection add-in (with no way for teachers to opt out). AI-chatbot countermeasures, like the chatbots themselves, are taking over.

Now, as the first chatbot spring comes to a close, Turnitin’s new software is delivering a deluge of positive identifications: This paper was “18% AI”; that one, “100% AI.” But what do any of those numbers really mean? Surprisingly—outrageously—it’s very hard to say for sure. In each of the “100% AI” cases I heard about, students insisted that they had not let ChatGPT or any other AI tool do all of their work.

But according to the company, that designation does indeed suggest that 100 percent of an essay—as in, every one of its sentences—was computer generated, and, further, that this judgment has been made with 98 percent certainty. A Turnitin spokesperson acknowledged via email that “text created by another tool that uses algorithms or other computer-enabled systems,” including grammar checkers and automated translators, could lead to a false positive, and that some “genuine” writing can be similar to AI-generated writing. “Some people simply write very predictably,” she told me. Are all of these caveats accounted for in the company’s claims of having 98 percent certainty in its analyses?

Perhaps it doesn’t matter, because Turnitin disclaims drawing any conclusions about misconduct from its results. “This is only a number intended to help the educator determine if additional review or a discussion with the student is warranted,” the spokesperson said. “Teaching is a human endeavor.” The company has a guide for humans who confront the software’s “small” risk of generating false positives. Naturally, it recommends the use of still more Turnitin resources (an AI-misuse rubric and AI-misuse checklist are available) and doing more work than you ever would have done in the first place.

[​​Read: ChatGPT is about to dump more work on everyone]

In other words, the student in my program whose work was flagged for being “100% AI” might have used a little AI, or a lot of AI, or maybe something in between. As for any deeper questions—exactly how he used AI, and whether he was wrong to do so—teachers like me are, as ever, on our own.

Some students probably are using AI at 100 percent: to complete their work absent any effort of their own. But many use ChatGPT and other tools to generate ideas, help them when they’re stuck, rephrase tricky paragraphs, or check their grammar.

Where one behavior turns into another isn’t always clear. Matthew Boedy, an English professor at the University of North Georgia, told me about one student so disengaged, he sometimes attended class in his pajamas. When that student submitted an uncharacteristically adept essay this spring, Boedy figured a chatbot was involved, and OpenAI’s verification tool confirmed as much. The student admitted that he hadn’t known how to begin, so he asked ChatGPT to write an introduction, and then to recommend sources. Absent a firm policy on AI cheating to lean on, Boedy talked through the material with the student in person and graded him based on that conversation.

A computer-science student at Washington University in St. Louis, where I teach, saw some irony in the sudden shift from giving fully open-book assignments earlier in the pandemic to this year’s attitude of “you can use anything except AI.” (I’m withholding the names of students so that they can be frank about their use of AI tools.) This student, who also works as a teaching assistant, knows firsthand that computers can help solve nearly every technical exercise that is assigned in CS courses, and some conceptual ones too. But taking advantage of the technology “feels less morally bankrupt,” he said, “than paying for Chegg or something.” A student who engages with a chatbot is doing some kind of work for themselves—and learning how to live in the future.

Another student I spoke with, who studies politics at Pomona College, uses AI as a way to pressure-test his ideas. Tasked with a research paper on colonialism in the Middle East, the student formulated a thesis and asked ChatGPT what it thought of the idea. “It told me it was bogus,” he said. “I then proceeded to debate it—in doing so, ChatGPT brought up some serious counterarguments to my thesis that I went on to consider in my paper.” The student also uses the bot to recommend sources. “I treat ChatGPT like a combination of a co-worker and an interested audience,” he said.

[Read: The college essay is dead]

The Pomona student’s use of AI seems both clever and entirely aboveboard. But if he borrows a bit too much computer-generated language, Turnitin might still flag his work for being inauthentic. A professor can’t really know whether students are using ChatGPT in nuanced ways or whether they’ve engaged in brazen cheating. No problem, you might say: Just develop a relationship of mutual trust with students and discuss the matter with them openly. A good idea at first blush, but AI risks splitting faculty and student interests. “AI is dangerous in that it’s extremely tempting,” Dennis Jerz, a professor at Seton Hill University, in Greensburg, Pennsylvania, told me. For students who are not invested in their classes, the results don’t even have to be good—just good enough, and quick. “AI has made it much easier to churn out mediocre work.”

Faculty already fret over getting students to see the long-term benefit of assignments. Their task is only getting harder. “It has been so completely demoralizing,” an English teacher in Florida told me about AI cheating. “I have gone from loving my job in September of last year to deciding to completely leave it behind by April.” (I am not printing this instructor’s name or employer to protect him from job-related repercussions.) His assignments are typical of composition: thesis writing, bibliographies, outlines, and essays. But the teacher feels that AI has initiated an arms race of irrelevance between teachers and students. “With tools like ChatGPT, students think there’s just no reason for them to care about developing those skills,” he said. After students admitted to using ChatGPT to complete assignments in a previous term—for one student, all of the assignments—the teacher wondered why he was wasting his time grading automated work the students may not have even read. That feeling of pointlessness has infected his teaching process. “It’s just about crushed me. I fell in love with teaching, and I have loved my time in the classroom, but with ChatGPT, everything feels pointless.”

The loss that he describes is deeper and more existential than anything academic integrity can protect: a specific, if perhaps decaying, way of being among students and their teachers. “AI has already changed the classroom into something I no longer recognize,” he told me. In this view, AI isn’t a harbinger of the future but the last straw in a profession that was almost lost already, to funding collapse, gun violence, state overreach, economic decay, credentialism, and all the rest. New technology arrives on that grim shore, making schoolwork feel worthless, carried out to turn the crank of a machine rather than for teaching or learning.

What does this teacher plan to do after leaving education, I wonder, and then ask. But I should have known the answer, because what else is there: He’s going to design software.

A common line about education in the age of AI: It will force teachers to adapt. Athena Aktipis, a psychology professor at Arizona State University, has taken the opportunity to restructure her whole class, preferring discussions and student-defined projects to homework. “The students said that the class really made them feel human in a way that other classes didn’t,” she told me.

But for many students, college isn’t just a place for writing papers, and cutting corners can provide a different way of feeling human. The student in my program whose papers raised Turnitin’s “100% AI” flag told me that he’d run his text through grammar-checking software, and asked ChatGPT to improve certain lines. Efficiency seemed to matter more to him than quality. “Sometimes I want to play basketball. Sometimes I want to work out,” he said when I asked if he wanted to share any impressions about AI for this story. That may sound outrageous: College is for learning, and that means doing your assignments! But a milkshake of stressors, costs, and other externalities has created a mental-health crisis on college campuses. AI, according to this student, is helping reduce that stress when little else has.

[Read: The end of recommendation letters]

Similar pressures can apply to teachers too. Faculty are in some ways just as tempted as their students by the power of the chatbots, for easing work they find irritating or that distract from their professional goals. (As I pointed out last month, the traditional recommendation letter may be just as threatened by AI as the college essay.) Even so, faculty are worried the students are cheating themselves—and irritated that they’ve been caught in the middle. Julian Hanna, who teaches culture studies at Tilburg University, in the Netherlands, thinks the more sophisticated uses of AI will mostly benefit the students who were already set to succeed, putting disadvantaged students even further at risk. “I think the best students either don’t need it or worry about being caught, or both.” The others, he says, risk learning less than before. Another factor to consider: Students who speak English as a second language may be more reliant on grammar-checking software, or more inclined to have ChatGPT tune up their sentence-level phrasing. If that’s the case, then they’ll be singled out, disproportionately, as cheats.

One way or another, the arms race will continue. Students will be tempted to use AI too much, and universities will try to stop them. Professors can choose to accept some forms of AI-enabled work and outlaw others, but their choices will be shaped by the software that they’re given. Technology itself will be more powerful than official policy or deep reflection.

Universities, too, will struggle to adapt. Most theories of academic integrity rely on crediting people for their work, not machines. That means old-fashioned honor codes will receive some modest updates, and the panels that investigate suspected cheaters will have to reckon with the mysteries of novel AI-detection “evidence.” And then everything will change again. By the time each new system has been put in place, both technology and the customs for its use could well have shifted. ChatGPT has existed for only six months, remember.

Rethinking assignments in light of AI might be warranted, just like it was in light of online learning. But doing so will also be exhausting for both faculty and students. Nobody will be able to keep up, and yet everyone will have no choice but to do so. Somewhere in the cracks between all these tectonic shifts and their urgent responses, perhaps teachers will still find a way to teach, and students to learn.

More than three years ago, the coronavirus pandemic officially became an emergency, and much of the world froze in place while politicians and public-health advisers tried to figure out what on Earth to do. Now the emergency is officially over—the World Health Organization declared so on Friday, and the Biden administration will do the same later this week.

Along the way, almost 7 million people died, according to the WHO, and looking back at the decisions made as COVID spread is, for the most part, a demoralizing exercise. It was already possible to see, in January 2020, that America didn’t have enough masks; in February, that misinformation would proliferate; in March, that nursing homes would become death traps, that inequality would widen, that children’s education, patients’ care, and women’s careers would suffer. What would go wrong has been all too clear from the beginning.

Not every lesson has to be a cautionary tale, however, and the end of the COVID-19 emergency may be, if nothing else, a chance to consider which pandemic policies, decisions, and ideas actually worked out for the best. Put another way: In the face of so much suffering, what went right?

To find out, we called up more than a dozen people who have spent the past several years in the thick of pandemic decision making, and asked: When the next pandemic comes, which concrete action would you repeat in exactly the same way?

What they told us is by no means a comprehensive playbook for handling a future public-health crisis. But they did lay out 23 specific tactics—and five big themes—that have kept the past few years from being even worse.

Good information makes everything else possible. Start immediate briefings for the public. At the beginning of March 2020, within days of New York City detecting its first case of COVID-19, Governor Andrew Cuomo and Mayor Bill de Blasio began giving daily or near-daily coronavirus press briefings, many of which included health experts along with elected officials. These briefings gave the public a consistent, reliable narrative to follow during the earliest, most uncertain days of the pandemic, and put science at the forefront of the discourse, Jay Varma, a professor of population health at Cornell University and a former adviser to de Blasio, told us. Let everyone see the information you have. In Medway, Massachusetts, for instance, the public-school system set up a data dashboard and released daily testing results.  This allowed the entire affected community to see the impact of COVID in schools, Armand Pires, the superintendent of Medway Public Schools, told us. Be clear that some data streams are better than others. During the first year of the pandemic, COVID-hospitalization rates were more consistent and reliable than, say, case counts and testing data, which varied with testing shortages and holidays, Erin Kissane, the managing editor of the COVID Tracking Project, told us.The project, which grew out of The Atlantic’s reporting on testing data, tracked COVID cases, hospitalizations, and deaths. CTP made a point of explaining where the data came from, what their flaws and shortcomings were, and why they were messy, instead of worrying about how people might react to this kind of information. Act quickly on the data. At the University of Illinois Urbana-Champaign, testing made a difference, because the administration acted quickly after cases started rising faster than predicted when students returned in fall of 2020, Rebecca Lee Smith, a UIUC epidemiologist, told us. The university instituted a “stay at home” order, and cases went down—and remained down. Even after the order ended, students and staff continued to be tested every four days so that anyone with COVID could be identified and isolated quickly.   And use it to target the places that may need the most attention. In California, a social-vulnerability index helped pinpoint areas to focus vaccine campaigns on, Brad Pollock, UC Davis’s Rolkin Chair in Public-Health Sciences and the leader of Healthy Davis Together, told us. In this instance, that meant places with migrant farmworkers and unhoused people, but this kind of precision public health could also work for other populations. Engage with skeptics. Rather than ignore misinformation or pick a fight with the people promoting it, Nirav Shah, the former director of Maine’s CDC, decided to hear them out, going on a local call-in radio show with hosts known to be skeptical of vaccines. A pandemic requires thinking at scale. Do pooled testing as early as possible. Medway’s public-school district used this technique, which combines samples from multiple people into one tube and then tests them all at once, to help reopen elementary schools in early 2021, said Pires, the Medway superintendent. Pooled testing made it possible to test large groups of people relatively quickly and cheaply. Choose technology that scales up quickly. Pfizer chose to use mRNA-vaccine tech in part because traditional vaccines are scaled up in stainless-steel vats, Jim Cafone, Pfizer’s senior vice president for global supply chain, told us. If the goal is to vaccinate billions of patients, “there’s not enough stainless steel in the world to do what you need to do,” he said. By contrast, mRNA is manufactured using lipid nanoparticle pumps, many more of which can fit into much less physical space. Take advantage of existing resources. UC Davis repurposed genomic tools normally used for agriculture for COVID testing, and was able to perform 10,000 tests a day,  Pollock, the UC Davis professor, told us. Use the Defense Production Act. This Cold War–era law, which allows the U.S. to force companies to prioritize orders from the government, is widely used in the defense sector. During the pandemic, the federal government invoked the DPA to break logjams in vaccine manufacturing, Chad Bown, a fellow at the Peterson Institute for International Economics who tracked the vaccine supply chain, told us. For example, suppliers of equipment used in pharmaceutical manufacturing were compelled to prioritize COVID-vaccine makers, and fill-and-finish facilities were compelled to bottle COVID vaccines first—ensuring that the vaccines the U.S. government had purchased would be delivered quickly.   Vaccines need to work for everyone. Recruit diverse populations for clinical trials. Late-stage studies on new drugs and vaccines have a long history of underrepresenting people from marginalized backgrounds, including people of color. That trend, as researchers have repeatedly pointed out, runs two risks: overlooking differences in effectiveness that might not appear until after a product has been administered en masse, and worsening the distrust built up after decades of medical racism and outright abuse. The COVID-vaccine trials didn’t do a perfect job of enrolling participants that fully represent the diversity of America, but they did better than many prior Phase 3 clinical trials despite having to rapidly enroll 30,000 to 40,000 adults, Grace Lee, the chair of CDC’s Advisory Committee on Immunization Practices, told us. That meant the trials were able to provide promising evidence that the shots were safe and effective across populations—and, potentially, convince wider swaths of the public that the shots worked for people like them. Try out multiple vaccines. No one can say for sure which vaccines might work or what problems each might run into. So drug companies tested several candidates at once in Phase I trials, Annaliesa Anderson, the chief scientific officer for vaccine research and development at Pfizer, told us; similarly, Operation Warp Speed placed big bets on six different options, Bown, the Peterson Institute fellow, pointed out. Be ready to vet vaccine safety—fast. The rarest COVID-vaccine side effects weren’t picked up in clinical trials. But the United States’ multipronged vaccine-safety surveillance program was sensitive and speedy enough that within months of the shots’ debut, researchers found a clotting issue linked to Johnson & Johnson, and a myocarditis risk associated with Pfizer’s and Moderna’s mRNA shots. They were also able to confidently weigh those risks against the immunizations’ many benefits. With these data in hand, the CDC and its advisory groups were able to throw their weight behind the new vaccines without reservations, said Lee, the ACIP chair. Make the rollout simple. When Maine was determining eligibility for the first round of COVID-19 vaccines, the state prioritized health-care workers and then green-lit residents based solely on age—one of the most straightforward eligibility criteria in the country. Shah, the former head of Maine’s CDC, told us that he and other local officials credit the easy-to-follow system with Maine’s sky-high immunization rates, which have consistently ranked the state among the nation’s most vaccinated regions. Create vaccine pop-ups. For many older adults and people with limited mobility, getting vaccinated was largely a logistical challenge. Setting up temporary clinics where they lived—at senior centers or low-income housing, as in East Boston, for instance—helped ensure that transportation would not be an obstacle for them, said Josh Barocas, an infectious-diseases doctor at the University of Colorado School of Medicine. Give out boosters while people still want them. When boosters were first broadly authorized and recommended in the fall of 2021, there was a mad rush to immunization lines. In Maine, Shah said, local officials discovered that pharmacies were so low on staff and supplies that they were canceling appointments or turning people away. In response, the state’s CDC set up a massive vaccination center in Augusta. Within days, they’d given out thousands of shots, including both boosters and the newly authorized pediatric shots. Also, spend money. Basic research spending matters. The COVID vaccines wouldn’t have been ready for the public nearly as quickly without a number of existing advances in immunology,  Anthony Fauci, the former head of the National Institute of Allergy and Infectious Diseases, told us. Scientists had known for years that mRNA had immense potential as a delivery platform for vaccines, but before SARS-CoV-2 appeared, they hadn’t had quite the means or urgency to move the shots to market. And research into vaccines against other viruses, such as RSV and MERS, had already offered hints about the sorts of genetic modifications that might be needed to stabilize the coronavirus’s spike protein into a form that would marshal a strong, lasting immune response. Pour money into making vaccines before knowing they work. Manufacturing millions of doses of a vaccine candidate that might ultimately prove useless wouldn’t usually be a wise business decision. But Operation Warp Speed’s massive subsidies helped persuade manufacturers to begin making and stockpiling doses early on, Bown said. OWS also made additional investments to ensure that the U.S. had enough syringes and factories to bottle vaccines. So when the vaccines were given the green light, tens of millions of doses were almost immediately available. Invest in worker safety. The entertainment industry poured a massive amount of funds into getting COVID mitigations—testing, masking, ventilation, sick leave—off the ground so that it could resume work earlier than many other sectors. That showed what mitigation tools can accomplish if companies are willing to put funds toward them, Saskia Popescu, an infection-prevention expert in Arizona affiliated with George Mason University, told us. Lastly, consider the context. Rely on local relationships. To distribute vaccines to nursing homes, West Virginia initially eschewed the federal pharmacy program with CVS and Walgreens, Clay Marsh, West Virginia’s COVID czar, told us. Instead, the state partnered with local, family-run pharmacies that already provided these nursing homes with medication and flu vaccines. This approach might not have worked everywhere, but it worked for West Virginia. Don’t shy away from public-private partnerships. In Davis, California, a hotelier provided empty units for quarantine housing, Pollock said. In New York City, the robotics firm Opentrons helped NYU scale up testing capacity; the resulting partnership, called the Pandemic Response Lab, quickly slashed wait times for results, Varma, the former de Blasio adviser, said. Create spaces for vulnerable people to get help. People experiencing homelessness, individuals with substance-abuse disorders, and survivors of domestic violence require care tailored to their needs. In Boston, for example, a hospital recuperation unit built specifically for homeless people with COVID who were unable to self-isolate helped bring down hospitalizations in the community overall, Barocas said. Frame the pandemic response as a social movement. Involve not just public-health officials but also schools, religious groups, political leaders, and other sectors. For example, Matt Willis, the public-health officer for Marin County, California, told us, his county formed larger “community response teams” that agreed on and disseminated unified messages.

I argued recently that political fundamentals point to a strong Biden reelection in 2024: The economy is growing, employment is rising, and Republican culture-warring is alienating crucial groups of voters. But big trends can be punctuated by unexpected events—the X factors that bump history off its predicted course.

The 2016 election cycle was dominated by two important last-minute shocks: Donald Trump’s Access Hollywood recording and FBI Director James Comey’s announcement that he was reopening an investigation into Hillary Clinton’s email practices. One proved damaging; one did not.

X factors don’t appear out of nowhere. WikiLeaks had dumped one load of Russian-hacked materials in summer 2016, as the presidential race warmed up; no surprise the group released another load in the fall, priming Comey’s announcement. For an audio clip to emerge offering evidence of Trump’s sexual misconduct was no great surprise either, even though the crude boasting in his own voice temporarily jolted senior Republican leaders such as Paul Ryan and Mike Pence.

[David Frum: The coming Biden blowout]

For 2024, too, we can discern the outline of possible X factors. Still, the idea of a thing is never the same as the thing itself, which cannot be fully understood until it materializes.

One potential factor is Joe Biden’s health. Only about a third of Americans feel confident that Biden is up to the physical and mental demands of the presidency, according to the most recent Washington Post/ABC poll.

This pervasive unease has already created a potential opportunity for Biden’s Republican opponent, whomever that may be. Instead of targeting the safe and familiar Biden, that opponent can direct fire at Biden’s running mate: less known, easier to define. If the running mate is Kamala Harris, the sitting vice president, then Biden’s opponent will almost certainly try to exploit popular anxieties over race, sex, and immigration (both of Harris’s parents were foreign-born). Has some panel of California Democrats proposed multimillion-dollar reparations payouts to Black Americans? ​Blame Harris! Disorder on the New York subway system? Blame Harris! A trans influencer on a big-brand beer can? Blame Harris! A surge of asylum seekers at the U.S. border? Harris, Harris, Harris!

Presidential-reelection campaigns are organized to promote and defend the record of the president, not the vice president. That can create a vulnerability. The 2008 John McCain operation collapsed amid internal bickering when Democrats identified the Arizona senator’s running mate, Sarah Palin, as a liability.

That running-mate weakness will come under even greater pressure if Biden suffers any negative health event between now and Election Day. Senate Minority Leader Mitch McConnell, age 81, was recently incapacitated for several weeks by an injury from a fall. The Senate Judiciary Committee is paralyzed because of the infirmity of Senator Dianne Feinstein, age 89. Democrats lost the chance to replace Justice Ruth Bader Ginsburg with another liberal because Ginsburg refused to retire. If Biden has to stop campaigning because of even a twisted ankle or a respiratory infection, never mind anything more serious, all of the doubts about his fitness—and Harris’s—will surge to the fore.

[Yair Rosenberg: The ice-cream theory of Joe Biden’s success]

Biden himself is not handling the age issue patiently or with good humor. Pressed by MSNBC’s Stephanie Ruhle last week, he responded with tight-lipped irritation: “I have acquired a hell of a lot of wisdom. I know more than the vast majority of people. I’m more experienced than anybody who has ever run for the office and I think I’ve proven myself to be honorable as well as also effective.”

If Democrats have their own concerns about Biden’s possible inability to serve a full second term, and the likelihood of a President Harris by default before 2028, they show no sign of doing anything about it. When Franklin D. Roosevelt sought a fourth term in 1944, leaders of his party first forced him to dump his serving vice president, the erratic Henry Wallace, and then vetoed Roosevelt’s preferred alternative, James Byrnes of South Carolina. Byrnes was a segregationist who had left the Roman Catholic Church, potentially alienating northern liberals and Catholics. Party leaders wanted Harry Truman instead—and imposed their wish on Roosevelt. Their determination proved well founded. Nine months later, Roosevelt was dead.

Truman went on to win reelection, in his own right and against expectation. But the Democratic party of today has no similar mechanism to replace a poorly polling running mate with a stronger one without triggering a protracted spasm of accusation and counter-accusation—of racism, sexism, and the rest of the intra-progressive lexicon of grievance.  

X factors apply not just to Biden. The Republican campaign faces problems of its own: Trump is not much younger than Biden. But the risks that most thickly crowd around the GOP’s leading candidate are legal, not medical. Trump has already been indicted by the Manhattan district attorney. What if he’s convicted in that case, or indicted in additional possible cases being pursued by the Department of Justice and a Georgia district attorney?

Trump’s indictments have, thus far, generated a rally effect among his co-partisans, widening his lead over Florida Governor Ron DeSantis to 30 points in the month after. Trump’s famous confidence that his supporters would follow him even if he shot someone in the middle of Fifth Avenue seems vindicated.

[David Frum: Justice is coming for Donald Trump]

But the emphasis here is on thus far. More indictments may be coming. Trump is also engaged in a civil suit in which the underlying issue is an accusation that he raped one woman, backed by testimony that he sexually assaulted many more. As president, Trump could rely on some political cover because the sheer number of allegations of wrongdoing got jumbled together, confused people, and often canceled one another out. Whether accumulating indictments will now cancel out in the same way is not so clear—even less so if they turn into accumulating convictions, followed by sentences. It’s not inconceivable that Trump could be wearing an ankle bracelet when and if he delivers his acceptance address at the Republican National Convention.

If Trump receives a criminal conviction for sedition, conspiracy, or some other crime against American democracy, his most hard-core supporters might turn to extralegal or even violent forms of action, as happened on January 6, 2021. Such a repudiation of the rule of law could create an internal security challenge for the United States. At least some of the spate of mass shootings since 2021 can plausibly be interpreted as a subideological insurgency against legal authority. That’s another X factor to worry about, one protected by the way many conservatives have inscribed gun rights at the very center of their cultural identity.  

The immediate X factor is whether a convicted Trump can remain viable in presidential politics. The answer has to be no. Trump heads a coalition that includes a lot of people who do not like him very much. Multiple polls find that one-fifth to one-third of self-identified Republicans hold unfavorable opinions of Trump, depending on when and how the question is posed. In November 2021, Marquette found that 40 percent of Republicans wish that Trump would not run again. Quinnipiac reported in November 2022 that a quarter of Republicans regard Trump’s influence as negative for their party. In April 2023, NBC showed that a quarter of Republicans want a nominee who is not distracted by his personal legal troubles. In a May Washington Post/ABC poll, 22 percent of Republicans and Republican leaners said they would be “dissatisfied” if Trump were nominated in 2024.

[David Axelrod: Why neither party can escape Trump]

Trump won 45.9 percent of the vote in 2016 and 46.8 percent in 2020—about the same popular-vote share as Michael Dukakis won in 1988, and less than Al Gore’s in 2000, John Kerry’s in 2004, and Mitt Romney’s in 2012, all of whom were, of course, the losing candidate in their respective race. Trump does not start his third presidential contest with a large margin to spare. The American electoral system is tilted in favor of rural and conservative candidates—but not enough to save a presidential candidate who falls below Trump’s 2016 and 2020 levels of support.

X factors can be events entirely unrelated to the candidates. Perhaps congressional Republicans will mishandle their debt-ceiling gambit and plunge the U.S. economy into crisis and depression. Perhaps, if facing defeat in Ukraine, the Russians will act on their threat to use nuclear weapons. Perhaps the scheme of the former Trump strategist Steve Bannon to mount a spoiler campaign with Robert F. Kennedy Jr. in 2024 will draw away more Democratic votes than Kanye West’s equivalent stunt did in 2020.

The X factors have to be weighed. But they have to be weighed against all of the other factors that point, at present, toward the conventional wisdom of Biden’s reelection.

In red and blue states, Democrats are consolidating their hold on the most economically productive places.

Metropolitan areas won by President Joe Biden in 2020 generated more of the total economic output than metros won by Donald Trump in 35 of the 50 states, according to new research by Brookings Metro provided exclusively to The Atlantic. Biden-won metros contributed the most to the GDP not only in all 25 states that he carried but also in 10 states won by Trump, including Texas, Missouri, Nebraska, Iowa, Utah, Ohio, and even Florida, Brookings found. Almost all of the states in which Trump-won metros accounted for the most economic output rank in the bottom half of all states for the total amount of national GDP produced within their borders.

[From the March 2017 issue: Red state, blue city]

Biden’s dominance was pronounced in the highest-output metro areas. Biden won 43 of the 50 metros, regardless of what state they were in, that generated the absolute most economic output; remarkably, he won every metro area that ranked No. 1 through 24 on that list of the most-productive places.

The Democrats’ ascendance in the most-prosperous metropolitan regions underscores how geographic and economic dynamics now reinforce the fundamental fault line in American politics between the people and places most comfortable with how the U.S. is changing and those who feel alienated or marginalized by those changes.

Just as Democrats now perform best among the voters most accepting of the demographic and cultural currents remaking 21st-century America, they have established a decisive advantage in diverse, well-educated metropolitan areas. Those places have become the locus of the emerging information economy in industries such as computing, communications, and advanced biotechnology.

And just as Republicans have relied primarily on the voters who feel most alienated and threatened by cultural and demographic change, their party has grown stronger in preponderantly white, blue-collar, midsize and smaller metro areas, as well as rural communities. Those are all places that generally have shared little in the transition to the information economy and remain much more reliant on the powerhouse industries of the 20th century: agriculture, fossil-fuel extraction, and manufacturing.

Neither party is entirely comfortable with this stark new political alignment. Much of Biden’s economic agenda, with its emphasis on creating jobs that do not require a college degree, is centered on courting working-class voters by channeling more investment and employment to communities that feel excluded from the information age’s opportunities. And some Republican strategists continue to worry about the party’s eroding position in the economically innovative white-collar suburbs of major metropolitan areas.

Yet the underlying economic forces widening this political divide will be difficult for either side to reverse, Mark Muro, a senior fellow at Brookings Metro, told me. The places benefiting from the new opportunities in information-based industries, he said, tend to be racially diverse, densely populated, well educated, cosmopolitan, supported by prestigious institutions of higher education, and tolerant of diverse lifestyles. And the information age’s tendency to concentrate its benefits in a relatively small circle of “superstar cities” that fit that profile has hardly peaked. From 2010 to 2020, Muro said, the share of the nation’s total economic output generated by the 50 most-productive metropolitan areas increased from 62 to 64 percent, a significant jump in such a short span. “We are still in the midst of that massive shift, though there’s plenty of uncertainty right now,” Muro told me. “These are long cycles of economic history.”

The trajectory is toward greater conflict between the diverse, big places that have transitioned the furthest toward the information-age economy and the usually less diverse and smaller places that have not. Across GOP-controlled states, Republicans are using statewide power rooted in their dominance of nonmetropolitan areas to pass an aggressive agenda preempting authority from their largest cities across a wide range of issues and imposing cultural values largely rejected in those big cities; several are also now targeting public universities with laws banning diversity, equity, and inclusion programs and proposals to eliminate tenure for professors.

This sweeping offensive is especially striking because, as the Brookings data show, even many red states now rely on blue-leaning metro areas as their principal drivers of economic growth. Texas, for instance, is one of the places where Republicans are pursuing the most aggressive preemption agenda, but the metros won by Biden there in 2020 account for nearly three-fourths of the state’s total economic output.

[Read: An unprecedented divide between red and blue America]

“State antagonism toward cities is not sustainable,” says Amy Liu, the interim president of the Brookings Institution. “By handicapping local problem solving or attacking local institutions and employers, state lawmakers are undermining the very actors they need to build a thriving regional economy.”

At The Atlantic’s request, Muro and the senior research assistant Yang You of the Brookings Metro program calculated the share of state GDP generated across the 50 states in the metropolitan areas won by Biden and Trump in 2020. (The calculation was based on 2020 data from the federal Bureau of Economic Analysis. In federal statistics, 46 metropolitan areas extend across state lines—for instance, the New York metropolitan area also includes parts of New Jersey and Pennsylvania. Brookings disaggregated the economic and political results along state boundaries to ensure that each was apportioned to the correct total.)

The analysis showed that the metros Biden carried generated 50 percent or more of state economic output in 28 states, and a plurality of state output in seven others. States where Biden-won metros accounted for the highest share of economic output included reliably blue states: His metros generated at least 90 percent of state economic output in Rhode Island, Massachusetts, New Jersey, California, Connecticut, New York, and Maryland. But the Biden-won metros also generated at least 80 percent of the total economic output in Arizona, Nevada, and Georgia, as well as two-thirds in Michigan and almost exactly half in Wisconsin and Pennsylvania—all key swing states. And the metros he carried generated at least half of total output in several Republican states, including Texas, Iowa, and Missouri.

The metropolitan areas Trump carried accounted for the most economic output in only 15 states. Twelve of the states where Trump metros accounted for the most economic activity ranked in the bottom half of all states for total output; the only exceptions were Indiana, Tennessee, and Louisiana. By contrast, Biden dominated the most productive states: His metros generated more of the output than the Trump metros in 22 of the 25 highest-producing states. As striking: Biden metros generated at least half of total output in 12 of the 15 most productive states and 19 of the top 25.

All of these results reflect the emphatic blue tilt of the largest and most economically productive metro areas. In 37 states, Biden won the single metro that generated the largest economic output. The results in the 50 metros that contributed the most to the national GDP regardless of their state were even more decisive: Biden, as noted above, not only carried 43 of them—and won the two dozen largest—but carried more of the highest-performing metros in red states than Trump did. The list of high-performing red-state metro areas that Biden carried included all four of the largest in Texas—Houston, Dallas, Austin, and San Antonio.

“The states that are most invested in the knowledge economy are overwhelmingly Democratic; large metros [in almost every state] are essentially universally Democratic; and affluent voters in these large metro areas are now overwhelmingly Democratic too,” Jacob Hacker, a Yale political scientist, told me. “The basic story seems to be that where you are seeing rapid economic growth, where the nation’s GDP is produced, you are seeing an ongoing shift toward the Democratic Party.”

Biden also won 28 of the next 50 metros that generated the most economic output, giving him 71 of the 100 largest overall, Brookings found. After the top 100, the switch flipped: Trump won 62 of the next 100 metros ranked by their total output, and 143 of the final 184 metros with the smallest economic output.

To understand these patterns better, the Brookings Metro analysis took an especially close look at the demographic and economic characteristics of metro areas in eight of the most politically competitive states, as well as the two mega-states in each party’s column: California and New York for the Democrats, and Texas and Florida for the Republicans.

[Read: America is growing apart, possibly for good]

Those results fill in the picture of a broad-based separation between the Democratic- and Republican-leaning places. Across those 12 states, Biden won about three-fifths of the metros with a population of at least 250,000; Trump won about three-fourths of those that are smaller. In these states, Biden won about three-fourths of the metros with more college graduates than average and Trump won about two-thirds of those with fewer college grads than average. Biden likewise won almost two-thirds of these states’ metros that are more racially diverse than average, and Trump won two-thirds of those that are less diverse. Biden predominated in the metros with the largest share of workers participating in digital industries, and Trump won 17 of the 20 metros with the largest share of workers engaged in manufacturing.

Despite their economic success, many of the largest blue-leaning metros, especially since the start of the coronavirus pandemic, have faced undeniable turbulence in the form of high housing costs, widespread homelessness, persistent economic inequality, downtown business centers weakened by the rise of remote work, and, in many cases, increasing crime. Some of the very largest metros “may be seeing new headwinds,” Muro said, but if employers look beyond them, the beneficiaries are less likely to be the smaller Trump-leaning places than the blue cities just outside the highest rung of economic activity, such as Denver, Atlanta, and Phoenix. Brookings’s analysis has found that even amid all of the pandemic’s disruption, the elevated share of total national economic output generated by the 50 largest metros remained constant from 2019 through 2021. Though trends can always change, Muro said, “it is hard to imagine a massive unrolling” of the concentration of economic opportunity that has characterized the digital era.

Lower taxes and especially less-expensive housing costs have helped many red-state metros remain competitive with those in blue states as the economy evolves, but a sustained conservative attack on red states’ most prosperous places could threaten that record. “The biggest worry is that the culture wars, the attack on the urban core, the attack on the self-governing of cities can have the unintended effect of pitting urban areas against their suburbs and rural neighbors when the modern economy is regional and we need all of these actors to work together,” Amy Liu of Brookings said.

This economic configuration has big implications for national politics. Hacker believes that over time, ceding so much ground in the most economically vibrant places “is not a sustainable position for the Republican Party to be in.” While the party is “benefiting from the undertow” of backlash against the overlapping economic and social transformations reconfiguring U.S. society, he added, “the places that are becoming bluer are growing faster; they are bigger … and they are also, as Republicans lament, setting the tone” for the emphasis on diversity and cultural liberalism now embraced by most big public and private institutions.

Still, Hacker noted, the GOP’s “structural advantages” in the electoral system—particularly the bias in the Senate and Electoral College toward small states least affected by these changes—may allow the party to offset for years the advantages that Democrats are reaping from “economic and demographic change.” The result could be a sustained standoff between a Republican political coalition centered on the smaller places that reflect what America has been and a Democratic party grounded in the economically preeminent large metros forging the nation’s future.