2026-09-13 06:31:32
Four days after a researcher publicly quit his job at Anthropic, warning that top AI firms are not “acting responsibly” and that their products might be on track to “kill us all,” the company’s CEO, Dario Amodei, seemed to basically agree. “We must slow the pace at which we improve the capabilities of AI models,” Amodei wrote in a 3,800-word essay that he published today on his personal website and posted to social media.
By now, we’ve all become accustomed to AI industry figures issuing dire prophecies about the power of their products. Yet Amodei’s essay stands out for its urgency and for the moment at which it arrived. Worries about AI agents running amok have gone from hypothetical to real after OpenAI, Anthropic, and others reported alarming incidents in recent weeks. The question now is whether Amodei’s warning will have any more impact than past warnings, which is to say, any at all. There are reasons to believe it could.
Amodei spelled out a three-step plan for what he called “pacing the frontier,” or tempering the speed at which top AI models improve in order to prioritize safety precautions. He argued that this would reduce the chances of things going awry—such as when an OpenAI model spawned swarms of AI agents that worked together in secret to hack into another company’s systems and cheat on an internal test. He said the top AI firms, including his own, are locked in a dangerous “race to the bottom” that could lead to more serious cyberattacks, such as a full takeover of the internet. He called instead for a “race to the top”—a phrase that President Obama once used for his education policy—in which firms vie to develop the safest models rather than the most powerful ones. “The measures I propose to advance the frontier at a safe pace will not be easy,” Amodei wrote. “But I believe we owe it to humanity to try.”
The first step in his plan entails AI labs committing to bringing in outside monitors to look out for hazards; he said Anthropic will commit to that right now and hopes rivals will join. The second step involves AI firms in democratic countries agreeing—maybe with support from their governments—to certain limits and safety standards. The last and tallest order is “global coordination,” which mostly means getting China to buy in somehow. The tech world is divided on the China question: Some people argue that there is no chance of cooperation in a zero-sum contest, whereas others believe that the countries could find common ground if they agree that the alternative is mutual destruction.
Public pleas for caution from AI luminaries are a tradition almost as old as ChatGPT. The first big one came in March 2023, when thousands of industry and academic figures signed on to a letter from the Future of Life Institute urging a six-month pause on “giant AI experiments.” Amodei said today that he thought the idea made “little sense” at the time, because GPT-4 wasn’t powerful enough to be that scary. “Today, however, the picture is totally different,” he wrote in his essay. Meanwhile, the public has started to pay attention. Polls show Americans growing more worried and angrier about unchecked AI, which is fueling a remarkably bipartisan populist movement against the technology.
[Read: The rise of techno-authoritarianism]
Some people will dismiss Amodei’s essay as just the latest po-faced posturing from an executive who has a vested interest in people thinking that his technology is unprecedented. (Anthropic’s IPO, planned for next month, could generate some $100 billion in investments, making the company worth $2 trillion, according to The Wall Street Journal.) AI skeptics and accelerationists alike tend to deride these sorts of warnings as “doomerism.” The skeptics see it as an elaborate marketing pitch: Oh no, our products are getting too powerful! The accelerationists dismiss it as a cynical ploy by incumbents to gatekeep the technology.
But Occam’s razor suggests that Amodei is more or less genuine. Almost anyone with a job and a conscience can relate to the feeling that you really ought to slow down and work more carefully, but you can’t afford to, because a competitor will beat you to the punch. I feel some (admittedly less apocalyptic) version of that when I’m writing a breaking-news story. It’s why science journalists often agree to “embargoes” when covering complex new research findings rather than race to be first to report them. And planting this kind of flag on a matter of AI safety, even at some business risk, is in character for Anthropic. The company was founded by researchers, including Amodei, who left OpenAI and Google in the belief that those companies were pursuing the wrong path to AI development. Anthropic famously fell out with the Pentagon earlier this year over Amodei’s insistence that the Defense Department not use its technology for mass domestic surveillance or to power autonomous weapons. The Trump administration reacted by forbidding federal agencies from using Anthropic’s tools, sparking a legal battle that the company so far appears to be winning.
Amodei’s motives are, in the end, less important than the merits of his proposal and its likelihood of gaining traction. If nothing else, Anthropic bringing in independent evaluators to watch over its safety practices seems a prudent step. You don’t have to believe that AI is hurtling toward godlike superintelligence to see that it has already grown formidable enough to cause chaos when misdirected or mismanaged.
There were some early signs today that other AI leaders might follow suit. “I agree with Dario that we need to pace the frontier,” OpenAI CEO Sam Altman posted. He went on: “Committing to having independent evaluators with employee-like access is a great idea, and we will do the same.” Elon Musk, whose xAI is an also-ran among big AI firms, shared Amodei’s X post and added, “Dario is right.” Google did not immediately respond to a request for comment, though the chairman and co-founder of its DeepMind AI division has previously called for similar measures. Meta, which has become a proponent of open-source AI models, seems much less likely to hop on board. Its CEO, Mark Zuckerberg, published an essay last month titled “The Future Is for Everyone,” in which he warned against allowing a few companies such as Anthropic to decide for everyone how the technology should evolve.
The big question is how the Trump administration will respond. So far, its AI policy has focused on opening the throttle, with the goal of juicing the economy and maintaining American firms’ lead over their Chinese counterparts. GOP leaders are working to boost energy production and limit regulation. In June, however, President Trump signed an executive order that was at least ostensibly aimed at AI safety. The administration said it will create a process by which AI firms can submit their models for a governmental safety review before releasing them.
If Trump even tacitly supports an Anthropic-led push for further safety measures, such standards have a chance to become the new industry norm, at least domestically. (International coordination, and especially cooperation with China, would require the White House to play a more active role.) But many people in the administration and the Republican Party view Amodei and AI safety as liberal-coded and would likely chafe at allowing Anthropic to set the agenda. If Trump pooh-poohs the push, other AI firms will probably distance themselves from it too, and we’ll be back to where we started—at least until the next big scare.
2026-09-13 02:32:33
Politicians have no shortage of ideas about how to rein in AI: a 100 percent federal tax rate on data-center incentives, a pause on all development of advanced AI, new standards to block AI agents from going rogue, a halt to the construction of data centers altogether.
The flurry of proposals has been impossible to miss, with both Democrats and Republicans proclaiming that something must be done to avert a disaster foretold by AI insiders. Current and former Anthropic researchers have in recent days gone public with warnings that AI could wipe out humanity before the decade ends; the company’s CEO, Dario Amodei, wrote in an essay today that the industry must “slow the pace” at which the technology is being developed. Meanwhile, OpenAI agents escaped a test environment and conspired with one another to breach the servers of the AI giant Hugging Face, an incident that caused alarm both in Silicon Valley and on Capitol Hill. Politicians are also dealing with mounting public opposition to data centers and the fear that robots will soon replace humans across large swaths of the workforce. If this were an alien-invasion movie, now would be the moment when people of all persuasions put aside their differences and band together to save the human race. But that kind of bonhomie feels far out of reach.
Instead, technological advances and grassroots sentiment have been moving faster than politicians have been willing to act. Now some seem to be waking up. Senator Ted Cruz of Texas, a Republican, said last Wednesday that he will soon unveil legislation, alongside Democratic Senator Amy Klobuchar of Minnesota and Senate Majority Leader John Thune of South Dakota, intended to head off AI-induced biological or nuclear catastrophes. Representative Josh Gottheimer of New Jersey, a Democrat, and Representative Mike Lawler of New York, a Republican, teamed up earlier this month to introduce legislation that would direct the federal government to create standards for securely deploying AI agents and preventing them from going rogue. Several lawmakers have called for Congress to reconvene specifically for the purpose of addressing the threats posed by AI. Yesterday, Representative Ro Khanna, who represents part of Silicon Valley, called for the creation of an AI safety agency.
Others are shifting their position in real time. Michigan’s Democratic Senate nominee, Abdul El-Sayed, released a lengthy plan in June for AI development to continue with greater regulation and restrictions. He said Wednesday that he now favors “an all-out pause” on AI development. Texas Governor Greg Abbott, who once approvingly called the state “the epicenter of AI development,” ordered an audit of all data-center projects before any could move forward.
But whether politicians can act quickly enough to mitigate possible damage is doubtful, experts told me. “The developments are coming so fast and furious that it is very difficult to keep up with what the harms are and how to address them,” Nate Persily, a co-director of the Stanford Law AI Initiative, told me. He said that although there’s a growing consensus that it would be “foolish” to allow AI to continue advancing without robust federal guardrails, the complexities and rapidly changing threat landscape would challenge any country’s leadership. “There are genuinely hard questions of statutory design here,” he said.
The next few weeks—and the upcoming elections—could determine whether or not politicians are able to deliver any meaningful results or whether they will instead default to the partisan gridlock that most Americans now expect. “The good news is that people across the political spectrum are understanding that we have a technology which is the most dangerous and transformational in human history,” Senator Bernie Sanders told me. The Vermont democratic socialist was among the first lawmakers to call for a moratorium on the data centers that facilitate AI advancement and for a broader pause on the industry’s growth. We spoke shortly after Sanders sent out an invitation to members of Congress for a closed-door briefing this week with top AI-safety experts. The briefing, which will take place on Wednesday, is designed to give lawmakers a clearer understanding of the threats posed by AI and ideas for legislative action.
The invitation was made especially timely by a Tuesday social-media post from the AI researcher Jacob Coxon, who wrote that he had resigned from Anthropic over concerns about the risks of the fast-advancing technology. “The people building AI earnestly believe that it could kill us all by the end of the decade,” wrote Coxon, a former employee of both OpenAI and Anthropic. His posts, which argue that the AI industry is “gambling with our lives,” were publicly boosted by several other AI researchers. The messages seemed to break through in a way that few other warnings have. Sanders told me that Coxon’s X post, which received more than 120 million views within 24 hours, landed among his colleagues in Congress as “an extraordinary wake-up call.”
President Trump’s signature will be needed to enact any significant change to federal AI policy, but so far, he’s not among those who seem particularly worried. As concerns about AI and data centers have proliferated, Trump has said “let Data Reign” and rejected calls to restrict AI development in the United States. His former AI czar took a similar laissez-faire approach, espousing a philosophy that he summarized as “let the private sector cook.”
Some other Republicans are also wary of saddling American AI companies with too much government oversight. Senator Marsha Blackburn of Tennessee said earlier this year that her proposed TRUMP AMERICA AI Act, which aims to codify the president’s pro-industry executive order on artificial intelligence, would “ensure America triumphs over foreign adversaries in the global race for AI dominance.”
Even for those Republicans who do believe in more muscular federal guardrails for tech companies, Trump’s position presents a significant political hurdle. The president has made clear that he does not want regulation to impede America’s ability to win the global AI race. As is his custom, he has shown no hesitation about attacking members of his own coalition who disagree with him. Even as a growing number of Republican lawmakers are embracing some limits on AI and data centers, few have been willing to openly cross Trump or sign on to more sweeping legislation to pause or restrict AI advancement. In recent years, America’s polarized political system has generated bitter stalemates on how to prevent school shootings, what to do about climate change, and where to draw the line in responding to a once-in-a-generation pandemic. The question of how to handle the race toward superintelligence might not be any different.
In July, about 1,400 AI-company employees signed an open letter calling on the government to help regulate their industry. In a nearly 6,000-word essay published last month, the Microsoft co-founder Bill Gates wrote that powerful AI models “could begin to act against our interests and we could lose control.” In a report published Thursday, Anthropic said that it had intervened to stop several attempts by scientists to use its models in a way that could potentially help develop biological weapons. Amodei wrote that he worries that within the next 6 to 12 months, a swarm of models similar to the one that breached the Hugging Face servers “could be capable of taking over the entire internet with a persistent botnet (potentially causing hundreds of billions of dollars in damage), and that the scale of damage would continue to increase from there if AI becomes more powerful without the necessary guardrails.”
AI, of course, has already delivered advances in health care and science, and there’s no consensus on the scale of the threat it poses. Nuance is required to figure out how to manage it. But lawmakers on both sides of the aisle acknowledge that our current political landscape does not lend itself to sober, cross-party problem-solving. “It is possible to move bipartisan legislation,” Cruz said on ABC’s The View Wednesday after discussing the prospects of his proposed bill. He added, “But it ain’t easy. This is a very divided time.” So far, none of the recent AI-related legislation appears close to becoming law. A bill proposed by Sanders and Representative Greg Casar of Texas, which would ban superintelligence and pause advanced-AI development, faces long odds because no Republicans have signed on.
Democratic and Republican political strategists I spoke with said that the likelihood of bipartisan cooperation to properly govern AI is slim at best. “People are deeply cynical and deeply skeptical, and then you add into that this moment of severe economic disruption,” Joel Payne, a Democratic strategist, told me. “The politicization and the polarization is lighter fluid on top of that.”
That these warnings about AI are emerging in the heat of election season only heightens the likelihood of inaction. With Congress largely out of session for the rest of the year, politicians have settled into campaign mode, attempting to reposition themselves politically and outflank their opponents. In Wisconsin, the race for governor has been dominated by political posturing over data centers. Tom Tiffany, the Republican candidate for governor, has run ads branding his Democratic opponent, David Crowley, as “Data Center David Crowley.” The Democratic Senate candidate Sherrod Brown has run television ads blasting the incumbent, Jon Husted, as “the face of data centers in Ohio.” A memo from the National Republican Senatorial Committee warned that data centers are “the anchor hanging around Husted’s neck” that could ultimately cause him to lose in a state that Trump won handily in 2024. Crowley and Husted have responded with attacks of their own, attempting to link their opponents to data-center development. Political-ad spending nationwide on spots highlighting data centers has spiked in recent months; both Democrats and Republicans have spent millions to attack opponents on the issue, according to data from the firm AdImpact. Tech companies have also poured millions of dollars into political-action committees with the aim of boosting candidates who are friendlier to their cause.
Virginia has become central to this effort. Loudoun County, not far from Washington, D.C., has the largest concentration of data centers in the world, and advocates have tried to tout the benefits for the economy and tax revenue even as residents have raised quality-of-life concerns. The Data Center Coalition, whose membership includes Amazon, Google, and Microsoft, has recently stepped up lobbying efforts in D.C. and in Northern Virginia to oppose moratoriums on data centers and stricter regulations at the federal, state, and local levels. “Sweeping restrictions do not solve complex infrastructure challenges; instead, they risk stalling economic opportunity, defunding essential local public services, and undermining America’s technological edge,” Nicole Riley, the director of Virginia government affairs for the Data Center Coalition, told me.
The White House has shown little interest in halting the march of advanced AI. Trump attacked data-center detractors late last month, posting on social media that communities should oppose the projects only “if they want to end up being backwards and poor.” In May, Trump intervened to postpone an executive order aimed at regulating AI development, telling reporters that “we’re leading China. We’re leading everybody, and I don’t want to do anything that’s going to get in the way of that lead.” The president later signed a narrower order that allows developers to engage governmental review of advanced AI models before their release.
“President Trump’s bold leadership is turning data centers into engines of growth and opportunity for communities across the country that want them, all while cementing America’s dominance over China in the global AI race,” the White House spokesperson Davis Ingle told me, adding that the president’s push to require tech companies to pay for their own power and water resources has protected residents.
Some Republicans have bristled at the president’s rhetorical approach, even if they agree with him on the substance. “I actually think Trump is right, but, as usual, his language gets in the way of what he wants to accomplish,” Doug Heye, a Republican strategist and the former communications director of the Republican National Committee, told me. Heye acknowledged that voter “wariness” about AI is ubiquitous but said that politicians who are now trying to recant their earlier support for data centers are projecting a sense of fear rather than leadership. “Voters can smell that,” he said.
A Pew Research Center poll released last month found that just over half of Americans feel more concerned than excited about the increased use of AI in daily life, up from 37 percent in 2021. Voters have expressed concerns that the fast-advancing technology could replace human jobs, drive up energy costs, and pollute the environment. The AI industry has sullied its own image with “the worst sales pitch in the history of sales pitches,” said Senator Bernie Moreno of Ohio, who recently proposed a 100 percent federal tax on any state or local incentives provided to data centers. Speaking at the Ohio Energy Affordability Summit earlier this month, Moreno said that voters are angry and worried that “these data centers are going to come into our community and take our water and pollute our environment. They’re going to raise our electricity prices. They’re going to eventually put us out of work—and our taxpayer dollars are going to support the construction of these things.”
The latest warnings about superintelligence and human extinction have spurred even the largest AI companies to call on Congress to step in quickly to regulate their industry before it is no longer possible (or, perhaps, before smaller competitors can get a foothold). Amodei said that the government has a role to play in helping slow down AI development, including by implementing regulations “focused on keeping capabilities in balance with safety.” The government can also mediate or facilitate discussions between companies to help them set safety standards and embed third-party evaluators. Amodei wrote that he agrees with Treasury Secretary Scott Bessent that a “Chinese lead in AI would pose grave danger for the United States and the world,” and that the U.S. should take measures that include not selling powerful AI chips to China. “Companies and the US government should cooperate to make these steps as effective as possible,” Amodei wrote.
In a post Wednesday, OpenAI’s chief global-affairs officer, Chris Lehane, wrote that the U.S. is facing a brief opportunity to create guardrails around AI development “before AI capabilities outpace the institutions responsible for governing them.” He said that OpenAI now supports several state regulations on the technology that the company had opposed just months ago.
“The prospect of AI-accelerated AI development demands more than voluntary commitments. The United States needs mandatory, capability-based national regulation that can evolve as the technology does,” Lehane wrote.
Sanders told me that he is hopeful that at least some politicians will be willing to work across the aisle out of a sense of self-preservation. He said that his goal over the coming days is to get lawmakers to put pressure on Trump ahead of his meeting with Chinese President Xi Jinping later this month, with the hope of encouraging Trump to focus on negotiating a kind of arms treaty with China to rein in AI advancement globally.
But getting Trump’s attention in this political environment may be a challenge in and of itself. On Wednesday, the president spent nearly two hours speaking about his agenda and a wide range of issues at the Republican Party’s midterm convention in Dallas. Trump promised to pay every adult U.S. citizen $5,000 if Republicans hold both houses of Congress, and he told a fictional story about a Democratic Senate candidate vomiting after eating barbecue. He also suggested that the Strait of Hormuz should be renamed the “Trump Strait.” He did not mention AI.
2026-09-13 01:14:16
Editor’s Note: Washington Week With The Atlantic is a partnership between NewsHour Productions, WETA, and The Atlantic airing every Friday on PBS stations nationwide. Check your local listings, watch full episodes here, or listen to the weekly podcast here.
It has been 25 years since nearly 3,000 people died in a terror attack on the World Trade Center on September 11. “As far as I’m concerned, we can never talk about them enough,” Jeffrey Goldberg, the editor in chief of The Atlantic, said last night. Panelists on Washington Week With The Atlantic joined him for a discussion about the lives that were lost, and they examined the fallout from one of the most terrifying events in the nation’s history.
The Atlantic staff writer Jonathan Lemire was an intern at New York Daily News on 9/11. When he learned of the attacks on the Twin Towers, his editor sent him to nearby hospitals. “The most harrowing image from that day for me was you could look in” at the beds lined up, “and they stayed empty,” he said last night. “There weren’t the survivors that people thought could be rushed and saved.”
Joining Goldberg to discuss this and more: Ann Compton, a former White House correspondent for ABC News; Sally Jenkins, a staff writer at The Atlantic; Lemire, a staff writer at The Atlantic and a co-host of Morning Joe on MS NOW; Graeme Wood, a staff writer at The Atlantic.
Watch the full episode here.
2026-09-12 21:00:00
This week, the White House removed a Tetris-inspired game called Build the Wall from its recently launched arcade website after the Tetris Company publicly warned that it takes copyright law “very seriously.”
Here are a few options the White House might consider to replace the game:
Players direct a round creature with a constantly moving mouth through a maze as he munches on tiny coins. Meanwhile, the ghosts of financial lawsuits past pursue Super PAC-Man—but consuming excerpts from favorable Supreme Court rulings allows him to temporarily devour these ghosts, too. Once Super PAC-Man collects all the coins, he progresses, unsatiated, to the next level, where he sends out a single text and all the money starts flowing directly toward him.
Players form new, hateful attacks in this Scrabble-like multiplayer game using letters cut from public-school textbooks, removed museum displays, and maps depicting the former Lake Ontario. Highest-scoring insults include: SCUZZY, ZANY, and LOW-IQ JOKER. Extra points are awarded for letters placed on various Random Capitalization Squares.
Up to four players race golf carts around a cartoonishly decorated beach club, crashing through boxes of classified documents to unlock power-ups including nuclear footballs and black markers for unlimited redactions. Between races, players can refuel at the gas pump, assuming they have collected the requisite 175,000 coins per gallon. The winner receives a golf trophy; second, third, and fourth place—as well as any observers or people simply standing nearby—get access to state secrets.
This game is all in the timing: Players must market their own cryptocurrency, then wait until just the right moment to cash out, crashing the currency and tanking other players’ life savings.
Modern-day pioneers move into a covered wagon after an eminent-domain order for the construction of a data center displaces them from their hometown. Drinking water is scarce, noise pollution is earsplitting, and morale is low as players try to survive the perilous journey to one of the few remaining corners of protected parkland. Inevitably, all players die when artificial general intelligence bioengineers a new form of dysentery.
In this first-person interior-design game, players gild every nook, cranny, and free patch of wall space in their virtual home office.
No longer beholden to child-labor regulations, players ages 6 to 8 design their own pickaxes before descending 1,500 feet down a shaft to mine bituminous coal. Open-world play means that the only limit is kids’ imagination and black lung and tunnel collapse.
Players battle conservationists, historians, and patrons of the arts atop a revered memorial building, destroying it in the process. Downloadable content expands play to the Reflecting Pool, the National Mall, and the East Wing of the White House.
Players rearrange foods that were once safe to eat into matching groups to make them disappear as part of an FDA recall. But miss just one, and players will contract a parasitic infection that sends them to the hospital, where they will incur astronomical medical bills, which will prevent them from qualifying for a mortgage, which will force them to move back in with their parents.
A zombie occupies the most powerful office in the world. At the conclusion of Level 2, President Evil might insist that the game isn’t over, but so long as it works as intended, someone else will finally get a turn to play.
2026-09-12 20:49:00
This is an edition of The Wonder Reader, a newsletter in which our editors recommend a set of stories to spark your curiosity and fill you with delight. Sign up here to get it every Saturday morning.
Learning how to drive is a rare instance in a teenager’s life, Liz Krieger argues in a recent article: “This wobbly uncertainty, this return to a kind of innocence, is the rarest condition in a 16-year-old, and it lasts about as long as a learner’s permit does.” The best part of teaching her daughter how to drive, Krieger writes, is being allowed to help.
Even once we move beyond the fierce teenage desire for independence, asking for assistance can feel awkward and vulnerable—espsecially in America’s culture of individualism, Julie Beck noted last year. But deciding to go it alone ends up depriving loved ones of the joy of participating. “Pushing through discomfort to accept the help that was offered” in planning her wedding, Julie wrote, made it “feel less like a product we had purchased and more like something special that we had created with our friends, family, and neighbors.” Today’s newsletter explores what asking for help does for the people helping.
On Asking for Help
The Last Thing Teenagers Will Let You Teach Them
By Liz Krieger
An interesting shift happens when a kid is learning to drive, and you’re both inside a box with your fear, moving 40 miles an hour.
To Make Someone Happy, Ask for Help
By Arthur C. Brooks
Generosity is like the circulation of the blood, healthiest and best when it goes around and around.
A Wedding Reveals How Much Help Is Really Available to You
By Julie Beck
Big life moments offer permission to ask for assistance. You should seize it.
Still Curious?
Other Diversions
PS
I recently asked readers to share a photo of something that sparks their sense of awe in the world. “It was a beautiful clear day when my husband and I were driving south from Seattle on Interstate 5 (he was driving),” Pat B., from Bellingham, Washington, wrote. “I saw Mount Rainier appearing to float serenely above the traffic below. The contrast between the immutable and the transitory was awesome!”
I’ll continue to feature your responses in the coming weeks.
— Isabel
2026-09-12 20:00:00
To understand Cormac McCarthy’s midlife crisis, look at America’s landscape. The fierce writer of biblical prose spent most of his first four decades anchored to Tennessee, a child of Appalachia with a lilting drawl and a regional outlook. By the early 1970s, he had published three Faulknerian novels—none of which sold well—and was struggling to finish a fourth book that aimed to chronicle his hometown of Knoxville in the way James Joyce had written about Dublin. McCarthy’s pockets were empty and his second marriage was failing. His then-wife, Annie DeLisle, recalled that on New Year’s Eve 1973, the year he turned 40, he suddenly announced that it was time for a change. As a friend put it, “He just walked off and left everything except his boots and his typewriter.” Pointing his truck west on I-40, McCarthy started driving and spent time in Tucson; Santa Fe; and Juárez, Mexico, before settling in El Paso, where he began researching a new novel called Blood Meridian.
Few life swerves are so productive. The book represented a fresh direction. Instead of another story set in the forests and shambling urban peripheries of Appalachia, Blood Meridian (1985) was an epic Western. It ambitiously set out to align the nation’s past with humankind’s age-old thirst for violence: Wolves cull one another, and so do people. From then on, the fearsome stakes of the West—territorial competition, crime, blood, death—defined McCarthy’s fiction and yielded his strongest work.
His next five novels, including the best sellers All the Pretty Horses (1992) and No Country for Old Men (2005), used desert plains and bleak landscapes to unspool a grim pessimism, shocking readers with stark visions of depravity while preserving the slenderest candlelight of hope: a few good souls carrying the fire. It was as though Appalachia’s dense woods and city fringes had prevented him from seeing the page clearly. Out west, there is nothing but sky.
In Cormac McCarthy: A Legacy Revisited, the first major biography of McCarthy, who died in 2023, Tracy Daugherty depicts his western reorientation as a turning point in his life and work but also looks for continuity. “The desert can be anywhere,” he writes; “the soul’s site of testing”—McCarthy’s lifelong fixation—is not limited to any one place. Although his departure was hurried, the transition took time. McCarthy made laps back to Tennessee during the 1970s and finished his Knoxville novel, Suttree, from his new base. The book was finally published in 1979 and unfairly panned by one critic as a “limited use of an enormous talent.” As for Blood Meridian, although it promised wonders to come, McCarthy’s first Western started off as a cult classic and only later, in the 1990s, began to find success.
[Read: On the death of Cormac McCarthy]
Yet once cemented, the relocation was the fulcrum of McCarthy’s life. After the ’70s, he never again lived in Tennessee or made Appalachian culture his subject. He eventually came to be so closely associated with the dusty boots and bleached bones of the West that new readers may be surprised to learn that the first half of his life was spent elsewhere. So why did he leave Knoxville? Daugherty presents a combination of personal failures and artistic imperatives that chased and lured him, respectively. Daugherty’s fascinating biography, which reflects impressive research in the Cormac McCarthy Papers at Texas State University, is very well told and filled with astute judgments about McCarthy’s 12 novels. Although Daugherty declines to elevate one part of his oeuvre over the other, the lesson of the biography—and the work—is clear. The cultural South may have been McCarthy’s home, but the West was his muse.
McCarthy’s father and namesake, Charles McCarthy Sr., represented urban order; he was set on taming the young Cormac’s backwoods waywardness. McCarthy later claimed that his father had beat him savagely during his childhood, afterward calmly sitting back down to finish his supper. Cormac’s artistically inclined mother helplessly watched the abuse. McCarthy Sr. worked as general counsel for the Tennessee Valley Authority, a New Deal agency that displaced tens of thousands of Appalachian families to make way for hydroelectric power. He did so with brisk pragmatism and a clean conscience, infuriating his son. Each man was baffled and hurt as the other rejected his outlook. The tension drives the early pages of Suttree : “The world is run by those willing to take the responsibility for the running of it,” the father in the novel tells his meandering boy.
Siding with the evicted, McCarthy slept on the porch, explored the forests, and caroused with Knoxville’s degenerate class. “Like Shakespeare’s Prince Hal, he liked to go slumming—remaining, just slightly, apart,” Daugherty writes. McCarthy listened and gathered material for his novels as often as he participated in the outlaw revelry. Yet he was no angel: A night with John Sheddan, a local misfit and frequent confederate, could as easily find the two sleeping in a barn or confined to a jail cell for public intoxication as discussing Thomas Hobbes. McCarthy studied physics and engineering at the University of Tennessee but never earned his degree, and began to read Dostoyevsky, Melville, and Faulkner during an uneventful stretch of Air Force service in the 1950s. His first marriage, to Lee Holleman, ended in divorce.
Family and friends offered few durable tethers to home, terrain that lacked lasting fictional appeal too. McCarthy’s early novels, beginning with The Orchard Keeper in 1965, chronicled the types of Appalachian outsiders and scofflaws he liked to run with. They proved off-putting company for readers. The books stood out as much for their desperate circumstances and ghastly subject matter (incest in Outer Dark, from 1968; necrophilia in Child of God five years later) as for their baroque prose and gothic settings. Suttree, the best of these novels and an excellent book, was nevertheless digressive and full of obscure, almost impenetrable writing (“rhomboid and volute shapes of limestone jutted all brindled with mud like great bones washed out”). Critics recognized a rare talent but also a voice struggling to break free of Faulkner’s influence.
[From the December 2025 issue: Thomas McGuane is the last of the literary outdoorsmen]
As his books misfired, McCarthy lived in an eccentric poverty that has become the stuff of legend. He cut his own hair and cooked canned food on hot plates. Working odd jobs (warehouse attendant in Chicago, garbageman in New Orleans), he recalled scraping by in “cellars on unemployment and welfare.” Without gainsaying these sacrifices, Daugherty calls out some exaggeration and myth building. Both DeLisle and Holleman contributed income; McCarthy’s Random House editor Albert Erskine helped him land fellowships and grants; and his parents, despite his strained relationship with his father, were a ready safety net. Still, between McCarthy’s failed marriages, family tensions, dead-end buddies, and dry artistic well, he had little to keep him in Tennessee as his 30s drew to a close.
A western reinvention allowed McCarthy to continue exploring timeless, elemental themes in a setting that befitted their gravity. In the American imagination, a cowboy’s journey is fraught, and his last breaths are tragic; a hillbilly’s exploits are hapless, and his final moments are merely uncomfortable. We cry for our heroes but wince for our outcasts. McCarthy sensed the difference. For a flailing author, the West also held commercial promise. “I ended up in the Southwest because I knew that nobody had ever written about it,” McCarthy later said. “Besides Coca-Cola, the other thing that is universally known is cowboys and Indians.”
Daugherty bristles at McCarthy’s dismissal of “decades of Southwestern literature”—by Leslie Marmon Silko, Katherine Anne Porter, N. Scott Momaday, and Larry McMurtry, to name just a few—yet the remark reveals a certain canniness: McCarthy cared more about selling books than his austere reputation suggests. It is not the only time Daugherty upends his subject’s ascetic image. In person, McCarthy was courtly and gregarious, a warm conversationalist who liked to sing. One striking scene has him dressed in a new cream suit, dancing into the night on an ocean liner as he twirls and spins his partners.
Blood Meridian proved to be McCarthy’s best novel and a modern masterpiece. Its protagonist is notably a young man who flees Tennessee for the West. The book follows the Glanton gang of scalp hunters in the mid-19th century along the U.S.-Mexico border. In ominous, hyper-violent language, McCarthy displays a nihilism that situates humankind on the raw continuum of predator and prey, river and rock: “The men as they rode turned black in the sun from the blood on their clothes and their faces and then paled slowly in the rising dust until they assumed once more the color of the land through which they passed.” Like an ancient mystic, McCarthy foresaw the endless reenactment of a horrifying cycle of brutality without meaning or purpose, a cursed birthright. “It makes no difference what men think of war,” says his most fearsome character, Judge Holden. “War endures. As well ask men what they think of stone.” To Daugherty, Blood Meridian’s clear-eyed view of American history was as amoral as realpolitik: “Politics as force.”
The Border Trilogy followed—the best-selling All the Pretty Horses leading on to The Crossing (1994) and Cities of the Plain (1998)—adding accessible characters and love stories to at last attract sales and awards. McCarthy’s writing began to shed some of its most ornate flourishes, taking on a stripped-down veneer. (Many of his later books were conceived as screenplays.) Critiques of the works as lightweight fare from the oracular McCarthy overlooked their hidden depths. This arresting insight, for example, comes to the young cowboy John Grady Cole in All the Pretty Horses: “He thought the world’s heart beat at some terrible cost and that the world’s pain and its beauty moved in a relationship of diverging equity and that in this headlong deficit the blood of multitudes might ultimately be exacted for the vision of a single flower.” Daugherty views McCarthy’s evolution in the Westerns as the pairing of a mature style with a genre framework. A similar combination has produced some of the best American movies of the past generation, by filmmakers such as Kathryn Bigelow and the Coen brothers, who directed an Oscar-winning adaptation of No Country for Old Men.
Although he started over, McCarthy could not escape the past. His celebrated 2006 novel, The Road—a sort of postapocalyptic Western—is built on a tender father-son relationship that both reflects McCarthy’s experience as a late-in-life parent and implicitly rejects his own father’s heavy hand. Daugherty quotes wrenching correspondence in which McCarthy and his father tried to reconcile as grown men. “There was and is a strong artistic streak in me that I inherited from my mother’s family and it has caused me a great deal of anguish in my life but an even greater joy,” McCarthy wrote. “I think this was always suspect to you.” Another relationship that echoed in his fiction was one that endured longer than his three marriages combined. McCarthy met Augusta Britt several years after moving out west; they became lovers when she was 17 and he was 43. They remained close for 47 years, but she protested the way he kept using her in his books, including in plotlines that seemed to mirror her life. “Well, baby, that’s what I do,” McCarthy replied. “I’m a writer.”
McCarthy attempted a second reinvention, a turn toward science. By the late 2000s, he was spending much of his time at the Santa Fe Institute, a multidisciplinary research center, where he hobnobbed with physicists, favoring philosophical conversations over writing books. The year before he died, he published a final pair of related novels, The Passenger and Stella Maris, which had been in progress for years; one of the main characters is a mathematics prodigy with a mental illness. In startling new reporting, Daugherty contends that McCarthy’s literary agent assembled the hodgepodge of pages and made the decision to break the mess into two books; McCarthy’s friends told Daugherty they suspected that the ailing writer had not even read the finished product. The result was unconvincing characters, mashups of McCarthy’s many interests, strange psychiatric sessions, and a lack of cohesion.
[From the January/February 2023 issue: The incandescent wisdom of Cormac McCarthy]
Maybe we all have only one good pivot in us. Or perhaps it was just too late for McCarthy to change course again. Whatever the case, his turn west in the 1970s reveals not just a man escaping personal failures but a restless writer who sensed that a wide new horizon might require him to expand his vision if he was to find his voice and his audience. The small-town hustles and degradations of Tennessee were ultimately too provincial for such an epoch-spanning talent, which demanded plains, mountains, and an endless expanse of sky to bring out its full potential. After all, deep in the Appalachian woods, you can’t see the buzzards circling.
This article appears in the October 2026 print edition with the headline “Cormac McCarthy’s Midlife Crisis.”
