Articles & Videos

CAR-T therapy engineers a patient’s own immune cells to fight cancer. Making these cels more resilient can make treatments more effective.

The environmental toll of the artificial intelligence boom continues to mount as tech companies use ever more power to run their data centers and enormous amounts of water for cooling. A new investigation by U.N. scientists warns that AI’s water use in 2030 will match the needs of 1.3 billion people, while its power use will be triple that of Pakistan, Bangladesh and Nigeria combined — countries with a total population of 650 million. “Most people understand AI as a digital technology, as a virtual thing, as something that is in the clouds,” says Iranian environmental scientist Kaveh Madani, director of the United Nations University Institute for Water, Environment and Health. “What we tried to do in this report was to remind people that there’s some physics to all of this.”

As an investor and customer, the UAE has deep ties to OpenAI.

Users are increasingly looking for value as well as capability.

It’s hard to think of a cohort of rich people in recent history as extravagantly exhibitionist as today’s tech billionaires. For the modal Silicon Valley oligarch, the life of easy luxury is not enough. The point of being unfathomably rich in our time is not simply to enjoy the fruits of extreme wealth, exert a crushing influence over the course of politics and public life, or hold up one end of the K-shaped economy: It’s to become the center of global attention, which is where real power in the digital era resides. The sheer visibility of the have-yachts today is staggering. Not only do we have to suffer their opinions elaborated at length on podcasts and in X posts; every squalid little detail of life at the top is out in the open now, freely available for public consumption. We’re intimately familiar with the ins and outs of Jeff Bezos’s time in space and his nuptials in Venice, and there are whole subindustries devoted to parsing Mark Zuckerberg’s T-shirt and jewelry choices.And then there’s Elon Musk. In the ranks of the billionaire exhibitionists, there’s no one who tries harder or squeaks louder. In 2022, Musk bought Twitter, later renaming it X, and promptly set about turning it into an online mess hall for right-wing lunatics; by 2024, he was posting on X an average of 60 times a day and sometimes up to 40 times every hour. At any given moment on X, Musk can be seen posting with the sweaty eagerness of a teen in the first flush of puberty: On a recent sampling, his feed included a signature mix of self-promotion (plugs for Grok, X’s in-house AI chatbot; the satellite internet service Starlink; and the self-driving Tesla), competition-targeting shitposting (a tweet claiming ChatGPT, OpenAI’s rival to Grok, is “pretending to be ‘unbiased’” scored the Musk retweet), C minus memes (a graphic showing four descending pairs of increasingly bloodshot eyes superposed with the words “Alcohol,” “Weed,” “Cocaine,” and “Monitoring the Situation”), transphobia (“People can pretend or dress up all they want and that’s fine, but they can’t force their mental illness to be my new reality!”), and white nationalist race panic (an approving “100” emoji slapped over the top of a tweet from a verified account claiming that “if White men become a minority, we will be slaughtered”), along with countless variations on the classic theme of middle-aged white guy whining about the state of the world today. A smattering of cruel and lazy jokes about trans people, some indignant quote tweets of garden variety racists like Stephen Miller and British far-right activist Tommy Robinson, a stray “Hmm” or “Troubling” in response to tweets about things like tech regulation, fertility rates, or anti-white racism? In Muskworld, that’s called a Monday.Musk’s presence on X can often feel like performance art—one of those continuous “bits” that have turned much of social media into a form of digital vaudeville. But where most extremely online victims of internet brain damage emote into the void, Musk is the rare oversharer who’s managed to turn the “based” and “epic” style of memeverse comedy into something more tangible. Dogecoin, the meme coin that has historically soared in value whenever Musk has tweeted about it, eventually became the inspiration for DOGE, the Big Government slash-and-burn operation that Musk led at the start of Donald Trump’s second term. In 2023, Musk committed $100 million in seed money for a new STEM-focused school in Austin that he chose to name the Texas Institute of Technologies and Sciences, or TITS: Every thumbnail on tits.academy, the still-unrealized university’s home page, now takes visitors to an “About Us” screen headed “Take a look at TITS” that includes an embedded YouTube video of Rick Astley’s 1987 pop classic “Never Gonna Give You Up”—an online prank of ancient vintage known as “rickrolling.” In August 2025, Musk announced a plan to launch a fully AI-simulated software company that could supplant all the functions and products offered by Microsoft; he called it “­Macrohard.” LOL! Not content with sitting on the world’s largest personal fortune, Musk seems desperate to be taken seriously as a comedian, too.Maybe there’s something else at work here. According to the authors of a new book, Musk’s transformation into a hyperactive super-troll expresses a serious purpose. “Musk’s online persona,” write Quinn Slobodian and Ben Tarnoff in Muskism: A Guide for the Perplexed, is often “misunderstood. Critics see immaturity or malice; fans see relatability or authenticity. Both fail to see that, in Muskism, trolling is infrastructure.” The point of these outbursts is for Musk to gauge whether he can bypass the ordinary institutions of grown-up accountability—the media, elections, quarterly earnings reports—and affect outcomes in the world of money, politics, and popular belief through the sheer force of his own personality. They are a stress test of society’s tolerance for the Muskian worldview.“Muskism” is the name that Slobodian, a historian, and Tarnoff, a technology writer, give his worldview: a kind of techno-maximalism in which autonomy for individuals and for nations is only achievable through an ever-deeper fusion between human and machine, and can best be guaranteed through the adoption of technologies offered by Musk’s own companies. Just as Fordism was “the operating system of the twentieth century,” Slobodian and Tarnoff contend, Muskism might provide the basis for a new consensus about economic and social life, the Fordism of our time. After all, Fordism, like Muskism, once seemed an intimate reflection of one man’s personality (The Washington Post defined it in a 1922 article as “Ford efforts conceived in disregard or ignorance of Ford limitations”). Only later did it come to mean the belief in mass production and standardization on the factory floor, high wages, and mass consumption as the chief motors of industrial growth. The clarifying and unsettling argument at the heart of Muskism is that, in a deglobalizing and increasingly digitized world, Musk’s vision of the future might win out. But is it possible to separate Muskism from Musk?Musk may need no greater reason to post than that it flatters his ego. He currently has 240 million followers on X—more than double the following of the second most popular account, that of Barack Obama. This grants him a kind of insurance against failure whenever he tweets, even for his weakest stuff. The “Monitoring the Situation” meme above, for instance, generated more than 900,000 likes, a return out of all proportion to the quality of the material presented—proof, if ever we needed it, that meritocracy in America is dead.Online ubiquity is also a moneymaking tool. Musk has long been a master of what Slobodian and Tarnoff call “financial fabulism”—the art of extracting ever-higher valuations and capital commitments from investors through confidence, showmanship, and a compelling view of the world to come. When raising money for online business directory Zip2, his first company, in the mid-1990s, Musk covered a normal PC with a big case and wheeled it around to venture capitalists to make them think his software ran on a supercomputer. Zip2 was still losing money when Compaq bought it for $307 million in 1999; Musk walked away with $22 million, and a valuable lesson in the power of his unstinting belief in the future. Musk has channeled this gift for financial fabulism into the digital world, which has compressed and supercharged the feedback loop between words and money. In 2018, he tweeted: “Am considering taking Tesla private at $420. Funding secured.” The market missed (or ignored) the weed joke in “420” and took him at his word; the company’s stock price jumped nearly 11 percent by the end of the day. In January 2021, he added #bitcoin to his bio on Twitter (as it was then known), and the cryptocurrency surged 20 percent within an hour. The shocking immediacy of cause and effect online, coupled with his embrace of Twitter, helped make Musk “as close to a free-money perpetual motion machine as you’ll ever see in finance,” Bloomberg columnist Matt Levine has written.But Slobodian and Tarnoff see in Musk’s online presence a deeper expression of his idiosyncratic vision of the future. From the mid-2010s onward, Musk became convinced that humans and machines were becoming one, and that success—for Musk as an entrepreneur, for humanity as a species—would involve accelerating this fusion rather than resisting it. Musk has often warned of a Terminator-style apocalypse from rogue artificial intelligence, but his solution is “not less technology, but more,” Slobodian and Tarnoff write: The risks inherent in the digital cognitive revolution mean we should “merge with AI,” Musk argued in 2016.All of Musk’s juvenile foolery stems from his “cyborg turn.” Modern Musk—the man as empire, the man as meme—is an embodiment of online culture.Under Muskism, the path to salvation lies in human-machine symbiosis. All of Musk’s juvenile foolery over the past few years—the insults, the weak memes, the hyperactive posting, the alliance with Trump, the dadly insistence that “I am become meme,” even the bit with the chainsaw at CPAC—stems from this “cyborg turn,” as Slobodian and Tarnoff call it. This turn, they write, “did not mutate out of one man’s cyberpunk delusions, but out of a bipartisan consensus about where America needed to direct its brightest minds and deepest investments.” Modern Musk—the man as empire, the man as meme—is an embodiment of online culture. Or as Slobodian and Tarnoff put it, he is “a hypervisible indicator species for the consequences of fusing the economy ever more tightly with the digital world.”Whether the future synthesis of human and machine emerges in perfect or imperfect form, few people will benefit from it more than Musk himself. His companies are building much of the infrastructure—the rockets, robots, satellites, collective intelligence networks, and brain-computer interfaces—that will shape the fusions to come. When he left his self-described political “side quest” in Washington last year, he explained that it was to return to his real quest, which is to build the future world of “humanoid robots and digital superintelligence,” to seed our “future machine descendants.”Slobodian and Tarnoff’s analysis of Musk’s cyborg turn is reassuring and discomfiting at the same time: reassuring because it suggests an intelligible strategy at the heart of Musk’s actions, making him an outlier at a time when most big decisions in politics—kidnapping Maduro, bombing Iran—seem guided by nothing more than the uncut essence of poster brain; discomfiting because, well, it also suggests we’re all about to become replicants welded to a Tesla bot or a SpaceX nozzle.Mechanization of the human is the logical end of a sequence of fusions that have defined Musk’s career. Social media has allowed him to merge, in some way, his own intelligence with the hive mind, to become one with the scroll. With SpaceX, he engineered a marriage of firm and government, grafting his enterprise onto the state, taking over some of the state’s functions, and exploiting its guarantees for private gain. While many in Silicon Valley profess to dream of liberation from the state, of wholesale exit from its mess of obligations and compromises, Musk’s business interests have pushed in the opposite direction, toward an ever-deeper coupling of public and private: His ultimate goal has always been power over the state rather than freedom from it.All technology businesses in the digital era arguably free-ride off the state—the internet and its foundational infrastructure, after all, were funded by DARPA, the Pentagon’s R&D arm. Yet few entrepreneurs have been as skilled at piggybacking off public resources as Musk has. Zip2 used GPS data from military satellites, for example, while Musk’s late-1990s online payments company X.com—which would later become PayPal—rested on the federally insured stability of the U.S. financial system. SpaceX—founded in 2002, now valued at $1.55 trillion, and preparing for a public offering that early reports suggest could become the biggest IPO in history—is perhaps the company that has benefited most from government resources.Speaking at Stanford the year after he launched the company, Musk acknowledged that NASA had already done the hard work “by spending the money to develop some of the fundamental technologies,” and now the door was open for the private sector to swoop in, build on those state-funded foundations, and monetize the gains. Twenty-three years later, SpaceX is firmly enmeshed with U.S. infrastructure. “By 2025, SpaceX accounted for 95 percent of all orbital launches in the United States and more than half of all launches globally—a position that made the Pentagon, NASA, and other government agencies deeply reliant on Musk,” report Slobodian and Tarnoff.Timing was key to this success. Musk founded SpaceX right as the United States launched the global war on terrorism, which collided with the neoliberal mania for privatization and outsourcing to create a bonanza for corporate defense contractors. But SpaceX’s elevation into an indispensable arm of the modern military, not only in the United States but for countries abroad, also reflected Musk’s real business acumen. Satellites had been around for decades when Musk announced plans for what would become Starlink, a satellite internet service, in 2015. Slow speeds were always a barrier to greater adoption of satellites for global communication. Musk’s solution was to bring the satellites closer to the ground, putting them into low Earth orbit—around 342 miles up, versus 20,000 miles in the air for traditional geostationary satellites. The only catch was that orbiting so close to Earth required a lot more satellites to ensure quality of connection. The result? Starlink now has more than 10,000 functioning satellites in orbit, around two-thirds of all active satellites: The fusion of SpaceX to the state has led the way, quite naturally, to the company’s colonization of the heavens.Tesla is yet another Musk vehicle that has benefited from state largesse. Musk did not create the company, but he invested in it and took over as CEO in 2008—right at the launch of the first federal tax credit program for buyers of electric vehicles; the company was saved from almost certain death by a $465 million Department of Energy loan awarded a year later. That loan helped Tesla survive the lean years that followed the collapse of the clean tech and renewables sector in the early 2010s—the result of the “shale revolution,” which touched off a surge in U.S. oil and gas production.But the real reason Tesla prospered through the downturn—eventually becoming the world’s leading EV automaker, a title it relinquished to Chinese firm BYD earlier this year—was that Musk insisted on sourcing and assembling as many of the company’s critical production inputs in-house rather than relying on the global market. At a time when manufacturing consensus was in favor of free trade, offshoring, and global supply chains, Musk seemed to have already begun to anticipate a deglobalizing, post-Covid world by embracing “vertical integration.” The real innovation here was the construction of the first Gigafactory in Sparks, Nevada—what Musk envisioned as a “truly gargantuan factory of mind-boggling size”—that allowed Tesla to make its own lithium-ion batteries. (Since then, fresh Gigafactories have sprouted in Buffalo, Shanghai, Berlin, and Austin.) When the world awoke, a few years ago, to a reality of reshoring, export controls, tariffs, and emerging trade blocs, Tesla was relatively insulated; Musk had built Tesla for a world in which self-reliance, rather than reliance on others, would become the key to survival.Musk’s fantasy of the future, the authors of Muskism propose, is “a fantasy of the factory,” stripped to its leanest and trimmed of human encumbrances. In 2016, Musk first mentioned his goal of creating an “alien dreadnought”—a factory with no human workers at all. Macrohard is perhaps the first serious attempt at realizing that goal: a company that relies for labor on AI agents. He’s not alone in this ambition: AI’s incursion into labor markets has already begun, and the dream of a fully post-human economy can be felt in every giddy LinkedIn post about the leaps in “productivity” that technologies like Claude Code are now making possible. Under Muskism, humans can either become obsolete or fuse with machines, becoming one with the factory itself.The sheer number of countries and institutions that rely on Musk’s technologies gives him a bullying leverage that he seems unafraid to flex. For many countries, the exercise of sovereignty brushes up against the whims of a single, unpredictable, and often quite petty man. In March 2025, Musk boasted that Starlink is “the backbone of the Ukrainian army,” and that “their entire front line would collapse if I turned it off”; when the Polish foreign minister, in reply, warned that his own country would be forced to look elsewhere for Ukraine if Starlink was an “unreliable provider,” Musk spat back, “Be quiet, small man.” Exchanges like this highlight the real problem for Muskism as a social project, which is that Earth is a place populated by humans not named Elon Musk, and that many of them disagree, often quite fervently, with his views and methods. Not everyone wishes to be transformed into a widget on the floor of the Muskian factory of the future.Humans are Musk’s greatest enemy, the source of his keenest humiliations and most obliterating rages: Who could forget the world’s richest man, on the verge of tears, telling Fox News in March 2025 that Minnesota Governor Tim Walz was “a creep,” and “a huge jerk,” all because Walz made fun of Tesla’s plummeting stock price at the height of DOGE-driven chaos in Washington? Empathy, Musk famously once said, is the “fundamental weakness of Western civilization.” That seems untrue, and Musk’s own direct involvement in politics has shown that a lack of empathy is no model for government.Slobodian and Tarnoff trace Musk’s attempts to correct a listing culture, to rewrite society’s operating code and clean it of the real bugs in the system: people. Twitter had helped propel Occupy Wall Street, Black Lives Matter, and #MeToo; under Musk, its replacement, X, would reinstate a public square built on precontamination hierarchies of class, gender, and race. Large language models promised a future of objective truth and universal information, but were trained on leftist propaganda; Grok would be refined in “post-training” to reflect Musk’s own anti-woke biases and priorities. The state was overrun with illegitimate claims on the public purse; DOGE would empower a set of very young software engineers to swiftly terminate grants and contracts and lay off public servants en masse.And yet, here we find ourselves today: On X, there are still as many jokes about Nazis as there are Nazis; Grok is not a serious competitor to ChatGPT or Claude; and DOGE accomplished little more than causing chaos and inflicting misery on federal workers. Its outcomes may have satisfied MAGA ideologues such as Russell Vought, whose express wish was to dismantle the federal bureaucracy by making working conditions unpleasant (“When they wake up in the morning, we want them to not want to go to work”). And in the midst of his DOGE spree, Musk seemed to share this wish, brandishing a chainsaw on stage at CPAC and bragging about cuts. But if, as Slobodian and Tarnoff argue, Musk had imagined DOGE as a program through which “engineers disciplined society like a factory floor,” it was a failure. Because, as they write, “society is not a factory.”Musk presents an intriguing contrast with Trump, his on-and-off-again political ally. Both are inveterate posters, both often sound deranged, both have a lot of money (Musk has considerably more, but much of it is illiquid, since it’s mostly held in stock in his companies). Trump is sui generis, a freak of nature whose coalition seems unlikely to survive his departure from the White House (assuming he ever leaves it). Vestiges of Trumpism will survive, but actually existing Trumpism? Only Trump can carry the brand. His interests are superficial: What animates him is a desire to enrich himself and elevate the status of his clique, even at the expense of America’s general welfare and standing in the world. That’s not a political realignment; it’s kleptocracy. And yet, despite all his shortcomings, despite his leaden approval ratings, Trump has twice been elected president. Musk spent less than six months in Washington before he was forced to skulk back to the Gigafactory. The argument that Muskism asks us to take seriously is that Musk, though less charismatic and less popular than Trump, is more symbolic of our era.Amid institutional breakdown and the overpoliticization of everything, Musk’s companies, his words, and his dogmas might represent a durable shift to a new common sense, a new model of political and economic order. Trump has pioneered a style, but Musk is a master of infrastructure, and the future is always shaped by matter rather than manner. The liberal international order is effectively dead, popular faith in institutions across the democratic world has cratered, and in the United States, the Trumpian wrecking ball has done serious—perhaps lasting—damage to the Constitution and the stately old politics of checks and balances. Muskism offers a set of solutions and arguments suited to the fragmented, paranoid, and militarized world in which we now find ourselves: Independence seems more appealing than integration at a time when multilateralism is at a historic ebb; genocide and wars of choice have normalized the exclusionary thinking and antipathy to empathy on which Musk’s businesses thrive; multipolarity and the return of great power competition invite a deepening fusion of technology and the state. Muskism is a tool kit for a world where concentration matters more than cooperation, where control trumps trust.Muskism-lite is not the Muskian way; the insults, the trolling, and the inhumanity are inseparable from Musk’s way of doing business.Still, a future defined by all-out Muskism is not inevitable. It’s not hard to imagine a greater role in years to come for a detoxified version of Muskism, characterized by a turn away from supply chains and globalization, a deepening digitization of daily life, and government by meme. But Muskism-lite is not the Muskian way; the insults, the trolling, and the inhumanity are inseparable from Musk’s way of doing business. There is little utopian promise in the future he portends; the Musk-led tech industry’s vision of collective life is a weird amalgam of sports, speculation, and collective dissolution into the jaws of the machines. Many people, as Musk himself has discovered in recent years, do not want this. Muskism may be one version of the future, but cyborgian oblivion is not our only choice.  

Trump and Sanders want the government to exact a public benefit from big AI firms. Others aren’t so sure.

CFTC Chair Mike Selig told Semafor he's weighing more lawsuits against states that try to regulate those markets.

As Wall Street gears up for SpaceX’s record-breaking IPO, regulators quietly granted brokerages an extraordinary exemption from rules protecting customers if markets turn volatile.

Europe’s AI buildout is orders of magnitude too small, a group of AI policy thinkers warned.