I. That Antichrist Nonsense, Explained
Twenty-second-century historians tracing the ascent of American technocracy from quirky garage tinkerers grooving on the Whole Earth Catalog to sinister oligarchs enacting Philip K. Dick prophesies may locate its apogee in four lectures that Peter Thiel (net worth: $29 billion), chairman of the data-mining giant Palantir and a co-founder of PayPal, delivered at San Francisco’s Commonwealth Club in September and October 2025. The subject was the Antichrist.
“In the seventeenth, eighteenth century,” Thiel explained, “the Antichrist would have been a Dr. Strangelove, a scientist who did all this sort of evil crazy science.” As Thiel spoke, dozens of protesters—some in devil costumes—marched outside bearing signs that said things like The End Is Near/ Palantir Is The Path/ Thiel Leads The Way. “In the twenty-first century,” Thiel continued, “the Antichrist is a Luddite who wants to stop all science. It’s someone like Greta or Eliezer.” Greta is Greta Thunberg, the Swedish climate change activist. Eliezer is Eliezer Yudkowsky, a Berkeley-based critic of artificial intelligence, or AI.
Class warfare doesn’t get more unhinged than this. Say what you will about the American plutocracy, it seldom frames its economic self-interest as a religious imperative. But even in its more innocent days, Silicon Valley inclined toward grandiosity, heralding not just a new technology but a new advancement in human consciousness. Now a prince of the technocratic elite was framing tech’s future prosperity quite literally as a battle against agents of Satan, with Thunberg and Yudkowsky cast as Gog and Magog. Suggest Thiel’s dominion could stand a bit more government oversight and he just might toss you into a lake of fire.
Thiel’s speeches prompted an appropriately horrified public response. But his was merely the most literal expression of a millenarian sentiment about the coming of AI that’s now conventional wisdom among tech barons. Thiel’s tactical error in applying the Book of Revelation to the digital future was to keep his message explicitly biblical. But other tech leaders have been secularizing the same narrative. “Technology is the godhead” is how the Columbia law professor and tech critic Tim Wu described this view to me, and Artificial General Intelligence, or AGI—AI that surpasses human intelligence in all dimensions—“is the Second Coming.”
The story goes like this. A war is coming between good (AI) and evil (government regulation). If AI wins, a New Jerusalem will dawn where human intelligence is eclipsed by intellectually superior computers that represent, figuratively if not literally, Jesus Christ’s return. Techies call this the Singularity. Silicon Valley executives can prepare for that great gettin’ up day by paying $15,900 to attend a five-day seminar at an establishment in Santa Clara County, California, called—I kid you not—Singularity University.
How glorious will be that dawn to come! “The first ultra-intelligent machine,” the British mathematician Irving John Good wrote in a 1965 essay that’s often credited with introducing the singularity concept, “is the last invention that man need ever make.” It follows that any who would impede AGI’s arrival, or attempt to control its development in any way, must be in league with the seven-headed Beast. Not a literal beast, mind you, but rather the notional one of government interference or (in more extreme versions of this tale) democracy itself. Either way, the meddlers must be stopped at any cost.
That’s more or less the story told by the Silicon Valley venture capitalist and Netscape co-founder Marc Andreessen (net worth: $1.9 billion) in the “Techno-Optimist Manifesto” that he posted online in October 2023. “We believe Artificial Intelligence is our alchemy, our Philosopher’s Stone,” Andreessen wrote, invoking a process to turn lead into gold and thence into an elixir of eternal life that, let me remind you, never existed outside mythology. Also: “We believe any deceleration of AI will cost lives. Deaths that were preventable by the AI that was prevented from existing is a form of murder” (italics mine). Andreessen declared war on such commonsense concepts as “sustainability,” “social responsibility,” “risk management,” and “tech ethics,” because they were part of “a mass demoralization campaign … against technology and against life.”
Andreessen may sound like a crackpot parading Sand Hill Road in a sandwich board, but in the tech world he’s hugely consequential. His venture capital firm, Andreessen Horowitz, is Silicon Valley’s richest, with $90 billion in assets. Previously a vocal fundraiser for Democratic candidates, in 2024 Andreessen spent $5.5 million to elect Donald Trump and helped Elon Musk recruit staff for his Department of Government Efficiency, or DOGE. Bloomberg’s Emily Birnbaum and Oma Seddiq reported in February that Andreessen Horowitz is “the first outside call that top White House officials and senior Republican congressional aides make when considering moves that could affect tech companies’ AI plans.”
The talk may be of a literal or figurative God, but what’s really at stake—as usual—is Mammon. Tech lords’ ferocious opposition to government interference reflects a collective financial investment in AI that’s quite literally unprecedented within the private sector. In February, The Wall Street Journal reported that the $670 billion to be spent this year developing AI by Meta (Facebook), Amazon, Microsoft, and Alphabet (Google) represents 2.1 percent of the U.S. gross domestic product. That’s slightly more than what the United States spent to build the railroads in the 1850s (2 percent of GDP), and considerably more than the amount spent to build the Interstate highway system (0.4 percent) or to put a man on the moon (0.2 percent). The only national investment the Journal could identify that represented a larger slice of GDP was the Louisiana Purchase (3 percent), which nearly doubled the size of the United States. That was in 1803, when GDP was a puny $488 million, not today’s $31 trillion. And unlike these earlier infrastructure projects, this year’s $670 billion investment in AI draws entirely on private-sector funds.
The stakes have pushed Silicon Valley into the arms of the GOP. As recently as 2020, the tech industry favored Joe Biden overwhelmingly against Trump, with 98 percent of its donations going to Democrats. The biggest tech contributor that year was Netflix chair Reed Hastings, who gave more than $5 million to Democrats. But by late 2025, the nonprofit Public Citizen found, nearly three-quarters of tech political spending went to Republicans, with Musk the biggest tech contributor; he gave $351 million to elect Republicans. Granted, Musk’s impact was outsize; nearly half of tech’s political spending came from Musk alone. But after Trump won in 2024, other tech chiefs fell in line, scrambling to attend Trump’s second inauguration, with four of them (Musk, Meta’s Mark Zuckerberg, Amazon’s Jeff Bezos, and Google’s Sundar Pichai) seated in front of Trump’s Cabinet members, a spot previously reserved for former presidents and the incoming president’s family. This digital Rushmore contributed a combined $26 million to Trump’s inauguration and planned “Golden Ballroom.” Overall, the tech industry kicked in $48.6 million. To you and me, that’s a lot of money. To tech companies, it’s pocket change.
Until recently, tech showed limited enthusiasm for Washington’s lobby scene. A decade ago, it ranked fourth among lobbying industries, spending less than half as much as Big Pharma, which ranked first, according to the campaign website OpenSecrets. But by late 2025, the last year for which data is available, tech had moved up to the number two spot, spending nearly three-quarters as much on lobbying as Big Pharma (which still ranks first). The reason was AI. “AI didn’t just increase its footprint in Washington,” noted Ashley Gold in Axios. “It ate tech lobbying whole,” and it added to tech’s existing lobbying on corporate consolidation, privacy, and free speech new topics like crypto, defense procurement, and the insatiable energy needs of data centers.
Even Silicon Valley’s own congressman, Democratic Representative Ro Khanna, frets about tech’s reach. “We need regulations that prevent companies from using AI to eliminate jobs to extract greater profits,” Khanna tweeted in November. “There should be a tax on mass displacement.” After Khanna endorsed California’s proposed onetime 5 percent wealth tax on billionaires, a partner at Andreessen Horowitz called Khanna an “obnoxious jerk” on X and committed to “voting him the fuck out.” Never mind that Khanna, who calls himself a “progressive capitalist”—and who as recently as last year collected $15,000 in campaign contributions from Andreessen Horowitz associates—is nobody’s idea of a radical; in 2024, govtrack.us rated him the sixty-seventh-most left-leaning member of the House. When tech peers out its window, it sees only enemies.
This was an industry that, a generation ago, came into being promising to disperse power and information to ordinary people. Now it fetishizes surveillance, misinformation, monopolization, and lethality as it races toward a Singularity that occasions, among thinking people with no financial stake in it, no small worry. How did it get so monstrous?
II. When Computers Meant Freedom
“Ready or not, computers are coming to the people,” wrote Stewart Brand in a December 1972 issue of Rolling Stone. (Carlos Santana was on the cover.) “That’s the good news, maybe the best since psychedelics.” Previously, Brand had been one of the novelist Ken Kesey’s Merry Pranksters, that youthful band of LSD-dropping nonconformists whose travels on a multicolored school bus Tom Wolfe famously chronicled in The Electric Kool-Aid Acid Test (1968). Brand had then moved on to create the Whole Earth Catalog, a wildly popular counterculture guide to living close to the land and far from the Man. Readers could learn where to buy the best cut beads, or how to build a geodesic dome, or what the I Ching was, or how to repair your Volkswagen Beetle. In a 2005 commencement speech at Stanford, Apple founder Steve Jobs would remember the Whole Earth Catalog as “sort of like Google in paperback form 35 years before Google came along.” Brand also created The WELL, an early online discussion forum still in operation.
When Brand wrote in 1972 about the coming democratization of computing, computers were still giant mainframes available only to governments, large corporations, and universities. Lewis Mumford, in his 1970 book, The Pentagon of Power, had called the computer “the omnipresent Executive Eye, he who exacts absolute conformity to his commands, because no secret can be hidden from him, and no disobedience can go unpunished.” Popular culture similarly cast computers as faceless villains in films like 2001: A Space Odyssey (“I’m sorry Dave, I’m afraid I can’t do that”) and Colossus: The Forbin Project (“You will say you lose your freedom. Freedom is an illusion”).
In his Rolling Stone essay, Brand was writing about a different kind of computer that was still a decade off—a personal computer that, as the former New York Times journalist John Markoff put it in his 2005 book What the Dormouse Said: How the ’60s Counterculture Shaped the Personal Computer, would stop “being dismissed as a tool of control” and instead be “embraced as a symbol of individual expression and liberation.”
The paradox of a computer-industry counterculture was evident from the start. Brand’s Rolling Stone piece was set in the heart of the Establishment—Stanford’s Artificial Intelligence Laboratory in the foothills of Palo Alto. But what Brand described was about 20 “raucous,” long-haired “hackers” (a term he had to define for his readers) playing a computer game called Spacewar! It was, Brand wrote, “the most bzz-bzz-busy scene I’ve been around since Merry Prankster Acid Tests.” Yet this boho playground was made possible by the government-created Advanced Research Projects Agency, or ARPA, which linked up 20 major computer centers around the country in what would evolve into the internet. Indeed, ARPA was part of the government agency the counterculture most hated—the Pentagon, then drafting young men to fight in Vietnam. And while material gain was not foremost for the early personal computer innovators, it was hardly absent. “They didn’t mind being rich,” Walt Mossberg, who wrote The Wall Street Journal’s “Personal Technology” column from 1991 to 2013, explained to me, “but that was not their principal thing.”
Still, Markoff argues in What the Dormouse Said, it was the more decentralized and free-spirited nature of West Coast computer culture that gave Silicon Valley the creative advantage in developing personal computers over the “more hierarchical and conservative” computer culture in the Northeast, where IBM, Harvard, and MIT resided. The historian Theodore Roszak, who popularized the term “counterculture,” wrote in 1986 that personal computing grew out of “a sort of primitive cottage industry. The work could be done out of attics and garages and simple means and lots of brains.” As late as 2009, the journalist Jeff Jarvis was still able to observe that “small is the new big” and “the Lilliputians have triumphed.”
Brand, who at 87 remains a techno-optimist, told me in an interview that much of that reality remains. “Bear in mind,” he said, “that the world’s encyclopedia is still Wikipedia, and the Internet Archive is doing a free archive of everything.” Both are nonprofits reliant on volunteer labor. Also, Brand said, “you have companies like iFixit which is pushing right to repair” by furnishing repair instructions free of charge. This last infuriates manufacturers, who want to charge consumers for repairs. (iFixit makes its money by selling repair kits and other tools.) “YouTube has completely replaced the Whole Earth Catalog,” Brand said, and “Open-source software [like Unix and Linux] is just as powerful as it’s ever been…. All of that is a democratization beyond what people were imagining as we went into the ’70s and ’80s.”
Tim Berners-Lee, who in 1993 made like Jonas Salk with polio and put his invention—the World Wide Web—into the public domain, takes a gloomier view. “Over the past few years,” Berners-Lee writes in his 2025 book, This Is For Everyone: The Unfinished Story of the World Wide Web, “I’ve fought to keep the web transparent, open-source, and freely accessible.” But with the advent of “the addictive forms of social media,” the majority of web traffic “is now concentrated in a handful of large platforms which harvest your private data and share it with commercial brokers or even repressive governments” (including, he might have added, in the United States). “Worse, authoritarian governments now use the web to spread disinformation and surveil their own citizens…. In the age of AI, these threats are more urgent than ever before.”
“Let’s not overstate it,” Kara Swisher, who’s been covering tech for three decades, told me. “I never thought they were liberal. You never knew what their politics were. It was a kind of libertarian-light if they thought about it at all.” Swisher’s 2024 memoir, Burn Book: A Tech Love Story, begins with the sentence: “As it turned out, it was capitalism after all.”
Of course it was. As far back as 1997, Thomas Frank, in his books Commodify Your Dissent and The Conquest of Cool, protested the co-optation of 1960s rebel youth culture by corporate America (a theme later recycled, minus the disapproval, by David Brooks in his 2000 bestseller, Bobos in Paradise). In Frank’s view, counterculture and corporate culture had forged an unholy alliance from the start:
The 1967 “summer of love” was as much a product of lascivious television specials and Life magazine stories as it was an expression of youthful disaffection; Hearst launched a psychedelic magazine in 1968; and even hostility to co-optation had a desperately “authentic” shadow, documented by a famous 1968 print ad for Columbia Records titled “But the Man Can’t Bust Our Music.”
The maturing tech industry’s posture of rebellion was most fully expressed, in January 1984, by Apple’s famous “1984” Super Bowl ad, directed by Ridley (Blade Runner) Scott, wherein a female athlete smashed with a sledgehammer a two-way telescreen on which George Orwell’s Big Brother was addressing the compliant masses. “On January 24,” the ad concluded, “Apple Computer will introduce the Macintosh. And you’ll see why 1984 won’t be like 1984.” The thrust was that the personal computer would liberate us from the UNIVAC-style corporate and governmental dominance Mumford feared. But its real-world inspiration was a more prosaic ambition to defeat giant competitors in the PC market like IBM and AT&T, which were already stumbling.
III. The March to Monopolism
As tech firms scaled up, Washington mostly left them alone, in large part because politicians had a poor grasp of their products; as late as 2004, President George W. Bush spoke of “the internets.” A rare exception was the Clinton Justice Department’s antitrust lawsuit against Microsoft, which Microsoft mostly lost in a 2001 settlement. More consequential, though, was Section 230 of the 1996 Communications Decency Act, which protected computer service providers from legal liability for third-party comments on their platforms, making it possible for social media giants like Facebook and X to flood the public with misinformation and toxic prejudice.
The slow death of journalism and consequent dumbing down of the electorate are largely the fault of Section 230. Look, I don’t dispute that giving five billion people a publishing platform was an intriguing social experiment. But 30 years on, the results are in. The experiment failed. If we repealed Section 230, that would kill off social media, at least in its present form. I’d be fine with that. But repeal may be a pipe dream, because it would be fought to the death by at least two social media barons (Elon Musk and Mark Zuckerberg) whose combined net worth approaches $1 trillion. Still, if asked to choose between Facebook and democracy, which would you pick?
Meta’s Mark Zuckerberg (net worth: $185 billion) made a great show for a while of curbing offensive comment on Facebook (and to some extent the company continues to do so). But during Trump’s first presidency, Zuckerberg gave a speech cloaking Meta’s financial self-interest in the free-speech protections of the First Amendment and spouting what by that late date (2019) was an obviously phony populism. “I understand the concerns about how tech platforms have centralized power,” Zuckerberg said, “but I actually believe the much bigger story is how much these platforms have decentralized power by putting it directly into people’s hands.”
That sounds reassuringly egalitarian, but some of these “people” are minors. According to a 2023 Pew survey, the proportion of American teenagers on Instagram was 45 percent for those aged 13 and 14, and 68 percent for those aged 15 to 17. In 2019, Meta’s Instagram briefly banned the use of “beauty filters,” or augmented reality apps that adjust selfies to make facial features better conform to conventional notions of beauty. Meta did this while it consulted 18 experts about whether beauty filters harmed adolescent girls’ self-image. The experts all answered: They sure do! Meta then ignored these findings and reinstated the filters. Zuckerberg later explained: “I think oftentimes telling people that they can’t express themselves like that is overbearing.”
Another group whose free speech Zuckerberg hesitates to curb is criminals. Last November, Jeff Horwitz of Reuters reported that an internal Meta document showed 10 percent of the company’s revenues for Facebook, Instagram, and What’s App came from ads for scams and banned goods. If Meta was at least 95 percent certain that the user was committing fraud, it barred that user. But if that certainty dipped below 95 percent, Meta punished the user by—get this—charging it a higher rate. “The idea,” Horwitz explained, “is to dissuade suspect advertisers from placing ads.” But if a fraudulent advertiser consented to pay the penalty, Meta effectively received a cut of that advertiser’s ill-gotten gains. A Meta spokesperson told Horwitz the internal document was “rough and overly inclusive.”
In a December follow-up, Horwitz and Reuters’ Engen Tham reported that Meta discovered that China accounted for about one-quarter of all ads on all its platforms pitching scams and various types of illegal content. In response, the company created an anti-fraud team focused on China. But “as a result of Integrity Strategy pivot and follow-up from Zuck,” an internal 2024 document said, the China team was shut down. A Meta spokesperson told Horwitz and Tham that the China team was always meant to be temporary, and that Zuckerberg’s directive to reduce fraud monitoring applied “all across the globe,” not just to China.
Meanwhile, over at X, Musk’s artificial intelligence chatbot Grok has opened the floodgates to sexualized deepfake images. Three million such images appeared on X over a recent 11-day period, according to the nonprofit Center for Countering Digital Hate, including 23,000 children. This coincided with record-breaking levels of audience engagement with X. Under public pressure, Musk disabled the feature for X but allowed it to remain for Grok posts elsewhere.
Tech’s enemy here is friction. Any impulse to play the responsible corporate citizen must be weighed against risk that a fix will make the product kludgy. Thus Apple, according to a lawsuit filed in February by the state of West Virginia, prioritizes privacy on its iCloud to so great a degree that it creates a haven for child pornographers who go undetected by the company. To address this problem, Apple created a detection tool in 2001. But the company withdrew it the following year after receiving criticism from privacy advocates that, according to the lawsuit, were “alarmist and unfounded.” The result, according to the lawsuit, is “a secure, frictionless avenue for the possession, protection, and distribution” of child porn.
In response, Apple said in a written statement: “We are innovating every day to combat ever-evolving threats and maintain the safest, most trusted platform for kids.” Apple also said that its Communication Safety feature blurs nude images. To that, West Virginia’s civil complaint answers that “users can proceed to view that content, and there is no mechanism to directly report illegal, explicit content, such as [child pornography] to Apple even with the Communication Safety feature enabled.” No court date has yet been set.
Don’t mistake tech’s hostility to friction for a sincere conviction that the customer is always right. Once a given tech company achieves sufficient market share, the quality of the customer experience tends to diminish in subtle ways. The writer Cory Doctorow calls this “enshittification.” In his 2025 book of that name, Doctorow explained that successful internet platforms start out being insanely helpful to customers. Once the customers are locked in, however, the platforms start to abuse them “to make things better for their business customers.” In the final stage, these platforms “abuse those business customers to claw back all the value for themselves.”
Here’s how Doctorow walks through this process at Amazon, which commands a stunning 40 percent of all digital retail sales and recently surpassed Walmart as the nation’s biggest retailer of any kind. Step One: Reel in customers through predatory pricing, i.e., selling below cost and discounting shipping. Step Two: Offer Prime membership. This is a “soft” lock-in because Prime customers pay shipping fees in advance. Step Three: Sell customers stuff they can’t use anyplace else. Audiobooks, movies, and most e-books and e-magazines can be viewed only on Amazon’s platform. Quit Amazon, and you lose them.
Amazon makes itself irresistible to merchants by initially paying full price for those items it sells below cost. But as merchants get more dependent on it, Amazon extracts ever more discounts from them (Step Four). These discounts attract more customers, giving Amazon even greater leverage (Step Five) to extract further discounts, and so on. Other Amazon extractions from merchants include taking a percentage of merchant sales on Prime and charging merchants for higher placement in search results.
Notice that Amazon does not put the financial squeeze directly on customers. That’s to conform to the still-predominant “consumer welfare” antitrust standard. But consumer welfare doesn’t go entirely unmolested, because in Step Six, Amazon’s various fees to merchants rise so high (45 to 51 percent, per Doctorow) that merchants adjust product prices upward, not just for Amazon but across the board. Thus Amazon, indirectly, pushes consumer prices up not just on Amazon, but also at Target, Walmart, and the corner hardware store.
What began, then, as tech empowering humanity degenerated into tech empowering predators of various kinds—and finally into corporate predation by the tech platforms themselves.
IV. From Hating to Being the Government
When did tech firms turn sociopathic? In many ways, the change was gradual as they, and especially their chief executives, acquired more money and power. “It just becomes easier and easier,” the veteran tech analyst Esther Dyson told me, “to kind of think everything good that happens to you is because you’re so smart.” As early as 2009, Thiel (who was never a liberal) declared publicly, “I no longer believe that freedom and democracy are compatible.” And in 2014, he wrote that “monopoly is … not a pathology or an exception. Monopoly is the condition of every successful business.”
“I see the 2010s as the turning point,” Wu told me. Initially, Wu writes in his 2025 book, The Age of Extraction, the web was built on public platforms equivalent to the town square—ARPANET, the National Science Foundation’s NSFNET, and private businesses tamed by antitrust actions (the phone companies and Microsoft). As newer platforms like Amazon and Google moved in, they initially presented themselves, Wu writes, “almost like corporate charities.” Amazon posted no profits for most of its first decade, and Google adopted the motto, “Don’t be evil.” But by the 2010s these and other private platforms were joining the rest of corporate America in prioritizing shareholder returns. In 2016, Google cut its tax bill by $3.7 billion by shifting most of its international profits to a Bermuda shell company. In 2018, Amazon paid no taxes at all on $11 billion in profits. That same year, “Don’t be evil” quietly moved from the top of Google’s code of conduct to the bottom.
Another change was the extraordinary power a handful of tech firms acquired. “Amazon determines how people shop,” The New York Times’ David Streitfeld observed back in 2017, “Google how they acquire knowledge, Facebook how they communicate.” That’s even truer today. In her 2024 book, The Tech Coup, Marietje Schaake, a former member of the European Parliament, calls this “de-facto governing power.” Tech firms have amassed actual governing power, too. Since 2008, Palantir has received $3.7 billion in government contracts; Microsoft, $5.8 billion; and Amazon Web Services (i.e., the cloud), $798 million—all primarily from the Defense Department. While Musk was laying waste to the federal bureaucracy in the winter of 2025, The Washington Post reported that his business empire was built on government contracts, loans, subsidies, and tax credits totaling $38 billion. Tech libertarians, it turns out, love to suck hard on the government teat.
Palantir, of which Thiel is co-founder and board chair, is the most obviously sinister of these firms, because, among other things, it supplies surveillance technology to Immigration and Customs Enforcement, which, according to one contract award, provides “increased efficiency in deportation logistics, minimizing time and resource expenditure.” A 2020 Amnesty International report concluded that “there is a high risk that Palantir is contributing to serious human rights violations of migrants and asylum-seekers by the U.S. government.” To this a Palantir representative replied, “We will not allow our software to be used for immoral or illegal purposes.” But at a February videoconference with shareholders, Palantir’s T-shirted, wild-haired chief executive Alex Karp could scarcely contain his glee as he said, “Palantir is here to disrupt and make the institutions we partner with the very best in the world, and, when it’s necessary, to scare enemies and, on occasion, kill them” (italics mine). By early March, Palantir was doing just that, picking bombing targets in the Iran war and seeing its stock climb 15 percent.
Tech’s final descent into unambiguous villainy was the result of three events during Joe Biden’s presidency: Lina Khan’s appointment as Federal Trade Commission chair in 2021; the advent of ChatGPT in 2022; and the election of President Donald Trump in 2024. Amazon and Meta lobbied against Khan’s nomination because she sought to reinvigorate antitrust enforcement, and after she was confirmed, both companies sought unsuccessfully for Khan to be recused from cases concerning them. ChatGPT’s introduction in November 2022 set off the arms race among Google, Meta, Microsoft, and other tech companies that resulted in tech throwing $670 billion this year at AI. And Trump, three days after his second inauguration, issued an executive order reversing what he later called “my predecessor’s attempt to paralyze this industry.” Trump also eased up on antitrust enforcement and within a year drove away the Justice Department’s antitrust chief, Gail Slater, not long after Hewlett-Packard and Juniper Networks hired two Trump allies to go over her head and settle an antitrust lawsuit on favorable terms.
At the moment, the United States has no federal law or regulation that broadly governs the development or application of AI. That gives the lead to the European Union, whose Parliament passed a tech-specific antitrust law in 2022 and an AI law in 2024 that prohibits, among other things, subliminal or manipulative distortion of human behavior. Washington’s main focus since Trump’s election has been to preempt state laws governing AI. The House-passed version of Trump’s “Big, Beautiful” reconciliation bill last year originally contained a 10-year moratorium on state regulation of AI, but that was dropped in the final bill. Instead, Trump plans to sue states that regulate AI. The rationale for preemption is that AI would be better governed at the national level, which is true. But there’s no reason to believe a Republican Congress will pass any meaningful AI regulation, or that Trump will sign any while he is in office.
There’s no end to potential AI practices that government should limit or prohibit. Your privacy should be protected (unless you’re a pedophile). You should always have access to a human alternative when a conversation with a chatbot goes blooey. You should always have a deep pocket to sue when AI harms you physically. Speaking of physical harm, extreme restrictions should be placed on “hyperwar,” wherein missile strikes and whatnot occur without human intervention (“kill them”). When Anthropic, the rare AI company that’s making some effort to behave ethically, tried to restrict how the Pentagon used its technology, it got barred from all government contracts on the dubious grounds that it posed a “supply-chain risk,” a designation it’s now fighting in court. The unavoidable lesson is that corporate responsibility is a poor substitute for government regulation.
Then there’s the question of job loss. White-collar America is waking up to its potential disposability, much as industrial workers started doing five decades ago. Back then the knowledge class shrugged and said: “Shit happens. Retrain.” Now brainworkers will take obsolescence more personally. Indeed, it’s happening already: The payroll company ADP reports demand for human software developers peaked in 2019; it’s been declining ever since. Financial analysts at Citrini Research call this “the human intelligence displacement spiral.”
As that spiral becomes more conspicuous, legal protections limiting job loss that economists once judged ridiculous will get a second look. Let’s hope any that we adopt are distributed broadly to jobs at all salary levels. Apart from the ethical imperative, doing so might drag the Democrats’ college grads back into a long-overdue rapprochement with the working class.
But to succeed, regulating AI will require standing up to a class of plutocrats more fanatically opposed to public accountability than any in history. The robber barons of the Gilded Age have gone down in history as the epitome of private avarice, but at least they believed in democracy (albeit as something to buy or sell). The tech lords, who match the robber barons’ greed, are weakly committed to democracy at best—and at worst, they’re millenarian nutcases who would dispense with government altogether. Suggest we slow the march to Singularity, and they’ll peg you as a literal or figurative devil. They’ve invested too much cash in their digital Second Coming to think otherwise.
Taming the tech lords won’t be a battle on the scale of Armageddon. But the stakes will surely be higher than we’re able right now to know. Democrats, and indeed all humankind, should prepare for a long and bitter fight, because this enemy is at least as crazy as it is rich—and it’s really, really rich.
