The Internet Has Become Too American to Trust

TECHNOLOGY The Internet Has Become Too American to Trust Trump and Big Tech have fused into a new kind of threat BY CORY DOCTOROW Published 6:30, MAY 27, 2026 As a species, we are extremely bad at dealing with slow-motion disasters. When the roof leaks, we typically put a bucket under the drip and then forget about it until the next time it rains. And if, next time, there are two drips, well, that’s just one more drip than the last time. No need to panic. And before you know it, every time it rains, we’re setting out dozens of buckets. And it’s not until the roof caves in, and the furniture is wrecked, that we take action. Because when the roof caves in, you have to move. So, spoiler alert: this is a climate change metaphor. This inability to act until it’s too late is a curse upon our species. It’s our existential risk. But that doesn’t mean I’m happy that, every so often, some mad emperor precipitates a crisis, which, in turn, precipitates change. I’m not happy that Russian president Vladimir Putin invaded Ukraine. But I’m also not sad that the resulting energy crisis lit a fire under the European Union’s energy transition, taking the bloc from ten years behind schedule on solarization to fifteen years ahead of schedule in just a few short years. It turns out that when you’re shivering in the dark, all the reasons for inaction fade into the background. The all-powerful fossil fuel lobbyists suddenly find a lot of doors slamming in their faces in Brussels, in Strausburg, and in Berlin. And that tedious neighbour who insists the solar panel you want to hang on your balcony will “spoil the aesthetic character of the neighbourhood”—that jerk can piss off all the way into the sun. No one is going to listen to them anymore. Once again, all things being equal, the world would be better off without mad emperors and their wars of choice. But if there is a mad emperor, and there is a war of choice, at the very least we should strive to salvage what we can out of their depraved and unforgivable acts of aggression. For avoidance of doubt, this is a statement about United States president Donald Trump’s lethal bungling in the Strait of Epstein.

And speaking of Trump, he is a machine for turning slow-motion disasters into unavoidable catastrophes. A chaos agent who spends every hour that God sends caving in roofs and getting people moving on important matters that they have neglected for decades. Like the American internet. We’ve known the Americans couldn’t be trusted to run our internet for decades—at least since 2005. That was the year an unassuming guy walked into my office at the Electronic Frontier Foundation (EFF) in San Francisco’s Mission District. His name was Mark Klein, and he had an incredible story. Recently retired from being a network engineer at AT&T, Klein brought a tranche of files from his old job that documented how his bosses had ordered him to build a secret room at the company’s Folsom Street office, and then insert a beam-splitter into AT&T’s fibre backbone in order to provide the National Security Agency with access to all of AT&T’s network traffic—warrantless, illegal access to the world’s communications. Yes, the world. Because the world sends its fibre lines across the ocean to make landfall in America; companies like AT&T provide interchange between these lines, serving as the global data hub rather than requiring all 200-odd countries run direct fibre links to and from every other nation. This reduces the number of expensive transoceanic fibre cables from tens of thousands to merely hundreds. I’ve spent twenty-five years at the EFF, the world’s oldest and most important digital rights group. And we hear from a lot of people with incredible stories, and not all of them are in possession of their senses. But every now and again we get someone like Mark. Indisputably sane, with a story so paranoid it sounds like a delusion. But Mark was telling the truth. So, we sued the NSA. We brought a series of cases that dragged on for years, capturing national attention. It led to a remarkable exchange in the United States Senate where Ron Wyden, ranking member on the Senate intelligence committee, point blank asked James Clapper, the US director of national intelligence, whether the NSA was engaged in mass surveillance of the sort described in our lawsuit. And right there, in front of the Senate and the C-SPAN cameras, Clapper just straight up lied.

Now, one of the people watching the C-SPAN feed that day was a young, idealistic NSA contractor named Edward Snowden. At that moment, Snowden lost all confidence that his bosses cared about the law or the constitution. And we know what happened next. Drip, drip. Klein in 2005. Snowden in 2013. And yet, today, we are still using the American internet—the surveillance-prone, easily interdicted internet. Because Klein and Snowden were individual leaks, we put our buckets out and hoped things wouldn’t get worse. Instead of migrating off the American internet, we tried to regulate the tech giants. We pretended that the problem with Mark Zuckerberg was that he was the wrong guy to be the unelected, permanent social media czar with total control over 4 billion people’s lives, rather than confronting the fact that no one should have that job. We don’t need Zuck to learn to be better at his job. We don’t need someone else in that job. We need to abolish that job. Because the ketamine-addled Zuckermuskian mediocrities that run our digital lives aren’t smart enough to be causes. They are effects. They are products of an enshittogenic policy environment that rewards the worst ideas of the worst people with the most money. Which means that if Elon Musk popped his clogs tonight, there would be an overnight succession battle amongst twelve equally horrible Big Ballses, and whoever emerged victorious would be indistinguishable from Musk himself. That’s because our policymakers created the enshittocene, not tech executives. Tech executives didn’t plan to enshittify their services because they don’t plan anything. That’s not their style. Like Trump, their strategy is to race across a river over the backs of alligators without losing a leg. These guys A/B tested their way into the enshittification playbook. You may have encountered my theory of that playbook by now. If not, here’s a quick recap. Stage one: the platform is good to end users, while finding a way to lock those users in. Think of Facebook. In 2006, Zuckerberg decided to open Facebook to the general public, dropping the requirement that you had to sign up with an .edu address from an American college.

But he had a problem. Everyone who wanted a social media account already had one on a rival service called MySpace. So, Zuckerberg made these MySpace users a sweet pitch. “Come to Facebook,” he said, “and tell me who matters to you—who you want to follow. And every time any of those people post something for public consumption, I will show it to you in an ad-free, reverse chronological feed. You will see the things you want to see and nothing else. And best of all, I will never spy on you. Not like MySpace, whose owner is an evil, crapulent, senescent, immortal, Australian, billionaire vampire named Rupert Murdoch.” You gotta hand it to him. That was a good offer. So, we flocked to Facebook. In doing so, we also fulfilled the second half of stage one: getting locked into the platform. There are lots of ways a platform can lock you in. Amazon sells us Prime and gets us to pay for a year’s worth of shipping in advance, so we’d be a fool to use some other service. Uber pissed away $31 billion worth of Saudi royal family money, losing $0.41 on every dollar for thirteen years, creating a lost decade for transit investment while cratering competing cabs, so it’s the only game in town. Google pays Apple over $20 billion a year not to make a competing search engine. But Facebook and other social media get to play lock-in on the easiest setting. Because we lock ourselves into social media. That’s thanks to something called the collective action problem. Which is how economists describe the fact that, even though you and your besties all hate Facebook, you can’t agree on when to leave or where to go next, and because you love your friends, leaving Facebook is a costly proposition. It has a high switching cost—another economist’s term, meaning “everything you have to give up when you switch from one service to another.” So long as you love your friends more than you hate Zuckerberg, you will stick around. So that’s stage one of enshittification. Be good to users, but lock them in, so that you can move on to stage two enshittification. Making things worse for users to make life better for business customers. To the advertisers, Zuckerberg said: “Remember I told these rubes that we wouldn’t spy on them?

Total lie. We spy on them from asshole to appetite. Give us a shockingly small amount of money, and I will target ads to them with exquisite fidelity. And because I’m such a conscientious craftsperson, I’m gonna throw in a whole building full of engineers who do nothing except hunt down ad fraud, so that when you give me a dollar to stick an ad in someone’s face, that face will have your ad shoved in it.” And to the publishers, Zuckerberg said: “Hey, remember when I told these suckers I would only show them the things they asked to see? Obviously a total lie. Tell you what. You post excerpts from your web articles to Facebook, with a link back to your site, and we will cram them down the eyeballs of our users. It’s a free traffic funnel you can monetize as you see fit.” So, the publishers and advertisers piled in, and they got locked in too. The publishers and the advertisers became dependent on Facebook. Which is to say, they became dependent on us. We took them hostage. And we took each other hostage. Which means it was time for Facebook to enter stage three enshittification. When they made things worse for everyone in order to benefit their stockholders and executives. Advertising prices soared, ad targeting fidelity plunged, and ad fraud reached farcical heights. In 2017, Procter & Gamble zeroed out its budget for so-called “programmatic” ads. The ads weren’t performing the way the company thought they should, so they tried an experiment. They went from spending $200 million per year to $0 per year, and they saw a 0 percent drop in sales. Because, to a first approximation, nobody saw the ads. It was all disappearing down the fraud hole. If things were bad for advertisers, they were downright terrible for publishers who found that they had to post longer and longer excerpts to Facebook just to be shown to the people who’d actually subscribed to their feeds, let alone getting boosted into the feeds of people who didn’t follow them. Eventually, publishers found that they had to upload the full text of every story, making Facebook a substitute for their website rather than a promotional tool to drive traffic.