Enshittification: Why Everything Suddenly Got Worse and What to Do About It

by Cory Doctorow

Facebook’s pivot was decidedly weird. Mark Zuckerberg addressed the world and said, “Look, I know I’ve spent the past decade insisting that the future would consist solely of you arguing with your racist uncle using a primitive text interface of my own devising. But I have had a revelation. It turns out that the future really will involve me converting you and everyone you love into a legless, sexless, low-polygon, heavily surveilled cartoon character in a virtual world called the Metaverse, which we ripped off from a twenty-five-year-old dystopian, satirical cyberpunk novel.”

That plan is shrouded in a lot of business-speak, but it cashes out to this: by making search results worse, Google could force us to run multiple queries before we got the information we were seeking, and make more money by showing us more ads with every search-results page. In the memos, Gomes—a two-decade veteran of the company—is palpably horrified by Raghavan’s proposal to juice search queries by making the answers to each query worse. Gomes made his bones at Google by overseeing the scale-up of Google Search to run reliably on ever-larger server arrays—in other words, Gomes played a significant role in developing Google’s two key competencies: search and scale. And yet the memos show how he is progressively sidelined by the company’s senior-most managers, who see no reason not to devalue the company’s flagship product to produce shareholder value.

It also tweaked new releases of its OS to deliberately break popular apps that competed with Microsoft products. (The spreadsheet Lotus 1-2-3 was the number one competitor for Microsoft Excel, hence Microsoft’s internal slogan “DOS isn’t done until Lotus won’t run.”)

The CEOs who do this got their MBAs at Darth Vader University, where the first lesson is “I’m altering the deal. Pray I don’t alter it any further.” It works with surprising consistency, and tech executives are so confident in the lessons of the Darth Vader MBA that they come over all affronted and hurt when their customers balk.

The giant teddy bear gambit is one of the most powerful forms of twiddling. It allows Uber to keep its algorithmic wage discrimination machine humming smoothly. As Veena Dubal has documented, the forums frequented by Uber drivers are full of posts from drivers who are certain that they are “good at Uber,” who boast of the giant salaries they bring home from driving. Dubal’s ethnographic work includes heartbreaking interviews with drivers who drive until they can’t keep their eyes open, sleep in their cars, get back on the road—and then blame themselves for the pittances they take home. They don’t understand that their indiscriminate desire to please the algorithm by taking every ride that comes their way is actually signaling that they are easy pickings and can be enticed into driving for sub-starvation wages.

What are we to make of industry claims to have made a selectively universal computer, which will only run the manufacturer’s preferred programs? It’s quite simple, really. When the Big Tech companies say, “It’s impossible to run code of your choosing on a computer we’ve sold you,” what they’re really saying is “It’s illegal to run code of your choosing on that computer.”

Doubtless, other makers of children’s gadgets have wished that they could get paid again when the gadget changed hands in the secondary market. The makers of SNOO aren’t smarter or even more evil than all those other executives. They’re just more capable—they have a device with a continuous internet connection that they can downgrade at will. They have DRM and IP law, which felonize anyone who disenshittifies SNOOs, say, by making an alternative app for it that restores all the features they’ve confiscated.

US Customs agents routinely seize and destroy shipments of Apple parts based on trademark tarnishment claims that boil down to something like this: “If you opened your phone after knowingly, deliberately getting it fixed by a third party who told you they’d be using refurb parts, and used a jeweler’s loupe to look at those parts, you would see a tiny Apple logo. Then, if the phone stopped working because the part was faulty, you would develop a negative association with our logo.”

This is why every company is so sweatily insistent that you use its app rather than its website. An app is a website wrapped in enough IP to make it a felony to install an ad blocker or any other modification that makes the product work better for you at the expense of the company’s shareholders.

This is an extraordinarily wasteful way to run a corporate recruiting system! The workers who “win” the acqui-hire game spend years working on a fake product, pulling all-nighters, and neglecting their families. The customers who love that product sink money and attention into it, not suspecting that it only exists as a way to demonstrate the team’s willingness and ability to wreck their lives in service to shipping code. The “investors” in these acqui-hires spend millions to get one successful acquisition, at very high risk.

But stock buybacks make share prices go up while decreasing the value of the company. Google was $70 billion poorer after its buyback program—nevertheless, its share price rose by 13 percent within hours of the announcement. To be clear, that announcement didn’t include any information to indicate that Google had found a way to make more money: Google hadn’t found more customers or discovered a way to charge its existing customers more. It hadn’t invented any new technology or launched any products. Instead, Google had removed $70 billion from its balance sheet, depriving itself of $70 billion that it could use to hire new smart people or fund its existing smart people to make new things that the rest of us would pay for.

Distorted economics: could go up or down?

Just as there is an entire genre of nonfiction books that have a phrase somewhere in them that reads, basically, “Midway through the production of this book, a once-in-a-century global pandemic struck, and everything changed,” there is also going to be a vast collection of books that contain something to the effect of “Midway through the production of this book, Donald Trump won a second term as president of the United States, and everything changed.” Sadly, this book is among them. Donald Trump’s election represents the ultimate triumph of enshittification in the political realm.

In 2017, a Yale Law Journal paper titled “Amazon’s Antitrust Paradox” leaped out of obscure legal circles and into the public eye. This is not normal. Law review articles are barely interesting to lawyers, let alone the broader public. What’s more, this wasn’t just any law review article; it was a law review article about antitrust, one of the dustiest, least-regarded, most abstract, and frankly most irrelevant areas of law, dominated by dull, mathematical models created by and for extremely specialized economists. The article’s author was a third-year Yale law student named Lina Khan. Four years later, Khan was the youngest-ever chair of the US Federal Trade Commission, the most powerful consumer regulator in the world. Khan is an extraordinary figure. Charismatic and quick-witted, she has a knack for cutting through the professional obfuscation that kept antitrust out of the public eye and connecting antitrust policy to bread-and-butter issues, like who you work for, what kind of health care you get, and how much you pay for eggs.

These laws also have structural separation provisions—these are rules that force companies to spin off or shut down divisions that compete with their business customers. The idea of structural separation is venerable, simple, and effective. Early US antitrust laws forced banks to spin out their investment arms, on the grounds that banks that owned companies that competed with the businesses that depended on them for loans would have an unstoppable temptation to cheat. If you own a pizzeria and the bank that loaned you the money to start your business also owns the pizzeria across the street, the bank can put you out of business anytime it wants—it can “loan” its own pizzeria enough money to sell pizzas below cost until you go out of business (at which point the bank’s pizzeria can jack up its prices). It can increase your interest rates when your loan rolls over. It can loan itself money to get through an economic downturn and deny the same loan to you.

This is what lawyers call a fact-intensive rule. It can be enforced only after extensive fact-finding, along with all that entails: comments from the platform explaining why the facts favor their case, counter-comments from you or your lawyer explaining why they’re wrong, and so on (and on and on).

For one thing, companies go under! “Smart” devices are almost universally designed without any fail-safes for this eventuality. Google bricked its first-generation Nest home automation hubs, which were often used to coordinate monitoring, access, and safety at remote weekend places, where they couldn’t be readily replaced; the automaker Fisker went under and bricked all its unsold inventory; and the med-tech company Second Sight went bust and bricked the artificial eyes its customers had wired into their optic nerves.

The only way to maintain a precise, theoretically grounded use of a term is to confine its usage to a small group of largely irrelevant insiders. Policing the use of “enshittification” is worse than a self-limiting move—it would be a self-inflicted wound.