Copyright New York Magazine
About three years ago, writer Cory Doctorow coined the term “enshittification” to describe a widely observed yet hazily defined phenomenon: the general worsening of online platforms in recent years. Google search declining in quality? Amazon overflowing with junky products? Instagram an endless scroll of doom? These are symptomatic of enshittification, in Doctorow’s telling. He describes enshittification-cursed products as possessing three distinct stages: good to users, good to business customers, and finally, good to shareholders — and thus increasingly unpleasant for everyone else. Doctorow, a longtime denizen of the internet who used to run the pioneering Boing Boing and is a special advisor to the Electronic Frontier Foundation — among many other ventures — hit a nerve. The term took off, and its application spread far beyond the tech world. Now Doctorow is out with a book-length treatise on the subject, Enshittification: Why Everything Suddenly Got Worse and What to Do About It. I spoke with him about whether the internet decay he describes was inevitable, how AI fits into his thesis, and Donald Trump’s unlikely role in a healthier online future. I’m going to start with something you wrote in the conclusion of your book: “The indignities of harassment scams, disinformation surveillance, wage theft extraction, and rent seeking have always been with us. But they were a minor side show on the old, good internet, and they are the everything and all of the enshitternet.” You also write that many of the same people who presided over that golden age of the internet are the people now in charge of the worse, modern internet — it’s just that the regulations aren’t there anymore to prevent their greed from overflowing. Do you think there’s any other way this could have gone? If different people had built this tech, if different laws had been passed in 1998 or whenever — would we be living in a completely different kind of world, or at least a different kind of internet? I absolutely think so. I don’t think it’s a matter of different people building it. I mean, it’d be nice if we had people better than Mark Zuckerberg. We can all agree on that. Really, my purpose with this book, in part, is to recover that lost history of the policy decisions made in living memory by named individuals, which had the extremely predictable outcome of producing this very enshittified internet. They were warned at the time that this was a likely consequence. They did it anyway. I think it’s worth remembering that this is not an accident; this is stuff that was completely foreseeable. If you reward people for doing bad things, you shouldn’t be surprised that they then do bad things. And if Congress doesn’t pass a consumer privacy law since 1988, we shouldn’t be surprised that people are invading our privacy. It’s quite amazing, really — in the years since the Video Privacy Protection Act of 1988, one of the biggest changes to our world is the rise and rise of working encryption, despite the fact that governments then and now have tried to stop it. Working encryption is kind of a miracle. We have less privacy than we used to, even though we now have a technology that allows us to scramble our messages so thoroughly that if you turned every hydrogen atom in the universe into a computer and asked it to do nothing until the end of the universe but guess the password, you would run out of universe a long time before you ran out of possible passwords. So the fact that we don’t have any privacy right now — it’s not technological. The technology cuts in favor of privacy. This is just entirely down to Congress not doing its job. I grew up in the nineties, when you might have one or two computers in the house that could connect to the internet. It was a very different world in terms of how we thought about being online. So I wonder how much you think this predicament we find ourselves in is just the product of the internet seeping into every aspect of our lives and a product of the bigness of the big tech companies, which are now so integral to all of it? I still think that’s kind of backward. If you look back to the debates over the Digital Millenium Copyright Act of 1998 — that’s the law that bans reverse engineering. And at the time, it applied to things like making sure you didn’t de-regionalize your DVD player, or that no one could sell you a Sega game unless it was pressed in their official factory where you have to pay a royalty to press the CD. These things were really quite petty. And yet at the time, if you read the briefs of people like Pam Samuelson, arguably America’s greatest copyright scholar, who’s on the Electronic Frontier Foundation’s board, and who organized mass resistance to this — she briefed exactly the consequences that we’re talking about now. Like, there will be digital computers in pacemakers and in cars and in our thermostats. And the idea that we’re going to say “it’s against the law to change how something works, if the manufacturer would rather that you didn’t” — what we’re saying is that everything is going to be off limits to our changing it, improving it, protecting ourselves from abuse over it. Those were the discussions we were having back then. They were very common. I’m a science-fiction writer, and the extremely plausible thing that science fiction writers were writing about back then was the idea that we’d have computers everywhere and they’d be in our pockets and in our bodies, and that we would put our bodies inside of computers and that an airplane would become a computer in a very fancy case. So none of this was unforeseeable. I really do think it’s a mistake to say “Well, we couldn’t have anticipated this, or it couldn’t have been better.” We don’t have a counterfactual — we don’t have a world where we didn’t allow for monopolization and didn’t disarm people who wanted to change how computers work, and gave absolute discretion to people who wanted to alter how the computers that they’d sold us worked even after we bought them, and so on. But even though we don’t have that counterfactual, I think there’s a very plausible case to be made that if we had allowed people to decide how the technology they used worked, that — given there’s a bunch of people are not very happy with how the technology they’re using today work — maybe some of them would’ve changed it. When we do have technologies people are allowed to change, they do it. Take Apple’s Facebook blocking. Apple gave everyone a little tick box, and if you click it, Facebook can’t find you anymore. And 96 percent of iOS users click the box. I think there’s a pretty good case that if there was a box like that for everything, we’d all take it. I assume the four percent who didn’t take it were Facebook employees, or drunk, or drunk Facebook employees — which makes a lot of sense to me, because I would be drunk all the time if I worked for Facebook. You could also say that anyone would tick a box that eliminated ads for everything, period. Which might be a problem for a lot of business models. We should think about these things as an equilibrium derived from a series of moves and counter moves. The way the market works isn’t that someone shows up and just takes all your money and gives you what they think you deserve. There has to be some source of dynamic pushback, right? If you believe in markets and price discovery and so on, there has to be a way to make a counter offer. What ad blockers do is they take everyone who is considering making their ads more invasive and more obnoxious and make them grapple with the possibility that users that see those ads, rather than tolerating them, will install an ad blocker and never generate ad revenue for them again. And that makes them think hard about what kind of ads they’re willing to put in there. You argue that one of the reasons we’re in this situation is that tech workers once felt more empowered to push back against the deterioration of their companies, but that they have less leverage now. What kind of reaction have you gotten from tech workers with your original essay and this book? Are you in touch with a lot of them? Oh yeah, I hear from them all the time. Mostly what I get is emails saying, “Tell me how to start a union.” The best time for them to have started a union is back when they were very powerful, but the second best time is now. There’s an old joke from Down East in Canada, that if you wanted to get there, I wouldn’t start from here. But you gotta play the ball where it lies. This is where we are. To what extent do you think AI is going to supercharge enshittification? When I swipe through my Instagram feed now, it’s like 80 percent AI slop. Certainly there’s a lot of AI slop, and I think that does make these platforms worse. I think to understand the economic foundations of AI, there’s a kind of paradox to having a monopoly. When you’re on your way up — when your company is growing — the market really loves it. It can see that growth and it can predict a future where you’ll have pricing power, and that gets priced in. The price to earnings ratio, P/E ratio of growth firms, is very, very high. If you have two comparable firms, one of which is mature and isn’t growing anymore, and one of which is growing, and they have the same turnover, and the same income per year, the one that’s still growing will trade at a massive multiple of what the one that has just reached its maturity does. And the thing is that when you are traded like that, your stock becomes highly liquid, so you can use it to buy companies and keep growing or hire people to keep growing. When Mark Zuckerberg pays someone a hundred million dollars to come work on AI for him, he doesn’t give that guy a hundred million dollars in cash. It’s all in stock. And Mark Zuckerberg can’t get dollars on demand. He has to get them from a creditor or a customer. If he tries making his own dollars in the Facebook offices, the Treasury Department will break the door down. But he can make as much stock as he wants — he just types zeros into a spreadsheet. So if you’re a mature company bidding against Facebook for growth, Facebook’s going to win every auction, because they can make more of the things they’re bidding with costlessly. They’re very afraid that when they become mature, they’re going to be valued as a mature company. They’re going to see a drop off in their share price; they’re going to see key employees who are hired with stock depart; they’re going to see an inability to start the process up again, because they can only buy companies for growth with cash and not with stock, since their stock is less liquid. It’s really dangerous for them. They don’t keep growing. They don’t just level off. They plummet. And so they’ve come up with all kinds of sweaty scams to convince you that there’s still growth left. It was pivot to video, it was NFTs, the Metaverse, shitcoin, Blockchain, Bitcoin, and now it’s AI. And AI is getting a little tired, so now it’s superintelligence. Which brings us to Instagram. So why are they putting slop in your feed? Because they want to convince the stock market that there is a giant AI market and that they’re going to take it over, so they’ll continue to be a growth stock. It’s the same reason that every button on your phone that you habitually press has been silently swapped for a button that summons an AI that you don’t want to talk to, and why it takes five swipes to get rid of it. Because someone’s KPI, their key performance indicator, which is how they get their bonus, is based on “Can you get someone to use AI for at least 10 seconds?” It’s the same reason that when you load a streaming video service like Netflix or YouTube, you have to handle the phone like it’s a photo negative. Because if any part of your body grazes any part of the phone, you’re immediately taken to a different video, and there’s no way to get back to the video you are watching. It’s because these streaming services want to convince Wall Street that they don’t have to keep investing in new content to keep you as a subscriber. They want to prove that they can recommend things to you successfully. So the KPI is to recommend a show to someone that they watch for at least 10 seconds, which means booby-trapping the entire screen. The whole thing is like one giant fat finger economy. They take away any way to go back to the thing you were watching so that you are quote unquote “accepting a recommendation” several times a day and watching it for at least 10 seconds. And they can just make the number go up. That’s what’s going on with AI. And to the extent that anyone is using AI and finding it useful, it’s also very enshittification-prone because AI is a black box by its nature. It makes a ton of errors. And distinguishing the instance in which the AI did something that was bad for you because it made a mistake and the instance in which the AI did something bad for you because that company is cheating and they periodically just scam you — it’s impossible to tell the difference. You’re saying these companies are pushing technology on people that don’t want it, and I think there’s certainly a lot of that going on. But many people also genuinely like to use AI. Millions and millions of people use ChatGPT, but even the example I gave of my Instagram feed — people are generating the AI slop, and it has hundreds of thousands of views. At what point do we blame the customers, at least to some extent, for what’s going on here? It would be amazing if people weren’t clicking and making AI by the hundreds of thousands, given that there has not been a single newspaper, website, radio program, or TV broadcast that has not led with the story of AI for at least five years. When you actually look at the amount of interaction that the median person does with AI, the number’s really low given all of that. Certainly the AI companies are not acting like we’d all like AI. They are jamming AI into our face as hard as they can. I was just talking to someone who’s in the CIO’s office at a Fortune 25 company, and Microsoft has just given them a rate card for the coming year for their Office 365 license, and it costs extra if they don’t want the AI. That’s not a company that thinks you want AI. I do feel a certain sweatiness, to use your word from before, with the way it’s being integrated into various apps and programs. I’m using a company to help me fulfill my last Kickstarter, and the person giving me tech support uses a chatbot to generate the tech support responses. And they’re gibberish. It’s so frustrating being on the other side of the email correspondence. Maybe the phrase we can put on AI’s gravestone is, “If you can’t be bothered to write it, why would I be bothered to read it?” I do use AI – it will transcribe this interview. And I sometimes use AI Overview, which I really hated when it came out. It actually does serve a purpose — it synthesizes information from multiple sources fairly well. I think if you use AI summaries for things that are low stakes, and where you can find out if it’s wrong really quickly, that’s fine. When it becomes something you trust automatically — the phrase is automation blindness. If it’s usually right, but sometimes wrong, it can be real easy to slip one past you. I’m in a lot of airports, and I used it last week to figure out where a lounge was. And if I’d gotten to the gate and the lounge wasn’t there, it wouldn’t have mattered. I also use AI sometimes for one-off file transformations, like when I needed to convert the subtitles for my Kickstarter video from one video service to another. It did a pretty good job, and it was pretty easy for me to tell if it had done a good job. That’s the kind of thing where I think it’s fine. The problem is no one is going to pay enough for any of those things to recoup a $700 billion CapEx on AI. You have a section in the book about solutions for enshittification, which involves open markets and regulation — the things you say used to be more prevalent. Does that imply that if we adopted just the right regulatory framework now, we could actually go back to the way things were in some sense? Is it really possible to put that toothpaste back in the tube? No, I don’t think it is. And I don’t want to go back, I want to go forward. I don’t want the old, good internet, because as much as I liked it, it was too hard to use, and my normie friends couldn’t use it. I like the thing that we did where we made it much easier for everyday people to use it, and lots of people got to join the party. I’m really glad about that. I’m not one of these Eternal September people, angry that normal people use the internet now. What I do think is that we can have the best of both worlds. There’s this idea that maybe the reason usable things are in walled gardens is that you need to have a walled garden to make them usable. But I think there’s another way of thinking about it, which is that they have more investment. And they have more investment because structurally, we made it possible for people to make walled gardens and then made it illegal to break down the walls. “We won’t make any technology unless it’s a walled garden” is a thing you say only if it’s illegal for people to break down the walls of your garden. And if it’s not, then you’re like “Obviously open development makes more sense.” The walled garden is more profitable only if I can ask the government to put people in jail for mucking around with my walls. It’s not like people are going to stop investing in electronic communications and e-commerce and all the rest of it. They’ll invest based on the contours of the market that are set by policy. One big counter example to big internet companies that have declined in quality is Wikipedia. Why has it worked so well compared to many other places? If I were a grad student right now, one of the research questions I think would be really interesting is these — they’re called benevolent Dictator for Life projects, these public-interest projects that were started by one person who thereafter said “I’m the final arbiter of how these things work. I’m a benevolent dictator for life. And however we structure it, I’m always gonna have a veto.” There are lots of those projects, and almost all of them are run by people who, like me, self-identify as left. And almost all of those self-identified leftists are like, “This is too important to let other people be in charge of. This can’t be a committee. There has to be a dictator.” Meanwhile, Wikipedia was started by the wonderful Jimmy Wales, a friend of mine, who is, among other things, a fire-breathing libertarian whose favorite author is Ayn Rand, and who one day woke up and said, “You know what? This is too important to be run by one captain/CEO. It needs to be a commune. It needs to be owned by its workers, and I’m going to hand over the reins of power.” In the same way that Tim Berners-Lee rolled out of bed one morning and said, “The web is too important for me to take out a patent on it. Everyone’s gonna be able to use it.” And the way Jonas Salk said, “The polio vaccine is too important.” He said that owning this vaccine would be like owning the sun,so he didn’t patent it. I’m not a “Great Man of History” guy by any stretch, but I think those people show us the downstream effect of being a real mensch when you start something, just a really solid person, and how it can create a durable culture where there’s an ethos of kindness and care. And when you combine that with a bunch of irrevocable legal and technology choices — they’re using open-source software, they’re using open-license content, anyone can clone Wikipedia if it gets enshittified — it means that anyone who’s contemplating enshittifying it has to think hard about what happens if it just gets cloned and they lose all that investment. All of that stuff — not one aspect of it, but collectively —comes together to make a set of extremely enshittification-resistant…I guess you’d call them equities. You came up with this concept three years ago. Have you seen changes since you first described it? Now that it’s a word on people’s lips, have you seen tangible results from your work? Well, I’d say the big change is not directly from this work, but it’s this work in combination with the stuff Trump is doing. When I was EFF’s European director, I worked in 31 countries. I was on the road 27, 28 days a month. I actually stopped plugging in my fridge because it cost me 10 bucks a month to keep my ice cubes frozen. And everywhere I went, I would talk to people about what kind of policy would make sense for helping them have the best possible internet for the people who lived in their country. And there were always ideas like, “We should be able to reverse engineer American tech and localize it and modify it and make it make local sense. We should be able to make our own add-ons for it to help us use it in a way that is best for our context,” and so on. But they said, “We can’t do that, of course, because the U.S. trade representative would hit us with tariffs if we did.” And so it’s quite remarkable for me — as someone who’s been adverse to the U.S. trade representative all these years, and watched as they really destroyed any hope of a good internet anywhere in the world in order to service American capitalist interests —that that office has been completely neutered by Donald Trump. It’s very weird that the source of enormous American advantage around the world is being killed by Trump. It’s wild to see the deliberate repudiation of this massive advantage that the U.S. has had in relation to the rest of the world. But you know what? I’ll take it. The rest of the world is really starting to feel urgency about the fact that they’re so reliant on American big tech, because Trump has made it clear that he views every other nation as a rival to the United States. And he’s made it clear that his tech platforms are aligned with him, and that they are an arm of American foreign policy but also an arm of Trump personally. And you have things like — the chief prosecutor of the International Criminal Court swore out a complaint against Benjamin Netanyahu over genocide, Trump denounced him, and then the next day Microsoft cut off his Outlook account and deleted everything: his contacts, his calendar, all of his working documents, his email archive, all gone at the drop of a hat. And they claimed it was because of an unspecified terms of service violation, but nobody believes that. But you’re saying this American aggression has some real upsides. At this point, the Europeans are starting to really get interested in what they’re calling EuroStack, which is clones of American tech platforms, open source and based in Europe. EuroStack is pretty cool and I’m watching them build it, and thinking, folks, you are gonna have to figure out how you get people out of U.S. stacks and into the EuroStack. No one’s going to copy and paste a million documents out of Office 365 and into a European equivalent for their government ministry. To do that, you’re gonna have to reverse engineer. You’re gonna have to scrape, you’re gonna have to jailbreak apps, you’re gonna have to do a bunch of stuff that right now you can’t do because the U.S. trade representative told you you need to have a law that banned it — otherwise they’d put tariffs on you. But if someone says, “Do what I say or I’m going to burn your house down,” and then they burn your house down anyway, and you keep doing what they told you to do, you’re kind of a sucker. You have the European Union building all this stuff, but it’s like they’re building housing for people in East Germany, except in West Germany. Sure, it’s beautiful housing, but you have to tear it down the Wall if you want people to move in. So I see just a massive opportunity on the horizon, and not just in Europe, but all over the world to start building a new, good internet outside of the U.S. One that’s multilateral, one that’s pluralistic, one that is enshittification-resistant from the start. It’s going to be accelerated by Trump, which is very weird. But as my friend says, when life gives you SARS, make sarsaparilla. This interview has been edited for length and clarity.
