What this covers
Ragebait, sponcon, A.I. slop — the internet of 2026 makes a lot of us nostalgic for the internet of 10 or 15 years ago.
What exactly went wrong here? How did the early promise of the internet get so twisted? And what exactly is wrong here? What kinds of policies could actually make our digital lives meaningfully better?
Cory Doctorow and Tim Wu have two different theories of the case, which I thought would be interesting to put in conversation together. Doctorow is a science fiction writer, an activist with the Electronic Frontier Foundation and the author of “Enshittification: Why Everything Suddenly Got Worse and What to Do About It.” Wu is a law professor who worked on technology policy in the Biden White House; his latest book is “The Age of Extraction: How Tech Platforms Conquered the Economy and Threaten Our Future Prosperity.”
In this conversation, we discuss their different frameworks, and how they connect to all kinds of issues that plague the modern internet: the feeling that we’re being manipulated; the deranging of our politics; the squeezing of small businesses and creators; the deluge of spam and fraud; the constant surveillance and privacy risks; the quiet rise of algorithmic pricing; and the dehumanization of work. And they lay out the policies that they think would go furthest in making all these different aspects of our digital lives better.
0:00 Intro 2:33 Sworn enemies of our tech overlords 4:14 What feels bad about the internet? 5:46 What do you say to people who think the internet is a good thing? 10:26 Defining Enshittification and Extraction 20:32 The Enshittification of Facebook 27:23 Competition and honor among thieves 32:23 Amazon and extraction 39:31 When does this become a public policy problem? 41:38 Anti-competition vs pro-scale 46:44 The decline of labor practices 1:05:32 Government and regulations (or lack thereof) 1:17:56 Solutions 1:24:53 Book recommendations
Read the full transcript here: https://www.nytimes.com/2026/01/13/opinion/ezra-klein-podcast-doctorow-wu.html
Watch more on @EzraKleinShow
Thoughts? Guest suggestions? Email us at ezrakleinshow@nytimes.com.
You can find transcripts (posted midday) and more episodes of “The Ezra Klein Show” at nytimes.com/ezra-klein-podcast. Book recommendations from all our guests are listed at https://www.nytimes.com/article/ezra-klein-show-book-recs.html
Source description (no synthesized summary yet).
Tech platforms have systematically shifted from providing user value to extracting maximum wealth through monopolistic control, enabled by regulatory capture and IP law expansion that eliminated competitive discipline; reversing this requires breaking monopolies, banning toxic business models, and restoring interoperability.
- Platforms operate in three stages: attract users with good service, lock them in, then extract value from both users and business customers while degrading service
- Four traditional market disciplines (competitors, regulators, workers, new market entry) have been systematically eliminated through consolidation, regulatory capture, and IP law weaponization
- Extraction manifests across all stakeholders—users pay with attention and data, sellers pay 50%+ of revenue, workers face algorithmic wage discrimination and surveillance
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Tim Wu argues that efficiency is good up to a point, but too much efficiency becomes corrosive to human bonds and solidarity. Just-in-time scheduling makes economic sense but destroys healthy families. The question of what kind of world you want to live in is more important than what kind of economy models work.
“There's a way in which efficiency housing is an interesting term in economics because in economics as in life you want things to be somewhat efficient. Mhm. But too much efficiency becomes truly inhuman. Uh I find this even in the the the very modest example of like personal productivity efforts, you know, it's great to have a to-do list. Mhm. If I really force myself under the scaffolding of a to-do list at all times, I I feel like I cease to be a human being and and and become a a kind of machine always just getting things done and responding to the emails and and and this is a place I I it was important to him when he said it raises the question of what kind of world do you want to live in? Because the truth is that I don't want to live in a maximally efficient world. I have other competing values. You know, a the competitive efficient market is good up to a point, and after a point, it becomes something corrosive to human bonds, human solidarity. Just-in-time scheduling makes sense from the perspective of economic efficiency, and not if you want healthy families in your society. And and and I I think being able to articulate that question of what kind of world you want to live in, not just what kind of economy works on models, I think is a is important and and often a lost political art, in my view.”
Tim Wu argues that platforms are foundational to understanding civilization because they are the core institutions that bring together buyers/sellers and speakers/listeners, and that every civilization has had them, making their structure crucial to what kind of society is built.
“This is a key concept in in my book, and I think key to understanding the economics of our time, is the importance of platforms, um, which are, you know, any space or any institution that brings together buyers or sellers, speakers or listeners. Every civilization has had platforms. I was in Rome a few weeks ago and you know, you go to the Roman Forum and there it is. It's all together. The buyers, the sellers, um they have the courts, they have where people gave their speeches. You know, they're kind of the core of every civilization. And at some level, why I wrote this book is I was I was interested in this question of what our fundamental platforms look like and how that reflects on the civilization we are building.”
Bell Labs conducted fundamental research and invented many critical technologies but never brought them to market; the internet being the most famous example of a Bell Labs invention that remained internally undeveloped until decades later.
“Bell Labs collected these you know, a lot of amazing people and then never let things come to market. The internet being probably the best example of it.”
When Apple gave Facebook users the power to opt out of spying via a checkbox, 96% of Apple users ticked the box, demonstrating that revealed preference for spying is overridden when users have actual choice and countervailing power.
“You know, when Apple gave Facebook users the power to tick a box and opt out of Facebook spying, 96% of Apple users ticked that box. So, the argument that Facebook users don't being mind being spied on, I think is blown out of the water when you actually give them a a way to express preferences.”
When browsers could legally use pop-up blockers (which were legal to create because browsers are open platforms), over 50% of users installed them—the largest consumer boycott in human history—demonstrating how legal interoperability enables users to opt out of harm.
“More than 50% of us have installed an ad blocker in our browser. Doc Searls calls it the largest consumer boycott in human history. And as a result, there is some moderation upon the invasiveness of what a browser does to you that is in marked contrast with apps because reverse engineering an app because it's not an open platform is illegal under American copyright law, violates section 1201 of the Digital Millennium Copyright Act.”
Giant tech platforms are increasingly spammed with ads and sponsored results, boost content designed to keep users hooked and angry, isolate and divide society, derange politics, and concentrate wealth among billionaires while relying on low-wage workers.
“What to do about these giant platforms increasingly spammed up with ads and sponsored results boosting content that will keep us hooked and angry, isolating and dividing us and deranging our politics and making a few billionaires ever richer, held up by an army of low-wage workers in warehouses and on delivery bikes.”
Regulation should involve making values judgments about what kind of world we want, using metaphors like a gardener pruning species: some decisions prevent blight (monopoly prevention), some are aesthetic (native vs invasive species), some are for native/small business health. Not all regulatory decisions can or should be neutral.
“Tim, one thing I was thinking about while I was reading your book was the metaphor you use of a gardener. That the way to think about economic regulation and antitrust and a bunch of the different buckets of solutions we're talking about is that it is like a gardener who is trying to prune certain species and and and plants from taking over their their garden. And the gardener has to make judgments and you know, the there are some decisions you make as a gardener where you don't want blight getting all of your garden and killing everything. But others are made for aesthetic reasons and others are made because you want to have native species and not invasive species.”
Doctorow describes 'shitty technology adoption curve': harmful technologies start with people without social power (prisoners, asylum inmates, refugees) then work up to children, high school kids, blue-collar workers, pink-collar workers, and finally white-collar workers and middle-class people.
“I have this notion I call the shitty technology adoption curve, right? If you've got a really terrible idea to that involves technology that's incredibly harmful to the people it's imposed on, you can't start with me. I'm a mouthy white middle-class guy with a megaphone and when I get angry other people find out about it. You have to find people without social power. And you grind down the rough edges on their bodies. You start with prisoners, you start with people in mental asylums, you start with refugee and then you work your way up to kids and then high school kids, blue-collar workers and pink-collar workers and then white-collar workers and it starts with like the only people who eat dinner under a CCTV are in supermax and 20 years later it's like, 'No, you were just dumb enough to buy a home camera from like Apple or Google or God help us all Facebook, right?'”
The healthcare system should care that nurses are underpaid through algorithmic discrimination because underpaid, exhausted, and skipping meals to make rent means nurses are inserting catheters while already at the physical/mental breaking point, making everyone worse off.
“And as they accumulate more consumer debt, their wage is continuously eroded. And I think we can all understand like intuitively why this is unfair and why as a nurse you might not want it, but also like, do you really want your catheter inserted by someone who drove Uber till midnight the night before and skipped breakfast this morning so they could make rent? This is the thing that makes everyone except one parochial interest worse off, and this is not a free-floating economic proposition.”
Doctorow disputes pure nostalgia, arguing instead that the question is whether technological systems that gave users control necessarily had to evolve into systems where computers have control over users, or whether there were alternative paths that were foreclosed.
“I'd problematize that? So um I I do think that like so we were both also science fiction readers back then and so I was pretty alive to the dystopian possibilities of computers at the time. So I wouldn't call myself optimistic. I would call myself hopeful and excited but not purely optimistic. And I'd also like to say that like Pax uh John Hodgman, nostalgia is a toxic impulse. And when I when I think about what I like about those days, it's not that I want to recover those days. It's more that I kind of dispute that the only thing an era in which people had lots of control over their computers could have turned into is one in which the computers had lots of control over them. That there is probably something else that we could have done.”
Tim Wu argues that even if some consumers lightly use the internet and find it useful, social media has had tough and increasingly negative effects on children over the last 15 years, which cannot be denied.
“Well, the the middle age thing was it used to be better, which is I'm I don't want to fall into that sort of situation. I just think the deal is not what it could be, and I think that, you know, maybe as a consumer who sort of lightly uses this, the internet is is still useful. But, if people I mean, I I have children, too, and you know, I think it's hard to deny that social media, um, has been tough on kids and has had all kinds of negative effects on that. And that really started accelerating over the last 15 years or so.”
Search results across Google, Amazon, and Spotify have become so filled with sponsored content that the speaker can no longer trust them as signals of quality, and this whole category of once-valuable search/ranking services has been 'shitified' through monetization.
“That when I go around the internet now, when I play something in a Spotify playlist or click on a song I like and move to the radio version of Spotify, or when I search something on Google or when I search something on Amazon, these used to be very valuable services to me. To search for something on Amazon and see rankings weighted by how popular the product is, how high the reviews are, right? That like I took the the waiting of the search as to some degree a signal of quality. Certainly Google the whole idea was that what comes first in search was, you know, built on page rank and it was going to be quality. And now there is so much sponsored content in every one of these results and it is so unclear what is what and who is paying for what and why I'm getting this song or that result that this whole industry, uh, or part of the industry that, you know, one reason I ended up on these platforms is because I trusted these results. And now I trust nothing.”
Joseph Schumpeter argued in 1911 that entrepreneurs were essential to economic growth and were unusual people willing to take risks. The US economy has thrived on having lots of such individuals who can start things, but the shift to centralized tech innovation has resembled AT&T in the 60s or IBM where innovation became centralized and group-driven.
“Joseph Schumpeter back in 1911 wrote a book about entrepreneurs basically. And he said, you know, these very unusual people who are willing to go out and start a, you know, take these kind of risks, they have some vision, they do this kind of thing. And uh he thought they were essential to economic growth, that they were these kind of unusual, almost like superheroes, and would do they do these things and go go and take these chances. The the United States economy in general has thrived because it has a lot of those kind of individuals and they can start things and I think we've aired too far in having all the brains under one roof. And you know, it's starting to remind me of of AT&T in in the 60s or IBM where they they sort of became much more centralized about innovation and big ideas would never be developed.”
Google has suffered from curse of bigness: almost all successful consumer-facing products failed or were shut down. Almost all their successes (video, ads, documents, maps, etc.) were acquired from others and operationalized rather than invented internally.
“When you look at Google, they've had like one really successful consumer-facing product launch, and that was in the previous millennium. And almost everything they made in this millennium failed, right? They they did not It either didn't launch or when after it launched they shut it down. Whereas their giant successes, um their video stack, their ad tech stack, documents, [snorts] collaboration, maps, um their uh navigation, server management, all of this stuff These are mobile, right? These are companies they acquired from someone else in operationalized.”
In 2022, two teenagers reverse-engineered Instagram and created OG app, which logged in as users, grabbed their Instagram feed, discarded ads and suggestions, and showed only chronological feed of followed users. Meta sent cease letters to Apple and Google, who removed the app due to DMCA prohibition on reverse engineering.
“In 2022 two teenagers reverse engineered Instagram and they made an app called OG app. And the way OG app worked is you gave it your login and password, it pretended to be you and logged into Instagram, it grabbed the session key, it grabbed everything in your Instagram feed, it discarded the ads, it discarded the suggestions, it discarded all of the stuff that wasn't a chronological feed of the people who followed you the people you followed rather that they had posted recently. Facebook or Meta sent a letter to Apple and Google who obliged them by removing the app because there is honor among thieves.”
Tim Wu experiences the current internet as a tool he cannot trust because it serves two masters—his interests and the platform's interests—making him unable to predict what he will get or how much time he will lose.
“I feel it's like a tool I cannot trust. You know, I feel like the tools I like in my life, like a hammer, you know, I swing it, it does something predictable. The internet seems like it's serving two masters. You know, I search for something, I get a bunch of stuff I don't really want and I don't really know what I'm getting. Um I go I want to write like one email or check one thing, I end up in some strange rabbit hole and reading and like 3 hours go by and I don't know what happened.”
OpenAI's threat to Google's business model was what motivated Google to release LLMs publicly; without this competitive threat, Google might never have commercialized transformers and LLMs.
“I think you have to take out of the picture the easiest ways of uh tamping down or eliminating challenges to your monopoly. I think that's been a really important thing in US tech uh since AT&T, since IBM, since Microsoft of keeping the main dominant market players uh insecure and force them to uh im- to succeed, to improve themselves as opposed to just buying off the competitors or excluding them. So, that's my third. [pause] I think you have to give Google some credit on LLMs specifically. They you were talking about transistors a a minute ago, but Google does the fundamental research in transformers and releases it publicly and and and creates in many ways the industry. But doesn't do anything with it internally until there's a competitor that threatens them.”
Within large companies, divisions can still 'bid against each other' but this is not a meaningful substitute for external competition. Stephen King's analogy: 'me and my wife promising to both bid against each other on the next house we move into' satirizes the claim that internal competition replaces external competition.
“They said, 'Oh, well, we'll still internally bid against one another with within our divisions for uh the most premium uh material and that will be exposed to discipline that way. And I love what Stephen King had to say about this when he testified. He said, 'That's like me and my wife promising to both bid against each other on the next house we move into.'”
Facebook's lock-in for users came from collective action problem: because users' friends were scattered across the network and users had rare disease communities, emigrant communities, and family connections on Facebook, they faced high switching costs despite the platform degrading.
“It's just the collective action problem, right? You love your friends, but they're a pain in the ass, and if the six people in your group chat can't agree on what bar to go to this Friday, you're never going to agree on when it's time to leave Facebook or where to go next, especially if some of you were there because that's where the people with the same rare diseases you were hanging out, and if some of you were there because uh that's where the people in the country you emigrated from are hanging out, and some of you were there cuz that's where your customers or your audiences or just that's how you organize the carpool for the kids' little league.”
A neoclassical economist might argue that algorithmic price discrimination is efficient because it lets firms charge Ezra a higher price if he values kombucha higher and charge someone else lower based on their valuation, finding market-clearing prices and achieving economic utopia.
“And I could imagine a a neoclassical economist sitting in my seat right now and saying, pricing becomes more efficient when it discriminates. That either market will be more efficient if it can charge, you know, Ezra a higher price for kombucha, if I'm getting that delivered, because of things it knows about me and my kombucha habits, and it charges somebody else a lower price because it knows they value the kombucha less, or a nurse a higher price and or a higher wage and a lower wage depending on their situation that in fact we're just getting better and better and better at finding the market clearing price and this is what economics always wanted, right? We're we're we're finally hitting the utopia of every person having, you know, the market clearing wage and the market clearing price.”
Only 7% of time spent on Instagram is spent on content from people you follow, with the remainder being algorithmic suggestions, representing the shift from showing users what they asked for to showing what Facebook calculates will keep them engaged longer.
“There's just the FTC versus Meta case. Tim was of course involved in that. And one of the statistics that came out during it is that only 7% of time spent on Instagram is spent on things your friends and family have actually shown you, things people you follow are showing you.”
The Netflix/Time Warner merger would violate antitrust guidelines because the guidelines require governments to block mergers between the number one and number three/four companies in a market.
“this is a place where if the antitrust laws are enforced uh correctly and fairly that um the merger uh the acquisition would be blocked. And I I'd say that this is not a particularly exotic situation in the sense that you have the number one, you know, premium streaming company wanting to buy the number three or number four. And if you do the numbers under the guidelines which the government issues to tell people when their mergers are uh presumably illegal, the result is that this is a presumably illegal merger.”
Utility regulation works not by dictating outcomes but by preventing certain harmful actions—electric utilities don't decide to take half a new invention's profits or favor certain manufacturers' toasters over others.
“the electric networks electric utility regulators are not perfect. On the other hand, if you think about the electric network, it has been an extraordinary foundation for people to build stuff on. You know, and the reason they're able to build on it is they don't think the electric network is going to take like half their profits if you invent the computer on top of it. Or they don't think that, for example, the electric network is going to decide that, you know, it likes Samsung toasters instead of like LGE, I don't know, whoever else's toasters, Zenith Zenith, something like that. So, they they don't discriminate between manufacturers on the electric network.”
Andrea Downing discovered that anyone could enumerate the full membership of any Facebook group whether or not they were a member—a serious privacy violation for a breast cancer gene previvor group making medical decisions—and Facebook refused to fix it, calling it 'a feature, not a bug.'
“Andrea discovered that you could enumerate the full membership of any Facebook group whether or not you were a member of it. This is hugely important to her friends there. She reported it to Facebook. Facebook said, 'That's a feature, not a bug. We're going to won't fix it. We're going to keep it.'”
Timothy Crook observed that maybe the problem is not breaking up big platforms, but that the entire attentional competition market is cursed by toxic business models (manipulation, addiction). Breaking up platforms won't fix this if smaller platforms use the same toxic models.
“But but I think this is a little [clears throat] bit of a narrowed example. As somebody who gets a huge number of press releases for all these pro-social [28:58] apps that are built to compete with Instagram and TikTok and all of them, apps that are meant to respect your attention, apps that are meant to be virtuous in a way these apps are not, and watches one after another after another after another basically go nowhere, get out-competed. [pause] But maybe the problem is that the entire thing that the companies are trying to do, whether there are two of them or 50 of them, is negative. Yeah, it's it's a really good point and a good question. You know, I think in the markets you're talking about, we have a serious failure to wall off, discourage, ban, um or ethically consider wrongful the most toxic for- ways of making money.”
Tim Wu draws a parallel to electricity networks: we accept that electricity is essential but still regulate it. We don't 'just put up with whatever choices they charge,' suggesting Amazon should be regulated as essential infrastructure to prevent arbitrary fee extraction.
“I would compare it to electricity. I mean, we'd all say electricity network is great. Um, we can't do without they provide this incredible service. But we really say, okay, we're just going to put up with whatever choices they charge. I don't think we would. And I think at some level once a market has settled at some point you got to call a limit.”
Cory Doctorow argues that even if people enjoy aspects of platforms like Facebook, they should not accept the premise that there is no alternative to having Mark Zuckerberg listen in on conversations; instead, the goal is to preserve good experiences while eliminating the surveillance.
“That if you're enjoying having a conversation on Facebook with your friends, which I stipulate lots of people do. I think that's absolutely the case and we should value and celebrate that, that you just have to accept that there is no way to have a conversation with your friends that Mark Zuckerberg isn't listening in on. And that to ask for otherwise than that would be like asking for water that's not wet. It's just like not possible. And what I'm militating for is not like you don't like that thing you like. It's like I like that you like the thing you like. I want to make it good and I also want to guard it against getting worse because just because it hasn't happened to you yet, uh it would be uh um I think naive to think that it would never come for you.”
Tim Wu argues that American business has increasingly moved to focus its efforts on finding points of extraction as a business model, as opposed to improving the product or lowering the price, seeking to find pain points where customers have no choice and taking as much as possible.
“And I think the idea of it comes from a sense, something I get from teaching at business school sometime, is that American business has, in my view, moved increasingly to focus its efforts on trying to find points of extraction as a business model, um, as opposed to say improving the product or lowering the price. You know, try to find the pain points where your customers really have no choice and then uh, take as much you can. Kind of like in a poker game when you go all in cuz you got the good hand.”
The question of when to move from 'this is their business model' to 'this is a public policy problem' requires recognizing when a market has settled and firms are running themselves as unregulated monopolies, which historically has been addressed in various markets but not tech.
“So, when does this move from this is just their business model and if you want to find something else, I go, yeah, go buy something on Walmart. Go buy something on Target. Go buy something on Best Buy. You can do all those. I've done all those. I just ordered a blender from Kohl's. Versus we've moved to extraction and we should see it as a public policy problem. Yeah, I think that's the a really great question. It's a kind of question we faced, um, I I think repeatedly in history when you start to have a business model start to settle down. You see less, um, real disruptive, uh, competition, um, possible.”
Tech has an exceptional property: any program adverse to user interests can be neutralized with a program beneficial to user interests, allowing new market entrants to exploit deliberately bad programs. However, IP law has made reverse engineering and modification illegal, eliminating this competitive check.
“There were new companies that could exploit something that I think is exceptional about tech. I'm not a tech exceptionalist broadly, but I'm an exceptionalist about this which is that every program in your computer that is adverse to your interests can be neutralized with a program that is beneficial to your interests and that means that when you create a program that is deliberately bad, you invite new market entrants to make one that's good, right? If you lock up the printer so it won't take generic ink, you just invite someone to not only get into the generic ink business, but get into the alternative printer firmware business, which eventually could just be the I'm going to sell you your next printer business. But what we've done over 20 plus years is monotonically expand IP law until we've made most forms of reverse engineering and modification without manufacturer permission illegal, a felony.”
Extraction is a technical economic term referring to the ability of an entity with market power to take wealth or resources far in excess of the value of the good being provided, and not only the value being provided but also its cost to provide it.
“So, extraction is actually a technical economic term that refers the ability of of any entity or any firm with market power, monopoly power, to take um wealth or uh other resources far in excess of the value of the good being provided. And not not only the value being provided, but also its its cost to provide it. That's the technical definition.”
A bright-line rule that platforms must allow users to leave but continue receiving messages from people they want to hear from would address hate speech problems by enabling exit, rather than trying to define and police hate speech (which happens 100+ times per minute).
“I think that um bright-line rules make a lot of sense, particularly where you have questions that have to be frequently adjudicated. The thing we really want to be asking before we ask any of these other questions is how often are you going to have to answer this question? So, lots of people are like, 'Oh, we should just ban hate speech and harassment on platforms.' Well, that's hard because not because we shouldn't do it, but because agreeing what hate speech is, agreeing whether a given act is hate speech, agreeing whether the platform took sufficient technical countermeasures to prevent it, is the kind of you might spend 5 years on and hate speech happens 100 times a minute on platforms. Meanwhile, if we said we're going to have a bright-line rule that platforms must allow people to leave, but continue to communicate with the people they want to hear from, then people who are subjected to hate speech, who are currently there because the only thing worse than being a member of a disfavored and abused minority is being a member of a disfavored abused minority who is isolated from your community, those people could leave and go somewhere else”
Acquihire is when a company purchases a startup not for its product but because the team has proven they can make products, then shuts down the product. When acquihires dominate an economy, it's a sign the economy is sick and incentives need to be rejiggered.
“An acquihire is when a company is purchased not for the product it makes, but because the team who made it have proved they can make a product, and then they shut down the product, and they hire the team. And acquihires are, I think, a leading indicator of pathology in tech and investment. An acquihire is basically a post-grad project where venture capitalists sink some money into you pretending that you're going to make a product. It's a science fair demo. In the hopes that the company will buy you, and in lieu of a hiring bonus will give you stock, and in lieu of a finder's fee will give them stock. But no one's trying to actually capitalize a product or a business.”
All optimization in algorithmic systems starts with surveillance, whether trying to make kids spend more time on platforms, manipulating consumers into desperation spending, or discriminating in hiring. Surveillance is the infrastructure of extraction.
“I think one way to frame this, rather than around efficiency, is around optimization. And I think that we can understand that for a firm, the optimal arrangement is one in which they pay nothing for their inputs and charge everything for their outputs. So, optimization, things are optimal from the perspective of the firm when they can discover who is most desperate and pay them as little as possible, or who is most desperate and charge them as much as possible. But, from the perspective of the users and the suppliers, things are optimal when you get paid as much as possible and are charged as little as possible. And so much of kind of the specific neurological injury that arises from getting an economics degree is organized around never asking the question sort of optimal for whom? I mentioned before that we don't have any privacy law in this country. One of the things that a privacy law would let us do is to become unoptimizable. All optimization starts with surveillance, whether it's things like TikTok trying to entice your kids to spending more time than they want to spend there, or whether that's advertisers finding ways to follow you around and hit you up with things that you're desperate for, or whether it's discrimination in hiring or in lending, all of this stuff starts with an unregulated surveillance sector.”
For almost any endeavor there's healthy competition and toxic competition. Healthy competition in attention markets is making great movies and TV shows. Toxic competition is manipulation and addiction. The US has taken a hands-off approach that doesn't try to direct positive competition, which is a giant mistake.
“And I also think this relates to our conversation earlier about competition and when it's good and when it's bad. Because, I think for almost any endeavor there's such a thing as healthy competition and such a thing as toxic competition. You know, I think this is We were talking about the attention markets earlier. What is good healthy competition in the attention markets? It's like making really great movies. New TV shows that people love. You know, podcasts that people want to listen to. Uh toxic competition was the stuff you're talking about. Essentially, different forms of manipulation and addiction. And we've had this kind of like hands-off, we cannot try to direct things in a positive direction. I think that has been a giant mistake.”
Optimal arrangements for firms (paying nothing for inputs, charging everything for outputs) are opposite to optimal arrangements for suppliers and users (paid as much as possible, charged as little as possible); much economic training creates 'neurological injury' by not asking 'optimal for whom?'
“for a firm, the optimal arrangement is one in which they pay nothing for their inputs and charge everything for their outputs. So, optimization, things are optimal from the perspective of the firm when they can discover who is most desperate and pay them as little as possible... But, from the perspective of the users and the suppliers, things are optimal when you get paid as much as possible and are charged as little as possible. And so much of kind of the specific neurological injury that arises from getting an economics degree is organized around never asking the question sort of optimal for whom?”
Historically, platforms faced consequences when doing things bad for stakeholders in four forms: competitor pressure, regulator pressure, worker resistance (due to tech labor scarcity and employee values), and new market entry through interoperability. These forces have been systematically eliminated.
“Broadly, my thesis is that platforms used to face consequences when they did things that were bad for their stakeholders and those consequences came in in four forms. They had to worry about competitors, but we let them buy those. They had to worry about regulators, but when a sector is boiled down to a cartel, they find it very easy to agree on what they're going to do and make their preferences felt because they have a lot of money because they're not competing with one another and they capture their regulators. They had to worry about their workers cuz tech workers were in very scarce supply and were very valuable and they often really cared about their users and they could really say no, I'm not going to enshittify that thing I missed my mother's funeral to ship on time and make it stick because there was no one else to hire if they quit and they were bringing a lot of value to the firm, but of course tech workers famously thought that they were temporarily embarrassed founders and they didn't unionize. They didn't think they were workers. So when the power of scarcity evaporated, they had not replaced it with the power of solidarity and so now you have 500,000 tech layoffs in three years and tech workers can't hold the line. And then finally, there was new market entry.”
When the Justice Department sued AT&T to break them up and forced them to stay out of computing and license their patents including the transistor patent, this led to people quitting to start semiconductor firms and eventually created US semiconductors and computing industries.
“When the Justice Department did a deal sued AT&T tried to break them up. And they forced them AT&T to stay out of computing forever. And also license all of their patents including the transistor patent and all kinds of people started quitting their jobs and saying I'm going to start a semiconductor firm and there lies the origins of US semiconductors and also frankly US computing without AT&T.”
Tech workers were historically a restraining force on platform enshittification because they were scarce, valued, and often cared about users, giving them leverage to refuse to build bad things. However, they viewed themselves as temporarily embarrassed founders rather than workers and didn't unionize, so when labor scarcity ended, they lost bargaining power.
“They had to worry about their workers cuz tech workers were in very scarce supply and were very valuable and they often really cared about their users and they could really say no, I'm not going to enshittify that thing I missed my mother's funeral to ship on time and make it stick because there was no one else to hire if they quit and they were bringing a lot of value to the firm, but of course tech workers famously thought that they were temporarily embarrassed founders and they didn't unionize. They didn't think they were workers. So when the power of scarcity evaporated, they had not replaced it with the power of solidarity and so now you have 500,000 tech layoffs in three years and tech workers can't hold the line.”
Instacart was found to be charging different customers different prices based on personal information known about them, which is distinct from Uber surge pricing or Amazon sponsored results and represents a direct move toward personalized extraction pricing.
“the Chiron on on CNN was about an investigation finding that Instacart was charging many different people many different prices. And so, the price you were seeing on Instacart wasn't the price, it's your price.”
By default, we don't pass laws or ethical codes about new technologies; instead, we let a 'trump card' of newness justify anything. But in the last 15 years, people haven't gotten to vote on whether they want children surveilled, targeted with ads, or subjected to addictive technology design.
“Well, the first thing I'd say is we should be making more of these kind of decisions about what we really care about and what kind of world we want to inhabit. I mean, one of the things that I think happens is by default um we don't pass any laws or have new ethical codes. I mean, ethics does a lot of work and we just sort of allow a trump card to to new stuff because it's new. And, you know, I I get that you don't want to ban everything new that shows up, but I feel that we have over the last 15 years or so sometimes just kind of taken a position that, you know, the people don't get to vote on this. I mean, a good example is everything to do with children. Um you know, I don't think there's a lot of people uh who think it's a great thing to surveil children and, you know, uh have targeted ads for children and know you know, and try to create addictive technologies for children. Uh you know, when I worked in government we tried to pass just basic even child privacy laws. We couldn't get a vote ever.”
Community banks have individual leaders in every congressional district who lobby their representatives, making them hard to regulate. This doesn't mean monopolies are easier to regulate; rather, both monopolies and cartels can resist regulation through concentrated power.
“Uh the one place where I do disagree with you, having covered a lot of different both cartels and rabbles, lobbying Congress, I mean, it's not easy to regulate the Association of Community Banks, for instance. When you have something where there are, in every single district, like individual leaders of the district who will come and lobby their member of Congress, it's really hard. I am not saying that monopolies are good because they make it easier to to regulate. I'm just saying that it doesn't solve the problem of the government runs on money and influence.”
Amazon's biggest profit engine is now advertising (sponsored results), where sellers bid against each other for search placement. Last year sponsored results were $56 billion in revenue and looking like they'll exceed $70 billion—more profitable than Amazon Web Services.
“The thing I would focus on is their what they call advertising, which may be familiar to you as as sort of the sponsored results that you get when you're searching. So, what's going on there is that sellers are bidding against each other, bidding down their own margins to get higher up in the search results. And that little trick, that's the one weird trick, has become this extraordinary cash cow. It's more profitable than Amazon Web Services, which is sort of surprising. Last year, it was 56 billion. Just paying Amazon for higher for higher rankings in the search results was 56 billion dollars.”
Netflix and Time Warner have historically been innovative, oppositional forces that take risks on new formats (sound film, television, prestige TV). If Netflix acquires Time Warner, these two competing innovators would be combined, losing the bidding competition for new shows.
“I I think that Netflix and Time Warner have frankly over their history been some of the most innovative, interesting outlets. Um and often in an oppositional role. You know, this goes way back but like, you know, Time Warner took a chance on on sound film back in the '20s. In the '50s they took a chance on on television which people thought was, you know, useless. And then prestige television early '00s with HBO and the golden age. So, they've taken a lot of bets. Netflix has done a lot of innovative stuff, really interesting obviously. And frankly, you want to talk about good tech over the last 20 years. How about, you know, not having to wait until your show comes on. Um that's a form of efficiency I can agree with and I think it would be a tragedy to have these two companies who are often so oppositional, you know, combined into one.”
Amazon was once a bookstore with a simple idea: be bigger and we'll sell more stuff. It then opened a marketplace and invested in fulfillment and search, becoming more successful than eBay around 2010 when seller take was below 20%, allowing sellers to build sustainable businesses.
“Um Amazon is, you may remember, was once upon a time a bookstore. I do remember that, actually. It's how old I am. And you know, their basic idea was be bigger and we'll sell more stuff. At some point they opened the marketplace, um the Amazon marketplace, which was different because it was a platform. In other words, it was a place that people could come and sell their stuff. At first it was used books, then it spread into other markets and they uh realized a few things. One is that fulfillment would be very important... And at that point I would say, you know, maybe around 20 10 or something like that, was fulfilling what I think, you know, what what call the the dream of of the internet age, which is a lot of people would be able to go in this place, you know, start their thing, make a lot of money. It kind of depends how you count. But you know, somewhere between 15 to 20 20%. Of the sales, yeah. Yeah, so if you sold like $100, they take $20.”
Amazon acquired diapers.com and other potential competitors to eliminate alternatives. Diapers.com was a direct threat because it could have built scale starting with a specific category (diapers for new parents) and could have challenged Amazon.
“They bought a couple companies that were potential threats to them, diapers.com for example. It might seem ridiculous, but diapers, you know, could have been a kind of a a way in to threaten them. Why don't you tell the diapers.com story for a minute? It's a kind of a famous story in Amazon, but I think it's worth telling. So, there was a platform launched to be an alternative to Amazon, and their thought was, you know, new parents, diapers, you know, every parent needs diapers delivered quickly. So, why don't we make that the beginning? In the same way Amazon started with books. And then Amazon saw this, thought it was kind of threatening, and in the strategy of the day just bought them.”
Facebook lured users in 2006 by pitching them against MySpace with the promise of privacy: 'We'll never spy on you. We'll just show you things you asked to see.'
“So Facebook uh really its big bang is 2006. That's when they opened the platform to anyone, not just people with a .edu address from an American college. And Mark Zuckerberg needs to attract users, and his problem is that they're all using a platform called MySpace. So, he pitches those users and he says, 'Look, I know you enjoy hanging out with your friend on MySpace, but nobody should want to use a surveillance-driven social media platform. Come to Facebook and we'll never spy on you. We'll just show you the things that you asked to see.'”
In stage two of enshittification on Facebook, advertisers are told 'We lied about not spying on users; we spy on them from to appetite. Pay us and we'll target ads with exquisite fidelity.' Advertisers and publishers pile in and become locked to the platform.
“So, think about advertisers. Advertisers are told, you know, 'Do you remember we told these rubes that we weren't going to spy on them? Obviously, that was a lie. We spy on them from to appetite. Give us pennies uh and we will target ads to them with exquisite fidelity. And uh so, the advertisers pile in, publishers pile in, too. They become locked to the platform. They become very dependent on it.”
The US has no federal consumer privacy law since 1988 (video rental privacy law). All other forms of privacy invasion and consumer rights violations are lawful under federal law, meaning data brokers can legally sell anyone's credit history, debt levels, and delinquency status to anyone with a credit card.
“And the reason for that is that the US government has not passed a new federal consumer privacy law since 1988, when Ronald Reagan signed a law that made it illegal for video store clerks to disclose your VHS rental habits. Every other form of privacy invasion of your consumer rights is lawful under federal law. And so, among the things that data brokers will sell anyone who shows up with a credit card is how much credit card debt is any other person carrying and how delinquent is it.”
Amazon delivery drivers have facial recognition sensors and van sensors watching them constantly, subjected to manual and automated analytics. Amazon was docking drivers for driving with their mouth open as a distraction concern, violating dignity and denying period breaks.
“If you look at drivers for Amazon, they have all these sensors pointed at their faces, sensors studded around the van. They're not given a long enough break even to deal with things like period hygiene and so women who drive for Amazon who go into the back of the van to deal with their periods discover that that's all on camera cuz that's all being recorded. All of this stuff is subject to both manual and automated analytics and at one point Amazon was docking drivers for driving with their mouth open because that might lead to distraction while driving.”
Privacy advocates like Johnny Ryan (Irish Civil Liberties Association) and Max Schrems (NOYB) have challenged GDPR enforcement and pushed the European Commission to allow browsers to reject all tracking by default rather than making users affirmatively click each cookie dialog.
“A few people um Johnny Ryan with the Irish Civil Liberties Association, uh uh Max Schrems with NOYB, this uh Nu- None of Your Business, this uh nonprofit European nonprofit. They dragged some of those cases to Germany. More importantly, they've got the European Commission to start modifying the way the law works. So, you can just hit You can tick a box in your browser preferences, and it can come turned on by default that says, 'I don't want to be spied on.' And then, they're not allowed to ask you. So, the answer's just going to be no.”
Democratic policymakers and lawyers have fetishized 'truly neutral rules' in economic regulation, seeking rules that don't require any judgment, when the real issue is having values about how the economy should work and being willing to enforce those values.
“there is a fetish in them for truly neutral rules. That what people always seem to be looking for is a rule that you don't have to apply any judgment on. You can just say if you get over the sign everybody knows it's bad. As opposed to actually having to say we have views about how the economy should work. We have views about how our society should work.”
Tim's third policy recommendation is constant pressure on dominant tech platforms through aggressive antitrust enforcement to keep them insecure in their positions, preventing them from easily buying off or excluding competitors.
“Number three for you? Number three, I I think we need a constant You know, I'm an anti-monopoly kind of guy. Constant pressure on the main tech platforms so that they stay, I guess, insecure in their position and aren't able to easily counter new forms of competition. I think you have to take out of the picture the easiest ways of uh tamping down or eliminating challenges to your monopoly.”
This surveillance and acceleration of work removes the human element from labor: instead of cashiers having time to chat with customers (creating relationship), they're accelerated so fast they can't even make eye contact. Tools can be used either to make work humane or to eliminate worker autonomy.
“You know if you give a a cashier a cash register instead of making them add up things on the paper you could give them the surplus to talk with the customers and have a human relationship with them or you could speed them up so that you fire 9/10 of the cashiers and you take the remainder and you make them work at such an accelerated pace that they can't even make eye contact. There were things in Corey's description there in his answer there that in my view we should just make a social decision to outlaw. Like I am willing to say politically like I want to vote for the people who think you can't eyeball surveil workers.”
The government should be willing to make decisions to help people in situations where they are not strong enough to deal with something much more powerful. This requires government to have 'backbone' and help people when they're powerless or distracted or don't have energy to deal with concentrated power.
“I think for people to start to believe in government again, it has to help us in situations where we are not strong enough to deal with something much more powerful or something that has a lot more time to think about it. I mean, it's like we're playing poker against experts. You know, at some point we need to get backbone and have government on people's side. Now I'm starting to sound like a politician. But I But I But I mean it. Like people say that, but really doing it makes making, you know, helping people when they are powerless or distracted or don't have energy to deal with things.”
Algorithmic wage discrimination through credit history surveillance is the result of specific policy choices taken in living memory by named individuals who were warned at the time that this would be the likely outcome and did it anyway.
“This is not a free-floating economic proposition. This is the result of specific policy choices taken in living memory by named individuals who were warned at the time that this would be the likely outcome, and who did it anyway.”
Cory's three policy solutions are: (1) repeal section 1201 of DMCA to legalize modifying owned devices; (2) create muscular federal privacy law with private right of action; (3) interoperability mandate for social media allowing users to leave and continue receiving messages.
“One would be getting rid of this anti-circumvention law. In In America, it's section 1201 of the Digital Millennium Copyright Act, and saying that it should be legal to modify things you own to do things that are legal, and that it shouldn't be the purview of the manufacturer to stop you from doing it. Another one would be to create a muscular federal privacy right with a um with a private right of action so that uh impact litigators like the Electronic Frontier Foundation as well as aggrieved individuals could bring cases when their privacy laws were violated. And I guess the third would be an interoperability mandate specifically for social media. So it would be a rule and we've had versions of this. The Access Act was introduced um I think three times, various versions.”
Tim's first policy recommendation is to ban the worst and most toxic business models that are out there—whether exploitation of children or extreme algorithmic price discrimination—with the confidence that government can identify genuinely harmful practices.
“So I'll say three things. So first I think we need the confidence to ban the worst and most toxic business models that are out there, you know, whether it's exploitation of children, whether frankly it's it's some of this um total absolute price discrimination you're talking about which may technically already be illegal.”
Terms-and-conditions capitalism, where regulatory burden is shifted to consumers to evaluate complex privacy policies and cookie choices, has failed to prevent surveillance and instead created an unlivable regulatory burden on users.
“I call this terms and conditions capitalism, where you just move the burden onto the the the consumer. So, Europe uh has put out some very sweeping rules that have given me the opportunity to individually decide which of the 303 cookies on every website I visit might be good and might be bad. Um similarly, nobody's ever, in my view, to a first approximation, read an iOS terms and conditions update.”
GDPR failed to prevent surveillance because: (1) European federalism allows tax havens like Ireland, (2) American tech companies pretend to be Irish to keep profits in untaxable grace in the Irish Sea, (3) GDPR enforcement for cookie pop-ups starts with Irish Data Commissioner who does nothing, allowing big tech to maintain surveillance through cookie dialogs.
“Just as a factual matter about where it comes from, what it permits, what it prohibits, and why it failed, cuz I I agree it failed. So you may ask yourself, how is it that GDPR compliance consists of a bunch of cookie compliance dialogues? And the answer to that is that European federalism allows tax havens to function within the federation. One of the most notorious of those is Ireland, and almost every American tech company pretends that it's Irish so that its profits can float in a state of untaxable grace in the Irish Sea. And because of the nature of the GDPR, enforcement for uh these cookie pop-ups, which are the uh progeny of the big American tech companies, starts in Dublin with the Irish Data Commissioner, who to a first approximation does nothing.”
An economy where people are constantly charged the maximum they are willing to pay in every situation is not efficient—it's oppressive and inefficient because it misallocates resources (too much money spent on one good versus education) and incentivizes monopolies to reduce output and raise prices.
“An economy full of nothing but maximized monopoly prices where people are in a position to extract a firm that is is uh, not it's inefficient for two reasons. One is too much money gets spent on that water versus, you know, other things like maybe pursuing an education. And second is that the entity that holds that much power actually has a impulse to reduce supply, reduce output. And therefore produce less of the stuff so that they can extract a higher price.”
Tim Wu argues that we have a highly polarized political structure made worse by social media, a problem with inequality that has gotten worse and is accentuated by thin margins for independent business, and a failed promise that the internet would democratize wealth-building and opportunity.
“I have children, too, and you know, I think it's hard to deny that social media, um, has been tough on kids and has had all kinds of negative effects on that. And that really started accelerating over the last 15 years or so. But, I think we have a highly polarized political structure, which is, you know, made worse by social media. I think we have a problem with inequality, which has gotten worse and worse and is accentuated by the fact that the margins are just so thin for independent business. And I also think this vision that it would be this, um, equalizer, leveler, this this technology that made a lot of people rich, not just a few people rich, that it was, you know, more or less not easy, but reasonable and and and a lucrative thing to do to start your own business, that it would sort of change some of the challenges of inequality and class, uh, structure in the United States.”
Veena Dubal coined the term 'algorithmic wage discrimination,' describing how hospitals use apps to hire nurses as contractors, and the apps buy nurses' credit history from data brokers, then use their debt level to charge a 'desperation premium'—offering lower wages to more desperate nurses.
“I I think we could do no better than to talk about nurses. Um and I'm going to be making reference here to the work of Veena Dubal. She's a legal scholar who coined a very important term, algorithmic wage discrimination. In America, hospitals preferentially hire nurses through apps, and they do so as contractors. So, hiring contractors means that you can avoid the unionization of nurses, and when a nurse signs on to get a shift through one of these apps, the app is able to buy the nurse's credit history. And the reason for that is that the US government has not passed a new federal consumer privacy law since 1988, when Ronald Reagan signed a law that made it illegal for video store clerks to disclose your VHS rental habits. Every other form of privacy invasion of your consumer rights is lawful under federal law.”
Cory Doctorow argues that competition is only valuable if it breaks monopoly cartels into rabbles, not competition to become 'the world's most efficient human rights violator.' The concern is not whether platforms compete but whether the competition is toward harmful or beneficial business models.
“So, let me say that I don't think competition is a good end to itself and I think it is absolutely possible to compete to become the world's most efficient human rights violator. The reason I like competition is because it makes firms into a rabble instead of a cartel.”
Many pro-social apps designed to respect attention are built to compete with Instagram and TikTok but get out-competed and fail, suggesting that without being able to modify platforms themselves, alternatives cannot effectively compete.
“I get a huge number of press releases for all these pro-social apps that are built to compete with Instagram and TikTok and all of them, apps that are meant to respect your attention, apps that are meant to be virtuous in a way these apps are not, and watches one after another after another after another basically go nowhere, get out-competed.”
Producer perspective: some people, particularly young people and parents, actually find TikTok and Amazon to be valuable and functional, using them happily despite criticisms of these platforms.
“Yeah, three three uh middle-aged guys who think the internet went wrong somewhere along the way. Uh when I was working on this episode with my producer, one of the interesting tensions behind the scenes was she doesn't think the internet is bad. She thinks Tik Tok is, uh, she said, 'a perfect platform.' Um, she's young kids and feels Amazon is a godsend for a young parent.”
Tim Wu argues that neoclassical economics assumes price is a method of discovering value, so if someone pays $70,000 for a pharmaceutical, that means they value it at $70,000 even if it seems extractive, but this ignores power dynamics and necessity rather than preference.
“And I mean, a lot of economists would say price is a method of discovering if you have a pharmaceutical, people are willing to pay $70,000 for it, that means they value it at $70,000 even if you think that is extractive.”
Tim Wu's counterargument to Doctorow: platforms like Meta have actually been under competitive pressure from TikTok, stole features from TikTok and Snapchat, and improved as a result. Saying this isn't enshittification but rather the magic of competition.
“I think if you had Mark Zuckerberg arise from his uh sarcophagus. I was going to say office, because I'm a more polite person. Uh here, he would say, 'We did this under competitive pressure. TikTok was eating our lunch. We stole a bunch of things from TikTok, and now we're doing better. We also stole a bunch of things from Snapchat and now we're doing better because in fact we are under a lot of competition and we are incredibly good at responding to that competition in ways that our user base responds to. This is not enshittification. This is the magic of competition itself'”
Enshittification is a label for both an observation about a characteristic pattern of how platforms degrade and why platforms degrade now, because something has changed—not that greed was newly invented in the middle of the last decade.
“Enshittification, it's really a label I hung on both an observation about a characteristic pattern of how platforms go bad, but I think much more importantly why they're going bad now cuz we didn't invent greed in the middle of the last decade. So, something has changed. I My thesis is that some exogenous factors are changed.”
Tim Wu argues that when users of Facebook spend hours wandering around on unintended content, buy things they didn't want, and have vulnerabilities exploited, it is not free—the user is poorer both in consciousness/control and in misspent money.
“When you use Facebook, you are constantly being mined for your time, attention, and data in a way that is extraordinarily valuable and that, you know, yielded something like 67 billion in profit last year. So, you know, things that feel free, is it free when you suddenly spend, you know, hours um wandering around random things you didn't intend to? Is it free when you end up buying stuff that you didn't really want and wonder why you get it later? Is it free when you feel that you've you know, had your vulnerabilities exploited? I would say none of that's free. You're poorer both uh in your own consciousness and in terms of what attention and your control over your life and your poorer probably in misspent money.”
Amazon's Choice badge is not algorithmic selection of best products, but a marketing tool where the algorithm prioritizes how many different Amazon services a seller uses (Prime, fulfillment, advertising). This makes the first search result on average 17% more expensive than the best match for your search.
“So, you're right that it is algorithmic, but the algorithmic inputs are not grounded primarily in things like quality or customer satisfaction. They're grounded in how many different ways you've made your business dependent on Amazon in such a way that every dollar you make is having more and more of that dollar extracted by Amazon. There's some good empirical work on this from Maria Mazucato and Tim O'Reilly where they calculate that the first result on an Amazon Amazon search engine results page on average is 17% more expensive than the best match for your search.”
We are collectively paying $70 to make search worse—the extraction of $70 billion from the search ecosystem manifests as degraded search quality.
“we are kind of paying $70 collectively to make search worse.”
Just because platform harms haven't hit a particular person yet doesn't mean they should not be worried about future harms, and accepting surveillance or manipulation on platforms should not be framed as inevitable or like asking for water that isn't wet; different design is possible.
“I'd say like just because this hasn't hit you, I I think unless you have a theory about why you are favored by these platforms, then you should at least be worried that this would come and I would follow up by saying like let's not fall into the trap of vulgar Thatcherism. You know, Thatcher's motto was there there is no alternative and I think tech bosses would like you to believe that, too... I want to make it good and I also want to guard it against getting worse because just because it hasn't happened to you yet, uh it would be uh um I think naive to think that it would never come for you.”
Interoperability would allow users to leave a social media platform and continue receiving messages from people they follow on the new platform, similar to switching phone carriers, with technical details solvable and tradeoffs worth it.
“you should be able to leave a social media network and go to another one and continue to receive the messages people send to you and reply to them the same way you can leave one phone carrier and go to the other. And there's a lot of technical details about what that standard looks like and how you avoid embedding um you know parochial interests of incumbents and so on. I don't think they're insurmountable uh and I think that the tradeoffs are more than worth it.”
People will make huge effort trying to move up a category they're in or pretend to be poorer to access lower prices, so perfect price discrimination efficiency is overrated because it depends on simplified models of what makes people happy.
“And the other thing I'll say is there's also a huge amount of effort people will make trying to move up a category they're in. Um and, you know, pretend to be poor. So it it I think it is overrated. Uh uh and relies on um I I guess it relies on overly simplistic models of what makes people happy.”
Human attention is extremely valuable and commercially exploited; the question of whether every second of your time in every space should be mined for attention at maximum value, even if it contributes to GDP, is important because your life is what you pay attention to.
“And human attention turns out to be quite commercially valuable. But do you ever do you really want every second of your time and every space you inhabit to being mined for your attention and its maximum value, even if that contributes to the, I guess, overall GDP of the economy? I mean, I'd like to have some time for my kids and friends in which no one's making any money. And you know, it's an example of a commodity that is very close to who we are. Um at the end of your days, you know, what your life was was what you paid attention to. And the idea that you can, with maximum efficiency, mine that at every possible moment, seems to me a recipe for a very bad life.”
Cory experiences the current internet as inherently designed to be bad and unchallengeable, contrasting with his early internet experience where he believed problems could be fixed and he might be the one to fix them.
“So, what I would do is contrast what happens when things aren't great now with how I felt about what happened when things weren't great before. So, I think when I was a larva on the early internet and I saw things that sucked, I would think, 'Someone's going to fix this and maybe it could be me.' Uh and now, when I see bad things on the internet, I'm like, 'This is by design and it cannot be fixed because you would be violating the rules if you even tried.'”
In the last 10 years, Amazon has turned to extraction by increasing fees, changing margins so many sellers pay over 50%, making Amazon prices no lower than brick-and-mortar alternatives, and preventing sellers from pricing lower.
“In the last 10 years, Amazon's strategy has just basically been for its marketplace to to turn the screws and increase the fees, change the margins so that many sellers are paying, you know, over 50% or more, you know, basically same as as a brick-and-mortar businesses. And Amazon prices are rarely any lower. They They actually have tried done a lot to try to prevent being priced, anyone pricing lower.”
Most tech workers don't work in Silicon Valley enjoying free kombucha and egg-freezing benefits; the majority drive for Uber or work in Amazon warehouses, many in factories in China with suicide nets, representing severe labor degradation across platform-dependent work.
“Yeah, I I mean, we could we could talk about the other tech workers, right? The majority of tech workers drive uh for Uber or for Amazon or work in a warehouse, and they certainly don't get like free kombucha and massages, and a surgeon will freeze their eggs so they can work through their fertile years. They're in a factory in China with suicide nets around it.”
For 90 different privacy laws addressing various harms, Congress could not even get a vote, and national security officials said it was a problem that it's too easy to spy on everybody, showing even security hawks support privacy regulation but political system won't act.
“And um you know, that's a problem for us as a national security issue. And we just could not get a vote on even the most basic anti-surveillance, which would suggest like, "If you download a dog-walking app, it shouldn't be just like tracking you and uploading every kind of information about you, that that should be illegal."”
Some surveillance and worker control practices (like facial recognition surveillance of workers) should be directly outlawed as policy decisions because we can collectively decide what kind of world we want, and other practices (like automated checkout) are harder to judge because there are real consumer benefits alongside worker harms.
“There were things in Corey's description there in his answer there that in my view we should just make a social decision to outlaw. Like I am willing to say politically like I want to vote for the people who think you can't eyeball surveil workers... I also get it. How do you think about weighing that, right? There's the stuff that is genuinely like grim and dystopian and maybe we should just outlaw. And then there is stuff like the just generalized automation in which there genuinely can be a consumer surplus from that.”
Boss wear and worker surveillance technology like Microsoft's productivity tracking allow managers to stack rank workers by mouse movement, typos, and words typed, and even show firms their performance relative to competitors in the sector, creating perverse incentives and leaking competitive information.
“So, the category that this that this falls into it's broadly called boss wear and there's a whole lot of different versions of it. Like if your firm buys Office 365, Microsoft will offer your boss the ability to stack rank divisions within your firm by like how often they move the mouse and how many typos they make and how many words they type. And then this is amazing. They will tell you how you perform against similar firms in your sector, which is like the most amazing thing I can imagine that that Microsoft is finding customers for a sales pitch that says, "We will show you sensitive internal information about your competitors."”
When you unshackle firms from the four forces of discipline (competitors, regulators, workforce, new market entry through interoperability), the same CEOs who would normally be constrained pull the 'enshittification switch' as hard as they can, and it goes from zero to 100 because the policy environment has lubricated it to function freely.
“And as a result, you don't have to worry about market entry with this incredible slippery dynamic character of technology. And when you unshackle firms from these four forces of discipline, when they don't have to worry about competitors or regulators or their workforce or new market entry through interoperability, the same CEOs go sit to the same giant switch on the wall in the C-suite marked "enshittification" and they yank it as hard as they can as they've done every day that they've shown up for work, and instead of being gummed up, it has been lubricated by an enshittigenic policy environment that allows it to go from zero to 100 with one pull”
People hate paying junk fees because they intuitively feel it's unfair, suggesting political opposition exists to extraction even if models don't capture it. The government called them 'junk fees' but had to avoid the president saying the word publicly.
“Yeah, I I agree, and I still there's, you know, there are some intuitive feelings, like people feel it's unfair, people don't like being ripped off. But people hate paying junk fees. The The original word for that, by the way, was fees. But there was inside government we felt we couldn't have the president say that.”
The Facebook feed has been denuded of things users asked to see, left with only the minimum calculated to keep them engaged, and this brittle equilibrium is disrupted when user tolerance reaches crisis point (e.g., after mass shooting livestream), prompting pivot attempts like the metaverse.
“And we've got a feed that's been, you know, basically denuded of the things we've asked to see. It has the minimum calculated to keep us there. And this equilibrium is what Facebook wants, but it's very um brittle. Because the difference between I hate Facebook and I can't seem to stop coming here and I hate Facebook and I'm never coming back, it can be disrupted by something as simple as a live stream mass shooting.”
Tim and Cory agree that breaking up monopolies is necessary but not sufficient for good regulation; the underlying problem is government dependence on money and influence. Both monopolies and cartels can corrupt government.
“Sure. I think we can agree on necessary but not sufficient. Yeah, so we can we can do that.”
Cory Doctorow and Tim Wu went to elementary school together in suburban Toronto at a school with 80 kids in one classroom where older kids taught younger kids and they were left to pursue self-directed learning.
“So I just learned that you both went to elementary school together. Yep. Yeah, that's true. In suburban Toronto, a a weird little school with like 80 kids. It was kindergarten to eighth grade in one classroom. Older kids taught the younger kids. We more or less were left to go feral and design our own curriculum.”
The early internet felt better before Twitter and Facebook went algorithmic, around the heyday of blogging when individual bloggers could reach 40,000 hits per day.
“When was the last year that the internet felt good to you? I think everybody has different answers to this. Mine I think go fairly far back, maybe to the heyday of blogging. words and getting up to 40,000 hits on his blog a day. At least before the moment when Twitter and Facebook went algorithmic.”
There is growing consensus something went wrong with the internet somewhere along its development, but no consensus on what to do about it.
“But even if there is this growing consensus that something went wrong with the internet somewhere and that it is driving our society somewhere we don't want it to go, there's not really a consensus of what to do about it.”
Current internet (2025) is characterized by anger, outrage, and AI-generated content, representing a broken promise of what the technology was supposed to deliver.
“But whatever your answer to it is, I have not found many people who think 2026, right now this internet with all of its anger and its outrage and its AI slop, this is what we were promised.”
Both Doctorow and Wu learned to program on Apple IIs in their early education and were excited about early computing, viewing it as the origin of their later advocacy against tech monopolies.
“Well, we loved tech at the time. I mean, we had we were early in on Apple II's. Mhm. And I frankly that's where it all started in a way. You know, both of our our our books have this kind of pining for a lost age and I think some of it is this early era of computing when we were just like so excited, so optimistic.”
Cory recommends Bridget Read's Little Bosses Everywhere: history of American pyramid schemes as central to current rot. MAGA movement members are often people predated on by scams and adopting toxic positivity from that experience.
“My second choice is a book by Bridget Read. It's a book called Little Bosses Everywhere. And Little Bosses Everywhere is a history of the American pyramid scheme. And it's an argument that the American pyramid scheme is kind of the it's the center of our current rot. And everywhere you look in the MAGA movement, you find people who have been predated upon by the kinds of scams that are characteristic of this and who have adopted the kind of toxic positivity that comes with it.”
The United States has been in a better place in other periods of American history, and while the internet is not the only cause of current conditions, it is part of what has made things worse in many dimensions.
“I think we've been in a better place in other periods of American history and I think the internet's not the only cause, but I think it's part of it.”
Recommended books: 'Small Is Beautiful' by E.F. Schumacher (questioning efficiency obsession); 'Manipulation: What It Is, How to Stop It' by Cass Sunstein (understanding allowed harms); 'Rise and Fall of the Great Powers' by Paul Kennedy (understanding imperial dynamics and cycles).
“Sure, I'd start with uh E.F. Schumacher's uh Small Is Beautiful, Economics As If People Mattered. And I say that cuz it, you know, targets this question of what kind of world do we want to live in? And you know, I think our efficiency obsession uh is taking us in one direction, and I think we should choose a different direction. A second book um is more recent. Cass Sunstein wrote a book on manipulation that uh I think has been um it is underrated and is really good for understanding what we have allowed to happen. It's called Manipulation: What It Is, How to Stop It. The last book, I guess this is where I got some ideas about uh tech platforms and the big picture is from Paul Kennedy, Rise and Fall of the Great Powers.”
Cory Doctorow recommends Sarah Weinman Williams' Careless People as an example of the Streisand effect: when a company suppresses something, it brings interest. Williams was a NZ diplomat who pushed for government job with Facebook and got it, but the job became a nightmare.
“So, my first pick is is Sarah Weinman Williams' book Careless People. Um and uh it's a great example of the Streisand effect that when a a company tries to suppress something, it brings it interest. So, uh Weinman Williams, she was a minor uh diplomat in the New Zealand diplomatic core, she became quite interested in how Facebook could be a player geopolitically. She started to sort of nudge them to give her a job as like an [pause] inter-gover- or an international governmental relations person. No one was very interested in it but she just sort of kept at it until she got her dream job. And then the dream turned into a nightmare.”