What the hell is “adversarial interopability”? Science fiction writer and prolific blogger Cory Doctorow thinks it’s going to set you free from Facebook, letting you take your data and pictures wherever you want to go. And he thinks surveillance capitalism is standing in your way.
Cory Doctorow, among many other sites, runs the daily blog CrapHound. Follow him on Twitter for some incisive and illuminating tweet threads.
Transcript
Ethan Zuckerman:
Hi everybody. This is Ethan Zuckerman. Welcome back to Reimagining the Internet. I am thrilled to have with us today my friend, Cory Doctorow. We could spend this entire podcast adequately introducing Cory. I will do a quick version. He is an acclaimed science fiction author, the author of Little Brother, Attack Surface, many other books. He is a prominent internet activist, works as a special advisor to the EFF, former co-editor of Boing Boing, now maintains a personal website, craphound.com, a link blog, pluralistic.net, publishes sort of amazing and remarkable Twitter essays.
Our producer Mike Sugarman has called him the most prolific poster we’ve ever had on the show. I think that’s actually kind of a low bar. He’s one of the most productive and influential writers on the web today. Right now he’s publishing the serialized ebook, How to Destroy Surveillance Capitalism on Medium and on OneZero. Cory, it is wonderful to have you here. Thanks for being with us.
Cory Doctorow:
It’s absolutely my pleasure. Thank you for having me, and for that very effusive introduction. I appreciate it.
Ethan Zuckerman:
So I want to start with this serialized ebook because How to Destroy Surveillance Capitalism is a topic near and dear to my heart. How should we destroy surveillance capitalism and can we start with Facebook?
Cory Doctorow:
Yeah. Well, I think to understand my program of the destruction of surveillance capitalism, you have to understand my critique of the idea of surveillance capitalism. So the term surveillance capitalism started with some leftist academics at the University of British Columbia who used it really as a critique of capitalism itself. But when Shoshana Zuboff popularized it first in articles in academic journals and then in a landmark, very long and interesting book with the same title, she really located the critique, not in capitalism, but in the surveillance.
And Zuboff is, it’s fair to say, a believer in capitalism and in capitalism’s ability to produce efficient allocations by aggregating all the decisions we make about what we’re willing to buy and what we’re willing to pay for them. And that somehow, through all of this, you get a kind of big, complex computation whose emergent property is a good allocation of all of our scarce resources from around the world. And her belief is that the surveillance capitalism industry, the big tech firms, have figured out how to use machine learning and surveillance, data gathering and surveillance, to bypass our rational faculties, so that instead of making choices about what we want to buy, they tell us what we’re going to buy. They tell us what we want.
And I think that there’s a plausible reason to believe this, which is that this is what the tech companies say they’re doing. When you ask why it is they should command such a giant premium for advertising, they say it’s because their advertising works really well.
Ethan Zuckerman:
But this is basically the Cambridge Analytica idea, that somehow through mastering surveillance, having all this incredible data on us, the reason these companies have such eye popping valuations is that they can make us do things that we don’t otherwise want to do. And I take it that you’re skeptical of this claim?
Cory Doctorow:
Yeah. I shorthand that hypothesis as Mark Zuckerberg built a mind control ray to sell your nephew fidget spinners. Robert Mercer stole it and used it to make your uncle a QAnon. And I think that there’s a really important logic inconsistency in swallowing this claim at face value because Zuboff points out absolutely correctly that the tech companies are extremely untrustworthy, that when you ask them which data they collect, they lie, when you ask them who they share it with, they lie, when you ask them how they use it, they lie. And yet she takes at face value this one claim, this claim that their products work really what well.
Ethan Zuckerman:
And this seems to line up a little bit with Tim Long’s book, Subprime Attention Crisis, where he basically says there’s no real proof that any of this works as well as we think it does. You could argue that this is essentially a speculative bubble where everyone has an interest in it increasing in value, but eventually this market may rationalize itself at some point.
Cory Doctorow:
Yeah. And I think that Zuboff’s blind spot comes from her trust in markets because her kind of tautological basis for assuming that these claims are true is that people wouldn’t spend all this money on advertising if it didn’t work. And you could make the same claim about say private equity, hedge funds and managed funds, more generally, right? Billionaires wouldn’t give trillions of dollars to hedge fund managers if they couldn’t outperform the market. And yet they can’t outperform the market. There are lots of instances of very sophisticated people who advertise or who give money to money managers and who don’t get anything for their money and are still convinced that they’re getting it.
And it’s not just advertising and whatnot. I mean, Steve Jobs drank smoothies to cure his cancer until he died. Being smart doesn’t mean that you are discerning in areas outside of your narrow area of expertise. And so, if you believe Zuboff’s hypothesis that the tech firms have found a way to bypass our critical faculties, then it’s kind of like there’s a meteor headed towards the earth and we have to do something to steer that meteor, to steer the comet, because we can’t tame it and if we try to break it up into smaller pieces, then all we get is like a killing reign of thousands of small comets that we could never hope to capture. And so Zuboff councils basically seizing control over these firms, perfecting Mark Zuckerberg is the way that I think of this.
The problem with Mark Zuckerberg is that the wrong person is the unelected unaccountable permanent social media Czar of three billion people’s lives.
Ethan Zuckerman:
If we only had a really good one who had a really good conscious, cared about things like mis and disinformation, then we’d be just fine. And let’s have auditions for this person to run our digital public sphere.
Cory Doctorow:
Well, and more to the point, we get the people who understand how this stuff works, and we get them to impose what, in antitrust land we would call conduct remedies on Mark Zuckerberg or his successor, so we make him do what he wants. The problem with this is the problem I think with market based analyses of political problems more generally, is that they leave out power. The thing that stands in the way of making good regulation of Facebook is not that Facebook is inscrutable and impossibly technical, and that all of its outcomes are the result of emergent properties of algorithms we couldn’t hope to understand. The problem is that every time we try to regulate Facebook, it out spends whole nation states lobbying to stop us from regulating it.
And then when we do manage to regulate it, it directs that bazooka of money at subverting the regulation. Zuboff says all you need to fix this is markets. If you’re not paying for the product, you’re the product. If we just charge money for access to Facebook, Facebook will be disciplined by markets, just look at Apple. And Apple, even if you think that Apple’s custodianship is good, and we can talk later about why I think it’s not.
Ethan Zuckerman:
About whether that’s-
Cory Doctorow:
Right. But even if you think it’s good, Apple was the prime mover of creating a regulatory arbitrage zone in Ireland. Apple was the company more than any other firm in the world that maintained the impolite fiction that its billions of dollars were in an untaxable state of grace somewhere suspended over the Irish sea.
Ethan Zuckerman:
So in some ways we’re putting Zuboff, we’re putting Tristan Harris, in this sort of camp of there is good capitalism, look at Apple, they’ve put their phone’s operating system at war with Facebook over privacy.
Cory Doctorow:
Right.
Ethan Zuckerman:
What we can do is change the market mechanisms and get to the point where a regulated Facebook with different pro market capitalist mechanisms can be good for all of humanity. Why is that not the case?
Cory Doctorow:
So the problem here is that it ignores the power relations. So if we’re going to bring Facebook to heal, if we’re going to bring any coordinated business lobby to heal, then we need to shrink it till it fits in a bathtub and then drown it. For so long, as firms are very large and they are able to collude on key policies, and remember Facebook and Apple agree on a lot more than they disagree about. They agree on tax shifting and intermediary liability and whole bunch of other stuff, some of which I agree with and some of which I don’t, but they have a lot more in common than they have as difference, especially when it comes to sort of employment law and unions and a bunch of other stuff that really matters, whether or not you should give billions of dollars to Republicans.
Those are things that they’re really aligned closely on and they diverge around the edges. And they do make a big deal out of it and sometimes it makes a difference. As you say, Apple adding opt out by default on its apps’ permissions was really big, but to characterize them therefore as the pitched enemies of Facebook is really the narcissism of small differences. Foundationally they agree, for example, that it’s legitimate for companies to grow by buying their competitors, not by making the best products possible.
You don’t solve this by charging money for access to Facebook. You solve it by weakening Facebook and the historic remedy for weakening Facebook would be the tools of antitrust’ breakup, merger scrutiny, unwinding mergers, certain conduct remedies, and so on. I actually think all of those are great ideas. I think that as a minimum standard, Facebook lied when they bought Instagram and WhatsApp and said that they would never merge their backends, that was part of the merger conditions. For having failed to uphold those promises, the punishment that they faced is that they were required to pay a fine. That’s stupid. A fine is a price. What they should do is they should have that merger unwound, and if it’s very expensive, they should absolutely have that merger unwound because that way you set an example to other people.
Otherwise, what you say is if you lie to regulators when undertaking a merger, all you need to do to make the lie stick, to avoid the real punishment, is just make it really expensive to undo the lie.
Ethan Zuckerman:
When we talk about this sort of solution, we hear this from folks like Elizabeth Warren, we hear this from other folks who are sort of looking at existing regulatory tools. Do you think we can handle this with existing regulatory tools or do we get into an area that you’ve written about that I found very, very helpful of adversarial interoperability? Is that one of the possible solutions here?
Cory Doctorow:
Yeah. That’s where I was headed because the problem with those tools is not that they don’t work. It’s just that they’re slow. Conduct remedies are really hard to administer. Like if you tell Facebook… Well, a good example is we told Facebook it’s not allowed to practice racial discrimination in advertising financial products anymore. To actually figure out whether they’re doing that turns out to be nearly impossible. And they actually sue independent scrutineers who try to figure that out, like New York University who they’re suing right now.
Ethan Zuckerman:
Laura Edelson. Yep. Absolutely.
Cory Doctorow:
Breakups are really slow. The first attempt to break up AT&T was 68 years before the actual breakup of AT&T. That’s a long time to wait. But there are other tools in our arsenal that are tech specific. And I want to speak here briefly about tech exceptionalism and the right kind and the wrong kind of tech exceptionalism. Because people who talk about the reasons for tech consolidation, the apologists for it, whether or not they’re critics, will lean very heavily on this idea of economics called network effects. And a firm enjoys network effects if it increases in value based on the number of people who use it. So you might have joined Facebook because your friends were there. Once you were on Facebook, that was a reason for someone else to join so they could hang out with you. So every time a user joins, it’s a reason for another user to join. But all tech firms have enjoyed network effects since the year dot and yet were not searching Alta Vista with our craze. There was something that made the tech sector-
Ethan Zuckerman:
Speak for yourself. I’m in a state of the art 1950s office building here and I’m actually comfortably seated on the couch of my cray.
Cory Doctorow:
I was going to say, you’ve got excellent furniture. Yeah. And the reason for that is the thing that is genuinely exceptional about tech, because tech concentration is not exceptional. We have concentration across the board. There’s concentration in beer and spirits and eyeglasses and shipping and finance and cheerleading leagues and the professional wrestling sector. These are not network effects driven efficiencies. These are just monopolies that bought all their competitors, just like tech did. But there is something exceptional about tech that made it very dynamic over the years, despite its advantages of network effects, which is the low switching costs of going from one tech platform to another.
And to understand this, you have to understand something very deep about computer science, which is that computers are universal, that we only know how to make one kind of computer and it can run every program that we can write. The Turing Complete Von Neumann machine. We can make some special purpose computers, like we could still probably go and build one of those electro mechanical ballistics tabulators, but all it could do is calculate ballistics tables.
Ethan Zuckerman:
Everyone actually gets excited about analog computers for about 15 minutes and then you sort of realize, except that I can simulate this on a digital computer.
Cory Doctorow:
Right.
Ethan Zuckerman:
Maybe we hit a point in the future where quantum computers are doing something significantly different, but we’re still at this point of Turing machines implemented on [crosstalk 00:14:14] architecture.
Cory Doctorow:
My understanding of quantum computing is thin, but I have never heard any advocate for quantum computing saying they would be less pluripotent, they would be less universal. So because of this universality, it means that you can always plug one thing into another thing. You can always simulate an old computer and a new computer. And indeed, there’s a slightly weaker version of this in the internet itself, that the internet is the protocol of protocols. Before the internet, we had fax lines and cable lines and so on. Now we just have the one wire and we reduce all those things to applications that we tunnel over the single protocol. And because we have a universal machine that’s connected to a universal network, anything can connect to anything and do anything it needs to do. Yesterday, I was marking up a Canadian regulatory investigation report where they were talking about interoperability and they said, “Interoperability should be acquired where technically feasible.”
And I was like, “What does that mean, technically feasible?” Like the only things that aren’t technically feasible for computers are things like calculating pie to the last digit or figuring out how to optimally pack a thousand objects in a finite space or like proof forcing all the possible keys for a well encrypted message. The technically possible is different when we talk about this stuff, compared to say, what would it take to put a diesel engine in a Tesla? That is technically possible, but it’s really hard. What would it take to run Apple Two Plus software on my state-of-the-art Linux machine? Really easy by contrast.
Ethan Zuckerman:
And so the point here is that if you want Facebook to be compatible with Mastodon, which for instance is a problem that I’m very actively working on right now in the hopes of building more open federated social networks. In the grand scheme of things, that is a much easier problem than making a Tesla work with a diesel engine. It’s a fairly trivial thing to do if Facebook allows you to do it.
Cory Doctorow:
Yeah. It’s a straightforward technical question. The point of doing that, the reason for plugging Mastodon into Facebook, for plugging new things into old things, is it lowers the switching costs and switching costs, along with network effects, are the other key element to understanding the growth of big tech and it’s dynamism before the current era. Because it’s so cheap to plug one thing into another thing, it means that you can switch from one thing to another thing without leaving anything behind. So to give you a concrete example, when I was a CIO in the 90s and early 2000s, we had a real problem with our Mac users in our PC environments. So we’d have like a designer in a shop that was mostly PC and the designer would use a Mac. And the problem was that Microsoft office was very bad for the Mac.
And so if you sent them a word document, they couldn’t open it, or if they could and they saved it, then you couldn’t open it. It was terrible. We eventually started buying PCs for the designers to sit next to their Macs just to communicate with the rest of the office. That was terrible. Then we started just buying PCs and throwing away the Macs and installing the Adobe Suite and Quirk on them. And that was the death nail for Apple. The way Apple saved its ass was that Steve Jobs had some of his engineers reverse engineer the Microsoft office protocol, the file formats. And they made iWork, which is pages, numbers and keynote. And now you just open and close those files. And actually, that brought Microsoft to the table. Microsoft, after that happened, stopped trying to obfuscate its file formats with every revision to fend off competition because Apple had signaled that it was willing to spend as much as it took to stay current with Microsoft’s file formats.
And they just submitted to standardization, which is when you got PPTX and docX and XLSX. The X stands for XML, their standard.
Ethan Zuckerman:
Right.
Cory Doctorow:
And what it did was it lowered the switching cost. If you were a Windows user who wanted to be a Mac user, before iWork suite, you effectively had to say goodbye to all the files that you’d previously made and all the people you used to exchange them with. That’s a very high switching cost to endure. What Apple did was they made the switching cost effectively zero. It was the cost of buying a Mac and some software. And they actually ran these ads, you might remember them, the switch ads.
Ethan Zuckerman:
I do remember them. Yeah, absolutely. Yeah.
Cory Doctorow:
Are you a Windows user? You can switch. That was the whole point. They made the switching cost zero. So what Facebook and its rivals have done, including Apple, in the year since, as they have become increasingly consolidated, as they have been able to gather together and arrive at a uniform lobbying position. So Microsoft and Apple were on opposite sides of the interoperability question in 2002. Today they’re on the same side of it. They’ve been able to extract monopoly profits and use those monopoly profits to influence policy so that today, that kind of interoperability, which is adversarial interoperability, Microsoft didn’t help Apple do this, Apple just did it on its own, that reverse engineering, scraping, bought guerilla warfare interoperability, is now effectively banned. It’s banned because of things like section 12.1 Of the digital millennium copyright act, which is a law that says that if there’s any DRM you have to remove to do the reverse engineering, that the tool that removes that DRM is a felony to produce punishable by a five-year prison sentence and a $500,000 fine.
They promulgated very expansive theories of computer fraud and abuse act. This is an ancient 1980s Reagan era cybersecurity law that was used to hound our friend Aaron Swartz to his death. And they claimed that it said that anytime you violated terms of service, you were a felon. That was narrowed by the Supreme Court last summer in Van Buren, in summer 2020, in Van Buren. They invented new theories like tortious interference theories related to terms of service. They created expansive portfolios of software patents, and they created a kind of thicket around their product lines. So if you were to reverse engineer the file formats used by Apple’s proprietary suite of programs, now the Apple books or Apple movies, or whatever iTunes is called this week, and make interoperable players that play those media files-
Ethan Zuckerman:
They’re going to come after you at this point.
Cory Doctorow:
They’ll bomb you till the rubble bounces. You will be a glowing crater when they’re through with you if you make that product.
Ethan Zuckerman:
So let’s explore that because I am actively in this space. I am actively playing around with these questions of software modification that can send you to prison. And the grad students who work with me probably would love your opinion on this. But think of it this way, I’m really interested in this idea that what we need in social media is an aggregator. We need a client that we have control over where we can decide our rule sets for what gets blocked and what gets promoted, our own sorting algorithms for it, and rather than having a Facebook app and a Twitter app and a WhatsApp app, I want to have an open source application where I can sort between all of those things and put them all together.
Cory Doctorow:
Yeah.
Ethan Zuckerman:
There are some platforms that actually make this pretty easy. It’s not hard to write a third party client for Mastodon, the open source project. It’s not actually that hard to write one for Twitter. It’s extremely hard to write one for Facebook, and I will say worse than Facebook is LinkedIn. LinkedIn has been the single worst organization that we’ve tried to work with on all of this. You could argue that this is data that is available to you through a web interface. You should be able to enter in your password and take the stream of data that would be coming to you on the web app, it’s going to be a stream of JSON, you should be able to interpret it and push it into something else. If I build something that then renders that and gives it control in your browser or in a mobile phone app, my assumption is that I will be blocked first by a set of technical constraints and then by a set of legal challenges.
Cory Doctorow:
Yeah, I think that’s true. So I think that my discussions with pretty senior people at the largest tech firms has made it abundantly clear to me that they are really reluctant to engage in technological warfare with people like you and that they would much rather use legal warfare. And it’s not because they don’t know how to hire engineers. They’ve got a building full of engineers and a building full of lawyers. They want the building full of lawyers to solve this problem for them. And the reason they want the building full of lawyers to solve this problem for them is because it presents a quantifiable risk that if they can go and say we’re going to spend X dollars in court over this period of time and then we’re going to put this guy out of business and we’re going to scare the hell out of anyone who might back another rival like his, or any hobbyist or ideologue who’s contemplating going on the same path.
Then they win a kind of permanent victory. Whereas guerilla warfare really favors the attacker here because for them to distinguish what you are doing from their own users, they have to be able to tease out the conduct of their attackers in this formal sense, you’re the attacker and their the defender, from the conduct of their users. And if you have three billion users, then you generate 3001 in a million use cases every day. There is nothing that you would do with your scraping and bots that is weirder than say the third sigma Facebook user is doing every day when they log into Facebook. And so you can take all of those users as hostages because everything they do to crack down on you, hurts them.
And this hurts Facebook where it lives because when you offer this alternative, and I would go beyond just aggregating this data, I think that communities should have their own federated ways of talking with one another where they can set their own community rules about what is and isn’t moderated and how it’s moderated and so on, and so I think that there’s lots of scope for this. Again, the problem isn’t that Mark Zuckerberg is the wrong guy to fill a three-ring binder with policies to regulate all speech among three billion people speaking a thousand languages in 150 countries. It’s that, that is a three-ring binder [crosstalk 00:24:31].
Ethan Zuckerman:
That’s an insane idea. Right. And then an idea in which we have multiple communities, which have multiple different rule sets where different people can moderate. So you and I are in very much the same place.
Cory Doctorow:
Yeah.
Ethan Zuckerman:
We want to see a world in which there are many networks, they run on different rules, sometimes you’re the one running them, sometimes someone else is the one running them. We know that it’s technically possible. We also know that it’s legally extremely fraud at the moment.
Cory Doctorow:
Right.
Ethan Zuckerman:
Let’s talk about the other folks who are-
Cory Doctorow:
How to do it. Yeah.
Ethan Zuckerman:
Let’s talk about how to do it, but let’s talk also about the huge set of people who are 20 or 30 years younger than you and I are, Cory, and who are convinced that they know how to do it because of a blockchain.
Cory Doctorow:
Okay. God, if we must, but let’s-
Ethan Zuckerman:
We must because people are investing in them and because people are paying attention to them. You who are someone who, for the most part, creates words and images rather than code or law is so potentially influential around this because you’re trying to shift the way that we think [crosstalk 00:25:37].
Cory Doctorow:
The norms.
Ethan Zuckerman:
But this is where I want to force you to say something about this weird normative shift that’s going on because… So look, we’ve had certain achievements, right? We’ve hit this point where you don’t have a Jaron Lanier saying everyone should just stop using social media. No one really believes that the vast majority of people are going to stop using social media. We also have people sort of agreeing that Facebook, this highly centralized social media environment that we’re in right now, has some significant problems with it. We’ve just spent 50 minutes talking about how those problems are rooted in a certain class of economics. And now we have enormous enthusiasm behind a wave of young tech people who are firmly persuaded that they can bring a new set of solutions to the table, make an enormous amount of money, and create something new and creative at the same time. Why are they wrong?
Cory Doctorow:
Well, I’m skeptical of Web 3 for a bunch of reasons. I think the most relevant one to the discussion we just had is it doesn’t solve the switching cost problem. If the problem is that you can’t leave Facebook, because other people can’t leave Facebook, and they can’t leave Facebook because you can’t leave Facebook, it doesn’t solve the collective action problem. And Facebook didn’t solve the collective action problem this way. When Mark Zuckerberg took Facebook from a college only product to one that the general public could use, they had a major competitor called MySpace. It was owned by a rapacious billionaire named Rupert Murdoch. He wasn’t going to play nice with them. And although Facebook was a better service than MySpace in every way, including privacy, which was their initial pitch, unlike MySpace, we won’t mine your data, the problem was that all your friends were on MySpace.
And so you could go to the empty place where none your friends were and enjoy the user interface or you could stay where they were. Well Zuck didn’t solve that problem with like international everybody quits MySpace day, right? What Facebook fielded was a bot that you gave your MySpace login and password to and it would import your waiting MySpace messages to your Facebook inbox and let you reply to them. So Web 3 doesn’t solve any of this. Web 3 leaves you in exactly the same place. Web 3 has the same problem in that regard as Mastodon. I use Mastodon. I like Mastodon. Mastodon is not going to make centralized social media go away for so long as there’s a high switching cost.
And so if you have no theory of lowering switching costs, then you are not part of the solution. I think more broadly, I could see a world of interoperable services where some of them were blockchain based and run by DAOs, but I am skeptical of the ability of the model for DAOs to provide public goods. And not all of these services would have to be public goods, but if that’s what you’re aiming to do is to produce a public good, then I think it will be a problem because the model for a DAO is to say that we are going to capitalize public goods, bring in the funds for public goods, by appealing to people’s self-interest, by making the governance tokens, the membership in the community, a speculative asset that can appreciate in value.
And I think that the normative basis for stable long-term public goods that say Eleanor Ostrom talks about in the work that won her the Nobel prize, is a feeling of solidarity and that self-interest and solidarity, they do coexist in all of us. But if your system promotes self-interest ahead of solidarity, then your system will eventually reach crisis points where you will have to choose, as a community, between self-interest and solidarity. And when those crisis points are reached, you will find that the solidarity is set aside in favor of the self-interest. I think that when you talk about public goods, markets produce solutions that are necessarily short-lived, that there will always be a crisis in which self-interest and solidarity are in tension. And to the extent that your system relies on market mechanisms on speculators, then they will always fall apart.
Ethan Zuckerman:
It’s been such a pleasure. Thank you so much Cory Doctorow. Pleasure to have you on Reimagining the Internet. Hope we’ll have you back at some point soon. Thanks for being with us.
Cory Doctorow:
All right. Thanks.