Science fiction has a longstanding love-hate relationship with the tech tycoon. The literature is full of billionaire inventors, sometimes painted as system-bucking heroes, at other times as megalomanical supervillains.
From time to time, we even manage to portray one of these people in a way that hews most closely to reality: ordinary mediocrities, no better than you or I, whose success comes down to a combination of luck and a willingness to set aside consideration of the needs of others. It’s easy to find such people atop our increasingly steep economic pyramid, but it’s very hard to find any who’ll admit it. There is nothing a successful person hates more than being reminded that “meritocracy” is a self-serving myth, a circular logic that says, “The system puts the best people in charge, and I am in charge, therefore I am the best.”
But while the powerful remain blissfully insulated from the bursting of the meritocratic delusion, public sentiment is increasingly turning against the ultra-wealthy, and in the most interesting way possible. Today, the commercial tyrant isn’t merely seen as a villain, but also as a fool – someone whose greatness is due to an accident of history and a vacancy of morals, not the result of a powerful genius gone awry.
It’s a distinction with a difference. If Facebook is Facebook because Mark Zuckerberg is a once-in-a-millennium genius who did what no other could, then our best hope is to somehow gentle the Zuck, bring him into public service, like a caged ET that government scientists either bribe or torment into working on behalf of the human race. That’s the constitutional monarchy model, the model where we continue to acknowledge the divine right of kings, but bind them to the material plane by draping the king in golden chains of office whose ends are held by an aristocracy that keeps the monarch from getting too frisky.
But if Facebook is Facebook because Zuck got lucky, if he just combined cheap capital with regulatory tolerance for buying out the competition and building a legally impregnable walled garden around his users, then we don’t need Zuck or Facebook. There’s plenty more where he came from, and all we need to do is withdraw the privileges that regulatory forbearance granted him. That’s the republic model, where we get rid of the king and govern ourselves.
That’s the direction that regulators are moving in now. Biden’s FTC chair, the incredible Lina Khan, came to prominence as a Yale law student with her paper “Amazon’s Antitrust Problem”, which makes a powerful case against the kind of vertical integration that Big Tech practices today. Her term in office thus far has been characterized by extraordinary skepticism of growth-by-merger and vertical integration more generally. As I write this in late January of 2022, she’s just announced that the FTC will seek comment on updated, 21st-century merger guidelines, and she’s in the midst of suing Facebook to break the company up.
So let’s assume that we’re entering a new world. A world where, say, the proposed merger between Microsoft and Blizzard that I woke up to today is blocked by the FTC as nakedly anti-competitive. Even moreso, let’s assume that we’re entering a world where the presumption is against vertical integration itself – where the bedrock is that companies must choose whether to be “platforms” or “platform users.”
In that world, Amazon could offer virtual shelf space to merchants, or it could compete with those merchants by making its own goods, but not both. Apple could have an app store, or it could make apps, but not both. Google could offer space for advertising, or a market for placing ads in advertising spaces, but not both, and so on – for Facebook, Salesforce, Microsoft, and all the other tech platforms that have business customers and compete with those customers.
This was once a commonplace of antitrust regulation. It’s called “structural separation.” We once banned banks from owning companies that competed with the companies they loaned money to, and railroads from owning freight companies that competed with the companies whose freight they hauled.
At root, structural separation is grounded in the idea that it’s nearly impossible to prove cheating (AKA “self-preferencing”) so we should just eliminate the incentives for it. It’s nice to think that a bank manager will determine our creditworthiness based on our individual merits and the intangibles captured in the idea of “character.” When a bank manager gives a loan to a local business that’s fallen on hard times but denies a loan to another business with similar financials, we generally accept the idea that the manager made a judgment call based on familiarity and trust (though sometimes we’re suspicious that “familiarity and trust” is code for “racism and bias,” of course).
But what if the business that the bank manager okays a loan to is owned by the bank, and the business the bank manager turns down is its direct competitor? Sure, the bank has a “firewall” between its lending arm and its commercial investment arm – but if you were the owner of the business who got screwed by that decision, would you be willing to assume impartiality as you packed up the ruins of your life’s work?
We don’t let lawyers represent both sides of litigation, and we don’t let judges hear cases involving their friends and family members. It doesn’t matter how much they promise to be impartial. We have an intuitive sense that the losers in any such arrangement would never be able to trust in the process’s impartiality.
Structural separation was an early casualty of the neoliberal era. When Bill Clinton signed the repeal of the 1933 Glass-Steagall Act in 1999, he removed the final barrier to banks competing with the businesses they lent to (it took nine years for this deregulation to give rise to the Great Financial Crisis of 2008).
Today, our regulators are once again ready to impose some structural separation on our giant tech companies, ordering them to fracture in ways that is supposed to remove the incentive to cheat.
I think tech needs structural separation, and that structural separation is the least invasive way of creating fairness and legitimacy. Take Google’s search ranking: if you search for an address, Google will prominently display a link to Google Maps, as well as other links relevant to that search result, from real-estate listings to USGS topographic data.
Now, if you were running a competitor to Google Maps, how could you establish – to yourself, or to a regulator – that Google’s link to its own Google Maps page was there because Google objectively determined that Google Maps was better than your product? How could Google prove that to you, or a regulator? We don’t have an agreed-upon measurement for the quality of an online maps result. We’re not gonna get one. There isn’t one.
Whether it’s Apple showing you its own music player ahead of a rival’s in a search of its App Store or Facebook linking to its own version of an article or video rather than one on the open web, the process will never feel fair. Companies can play the game, or they can be referees, but they can’t play in a game that they are also refereeing. A judge can’t hear a case involving her mother, and a lawyer can’t represent both sides of a dispute.
But while it’s obvious to me that structural separation is necessary, it’s not obvious to me where the separation lines should run. I know for sure that we can’t trust tech giants to choose them, obviously. I mean, obviously.
Like, we should absolutely ignore Google’s transparent “Alphabet” ruse, in which it spun out many of its divisions into separate companies and grouped them under a single umbrella company, in a bid to create pre-scored perforation lines that regulators could tear along when they finally came for the company.
An even cursory glance at Alphabet shows you that more-or-less all the companies that make money are still called “Google,” while the other letters of the alphabet are mostly taken up by follies and also-rans, relatively infinitesimal companies that either make very little money or lose moneysometimes a lot of it. Alphabet is like a kid saying, “Okay, okay, punish me if you must, but please, whatever you do, don’t deny me my delicious boiled liver or these nutritious Brussels sprouts!” No dice kid; we’re taking away your ice-cream.
But, as the economist Ramsi Woodcock has pointed out to me, you can’t tell a company that it’s not allowed to have any vertical integration. A company is vertical integration – that’s it’s point. We accept that an accounting company might have its own receptionists and cleaners, even though accounting and cleaning and reception are not immediately related.
Even in the case of a platform, there’s no bright line between anticompetitive and procompetitive integrations. Imagine that we tamed Amazon, made it a respecter of human rights and workers rights and the climate and the rights of local businesses and communities – we’d still have a use for Amazon. How should it be separated?
Some calls are easy: split off Prime studios and the publishing arms, which compete with Amazon’s business customers. Same for Amazon’s in-house clothing, furniture and other products.
But what about warehouse automation and fulfillment? These clearly offer a significant competitive advantage to Amazon as a platform operator, and make it harder for other companies to compete to offer platform services. What’s more, Amazon could use its warehouses to self-preference, in fact, it already does, penalizing platform sellers who choose to warehouse and fulfill their own products, rather than paying Amazon to do this for them. Sure, Amazon might say it gives preferential search results to businesses that use its warehouses because it can be sure that those items will be delivered more efficiently and reliably, but it also benefits every time it makes that call. It’s the referee and it’s playing the game.
This is a thorny problem, to be sure. The bright-line cases are easy enough to lay out, but these liminal cases multiply the more you think about them, and I haven’t yet seen any convincing explanation of how we deal with them.
If I were playing for Big Tech, this is the fracture line I’d drive my wedge into. But Big Tech doesn’t want to go anywhere near that line, because the last thing Big Tech companies want to do is get us thinking too hard about which parts of their businesses they own and which parts they outsource.
Facebook says it has to gobble up virtually every VR company or it can’t build the metaverse. But it also says that it’s inconceivable that it could operate its own moderation shop, and that it’s only natural – inevitable, really – that moderation be outsourced to traumatized, low-waged subcontractors who do a terrible job. Amazon says it has to buy up every part of its supply chain – but also that its drivers and warehouse workers must be procured through staffing agencies it decidedly does not own. Apple can buy 90 companies a year, but also can’t figure out how to own the factories where its phones are made, which would allow it to decisively prevent slave labor in their production.
In general, tech companies want us to think that it’s literally impossible for them to be vertically integrated with the parts of their business that impose costs on the rest of us – the divisions that pollute, or maim workers, or oversee fraud. At the same time, they insist that it’s impossible for them not to be vertically integrated with the parts of their supply chain that prevent competitors from emerging.
Perhaps we don’t just need a doctrine of structural separation, but also of structural integration: a requirement that the business functions that harm the rest of us when they go wrong be kept in-house, so that the liabilities from mismanaging those operations end up where they belong.
Cory Doctorow is the author of Walkaway, Little Brother, and Information Doesn’t Want to Be Free (among many others); he is the co-owner of Boing Boing, a special consultant to the Electronic Frontier Foundation, a visiting professor of Computer Science at the Open University and an MIT Media Lab Research Affiliate.
All opinions expressed by commentators are solely their own and do not reflect the opinions of Locus.
This article and more like it in the March 2022 issue of Locus.
While you are here, please take a moment to support Locus with a one-time or recurring donation. We rely on reader donations to keep the magazine and site going, and would like to keep the site paywall free, but WE NEED YOUR FINANCIAL SUPPORT to continue quality coverage of the science fiction and fantasy field.
©Locus Magazine. Copyrighted material may not be republished without permission of LSFF.