A Murder Story: Whatever Happened to Interoperability?

Tech companies once dealt with competitors by attempting to develop superior products. Then they discovered they could simply buy them.

by Cory Doctorow

This article first appeared on The American Conservative, October 5, 2023

Tech exceptionalism is a sin. Tech leaders aren’t especially smart; they’re not even especially evil. They’re just ordinary mediocrities, no smarter or more cunning than you or I. Yet today’s generation of tech bosses have shrunk the internet down to “a group of five websites, each consisting of screenshots of text from the other four.”

What happened? How did a clutch of utterly undistinguished businesspeople convert the dynamic, fast-changing world of tech into a sector so inbred that it’s practically got a Habsburg jaw?

Like tech itself, tech concentration is unexceptional. Tech companies once dealt with competitors by attempting to develop superior products. As antitrust enforcement was progressively neutered—tentatively by Carter, aggressively by Reagan, and with varying degrees of enthusiasm by each successor administration, Republican and Democrat—tech companies were able to simply buy their competitors. 

Take Google, a company with a single blockbuster product: its quarter-century old search engine. Virtually every other in-house product Google has developed since—its video platform, its numerous social media networks, its RSS reader, and its “moonshot” smart cities division—has failed. Google’s successful products are—with few exceptions—other people’s products, which it bought and operationalized. From mobile to ad-tech, server management to maps, document editing and collaboration to satellite imagery, Google is a company that bills itself as an idea factory, but is best understood as a deep-pocketed, self-sabotaging investor that buys the products it is incapable of designing and fielding on its own.

Google isn’t exceptional here. Facebook identified WhatsApp and Instagram as threats when its users started jumping ship for the new rivals. The company wrote a couple of checks and acquired these two nascent rivals before they could grow to challenge the company’s dominance. As Mark Zuckerberg says, “It is better to buy than to compete.”

Apple, too, is a buying-things company: In 2019, CEO Tim Cook cheerfully boasted to Kara Swisher about the 90 companies Apple had acquired that year. Cook brings home a new company for his board more often than you bring home a bag of groceries for your family.

Not every startup wants to be agglomerated into a Big Tech borg. When a tech giant finds a firm that isn’t willing to sell out, they use their investors’ cash to force them out. Take Diapers.com: When they refused to sell to Amazon, Amazon responded by selling diapers below cost, lighting $100 million on fire before buying the suffocated husk of Diapers.com for pennies on the dollar. 

Again, this isn’t exceptional. Most industries have indulged themselves in a 40-year-long orgy of mergers and acquisitions, transforming virtually every sector into an oligopoly ruled over by a cartel of five or fewer companies: athletic shoes, beer, cheerleading, defense contracting, book publishing, eyeglasses, semiconductors, bottlecaps, professional wrestling...the list goes on and on. Whole supply chains have consolidated into fistulated masses: The drugs produced by the pharma cartel are overseen by the pharmacy benefits management cartel and retailed through the pharmacy cartel.

But actually, tech is a little exceptional.

Digital technology is exceptional in a deep, technical sense. The only kind of digital computer we know how to make is the “Turing-complete, universal von Neumann machine”—a device that can run every valid program. All computers are, at their deepest level, equivalent, able to run any app we can write. That universality has been a boon to tech competition. 

The flexibility of digital systems means that a new market entrant can always make a compatible (“interoperable”) add-on, mod, or accessory. When IBM gouged its customers on mainframe hard-drives, innovators like Memorex jumped into the market with “plug compatible” drives that sold at a fraction of the price.

Microsoft’s wildly unreliable Office for Mac product turned reading and writing Word, Excel, and Powerpoint files into a crapshoot for Mac users, whose files were frequently unreadable by their Windows-using colleagues. Apple responded by reverse-engineering the Office file-formats and releasing the iWork Suite, whose applications—Pages, Numbers and Keynote—could perfectly read and write those Microsoft files. Thereafter, Mac users could collaborate seamlessly with Windows users, and—crucially—Windows users could switch to the Mac without giving up their files or their compatibility with the Windows users who comprised the majority of the market.

For Facebook, the infinite malleability of the digital world was key to its triumph over the incumbent MySpace. Facebook promised dissatisfied MySpace users a superior, surveillance-free alternative, but Facebook understood that no matter how much users disliked MySpace, they loved the friends they had on MySpace more. “Come to Facebook and sit alone in an empty virtual room, admiring the clean user-interface” was a poor pitch.

So, Facebook used interoperable tools to let ex-MySpace users eat their cake and have it too. Facebook provided those MySpace users with a “bot,” an automated program that used the user’s login and password to impersonate that user to MySpace, scraping the user’s waiting messages and putting them in their Facebook inbox. Replies that users typed in on Facebook were then autopiloted back to MySpace. Thus, MySpace users could become Facebook users without having to convince their friends to make the leap with them—and without having to give up on friends who weren’t ready to leave.

This digital flexibility and the interoperability it enabled was a perennial check on incumbent dominance. Any time an incumbent tightened the screws on its customers—raising prices, lowering quality, converting included features into pricey add-ons—an interoperator could sell that customer a screwdriver to loosen the screw again.

What’s more, these interoperators enjoyed the “attacker’s advantage.” In security, the defender—say, an incumbent hoping to prevent its products from being reverse-engineered by an upstart—needs to make no mistakes. The attacker—say, a would-be interoperator painstakingly decompiling a popular product with an eye to offering an after-market improvement—has only to find and exploit a single defect in the defense.

Tech products have always enjoyed explosive growth thanks to the network effects that are endemic to the sector. Every Facebook user signs up to be with the users who are already there, and then becomes a lure for other users who want to hang out with them. Every iOS user is a reason to be an Apple app developer; every new app is a reason to become an iOS user.

But tech’s network-effects-driven growth have always been held in check by that inescapable universality. The second a company used its network-effects-drive scale to start shifting surpluses from customers to its shareholders, an interoperator popped up to offer those customers a better deal. 

That was true when IBM gouged on hard-drives; it was true when Microsoft used its scale to visit pain upon Mac users; it was true when MySpace tried to hold its users hostage. Wherever an incumbent used lock-in to extract super-normal rents, a new market entrant popped up to offer those customers an alternative. The more invasive your ads are, the more the ad-blocker company can raise in the capital markets.

The universality of digital platforms meant that every platform that became greedy thanks to its explosive growth was pruned back to size through the good graces of low switching costs delivered by interoperability, which either lowered the friction associated with going from one product to another, or simply modified the product the manufacturer delivered to remove extractive antifeatures.

Tech was forever a dynamic industry, where mainframes were bested by minicomputers, which were, in turn, devoured by PCs. Proprietary information services were subsumed into Gopher, Gopher was devoured by the web. If you didn’t like the management of the current technosphere, just wait a minute and there will be something new along presently. When it came to moving your relationships, data, and media over to the new service, the skids were so greased as to be nearly frictionless.

What happened? Did a new generation of tech founders figure out how to build an interoperability-proof computer that defied the laws of computer science? Hardly. No one has invented a digital Roach Motel, where users and their data check in but they can't check out. Digital tools remain stubbornly universal, and the attacker’s advantage is still in effect. Any walled garden is liable to having holes blasted in its perimeter by upstarts who want to help an incumbent’s corralled customers evacuate to greener pastures.

What changed was the posture of the state towards corporations. First, governments changed how they dealt with monopolies. Then, monopolies changed how governments treated reverse-engineering.

The Apple II Plus hit the shelves the same year Ronald Reagan hit the campaign trail. While it was Jimmy Carter who pulled the first tentative Jenga blocks out of the antitrust enforcement system, Reagan went at it with gusto, tearing them out by the fistful. 

The theories of Robert Bork and his Chicago colleagues gained currency. Their new orthodoxy held that Congress had never intended to police the emergence of monopolies per se, but only to rein in those rare monopolists who abused their market power to raise prices or lower quality. This “consumer welfare” theory of antitrust enforcement gradually won out over historic notions of the inherent risks of large corporations and concentrated sectors.

With the drawdown of antitrust enforcement, industries across the board dwindled to oligopolies, but tech got there first and faster. The capital markets proved exceedingly bullish on subsidizing formerly illegal tactics, such as predatory pricing, acquiring and extinguishing nascent rivals, and mergers to monopoly between major competitors.

Shareholders lined up to subsidize below-cost offerings. Amazon spent a mere $100 million destroying Diapers.com. But Uber’s investors financed a bonfire of $31 billion, losing $0.41 on every dollar it took in for the first 13 years of operations. Investors snapped up shares of companies like Google, whose ineptitude at creating successful new products was counterbalanced by its ability to buy other peoples’ companies.

As tech diversity declined, the industry's ability to capture its regulators increased. 

The early, chaotic years of the consumer internet were full of policy setbacks for tech companies. Recall the Napster Wars, when the entertainment sector—dominated by fewer than a dozen companies then, a number that’s halved in the ensuing quarter-century—comprehensively trounced the tech sector. Bedrock case law crumbled, like the “Betamax” decision (Sony Corp. of America v. Universal City Studios, Inc.) that immunized the manufacturer of any technology that was “capable of sustaining a substantial noninfringing use” from copyright liability. 

Congress and state houses introduced strings of bills intended to limit the tech sector’s ability to make new tools without permission from the entertainment company. Michael Powell’s FCC passed the “Broadcast Flag” rule, which required every single digital device to be fitted with technology that would disable file-transfers if a file was marked with a copyright restriction flag (the Broadcast Flag rule was struck down by the D.C. Court of Appeals in 2005).

Then, as now, the entertainment sector was a pipsqueak when compared to tech. By every measure—number of users, number of sales, total profits, profit margins—tech was bigger and (theoretically) more powerful than entertainment. But the difference was that all the power in the entertainment industry had been boiled down to a cozy cartel of frenemies who were capable of forming a united front in policy fights. By contrast, tech was a squabbling rabble. In every policy fight, some tech companies would defect, taking the entertainment industry’s side, muddling the message that regulators, courts and lawmakers heard. 

That disunity provided an opening in tech’s defense that entertainment drove straight through, again and again.

But there’s more than one way to solve a collective action problem. The tech sector’s solution was brutally simple: As tech companies merged and re-merged and re-re-merged, its policy goals grew less fractured. Tech companies that had once rudely shouldered their way into the market by reverse-engineering their rivals’ products—Apple, Google, Facebook—suddenly got DMCA-happy and began suing and threatening rivals who tried to do unto them as they did unto their own forebears. When Apple reverse-engineered Microsoft Office to make iWork, that was progress: If you were to try to do the same—say, to make a platform that could run iOS apps and play back media from the iTunes Store—Apple’s lawyers would nuke you until the rubble glowed. 

Everywhere questions of tech policy arise—U.N. specialized agencies like WIPO, standards bodies like the World Wide Web Consortium, expert agencies like the FCC, intergovernmental bodies like the E.U. Commission—tech arrives in lockstep to sing from the same hymnal. Their self-serving rhetoric is no longer checked by other commercial actors who speak with the authority of firms that serve millions of customers. 

To the extent that a counternarrative appears in the record, it comes from minuscule firms, hobbyists, activists, and academics. For regulators—especially regulators drawn from the ranks of large tech firms—giving them what they want is a no-brainer.

This policy unity has produced a regime where everything that is not prohibited is mandatory. Tech companies get to violate privacy laws where they exist (Europe) or prevent their passage altogether (Washington). They bypass labor law with the fiction that when an app is your boss, you are transformed into a “small business.” They circumvent consumer protection law, arguing that a pay-to-play system that fills our search results with deceptive lookalike products that are inferior to the goods we’re seeking is a form of “advertising.”

Meanwhile, we are prohibited from taking countermeasures. A worker can’t reverse-engineer the Uber app to compare wages with other workers and algorithmically determine when a job offer is a lowball. A consumer can’t install an ad-blocker in an app without risking felony prosecution for violating the DMCA. An entrepreneur can’t mod the Amazon app to remove all the paid results and display comparison prices from rivals. An investigative journalist can’t scrape grocers’ websites to build the case that large firms are colluding to raise prices while blaming inflation.

Interoperability wasn’t killed by the reverse-engineer-proof computer nor by the lack of market demand for aftermarket modifications. It was killed by regulatory capture, arising from market concentration, arising from the bad idea that monopolies are efficient sources of consumer welfare, rather than deep-pocketed rent-seekers who leverage their scale to extinguish rivals, corrupt our politics, harm workers, and gouge customers.

This article is part of the “American System” series edited by David A. Cowan and supported by the Common Good Economics Grant Program. The contents of this publication are solely the responsibility of the authors.

Cory Doctorow (craphound.com) is a special advisor to the Electronic Frontier Foundation (eff.org) and the New York Times bestselling author of dozens of books. Follow his newsletter at pluralistic.net.