A historical accident made Massachusetts a lab for studying how tech can serve monopolies, and the moves, countermoves and counter-countermoves show how businesses, tinkerers, governments and the public can liberate themselves from seemingly all-powerful monopolists.
It all starts with Right to Repair. Companies love to monopolize the repair of their products. If the only place to get your broken stuff fixed is at its manufacturer’s authorized depots, the manufacturer can move all kinds of value from your side of the deal to their own. Like, they can force you to buy their original parts, rather than cheaper replacements. They can charge you uncompetitive rates for repairs, based not on the cost of labor and parts, but on how much they think you’ll pay before you just buy something else. Best of all, a repair monopoly lets manufacturers decide when your stuff is “beyond repair,” whereupon they can offer you a generous “trade-in rebate” if you buy a new gadget.
(No coincidence that Apple leads the anti-repair movement; CEO Tim Cook warned his shareholders that the biggest threat to the business was people fixing their phones rather than replacing them.)
But monopolizing repair is hard. Even if a manufacturer refuses to supply repair manuals, companies like iFixit will step in, do teardowns of your products, and make their own documentation. If you gouge on replacement parts, companies will reverse-engineer the parts and make their own alternatives. These factors keep businesses honest: if they abuse their customers, they create a market for rivals who’ll treat their customers better.
But American innovation is alive and well in the anti-repair sector. Canny corporate lawyers have devised clever ways to abuse copyright and trademark to fight independent repair. Like, if you etch a minute company logo into all the parts in your gadgets, you can use trademark to fight refurbishment and re-importation of parts harvested from scrapped items.
Even more powerful is DRM. Thanks to Section 1201 of the 1998 Digital Millennium Copyright Act, making tools to bypass a “copyright access control” is a felony. Software is copyrighted, so if your product has a chip in it, you can wrap it in a thin layer of “access controls” (DRM). Anyone makes a tool to bypass that DRM — say, to extract diagnostic information or activate a new part — commits a felony.
A company that puts cheap microcontrollers into its gadgets can make it a literal crime to reconfigure your own property so that it serves you, rather than the company’s shareholders. The falling price of chips (notwithstanding our current supply-chain blip) encouraged manufacturers to deploy this strategy to monopolize all repair.
The automotive sector was an early adopter of these dirty tricks. Car companies hate independent mechanics and third-party parts companies and have waged war on them for decades. By adding DRM to our cars, the auto makers found a way to block third-party parts, and to prevent independent mechanics from interpreting diagnostic messages.
By 2012, the people of Massachusetts had had enough. They put a question on the state ballot that year and 75% of Bay Staters voted for a Right to Repair rule that required Big Car to supply independent mechanics with the information needed to interpret the diagnostics that traveled over their cars’ wired network (the “CAN bus”).
But the car makers weren’t going to give up on the dream of making indie mechanics extinct. They shifted their diagnostic messages from the CAN bus to the cars’ wireless networks, which were not covered by the Right to Repair law.
Eight years later, in 2020, Bay Staters once again put a Right to Repair question on the ballot. The car makers pulled out all the stops to fight it, including a surreal scare-ad campaign that claimed that allowing independent repair would lead to women being stalked and murdered:
The car makers went to court to fight this (naturally), but that was just the start of their fuckery. Manufacturers like Subaru and Kia found a truly disgusting way to punish the people of Massachusetts for having the audacity to demand the right to choose their own mechanics: they are disabling advanced features on cars sold in MA:
If you buy your Subaru or Kia in Boston, you lose the ability to remote-start your engine on cold mornings. Your car won’t send you oil-change reminders or tire-pressure warnings. You lose access to emergency assistance services (I was a happy Kia owner — until I read this).
This is just more of the fuckery that started in 2012, when car makers shifted their diagnostics to the wireless network. It’s a winning strategy. If car makers can maintain a legal barrier to repair, they can force independent mechanics out of business and scare investors away from parts manufacturers. Do it long enough and the car makers won’t need to fight right to repair, because there won’t be any indie fix-it shops left — all the mechanics will have changed careers or taken jobs with the big auto-makers.
All of this is an object lesson in the limits of interoperability in fighting monopoly. Interop — the power to plug a new product or service into an existing one — is a powerful anti-monopoly tool.
To understand why, you have to first grasp the fatal flaw in the arguments in favor of tech monopolies. Antitrust law (theoretically) bans companies from trying to secure monopolies, but if a company just happens to get a monopoly by being the best at what it does, antitrust says we should leave it alone.
So Big Tech’s apologists like to claim that tech is a “natural monopoly” — a “winner-take-all” market that inevitably results in a single company dominating each tech subindustry. They attribute this natural monopoly to “network effects” — that’s when a product gets better as it adds customers. You probably joined Facebook because your friends were there, and once you joined up, you became a reason for more people to join. That’s “network effects.”
Tech companies definitely enjoy network effects, but that’s not the whole story. You see, digital technology is configurable in a way that other technologies are not. At a fundamental level, every computer we know how to build can run every program we know how to write. The inescapable universality of technology means that we can’t just think about network effects; we also have to think about “switching costs.”
Switching costs are another idea from economics: switching costs are (unsurprisingly) the costs you have to bear when you switch from one product to another. Switch from Facebook, lose the ability to talk to the customers and communities and friends you leave behind. Switch from iOS, lose your apps and media. Switch from HP, throw away all the ink you’ve bought on your subscription plan (and maybe pay a fee to cancel the subscription).
All of these switching costs may seem technological, but they’re actually legal. The universality of computers means that you could absolutely switch from iOS to Android and keep running your apps in a virtual machine. There’s no technical reason you can’t install modified HP printer software that lets you use third-party ink. There’s no technical reason you can’t leave Facebook but continue to participate in the messaging and groups you left behind:
The reason you can’t do these things is that it’s illegal. The reverse-engineering, scraping and other guerrilla tactics you need to accomplish these things without the manufacturers’ cooperation put you at risk of prosecution under cybersecurity, copyright, trademark and other laws.
The antitrust reform movement has cottoned onto this and there is a groundswell of support to force the tech companies to open up their platforms to competitors. In the USA, there’s the ACCESS Act:
And in Massachusetts, there are Right to Repair laws.
Like I said, Massachusetts is the lab where these ideas are playing out. They’ve been at it for a decade. We should study the monopolists’ countermoves in Massachusetts because they are a playbook that will be used by other monopolists if we pass, say, the ACCESS Act.
Back in 2012, Big Car was forced to open up its wired network, so it moved all the useful network traffic to its wireless net. If the ACCESS Act forces Facebook to expose an API so third parties can interoperate with it, we should expect Facebook to pull a similar stunt.
They could claim that they are “refactoring their data-structures” to provide security, or innovative features, and feign remorse and surprise that this move means that the API is no longer useful for interoperators who supply ex-Facebook users with access to their friends and communities.
And yeah, we could go to court to fight them, or we could pass another law to try to prevent this kind of fuckery, but in the meantime, the fate that befell Massachusetts mechanics awaits interoperators. Neither small co-ops, nor tinkerers’ hobby-projects, nor startups will be able retain users, programmers or funding for their federate refuge from Facebook if it can’t hook up to Facebook. As Facebook ties up enforcement action in court or Congress, these projects and businesses will fail. When (if?) we force Facebook to stop cheating, the pool of interoperators will be much smaller. And like the car-makers, Facebook can create a fuckery stockpile, and roll out a new tactic every time an old one is banned.
But remember computers are intrinsically universal. Even if manufacturers don’t cooperate with interop, we can still make new services and products that plug into their existing ones. We can do it with reverse-engineering, scraping, bots — a suite of tactics we call Adversarial Interoperability or Competitive Compatibility (AKA “comcom”):
These tactics have a long and honorable history, and have been a part of every tech giant’s own growth. That’s given Big Tech companies a deep and urgent appreciation of how effective these tactics are at overthrowing dominant firms. First, you make a name for yourself by selling cheap ink that works with a monopolist’s printers, then you leverage that name-brand recognition to go into the printer business! Unsurprisingly Big Tech now hates comcom, claiming that when they used guerrilla tactics, it was part of the legitimate progress of the industry, but when others do it to them, that’s piracy, hacking — digital terrorism!
Imagine if comcom had been in the mix back in 2012, when the automakers broke Massachusetts right to repair by shifting service messages to their cars’ wireless networks. If breaking DRM was legally permissible, Bay Staters wouldn’t have had to wait eight years for a new ballot initiative. Instead, a couple of smart MIT kids could have hacked those wireless service messages, designed a decoder for them, manufactured a gadget with a $3 bill of materials, and sold it for $50 to every independent mechanic in Massachusetts (and the US, and the world!). They could have raised capital to do this, promising a subscription revenue stream that would cover updates that kept the gadget current with new models of cars and new scrambling schemes.
The auto-makers could fight back, but they’d be at a serious disadvantage. If they pushed updates to their cars (or revised the systems on new cars coming off the line), they would have to ensure that their authorized service centers were current. The added complexity would piss off the mechanics paying for the companies’ stamp of legitimacy, and their customers. The harder the auto-makers fought, the more mechanics would give up on being authorized and buy the indie diagnostic gadget instead — and the more drivers would switch to indie mechanics.
Meanwhile, our MIT kids could expand into ancillary services — like a price-comparison tool to help indie mechanics source cheap third-party or refurbished parts, or third-party warranties for drivers, or… Well, you get the picture.
If auto-makers knew that committing a fuckery would immediately spark this kind of comcom guerrilla warfare, there’s a good chance they wouldn’t commit the fuckery. Large companies hate competition, but they hate uncertainty even more. Providing a standard interface for interoperators may not be their preferred outcome, but at least it’s orderly and predictable. Comcom represents an unquantifiable risk to large businesses whose shareholders get really angry when unexpected things happen.
So in theory, just the threat of comcom might dampen manufacturers’ fuckery inclinations. But who knows? Some of these companies might decide to roll the dice, driven by spite or hubris. When that happens, it’s MIT kids to the rescue, and a market for cheap guerrilla comcom gadgets that keep the independent sector going while lawmakers and regulators go through the slow, plodding business of fixing the law.
This isn’t just a theory, it’s one with an evidentiary basis. This is how previous compatibility wars actually played out. Back in the early 2000s, Microsoft was kicking Apple’s ass by undermaintaining Office for the Mac. Windows Office docs and Mac office docs were only barely compatible, and many was the time when opening a Windows Excel or Word file on a Mac would permanently corrupt it.
I was a CIO back then, overseeing heterogeneous Mac/Windows networks, and we actually started giving our Mac-using designers a second Windows computer so that they could reliably share Office docs with their Windows-using peers. In some cases, we just bought Window versions of Quark and Photoshop and got rid of the designers’ Macs.
Apple didn’t solve this problem by convincing Microsoft to fix its Mac products. Instead, Apple reverse-engineered the Windows Office file-formats and created iWork (Pages, Numbers, Keynote) a Mac productivity suite that could seamlessly read and write those files.
That was the moment when Microsoft finally stopped fighting standardization of its Office formats, and started cooperating with the project of moving them to a universal, XML-based successor (docx, xlsx, pptx). Why not? All the obfuscation they had thrown into Office represented a major engineering headache. Maybe the cruft was worth it if it helped Windows kill off the Mac, but once Apple committed the resources to untangle Office’s spaghetti code, all the benefits to Microsoft disappeared.
Today, the comcom tactics that saved Apple are largely off the table, thanks to all the legal restrictions that have been bought and erected by tech giants, including Apple. Like, if you try to make a runtime for iOS apps, Apple will 100% sue you:
Corellium, the company that built that iOS runtime, squeaked out a very narrow victory. A broader version of their product — aimed at a mass market, not just security researchers — would have a much harder struggle in court.
Comcom and mandatory standards are complementary. Mandatory standards provide new interoperators with a reliable, low-cost way to build new products and services. Comcom keeps tech giants’ cheating impulses in check, and also offers interoperators a fallback position when an exec decides to cheat anyway.
We can get mandatory standards through a law: pass the ACCESS Act or DMA. But comcom is harder to revive. Ideally, we could reform all the laws that block comcom — patent, copyright, cybersecurity, contract, etc — to protect interoperability. But that’s a lot of slow lawmaking, and there’s no guarantee that the high-priced corporate lawyers who figured out how to twist those laws to block comcom won’t come up with more gambits.
Another possibility is to pass an “interoperators’ defense” — something like an Anti-SLAPP law, where an interoperator accused of an offense under any statute could use a streamlined court process to show that they’re engaged in legitimate interop — improving competition, security, privacy, accessibility, usability, etc.
We could use the ongoing antitrust cases against the tech giants to appoint “special masters” — independent overseers who’d have to bless any legal action taken by the company against a competitor, only after verifying that it wasn’t a pretext for blocking interop.
We could use government procurement rules: if you want to sell a car to the government, you have to promise not to sue rivals that make it possible for independent mechanics to fix that car, even if they bypass your DRM. A similar rule could require, say, Google to promise not to attack rivals who made interoperable ed-tech that plugged into Google Classroom.
All of these would leave the companies with the legal right to fight actual bad actors — companies that stole user data or enabled fraud. Even more important, though, is the long-overdue creation of a federal privacy law in the US. That way, we could have a single standard that applied to both Big Tech platforms and the scrappy interoperators that offer liberation from their walled gardens:
After all, even companies that claim to stick up for their users’ privacy — like Microsoft and Apple — are highly selective about when that privacy is worth defending. So long as the switching costs are high enough, those companies can be relied upon to put their shareholders’ interests ahead of their users’ privacy:
The antitrust world has really started to think hard about the power of interop to smash monopolies in rapid and durable ways. An excellent new Yale Law Journal paper from Herbert Hovenkamp, the dean of American antitrust legal scholarship, delves into the legal basis for an interop remedy:
I’m really hoping that as the idea of interop spreads through antitrust circles, that these game-theoretical analsyses of the ways that different kinds of interop interact with each other to create durable anti-monopoly measures enters the discourse.
