Daniel's been circling right to repair for a while now — it came up in the laptop modularity episode, the repair-tooling one, and most recently when he was trying to figure out whether he could fix a monitor power board himself. He wants the full treatment, and the timing's good: the EU's Right to Repair Directive, twenty twenty-four slash seventeen ninety-nine, hits its member-state transposition deadline on July thirty-first, which is four days from now as we record. He's got four things he wants from us.
Four things. Okay.
First, the actual history — not the vibe, the lineage. Magnuson-Moss in the US, the Massachusetts automotive ballot initiative, John Deere and tractor firmware, the DMCA Section twelve-oh-one exemption battles, and the organisations that turned teardowns into policy. Second, which manufacturers have genuinely gotten behind this and which have only appeared to. He wants names, receipts, and the distinction between real commitment and PR. Third — and this is the one he finds interesting — where's the legitimate limit? What are the real safety and security cases, not the manufactured excuses, and how do you tell them apart? And fourth, briefly, has any of this actually worked? The repairability indices, the battery regulation, whether devices are measurably more repairable or whether the metrics are being gamed.
That's a lot. And the third one especially — he's not wrong that most coverage just picks a side and stays there.
He also said explicitly he wants the steelman of the manufacturers' position, and an honest account of where repair advocates overreach. Not a cheer.
Good. Let's start with the history then, because the history actually explains why the third question is hard. Most people think right to repair started with iPhones and glued-in batteries, but the legal bedrock predates smartphones by decades. Magnuson-Moss Warranty Act, nineteen seventy-five. Federal law. It said if you offer a warranty on a consumer product, you can't tie that warranty to the use of a specific brand of part or service unless you provide the part or service for free — or you get a waiver from the FTC. The practical effect was that those "warranty void if removed" stickers? They were never actually enforceable. Companies kept putting them on anyway because consumers didn't know, and the FTC didn't enforce aggressively for years.
So the law existed, but it was sleeping.
The FTC woke up in twenty eighteen, sent warning letters to six companies — Nintendo, Sony, Microsoft, ASUS, HTC, and Hyundai — telling them their warranty-void stickers and policies violated the Act. That was the first real shot. But the right-to-repair movement as a named thing didn't start with consumer electronics at all. It started with cars.
Massachusetts, twenty twelve.
Massachusetts twenty twelve. Ballot Question One, the Motor Vehicle Owners' Right to Repair Act. Passed with eighty-six percent of the vote. What it required was that manufacturers selling cars in Massachusetts had to make the same diagnostic and repair information available to independent repair shops and owners that they gave to their franchised dealers. The car companies fought it, lost at the ballot box, and then — this is the key strategic move — rather than fighting state by state, they signed a national memorandum of understanding in twenty fourteen. Agreed to apply the Massachusetts requirements nationwide. Two reasons: one, complying selectively is more expensive than complying everywhere; two, they didn't want fifty state-level fights.
Which is the structural point Daniel flagged. The strictest state becomes the de facto national standard. It's not conversion, it's arithmetic.
And it still worked. That template — state-level ballot initiative or bill, manufacturer resistance, then a national MOU or a big-state law that becomes the effective standard — that's the playbook the entire consumer electronics movement adopted. The automotive win was huge because it proved the model.
Then agriculture took it somewhere unexpected.
John Deere. This is where the movement got a political constituency it didn't have before. Farmers running cracked Ukrainian firmware — no, really, firmware from Ukrainian forums — to bypass the diagnostic locks on their own tractors. You've got a quarter-million-dollar combine sitting in a field during harvest, it throws a sensor code, and you can't fix it yourself because the software won't let you clear the code or even identify which sensor is bad without a dealer technician driving out. The dealer might take days. The crop doesn't wait.
That's the thing that turned a niche technical grievance into something with actual rural Republican constituencies behind it.
Right. You've got farmers — not exactly a demographic that defaults to regulatory activism — showing up at statehouses demanding the right to fix their own equipment. Nebraska, Missouri, multiple farm-belt states introduced bills. Deere's initial position was that farmers didn't own the software in their tractors, they licensed it, and modifying it violated the DMCA. That argument landed terribly. The phrase "you don't own your tractor" became a PR disaster.
Which brings us to the DMCA angle. Section twelve-oh-one.
The anti-circumvention provision. Nineteen ninety-eight Digital Millennium Copyright Act. Section twelve-oh-one makes it illegal to circumvent technological protection measures that control access to copyrighted works. The Copyright Office has the authority to grant temporary exemptions every three years through a rulemaking process. And for years, repair advocates had to go back every three years and argue that circumventing a software lock to repair your own device should be legal. Every three years. If they lost, the exemption could expire.
So the legality of repairing your own phone was contingent on a triennial copyright proceeding.
Yes. That's insane as a policy structure, but that's how it worked. The exemptions were granted — tractors, smartphones, consumer devices — but the fact that it was a recurring battle rather than a settled right is what drove the push for statutory fixes rather than relying on the Copyright Office. The twenty eighteen music modernization bill actually made some of these exemptions permanent for certain device categories, but the structural problem remained.
And the organisations driving this?
iFixit started as a teardown website — take apart the new iPhone, photograph the guts, score it for repairability. Kyle Wiens and Luke Soules founded it in two thousand three. They built a massive repair guide database, sold tools and parts, and gradually became a policy actor. They publish position papers, they testify, they grade legislation. The Repair Association is the other major one — that's the trade group, lobbying arm, model-legislation shop. Gay Gordon-Byrne has been their executive director and the public face of a lot of the state-level fights. iFixit provides the cultural visibility, the Repair Association does the legislative grind.
And they've been effective.
They've been remarkably effective for a movement that didn't exist as a named thing twenty years ago. But effectiveness doesn't mean every argument they make is right, which gets us to Daniel's third question.
Before we get there — manufacturers. Who's real and who's performing?
Framework is the obvious genuine case. The entire company is built on modularity and user-replaceable components. They publish schematics, they sell every part, they design for repair from the start. That's not compliance, that's identity. Fairphone is the European equivalent — modular smartphone, fair materials, long support life. Valve with the Steam Deck is interesting because they didn't have to do any of it. They published repair manuals, partnered with iFixit to sell official parts, and designed the Deck to be opened with standard screws. There's no regulatory pressure on gaming handhelds specifically. They just did it.
Apple is the more complicated story.
Apple went from actively opposing repair bills — lobbying against them in multiple states, arguing security and safety risks — to launching Self Service Repair in twenty twenty-two, to publicly backing California's Senate Bill two forty-four in twenty twenty-three. That's a whiplash arc in about eighteen months.
Conviction or regulatory arithmetic?
I think it's mostly arithmetic, but arithmetic with a strategic read. California's bill was going to pass with or without them. By endorsing it, they got a seat at the table on the specifics — what parts had to be available, what documentation, what pricing. They also got ahead of the "strictest state becomes national standard" dynamic. If California was going to set the de facto national rule, better to shape it than fight it. But I don't think that makes it fake. The outcome — parts and manuals available to consumers — is real regardless of the motive.
Microsoft is the other interesting reversal.
Microsoft moved after a shareholder resolution, not a law. In twenty twenty-one, As You Sow, a shareholder advocacy group, filed a resolution asking Microsoft to study the environmental and social benefits of making devices more repairable. Microsoft agreed to commission an independent study, and the study came back saying — surprise — repairability reduces carbon emissions and e-waste. By late twenty twenty-one they'd committed to expanding repair options, and they've since released parts and tools for Surface devices through iFixit. That's a different path — investor pressure rather than legislative — but the result is similar.
Samsung is the cautionary tale.
Samsung announced an iFixit partnership in twenty twenty-two — official parts, repair guides, the whole thing. It lasted about two years. In twenty twenty-four, the partnership ended. iFixit's side of the story was that Samsung imposed restrictions that made the program unworkable — high parts prices, bundling requirements that forced you to buy a screen and battery together when you only needed one, limits on how many parts independent shops could order. Samsung's side was that iFixit wasn't meeting quality standards. The truth is probably somewhere in the middle, but the takeaway is that these voluntary partnerships are only as durable as the manufacturer's enthusiasm, and enthusiasm wanes when service revenue is at stake.
And John Deere is the bad-faith endpoint.
Deere signed a memorandum of understanding with the American Farm Bureau Federation in twenty twenty-three. Promised to make diagnostic tools, manuals, and software available to farmers and independent repair shops. It was hailed as a breakthrough. Then in twenty twenty-four, the FTC sued Deere anyway, alleging that the company had engaged in unfair practices by restricting repair access and that the MOU didn't actually deliver what it promised. The FTC complaint pointed to high software subscription costs, restrictions on what the diagnostic tools could actually do, and ongoing barriers that made the MOU effectively hollow. That's the lesson: voluntary commitments without enforcement mechanisms aren't worth the paper they're printed on.
Which is why binding law matters. But binding law has to draw lines somewhere, and that's the hard part. Where's the legitimate limit?
Let's take the safety cases first, because some of them are real. Lithium cell puncture is not a hypothetical. A punctured lithium-ion cell can enter thermal runaway in seconds — temperatures over five hundred degrees Celsius, toxic gas release, fire that doesn't go out with a standard extinguisher. If you're prying a swollen battery out of a phone with a metal spudger and you nick the pouch, you can have a very bad day. The residual voltage in a switching power supply — which Daniel just dealt with on his monitor board — the primary-side capacitor can hold a charge of several hundred volts for days after it's unplugged. Touch the wrong point and you get a shock that can stop your heart. EV traction batteries operate at four hundred to eight hundred volts. Airbag pyrotechnics are explosive devices. Medical devices — pacemakers, insulin pumps — a failed repair can kill someone.
These are not manufactured excuses.
They're absolutely not. And a manufacturer who sells a repair kit that enables a consumer to open a device containing these hazards is taking on real liability. The question is whether that liability is manageable, and on that point, the car analogy is strong. People replace their own brake pads. Brakes are safety-critical — if you do it wrong, you can kill yourself and others. And yet we don't ban auto parts stores from selling brake pads to consumers. We rely on a combination of clear instructions, warning labels, and the fact that most people who aren't competent to do the job know they aren't competent and take it to a professional. The legal system did not collapse.
So why is a phone different?
Partly it's not — the liability argument is often pretextual. But partly it's that the risk profile is different. A failed brake job is usually obvious before catastrophic failure — you hear grinding, the pedal feels soft. A lithium cell that's been slightly damaged during a DIY screen replacement may not show symptoms for days or weeks, and then it lights up in someone's pocket. The failure mode is less detectable and the consequences are more sudden. That said, the right regulatory response isn't to ban repair — it's to mandate clear warnings, design batteries to be replaceable without prying, and let consumers make informed choices.
The security cases are trickier because they're mixed.
Parts pairing is the flashpoint. When you replace a component in a modern device, the device checks whether that component is cryptographically authenticated. If it's not, the device may refuse to work with it, or degrade functionality. Apple's been the most aggressive — replacing a screen on a newer iPhone can disable True Tone and auto-brightness, even if the replacement screen is a genuine Apple part taken from another identical phone. The device knows it's not the original and penalises you.
And the justification is security.
For some components, the justification holds. A biometric sensor — Touch ID, Face ID — that authenticates payments and unlocks the device, there's a real argument for cryptographic pairing. If you could swap in a compromised sensor that captures and exfiltrates fingerprint data or facial scans, that's a genuine attack vector. The pairing ensures the sensor is the one Apple installed and trusts. But a battery? A screen? A camera module? None of those process authentication data. Pairing them doesn't protect anything — it just prevents third-party repair.
And the industry's mistake was using the strong argument to defend the weak position.
By lumping everything under "security," they eroded their own credibility on the cases where security actually mattered. Oregon's right-to-repair law, passed in twenty twenty-four, was the first to directly address parts pairing. It bans the practice for most components but explicitly carves out security-related parts. The twenty twenty-six model legislation from the Repair Association follows the same structure — parts pairing is prohibited unless the manufacturer can demonstrate that the pairing is necessary for a specific security function. That's how well-drafted law draws the line.
The certification problem is the boring one that nobody talks about.
And it's legitimate. If a device has an IP sixty-eight water resistance rating, that rating was certified in a lab with the device assembled to factory specifications. Nobody can guarantee that a home reseal restores that rating. If you drop the phone in a pool after a DIY battery replacement and it dies, whose problem is that? Same with RF type approval — if you replace a wireless module with a non-identical part, the device may no longer comply with its FCC or CE certification. The manufacturer doesn't want to be on the hook for a device that's technically operating illegally.
But again, we solve this with disclosure, not prohibition.
Right. The honest manufacturer position is: "We'll sell you the part, we'll give you the manual, but we're clear that the IP rating isn't guaranteed after you open it, and you assume that risk." That's reasonable. The dishonest position is: "We can't sell you the part because you might break the IP rating." That's pretext.
How do you tell the difference?
A few tells. If the manufacturer claims a safety or security risk but refuses to specify what the risk is or which component creates it, that's a red flag. If they oppose a bill that already has cybersecurity carve-outs and tiered access provisions, they're not worried about security — they're worried about service revenue. If they claim independent repair will lead to widespread battery fires but can't point to data from jurisdictions where repair is already legal, they're bluffing. And if they support repair legislation in one jurisdiction while fighting it in another, they're doing arithmetic, not principle.
Tiered access is the other thing well-drafted law handles.
The EU directive does this. It distinguishes between consumers and independent professionals. Some repairs — the ones involving mains voltage, high-capacity batteries, safety-critical components — are restricted to professionals. Consumers get access to the simpler stuff. Documentation and parts mandates apply to both tiers, but the physical repair permissions differ. That's sensible. It acknowledges that there's a line without using the line as an excuse to block everything.
So has any of this actually worked?
France introduced a repairability index in twenty twenty-one — a mandatory score from zero to ten on smartphones, laptops, washing machines, and a few other categories. Manufacturers self-score based on criteria like parts availability, price of parts, and ease of disassembly. The scores are public, displayed at point of sale. The EU is moving toward a similar system at the union level, and the battery regulation — which applies from twenty twenty-seven — requires that portable batteries in consumer devices be user-replaceable without special tools.
But are the metrics being gamed?
Some evidence of it. Manufacturers can score well on the French index by making parts available but pricing them so high that repair is economically irrational. A screen that costs ninety percent of a new phone technically satisfies the parts-availability criterion but doesn't actually enable repair. There's also the question of whether the scores influence purchasing — the data is mixed. Some surveys show consumers say they care about repairability, but price and features still dominate actual buying decisions.
So the scores exist, but the behavioural shift is unproven.
The behavioural shift at the consumer level, yes. At the manufacturer level, the shift is more visible. Devices are getting more repairable, but it's hard to disentangle regulatory pressure from other factors. The EU battery regulation's user-replaceable requirement is forcing redesigns — that's a direct causal link. But some of the improvement we're seeing is companies anticipating regulation rather than responding to it. And some of it is companies discovering that modular design has supply-chain advantages — easier to repair means easier to refurbish, which means better margins on trade-in programs.
Which brings us to the honest verdict. Where do you land?
The movement has been remarkably successful at changing the legal landscape in a short time. The automotive template worked. The state-by-state strategy worked. The EU directive is real and binding. Parts and manuals that were unavailable five years ago are available now. That's all genuine progress.
And the overreach?
The overreach is mostly rhetorical. The claim that all repair restrictions are indefensible and that every device should be fully user-serviceable — that's not serious. There are real safety boundaries, real security boundaries, and real cases where the certification chain matters. Pretending otherwise gives manufacturers an easy target. The stronger version of the argument — the one that's actually winning — is that restrictions should be proportionate, evidence-based, and not a pretext for capturing service revenue. That version is harder to dismiss.
The steelman of the manufacturers' position is that they're caught between competing obligations. They're required to make devices safe, secure, certified, and durable. They're also being told to make them repairable. Those goals can conflict, and they bear the liability when something goes wrong. The best version of their argument isn't "we don't want to" — it's "here are the specific conflicts, let's solve them together rather than legislating in ignorance of the engineering."
And the reply that still holds is that we've been here before. Cars had all the same conflicts — safety, liability, certification, security — and we worked it out. The sky didn't fall. The difference now is that the manufacturers who are resisting have a service-revenue line item they're protecting, and that's not a safety argument. It's a business-model argument dressed in safety clothing. The laws that are passing are getting better at seeing through the clothing, and that's the real story — not that right to repair won, but that the legal frameworks are getting more sophisticated at distinguishing the real constraints from the fake ones.
One open question I keep coming back to: the EU battery regulation requiring user-replaceable batteries from twenty twenty-seven — that's going to force a hardware redesign across basically every smartphone on the market. We're going to see whether the industry fights it to the last minute or actually embraces it. And if they embrace it, whether they try to implement it in the most annoying way possible — technically replaceable but practically awful — or whether they actually make it good. That'll tell us more about where this is going than any legislative scorecard.
The compliance quality will be the real test. A user-replaceable battery that requires a heat gun and three specialised pry tools technically meets some readings of the requirement. The question is whether regulators accept that or push back.
Thanks to our producer Hilbert Flumingtop.
This has been My Weird Prompts. Find us at my weird prompts dot com or email the show at show at my weird prompts dot com.
We'll be back soon.
