If you give something away to repair you allways have to trust them. If they are evil, there are almost unlimited ways to manipulate and plant spyware etc. Apple devices might indeed be harder to manipulate, but I doubt an expert will have much trouble with it, if he has physical access to it.
So in the end, by locking it more down, you are just locking yourself more to apple. If that's what you want, fine. But don't expect to be really more secure because of it.
I got confused about this initially too, but that's not relevant to the argument being made here. The argument isn't about repair -- that's something you can always avoid. The argument is about others (thiefs, etc.) messing with security-related components and bypassing security. He doesn't want that to be possible.
But it is and allways will be. You can make it harder, yes, but in my opinion the benefits are way smaller than the disadvantages.
Really the crux of this is whether this is best answered by the markets like you say, or by regulation. I see plenty of issues with it being answered by governments, but I have absolutely no faith in the market coming up with a good solution for this, as the players have too much power and the consumers are uneducated. I worry, of course, that governments will stick their oar in but paddle in the opposite direction.
A trusted hardware stack isn't one that's been locked down by a company, it's one I can modify and control.
In short, I trust that Apple will control any hardware I buy from them, and I trust Apple will always work in Apple's best interests. I cannot trust that Apple's best interests will always align with my own. I cannot trust that Apple's best interests currently align with my own.
> I think a well crafted system might be able to square this via dual chains that an owner could pick on purchase and appropriate manufacturer liability mitigations. So for example you could choose "only Apple may ever modify the trusted core of this device" or "anyone with the key can" and in each case Apple would utilize a separate root private key.
Only if I could revoke the manufacturer's ability to modify my hardware.
> In the latter case Apple would also be relieved of all liability for any damage ever done due to a 3rd party Touch ID repair
This is obvious.
I agree with the rest of your post.
There is a distinct pattern in the tech and engineering world of people who assume that what they want is what everyone else wants, and if they don't want that they don't know what is good for them or they've been duped.
Then they get presented with a cogent argument like yours and realize their neat world view isn't so clean cut after all.
Why doesn't encryption sidestep this issue for you?
The issue here is what happens when someone gets a hold of your phone and tries to get your data off it. This is NOT even remotely the same threat as someone secretly getting your phone, secretly opening it and installing a backdoor touchscreen reader, then sealing it back up perfectly and leaving it somewhere for you to find it with no trace without you realizing anything has gone wrong. The latter is far harder to pull off and far less likely to occur. Even CBP officers who take your phone wouldn't do this sort of thing. I don't really know of any evidence that this is the threat that Apple has been trying to mitigate, let alone that they have had any success in doing so. It's a non-argument as far as the discussions about Apple's existing security measures go.
As far as I can imagine, the only theoretical user benefit comes from being able to protect the device even with user cooperation, assuming he's being forced or a danger to the product. However if it's possible to circumvent the security mechanism anyway through a security bug in the implementation, this becomes a moot point. The consistent availability of jail breaks for iOS shows that the system is inherently insecure and considering the vast attack surface that's not a surprise either. There is no guarantee against the stack having been compromised.
The real leverage from the locked hardware is very much on Apple's side - they force the user into their ecosystem after purchasing the device to apply arbitrary restrictions and extract additional revenue from purchasing applications or media.
Regarding Touch ID, the sensor input (fingerprint) has been shown to be forgeable without professional tools shortly after each revision was released. The technology is insecure out of the box and my original point applies here too. You'd have to argue why the user shouldn't be able to trust the new sensor after authenticating with the primary method (pass code) and how that substantially improves device security.
Liability is btw. hardly a concern, the EU has had implied warranty that doesn't get voided by 3rd party repairs for a long time. Damage from repairs isn't really any different from just dropping the device and thus can be handled equally.
Point (2) is not well justified either. For once every battery has a connector as it's fragile and dangerous - including Apple's. You can't run it through the reflow oven with the main board and you generally don't want to add it to the assembly at the same time. The argument about repeated connector uses is absurd, we're talking about very few replacements in the product lifetime. There are plenty of cheap low footprint options that are specified for at least several uses and almost(?) every battery uses one of them.
There have been high end phone that supported toolless battery swapping in the past without showing huge tradeoffs for it. Going beyond toolless many phones can be fixed without being destroyed in the process - no tradeoff whatsoever. Not using excess amounts of irreversible glue or substituting some with screws goes a long way. It can be assumed that the actual reasoning is more along the lines of not caring to save some marginal cost or deliberately preventing repairs.
Using less glue or a slightly different design is no dead matter, other means are hardly comparable to manufacturing the high tech components. For recycling it is fundamentally important to be able to separate the components easily and any extended lifespan is vastly more beneficial than recycling to start with. The presented argument about any environmental advantage for unfixable designs is exceptionally weak.
1: Security. The fact is that there is an absolutely enormous (and growing) amount of value in a fully trusted hardware stack. Apple has taken a lot of flack from "right to repair" over issues like not being able to "repair" the Touch ID, but I haven't seen any great answers over how to allow any random 3rd party to repair Touch ID without also meaning any random malicious 3rd party (including legitimate repair shops in some scenarios) could replace it with something hostile too. I actively do not want anyone but Apple to be able to mess with the hardware authentication chain on my iPhone, even though I do support legislation to (and have written to my reps in favor of) require manufactures to allow users to additionally add their own master keys for running software on top of it. Having the option to minimize the number parties who can alter your hardware is something that has value.
I think a well crafted system might be able to square this via dual chains that an owner could pick on purchase and appropriate manufacturer liability mitigations. So for example you could choose "only Apple may ever modify the trusted core of this device" or "anyone with the key can" and in each case Apple would utilize a separate root private key. In the former case it'd be like right now where Apple never shares that, in the latter they'd offer signed leafs to anyone who asked so any 3rd party could make the repair (and sign off on it) instead. In the latter case Apple would also be relieved of all liability for any damage ever done due to a 3rd party Touch ID repair, though I suppose in practice they might still face legitimate PR damage (but that should be able to be minimized if it's clear enough that the owner willingly gave up protections).
2: Active/Dumb Matter ratios and dynamic flexibility: The other issue is that "repairability" in many of the formulations I see in HN and other tech community discussions comes with a lot of costs of its own that aren't address. A trivial, super common example is the battery. It's hard to go a single mobile phone discussion without somebody complaining about how they want the trivially replaceable batteries of yore. However, there are fundamental tradeoffs there in terms of actual basic physics. A user swappable battery has to have physical connectors capable of repeated usage. It cannot be soldered on. The battery itself must have its own safe case. The phone case must be altered to allow the swapping which requires structural compromises vs a solid piece. The battery must generally be a reasonably simple geometric shape, rather then whatever arbitrary shape the manufacturer wants to maximize volume in relation to other facts of the internal design.
All of which ultimately boils down to lots of extra matter that isn't battery or phone. There isn't any way around that with current technology. A smartphone using a non-swappable battery will always fundamentally have more single capacity, be lighter, or both, and of course can still just utilize an external battery pack if bulk is not an issue and more between-charge run time is necessary. Swappable has its own use case, but it's not a slam dunk at all nor is it a conspiracy nor are regular consumers "stupid" for preferring a different balance of trade offs.
This same sort of tradeoffs happens elsewhere, where more integration allows for fundamentally more efficient and superior performing designs at the cost of making it harder to repair. And this too is an environmental issue: across billions of phones, extra dead matter adds up all by itself.
----
I'm not saying that there isn't abuse, that complaints about John Deere and the like aren't justified, or that there is no room for better rules. But particularly with electronics a lot of these questions really do seem like they're best answered by the markets, and that legislation should first be focused purely on making those more efficient (by ensuring that externalities from total lifecycle costs are all in the price up front for example). Any legislation directly interfering should be treated cautiously and given a lot of careful consideration. I'm worried that much of the movement I've seen so far is too focused on its own narrow use cases rather then the general population, and not giving a sufficient level of thoughtfulness to how different areas of the overall world market face different use and threat scenarios. A lot of tech people have a bad habit of calling incredibly smart people "dumb" for sharing different priorities wrt technology, and failing in turn to consider whether their cases can be addressed without it being a zero-sum game. I don't want to see the proverbial baby thrown out with the bath water here is all.