Nothing except it being fantastically illegal. Unlike most other products, car owners are actually legally protected (in the US). The manufacturer can’t keep you from working on or modifying your car. They can’t require you to use their authorized service centers, or to use manufacturer approved replacement parts, and they can’t refuse to honor your warranty for doing any of that unless you caused the problem.
They could sue you for hacking to enable some paid feature. I don’t know if they would succeed. But they can’t (legally) just brick your car remotely.
Because I have actually worked in vehicle R&D, I would never buy a Tesla. But rather than get into that, which is a separate issue, I would suggest that the Intel case is rather different, as were the ss/ds floppies.
Intel’s problem was yield. A lot of dice had bad coprocessors. In those days lithography wasn’t what it is today, nor were clean rooms. Binning for the same part was necessary. Rather than dispose of parts with good cpus and bad coprocessors, design to disable faulty coprocessors with a laser and then sell them as the non-FP version. I think the independent coprocessor was a different design due to the need for I/O, so they couldn’t reuse dice with bad cpu and good fpu.
The same with floppies. During the production of the film used as the disc, imperfections could occur on one side but not the other. Solution: use the material with faults on one side for ss floppies. As with the 486 this meant less wastage and so a more competitive product.
The level of flaw at which a disc surface is rejected means that many discs would have zero flaws - it’s not worth trying to isolate down to the 8 inch or 130mm level. However, many would not and so using them as ds could result in data loss when you least expected it.
Tesla however…well, let’s just say I am unsurprised. It’s basically the IBM mainframe technique; deliver machine able to be uprated by a jumper change because it isn’t worth having two kinds of motherboard (since, unlike a 486, a defective motherboard can simply be reworked.) Ethics? It’s not as if buyers don’t know about it, but some of them, from their postings on tech websites, seem to treat Tesla as a religion.
Considering Tesla have done something positive here, it seems a touch odd to jump on them for their “evil” ways. My VW has a bunch of new features I added by tweaking the code, including some VW would require me to pay for if they did the tweaks. I spoke to one of the service guys and he advised they basically don’t care unless I brick my car, at which point they reset the car and I can start again. The potential for evil exists, but seems pretty limited, all up.
Now if you change the ECU envelope and blow up the engine, good luck with your warranty claim. If you unlock features for which Tesla charge $000s, you might get a slightly different attitude.
My response was actually meant to point out that I don’t think Microsoft is a good analogy to the Tesla case. If Microsoft built their own computers, they would be daft not to sell them pre-installed with Windows because then users would probably install another OS (or pirate Windows). I mean Apple doesn’t sell its devices with anything else (and will even make it hard for people to run anything else on it).
But still, yeah, I think it’s different: Some programmer who was employed or contracted by Tesla was paid to put in the switch to turn off functionality that a user would benefit from, so the cost of the product should actually be higher. On the other hand, Microsoft would have to employ or contract someone to install Windows on these computers which would likely be more expensive than a computer without it.
On the other hand, I could imagine Microsoft selling computers and selling them either as a “user-friendly” standard edition or a “hacker-friendly” premium-priced one. The standard edition comes with Windows pre-installed and somehow sabotaged/DRM-ed so you wouldn’t be able to install any other OS on it without breaking warranty. The premium one comes with nothing and no DRM. The premium one should actually be cheaper because less work has gone into producing it. Same thing with Tesla and, yes, I perceive it to be morally shady.
Wait. What? Intel was morally superior by spending more to physically damage the chip?
Surely we’re now into “your brain’s intuition doesn’t make moral sense” territory here.
Honestly, I don’t understand the whole “labour theory of value” business.
I understand the the instinct - it’s like “this book can’t be worth the money - it’s too light” or “this book was good until I learned the author wrote it in 3 months” or “if the other person made money on a deal, then I must have lost money”.
But it barely ever made sense, and in modern technologies, it doesn’t make sense at all.
How about we look at how the business practices affect the pricing and product (and the employees) and judge on that rather than trying to decide what the ‘correct’ price is? Especially when we’re talking about luxury goods here.
Who said anything about moral superiority? I said they repurposed parts that would have otherwise been binned. (See @Enkita’s post for much more detailed explanation on this.)
Your phrase “I’m not sure this is an entirely fair comparison” suggested (to me at least) that there was a moral problem with one practice and not with the other. (i.e. it was fair to excoriate Tesla and not Intel.)
If I misunderstood and you weren’t condemning Tesla’s practice (which I consider almost functionally identical to Intel’s), then I offer my apologies.
It’s not a fair comparison. Intel repurposed “bad” chips into “good” ones they could sell them at a lower price. (Something they still do today.) That seems like a win for Intel, and win for consumers. There’s no “software fuckery” at play here as there’s actual differences in the hardware.
Tesla is selling identical parts with the only difference being a bit set in software that is artificially limiting functionality (and they will be happy to DMCA you if you try to “fix” it yourself).
I’m no Tesla or Intel apologist but I am poking a hole in one of Cory’s false equivalencies.
Everything I remember from the era (albeit this would be from distributors, not insiders) was that while they did so, the vast majority of the chips so sold had never been tested (at least after the initial FPU manufacturing difficulties).
I remember being annoyed at the time because I felt their practice pretty much had the exact same modus operandi as IBM’s deal of unplugging the jumper. However, because they had a slight cover story by tossing a few failed FPUs into the batch and, more importantly, they physically destroyed the traces, they basically tricked our brains into saying “this is fine”.
I know hacking our instincts is good marketing. But I still get slightly annoyed when I see it working.
The problem with that is that some single-sided floppy drives had the head on the top side of the disk and others on the bottom side, so even a single-sided floppy disk had to be good on both sides unless you were very sure that it would be used only on one specific type of floppy drive.
I still own a special hole punch that will punch a write-enable hole in a 5.25" floppy disk casing so you get to use the other side. I wonder why I have kept this for such a long time. Perhaps I’m sentimental.
The traces had to be opened as otherwise the unused fpu would draw power and interfere with bus operations.
I doubt that the distributors understood how statistical process control works. Basically you do detailed tests on a few dice per wafer and fewer tests on all devices. If more than a certain number of fpus on the sample devices fail, at some point it becomes more economic to scrap all the fpus without individual testing. The spreadsheets we were using around that period to deal with yield analysis and tradeoff were pretty interesting.
Why yes, I do know an ungodly amount about ISO 9000 and I did do a thesis on statistical process control methodology!
Out of curiosity I checked to see what Wikipedia says, and I found:
"Single sided disks were coated on both sides, despite the availability of more expensive double sided disks. The reason usually given for the higher cost was that double sided disks were certified error-free on both sides of the media. "
I don’t think the author fully understands the manufacturing issues, but basically that’s what I said.
I would be very surprised if you were right because the underside of a single sided disc is not guaranteed to be capable of data storage.
Obviously though, since the medium needs to be coated on both sides to provide abrasion resistance, it makes sense to make double sided media and then use rejects where possible to make s/s discs. (In the days of film photography the plastic film was also coated on both sides though the non-sensitive side had a protective anti-halation layer. If this got scratched by the pressure plate it didn’t matter because it was removed during processing.)
I’m talking about enabling similar features to those that Tesla lock behind pay walls, self-driving functions, light and functionality changes, etc. Even ECU modification won’t result in warranty declines, by all accounts. VW and Audi charge thousands for some of these features too. I fail to see how it is different.
If you think that e.g. the different BHP ratings of different part numbers of VAG engines of the same capacity is simply due to an ECU change, I can only imagine that you are a software person who doesn’t know a lot about engines.
In which way? How are they polluting the world in order to sell more cars?
To be fair, it’s a standard practice where hardware and software meet. I’m no Musk fanboi, but to call out Tesla for this practice when it’s common from flashlights to defibrillators might be a a bit sensationalist.
I’ve been on the inside of multiple companies on the intersection between hardware and software. Inevitably, when it’s time to scale those products, the time will come when it’s more efficient to mass-produce the same hardware and differentiate by firmware. Differentiation is important to serve multiple customer groups, who often are willing or unwilling to pay more for certain features. BOTH GROUPS pay less for the product when common hardware can be leveraged for scale. It’s especially true with heavily regulated products like cars (or even more so, pacemakers) that have to go through expensive testing for every physical configuration change, but not for software differences.
