And yet, when I had to return my M1 Macbook Pro 3 weeks back because of dead pixels, the dude at Best Buy told me that in order to wipe it, they had to update the OS because the version of the OS that came installed had a bug that would make the machine unbootable if they tried to do a factory reset from that version. So it had to download ~3gb to update the OS, just to wipe itself.
I mean, bugs are most likely to be found in rarely used scenarios (like wiping) and also in the version of software that actually gets shipped with hardware, because it's got a non-negotiable tight release deadline. (Which is why many devices force an OS update from the internet before you can even use it for the first time.)
Of all the places for a crazy bug like this to be found, this seems like one of the least bonkers. :) And the excellent thing about Apple is that if you did brick it, Apple would just exchange it for you for free since it's their own defect.
I bought a non-M1 Macbook a decade ago and there were absolutely zero issues wiping it from stock. That was one of the first things I did, and everything performed as expected each time. How did things progress so far backwards in stability?
> I bought a non-M1 Macbook a decade ago and there were absolutely zero issues ... How did things progress so far backwards in stability?
We are talking about a system which just underwent a huge architectural change as well as a new major release of the operating system. I think it is unreasonable to expect the same level of stability as a decades-old architecture only a couple months after release.
You can see in this case the issue is apparently already patched. It seems only with the very first software to have ever shipped on the device would you experience a bug such as this.
On the contrary, given the huge risks being taken here, you would think that the recovery/reimage solution would be the thing that they would test the most!
On the other hand, downloading 3 gb to wipe doesn't really sound that bad concidering the way to do that used to be to download the entire OS and format the disk
Why would you download anything? It seems possible to have a DVD/BluRay or USB disk with the golden copy on it, or the OS itself should be able to restore itself to factory settings.
Apple leaks talked about how they maintained a “Marklar” x86/x64 release branch for MacOS for years leading up to the official launch. I was hearing about them during the G4 desktop era.
The M1 benefits from its relation to the silicon in iPhone and iOS being based on MacOS but there are major differences. It’s a revolutionary mobile computing platform. So far the biggest problems have been a few software glitches that can quickly be patched over the internet. This is a trivial set of issues to trade for the power / performance gains.
I’m looking forward to the newer 16” models coming out. I still need Windows on my computer for my workflows though so if the virtualization isn’t even beta quality I’ll need to be an Intel holdout for a while longer.
> I still need Windows on my computer for my workflows though so if the virtualization isn’t even beta quality I’ll need to be an Intel holdout for a while longer.
I’m running the beta of Parallels on M1. Windows works fine, though Parallels itself is still a work in progress. The issue is only licensing with MS, who currently don’t license Win for ARM independently. Then it’s just a question of whether you’re willing to enter a licensing grey area, similar to when you want to pay for content but it isn’t available in a legal format (thus, a person turns to a torrent). This is not an endorsement, just an observation.
Apple mentioned the secret x86 build of OS X during the keynote when they announced that transition [1], I think they alluded to it in one of last year’s keynotes too. That said, you would expect them to have booting, installation, and wiping as one of the first things built in the macOS-on-ARM porting process.
Ah, but once a deadline is set, it is very hard to un-set. Apple does not exist in a vacuum, it communicates its schedules to others and huge efforts are made by many people and companies to meet that deadline; it is not a trivial thing to change.
It's a matter of priorities. Apparently recovery and reimaging scenarios are not considered priorities. Personally, I want those to be absolutely rock solid.
That’s a conclusion that does not follow from your premises. I’m sure it was tested. A lot. But here’s the thing: when you ship anything of sufficient complexity where the quantities involved are measured in millions, there’s no such thing as a “small problem” or “edge case”.
Apple will test and fix bugs they find, as much as humanly possible, within the constraints of execution. You don’t just decide to launch on a whim, it’s baked in 3 years previously, with tens of thousands of people working towards that one goal; with supply ramps for multiple other companies arranged and enabled; with well-known public launch dates that can realistically only give you a few days wiggle room. The fact that they do this at all is a breathtaking success and a testament to the business - this applies to any such at-scale business, not just Apple.
So they’ve covered 99.9% of all boot issues before launch. As the CEO, do you go ahead ? Or do you miss the launch date, possibly invite legal action from your supply chain or worse, a critical manufacturer folds because of cash flow, and do you risk the reputation and stock price hit of a company as large as Apple “swinging and missing” in the press ?
I think it’s pretty clear what the correct choice is, and even though 0.1% of those millions of devices still adds up to a sizable number of complaints, you’re still way ahead of the game. And you get to keep that well-oiled machine moving forwards rather than stalling.
Apple has priorities. They may not be your priorities. If they differ sufficiently, you should go elsewhere, and if sufficient people agree with you and do the same, Apple will realign its priorities. I wouldn’t hold your breath though.
> How did things progress so far backwards in stability?
Because they progressed so far forward in security. Design is an exercise in priorities and trade-offs. The ability to wipe your Mac is now second fiddle to the ability to secure your Mac.
I disagree. The most important part of a car is the brakes. Likewise, the most important part of a complex programmable device is the capability to slick and reinstall to restore a known good working state.
Shipping something that cannot uninstall/reinstall cleanly is shipping a not even half functional base device. Ubiquity of Internet access is not a crutch that can be counted on at all times. The fact that modern hardware vendors are starting to assume that is disturbing, and likely another symptom of companies slowly trying to acclimate users to not really "owning" what it is they bought.
It sounds like the M1 literally cannot be re-imaged from external media though, because the internal copy of the OS is so locked-down. Which fundamentally seems like a very Apple sort of disregard for providing tools for people to deal with failure modes. It also ensures the people actually solving the problem and the people who have the problem are never in the same room and never talk about what's going on.
It can[0], it's just that the partition holding 1TR and the second-stage bootloader (which is almost but not quite macOS) isn't overwritten. Just like how PCs that store UEFI variables on the SSD don't overwrite those when wiping the OS (as OS installers are now smart enough not to wipe those).
As opposed to "you need a special piece of hardware" it's "you need a readily available piece of hardware with the right affordances".
Doesn't sound particularly bad, especially when compared to the situation in other industries (e.g. cars and all the special proprietary OBD stuff they do).
I suppose the T1, T2 and Touch Bar are, and with the M1 those parts are integrated on the SoC. But the runtime OS is still macOS and like on x86 that is not very iOS. The best description would be 'both', as the embedded platform is added instead of displacing the existing 'computer' platform. On M1 it is closer, and much more integrated, but still very distinct.
I don't think iOS has any similarity to booting an x86 CPU by faking an SPI Flash chip, compiling the IBB, ME and BUP code on-the-fly as needed and validating it externally before feeding it to the ROM inside the CPU for loading and microcode installation. But the T1, T2 itself is similar to iOS, as is the initial M1 boot stage.
>I mean, bugs are most likely to be found in rarely used scenarios (like wiping)
Some would posit that such scenarios should be some of the most important, because when you need to do such recovery, such scenarios can become immensely vital.
A dude at Best Buy didn't know the most efficient way to wipe one of the products they sell?
I can't say I'm shocked.
Apple's free Configurator 2 utility is much quicker than the Windows 10 "Reset this PC" functionality, and will factory reset an M1 Mac even if the system is unbootable.
The version of the OS that shipped with them has an unfortunate bug when reinstalling. IIRC it doesn't brick them, but it won't get passed the final stage when you create your account.
In any case, the solution would be "download a newer version of the OS and make some installation media like a usb".
Luckily, the recovery OS can still connect to wifi, so you can skip the whole "make an installation media" step. You just download the new installer and run it.
If you never ran software update and bought it the first day it was available then they are correct. If you ever updated the OS - even just once - then you had the patch for the recovery mode that had the bug.
But even with the bugged first version of macOS, the laptops are still recoverable - you just need another Mac with Apple Configurator 2 to get around the bug. Something you would think the Geek Squad should be able to handle (yes, that was sarcasm).
All you really needed to do was boot into recovery mode, launch disk utility then nuke the disk. If they want to re-install the OS, let them. I ended up taking my Air back to Apple because it turned out I really do need more than 16GB for RAM for a few things (shoot!) and I just nuked the drive. They didn't even bat an eye.
My new M1 is full of bugs and at this point after being downvoted here for stating the facts I do not care wether those are software bugs or hardware issues. It's the buggiest machine I ever had.
It even became not bootable in the first day. And I mean completely not bootable. You can't boot it from external usb stick. You can't do anything without 'another Mac' to 'revive' it as they call it. (see here for description: https://news.ycombinator.com/item?id=26203799 and I have described more of the bugs in this thread)
I can't say this about MacBookPro of 2009. This one worked. But the model of 2015 just have fallen apart.
(You can dig in my comments how. https://news.ycombinator.com/item?id=25809097)
This M1 became not bootable because their Disk Utility in recovery options become unresponsive at some point
and then it showed black screen containing exclamation mark with url where they tell you that you need
another Mac to revive this one. Somehow not this black screen with exclamation mark nor the web pages it points to did not mention that there is another boot option that is discussed here. Look for idiotic explanations for that fact around this thread. And see how I've been downvoted for stating that fact.
"Very fast" some say? I still wish to see it, because so far it feels just like the one I had before or slower! I am not talking about benchmarks, I am talking about day to day experience. Example: Screen shot (cmd-shift-4) takes ~1-2 seconds for cross-cursor to appear.
The saddest thing that all of my attempts to communicate about the problems are ignored or facing downvotes
instead of digging into it and trying to find out why it happened which was my initial intent and what I was expecting to happen at least here.
Instead I observe some bs like "oh, I never heard about any problems or let's ignore those problems because perhaps it's your hardware". But even if it is?
It's still an issue! BTW I ran apple diagnostics and it shows "no issues found". So I've seen no evidence for any hardware issues so far.
With such attitude when attack goes toward the messenger and not toward the problem It's hard for me
to expect a brighter future with that or that those issues would be fixed.
They perhaps would simply pretend that nothing has happened and everything is perfect. And at this point I am getting sick of this BS. If this bullying mechanism would not be removed from HN I see no reason to be reasonable here or trying to contribute
No one is ignoring you got a dud machine. The problem is that contrary to all the evidence posted otherwise, you keep insisting all m1 Macs have your issues. They do not. The many many reviews and other m1 users show that your experience is not normal (cmd-shift-4 is instant on every Mac I've ever used, including the m1).
The day 1 boot issue is fixed if you have applied any update since then. You never responded if you took your machine to Apple and see what their hardware diagnostic tool said. But, it's clear you're angry you got a dud machine, and instead of replacing you want to complain. That's fine.
It is always a possibility but even in that case it's the quality issue. Who cares whether it is software or hardware ?
It could be hardware but how anyone can be sure about it? At least Apple diagnostics says: no issues found. I have not seen direct evidence of that so far. All bugs are perfectly explainable by other reasons. But even if it's a dud machine, it's still the buggiest one I ever had. Why should I care it's software or hardware.
>You never responded if you took your machine to Apple and see what their hardware diagnostic tool said.
I am stuck with covid situation in another country. There is no Apple here, no "their hardware diagnostic tool" and there is no way to find out what it says. I need to work and this Mac is the only one I have, so sending it somewhere for two or four weeks is not an option. I need to work on my software I develop.
>you keep insisting all m1 Macs have your issues.
Where I have said it at least once? I was just sharing facts of my experience and my feelings about the situation. I suggested it can be a tendency because my experience tells me so but I never claimed it as a fact. Actually I expected to hear what wrong can be in the software part and I've got zero answers of that sort. This is what I call "not listening". Expressed shortly in: "oh it's just a dud machine, let's ignore it".
>The many many reviews...
I've seen many many reviews and rarely one of them manages to spot all issues that I am aware of. I am not talking about this M1, it's usually like this with many many products. They do not even mention problems I notice. This the quality of those many many reviews. Show me one that goes thoroughly throw all problems of some product. Usually they miss issues and honestly I get used to it. It's just me? I didn't think it's such a big secret that reviews usually have poor quality.
>your experience is not normal (cmd-shift-4 is instant ...
who measured it properly? what 'instant' even means? I didn't see one measure in answers here. Among other problems I have mentioned it's the only one that could be related to some 'dud machine'. Boot problems turns out to be known and not only on my machine. Garage band problem I believe related to bluetooth and some bug there and I suspect it is also not only on may machine, because I've seen many people voted up this, while (to my surprise) some downvoted even that. I mean why? Even if you do not have some bug or failed to reproduce it it's better to be voted up to check it and fix it. When I write software I would better check possible bug rather then ignore it which appears as insane thing to do.
But the main point is not those bugs I've mentioned. Those were presented just for example. The point there was that there are a lot of bugs that I _didn't_ mention. A lot of bugs are traveling from release to release and nobody cares to fix them like those in Disk Utility. They are not too hard to find. One of them made my machine not bootable. Others just annoying but less destructive. Still tbey are there. I am not talking about overcomplicated things. I am talking about basic things like formatting the usb drive, where Disk Utility fails to format it just because it fails to unmount it and it fails to unmount it just because it is in use by the very Disk Utility for example. I am talking about this level of bugs in very essential part of Mac software.
>it's clear you're angry you got a dud machine, and instead of replacing you want to complain. That's fine.
Yes, at this point I am angry because I paid money and getting shit doesn't work. But even this is not why I am angry. I am angry when BS starts to prevail instead of a reason, facts and logic and when it starting to prevail in forum like this one that really is upsetting.
>you got a dud machine
we still do not know that for sure.
>and instead of replacing you want to complain.
I can't replace it in current situation and if it's dud I'll have to work with what I have, at least until I'll fix my second machine. The thing is I do not want to complain honestly it doesn't help me too much. My intent was to tell the truth , to tell exactly what have happened, to make other people aware that such problems do exist so they can try to avoid them or or get them fixed. And I am getting angry that even that is impossible to do because of downvoting. There should be no downvoting on this site for stating the facts.
This is very different from how have described my situation and intents and unfortunately it could easily fall into category of 'not listening' But at least you've managed to tell what you think instead of downvoting and I am grateful to you for that.
"So it had to download ~3gb to update the OS, just to wipe itself."
Similarly, if one wants to replace Windows 10S with Linux, have to either have or create a Microsoft account and then have to upgrade to Windows Home via a large download ...
I both looked for and searched for a way to do so without a Microsoft account about a month ago - could not upgrade to Windows 10 Home without it and could not remove 10S without upgrading it first. I had downloaded a copy of Windows directly from Microsoft but could not use it because the Windows 10S was preinstalled by the OEM and did not come with a windows key code.
Do you really believe one of the most valuable companies in the world that just on-the-fly transitioned half their product line to a different CPU architecture, released their flagship desktop OS for the new architecture and thrown in a compatibility layer that works like magic for good measure, somehow doesn't understand how computers work?
Whatever they did, they did on purpose. The good and the bad.
Well, they did everything on purpose… with the purpose of satisfying business constraints – shipping a shiny end product within the deadline, and nothing else.
So they quickly repurposed all the iDevice stuff, minimally expanding the boot process to be more open and general-purpose-computer-like.
They could've built the SoC differently (with standard Arm components – GIC, SMMU, PL011… – instead of custom ones), could've used UEFI+ACPI, storing the firmware on a typical SPI flash chip, and so on. But that was not done because that stuff doesn't do much for the business objectives of the company, since what's being marketed is the full experience with macOS, not an interoperable standard platform for your favorite Unix. And so they went the easy, lazy, proprietary way.
Do you really think the engineers at Apple are gods incapable of mistakes? The MacOS division in particular has been a prime example of "don't upgrade until you have to" with how buggy each new update has been.
> Do you really think the engineers at Apple are gods incapable of mistakes?
Read the comment the parent is replying to. It implies that the cruft of legacy architectures is to support these scenarios and Apple didn’t realize that. Both things are laughably wrong and the parent pointed out why the latter is so ridiculous.
I love that in this thread we have one fanboi defending apple about this by saying that bugs happen and that they're often found in the least tested software parts (fucking reinstall? Not tested?)
And then we also have someone complaining they just did this on a whim and that they did it because they're basically gods and how dare you question them.
And then all of this conversation is based on the assumption of a bug someone heard about from a Best Buy employee. We’re arguing about a scenario (re-install wasn’t ever tested) based on a shitty game of telephone.
AP was clearly making the link between apples "on the fly" transition and software rewrites by newcomers to a system that will often make the same mistakes that the old system had already figured out solutions to.
Nobody claimed apple doesn't understand computers, thats your mischaracterization of the AP.
What “cruft” exactly that Apple removed do you think would have prevented this bug? Clearly it couldn’t have been too big a deal as this bug was quickly fixed.
Do you genuinely think “bugs might happen” is a legitimate reason for a computer manufacturer to leave on the table the absolutely massive perf and perf per watt gains Apple has demonstrated with the M1?
Indeed. Whenever you see {thing that shouldn't need to be there}, you should think very hard about why it could be there.
Sometimes, it's because {situation that no longer happens / is relevant}.
Sometimes, it's because {situation no one thinks would happen, but does with non-zero frequency in large deployments}. E.g. road debris forming a high force, sharpened spike directly into Tesla battery packs.
So this means modern macs essentially have 4 copies of OS X installed? One regular, 2 recoveries, and whatever runs on their management controller(Bridge OS? Does that still exist on M1 macs?) and the touch bar.
Seems to work out fine for them if it creates resilience and makes code-sharing/interoperability easy.
> the other 2 are iOS kernel for touchbar and controller
On Intel yes, as it's managed by the T2 chip which runs bridgeOS (which IIRC is based on watchOS rather than iOS, but it's XNU all the way down regardless). On M1 chips the controller functions have been integrated into the M1 chip, and I was under the impression that the touchbar was too; i.e. it's now controlled directly by macOS.
“ The Secure Enclave Processor runs an Apple-customized version of the L4 microkernel. It’s designed to operate efficiently at a lower clock speed that helps to protect it against clock and power attacks. The Secure Enclave Processor, starting with the A11 and S4, includes a memory-protected engine and encrypted memory with anti-replay capabilities, secure boot, a dedicated random number generator, and its own AES engine.”
iOS still has an XNU kernel. L4 is a different beast entirely and being used for a different role here.
Thanks! It’s interesting to read about the history of L4, which Wikipedia claims developed out of dissatisfaction about XNU predecessor Mach’s performance.
M1 Macs do not have a T2 coprocessor like Intel Macs. The Touch Bar on M1 systems is run by macOS directly. macOS also handles talking to the Secure Enclave.
I have a theory that one of the reasons why M1 isn't particularly well documented (if at all) is because Apple themselves are still getting to grips with it.
If their own tech support didn't know that does kind of support my theory
> Don’t ever format a drive of M1 Macs from recovery mode
> Because Apple doesn’t put a warning alerting that it doesn’t really format a drive and Big Sur installer is stupid and it doesn’t see it as a “clean” drive but it see the “Data” Volume and it tries to find a user to authorize the installation, but since the drive has no users, it fails.
I believe you could intentionally make it not bootable using Disk Utility if you delete the firmware partition (thus requiring a DFU mode recovery) but that is an expected behavior
Presumably it's a very reproducible bug during very basic operation but with easiness of getting 'healthy' supply of COVID while accessing another Mac in the repair-shop I would rather not play with it to find out more unless Apple sends me another 20 Macs and pays a lot for this headache.
Maybe it is easily reproducible, and that would certainly be concerning (at least until the issues are patched). But there's no reason to believe that based on one person's singular occurrence of the bug.
I hope you get a chance to repro the issue eventually so that we can all benefit. Eventually it will be possible to do the DFU restore from non-OSX systems using idevicerestore, so that should make the testing more convenient.
Honestly I do not observe any 'concerning' here, I've just get downvoted for sharing facts. I hope I would not 'get a chance' to reproduce it because that would mean my work would stop again. I could easily reproduce it with another Mac but I don't have one. And my previous one have gracefully failed completely.
( https://news.ycombinator.com/item?id=25809097)
>But there's no reason to believe that based on one person's singular occurrence of the bug.
Some times it depends which person it is. With all modesty I can observe tendencies with Apple bugs for some years now. There are very obvious for me reasons why those bugs appear and why they would appear more. I've got too much of bugs and I've started to talk about some of them because I just can't stand it. Apple should be Way better then this and for me it seems like no one really cares too much. I do care. Product should be done with some love for goodness sake...
> With all modesty I can observe tendencies with Apple bugs for some years now. There are very obvious for me reasons why those bugs appear and why they would appear more.
So then there are really two points you are trying to make here: M1 has a buggy boot process, but really more importantly is that Apple's product quality has been declining as a general trend.
Maybe the latter is true, I don't know (I've never owned an Apple computer prior to the M1). And I don't want to diminish that issue. But I don't think it is fair to assume everyone will have the bugs you had just because you believe their product quality is declining. Of course if you assume every bug you face is a typical occurrence then it will only reinforce your existing notions about the product quality, without really helping anyone else to get informed about the true state of the product.
>without really helping anyone else to get informed about the true state of the product.
I believe this is exactly what I am trying to do:
Helping anyone else to get informed about the true state of the product.
And I want it to be fixed and gladly report about it.
> What you then get is every bit as good as regular 1TR, with one significant exception: you can’t set the system security state using the Startup Security Utility. Apple explains that this is because “LLB [Low-Level Bootloader] doesn’t lock an indication into the Boot Progress Register saying it is going into recoveryOS”. But for all other purposes, this is just as good as 1TR, and is identical.
What does this mean? Do you have more rights than you usually would, less rights, or the same?
What was the rational for Apple making the M1 Macs so difficult (for lack of a better way to say it) than the older macs. Example is currently you can't easily boot from an external drive [1] a feature that many people found very useful. [2]
The answer to the question is not 'oh sure for security reasons'. I am wondering how making it so difficult for what I would call the majority of power or experienced or even regular users to do certain things on an M1 Mac. That is what is gained (so what is it?) vs. what is lost.
Who threat is the audience protected against by doing it this way?
Or is this just an artifact of the M1 Chip in some way and not something added after the fact?
Hardware Root of Trust means you have to assume an external boot device is compromised. Apple have their mechanisms for securing each step in the chain which is they enforce it out of the box. For 95% of users this is fine as it's only advanced experts who would need to use other booting technology after disabling HRoT chain. Apple makes it clear in big bold letters that you're compromising security by using alternative booting technology.
> Apple makes it clear in big bold letters that you're compromising security by using alternative booting technology.
Can you give an example(s) of how you are compromising security?
If I make my own boot drive and boot from it how am I compromising security?
I can think of scenarios like the following but the way I see it this would be a corner case:
"I somehow social engineer making a clone of a person's (say my wife/coworker (not computer literate)) hard/ssd drive. I then figure out a way to make their machine only boot from that drive by simply choosing that as the startup disk. I have installed some malware on that external drive that grabs or does something nefarious". In that case though with that kind of access I could have done similar on their internal drive since I have access to it already."
My point is why do we assume that the external boot drive is compromised if it is being created by the person who owns the hardware? This seems like some kind of government or large company protection applied to the entire apple customer base. Why couldn't the device ship with the ability to turn off this protection?
HW RoT prevents you from being able to replace OS binaries with compromised versions. How are you going to prove your external boot drive isn't compromised without HW RoT? HW RoT reduces the attack surface available and gives a more secure baseline but sure it can be compromised. But an external drive is the Wild West. This is the reason black sites don't allow support people to bring their own tools onsite as you have to assume they're compromised.
But assuming that the internal disk data is encrypted and unavailable to the external OS, what does it matter if it has modified kernel modules? The worst it can do is scribble garbage over the cipertext, which is unfortunate but not a security risk.
Apple traditionally advocated discoverable (without the documentation) features, particularly for cases where the user might be experiencing a problem.
It is fascinating how "good UX" has evolved over time.
We seem to collectively forget a lot of things with the Mac experience. You know how many troubleshooting required to zap the pram? Option-cmd-p-r on startup. Was that good Ux that people could guess on their own?
I'm familiar with Option-cmd-p-r as well as other enumerable startup sequences. That was kind of the joke with Apple. Good UX principles were very much a "do as I say, not as I do" kind of thing. ;-)
With Apple, all you get is either an Apple logo, or some obscure symbol when something does go wrong. It's a huge difference from PC BIOS that is far more informative and instructive.
It reminds me of the Douglas Adams quote: "The major difference between a thing that might go wrong and a thing that cannot possibly go wrong is that when a thing that cannot possibly go wrong goes wrong it usually turns out to be impossible to get at or repair."
Sure, it's just when I've got exclamation mark screen, apple didn't tell me that I have another boot option. Instead it told me to get access to another Mac with possible 'healthy' supply of COVID. If that is not 'hidden' then what it is? How would you describe it?
Absolutely. This is why we will see very soon detailed
instruction about second boot option instead of a black screen with exclamation mark. But if not, perhaps it wasn't a mistake.
One could only hope.
So the web page mentions Apple Configurator, firmware
and also for complete details provides link to
"Revive or restore a Mac with Apple silicon" where the process of reviving and restoration is described with all details
but
somehow the other option of booting the Mac is not mentioned to filter random non-technical users from technical users/experts?
I presume (following your logic) "Revive or restore a Mac with Apple silicon" doesn't mention it by the same reasons: To filter non-technical users.
So why the second boot option was not mentioned here[1] nor there[2]?
Mistake, I presume?
I've got my M1 Mac not booting because of the bug in Disk Utility and I really could've use second boot option. It would've saved me from the effort to reach 'another Mac' endangering myself with COVID exposure.
Now, Out of curiosity already. What is factually incorrect in the previous post? Or since everything is correct it's easier to simply downvote for bullying reasons by Apple employees? A small warning The more bullying the more chances of special film about the topic, so bullying would no work.
It leads me to believe there are more 'hidden' things there. Honestly I do not even consider it a 'personal' computer because of this. It's a 'terminal' at best and even in that role I have no idea how secure it can be.
Macs have always had tons of these (such as the command-option-p-r to reset the NVRAM, option to show the list of bootable drives, I can’t remember which one to invoke OpenFirmware, etc).
Well almost. To make it really work in my case I had to restart Apple Configurator, because otherwise it would not see the other Mac. You can imagine how many boots and attempts you need to discover this.
When I think about it now it could be I accidentally have entered to the second boot option and only this way it managed to put it in DFU. Hard to say now.
Almost every time I hear something new about the boot loader on these M1 Macs it's always something that on it's own would be a huge deal breaker.
It's not even the weird ISA that causes this, Apple's PowerPC macs used standardized firmware with documentation. All of the crap they do to prevent their users from tampering with their own machines just makes them unusable.
Not being able to quickly and easily whipe the machine when something goes wrong means it's practically unusable: There will always be accumulated state and it's much harder to safely get rid of.
If you deem Big Sur entirely unusable, I guess so. It has a few quirks that I don't like (mostly related to retarded fullscreen handling), but it has a bash terminal, can run Goland and is the only computer I have had any luck using my in ear sony head phones on.
At this point windows has a "bash terminal" and also comes with openssh preinstalled (which is all most people I know use their macs for anyway.)
Furthermore, Apple stopped updating the bash in OSX and at this point it's crazy out of date. They even changed the default shell to zsh because of this.
I think if the NSA want in they don't need to bother with stuff like this, and if Apple were forced to compel they wouldn't be documenting anything related to it
If we're spitballing tinfoil type stuff, I'd guess it's in the NSA's interest to have American OS's like MacOS be as secure as possible. I'm sure they have access to any of Apple's certificate chains, cloud data, and could decrypt a Macbook as well as Apple could, but I don't see known software backdoors being desirable.
you would think that a sane national security strategy is to maximize security of every citizen, but somehow nsa prefers to undermine everyone else's security to have a chance to hack anyone and everyone
That indeed makes sense, but sometimes one hand doesn't know what the other does. Also not all decisions are rational (actually it's the other way around). A better example for this is the FBI tho, since they really hate any security mechanism that plebs can use.
I'm not saying that the NSA has any involvement in this situation, but clearly their strategy is sometimes to undermine security. It's not their whole schtick, because of NOBUS, but still part of it.
That and they've always been fairly strongly linked in the past anyway - the kind of sexy counter-intelligence has always greatly benefitted from a good offense.
Edit: It seems downvoters know something about the patriot act that I don't, because last time I checked it still forced US companies to comply with requests from US secret services. If you know more please explain rather than downvote.
I would prefer one that works! How on earth they have managed to make this thing with 'recovery partition' to become 'not bootable'???
They officially say it can become 'not bootable' if the interrupt occurs during system update. BUT I've got it
'not bootable' without any power interrupts during the very first day.
After simple reinstall attempt without any power interrupts this sh..t didn't boot at all !!!
It showed black screen with big exclamation mark and url to support which basically says ... go find another Mac to boot this one ... (https://support.apple.com/en-us/HT211868).
Thanks Apple! Very nice advice with COVID situation
I would really prefer simple booting from external drive at this point.
This is the buggiest thing I EVER had. Even the screenshot doesn't work properly. It reacts to(CmdShift-4) after ~2!!! second delay while on previous Mac it was instant.
"Very fast" some say? I still wish to see it, because so far it feels just like the one I had before or slower! I am not talking about benchmarks, I am talking about day to day experience.
And there are bugs bugs and bugs. I have stopped counting honestly. Just another one as example: Playing something in Garage Band and going to settings gives dialog saying sound engine was not able to process ...
I never saw something like that before.
And of course enormous amount of useless dialogs with especially jumpy animation just to annoy you enough. It reminds micros..t more and more. I believe this particular 'dialogs feature' have been introduced in Vista? But Vista had another beautiful feature - erasing it and install something that works. Unfortunately here we do not have even this, because it simply would not boot !
Goodness, my feeling is: Can Apple make at least something these days that 'just works' ?
> This is the buggiest thing I EVER had. Even the screenshot doesn't work properly. It reacts to(CmdShift-4) after ~2!!! second delay while on previous Mac it was instant.
That's not an issue I've experienced. Screenshot works instantly for me (both CmdShift-3 and CmdShift-4).
> "Very fast" some say? I still wish to see it, because so far it feels just like the one I had before or slower! I am not talking about benchmarks, I am talking about day to day experience.
What are you running? I'm comparing my M1 air to two a year and old MBP (which admittedly has half the RAM), and it flies in comparison; both in perceived and actual performance. The MBP has consistently latency in the UI animations, which makes using it a pain; and the fan constantly blares at the slightest hint of CPU activity.
> Goodness, can Apple make something that 'works' these days?
For most people, it seems to be working for them. So it looks like they can make something that 'works', but you've apparently run into an extreme edge case alongside some poor luck.
> That's not an issue I've experienced. Screenshot works instantly for me (both CmdShift-3 and CmdShift-4).
I have seen it take a few seconds to appear on the Desktop, but I thought that was the new Mac OS thing: the little thumbnail of your screenshot appearing during that delay in the lower right of the screen, not a hardware-specific thing.
No, In my case I meant, pressing Cmd-Shift-4 takes ±1-2 secs for cross-cursor to appear before you actually take the screenshot. Now I've checked taking with screenshot of Cmd-Shift-3 and it 'Takes' it after some delay. Not instantly like before. It saves it with additional delay like you describe, but it's something else. Still dumb and annoying in my opinion but I believe you can turn it off somewhere in GUI or with the command in terminal.
The next screenshot you take can be faster. But not always. There is some loading happening I believe before the first one. Than it works fast but Then it unloads something sometimes and sometimes not. I do not think it's a hardware problem like some say.
For anybody wondering about the terminal command, you can disable the thumbnail and make it instantaneously save with: defaults write com.apple.screencapture show-thumbnail -bool false
I can't speak to the rest of the slowness. But that thumbnail sure is annoying.
I think the parent is saying you’re having bad hardware luck manifesting as glitches in software. I just tried what you described with Settings+GarageBand and was not able to reproduce on my M1 MBA.
> you’re having bad hardware luck manifesting as glitches in software
This was precisely my experience back in December.
I purchased a stock M1 MacBook Pro as well as a customized M1 MacBook Pro, so I could set up the OS as I wanted and migrate Data and settings once the customized Device arrived.
The first device behaved as expected: fast performance, responsive UI, cool operating temperature.
The second device, however, had issues with UI latency, beachballing Finder operations, and inexplicable slow downs.
I tried installing and reinstalling and remigrating the OS, eventually discovering that one or both of my users would be unable to even login to the system. Further analysis revealed that the communication between the T2 chip and the hard drive (or some aspect of the trust a computer system) was not operating properly.
I'm not technical enough to detail the exact problem, but it was most certainly a hardware issue or a hardware issue related to the incomplete erasing and reformatting of the M1 MacBook Pro's hard drive.
I returned both devices, waited a couple of weeks, and purchased a single customized MacBook Air.
This MacBook Air in my experience actually outperforms both previous M1 MacBook Pros. I don't have benchmarks nor do I have the other devices to compare directly, but it seems that this MacBook Air is much more ready for prime time then my mid-December MacBook Pros.
>Further analysis revealed that the communication between the T2 chip and the hard drive (or some aspect of the trust a computer system) was not operating properly.
From my experience there is no point dealing with issue on brand new Macs. The best is to send it back if you’re within the 14 days return window, or take it to a store to get it fixed. It’s not as smooth as it used to be because there is so much in M1 devices that’s different from previous generations, and support people are not necessarily used to it yet. But customer support is consistently great, all things considered.
The reverse is also true. People think if they experience a bug, then everyone experiences the bug and it's the most important bug ever.
Everyones setup and use case is unique. All the many interactions can lead to fairly unique bugs. Even if two people think they see the same bug, it could be two different bugs causing the same symptom.
Let's take the bugs around re-installing that have been popping up on HN lately. How many people buy a brand new machine and the first thing they do is wipe it and re-install? I don't personally know anyone who does that, it's certainly not something I've ever done with any of my Macs over the last 20 or so odd years. So while I think bugs around re-installing are important and need to get fixed, they are not something that a majority of users will run into this early in the m1's life.
In this case I doubt "two HN users doing daily development and production on a new Mac" are so unique as to be incomparable.
And arguing that certain users are in a minority such that they can be ignored - without hard proof mind you, and keeping in mind that knowledge work is squarely the demographic for people requiring high performance processors in the Apple space (or else they wouldn't push that angle so hard in their advertising) - when the functionality was stable on previous platforms much older than the so-called cutting edge, is justifying a regression in functionality by blaming user expectations. This is consumer hostile.
> In this case I doubt "two HN users doing daily development and production on a new Mac" are so unique as to be incomparable.
I would have argued the opposite. Every developer I know, myself included, tweaks every last aspect of their machine to their exact liking - even on a Mac.
And I didn't say they should be ignored, and said the opposite.
> So while I think bugs around re-installing are important and need to get fixed
The point is it's a bug impacting a subset of users and needs to get fixed. IMO, it's not one worthy of multiple HN front page posts, but here we are :)
Clearly it is important to the people posting the issues, and people upvoting. People, which are some of the main target audience of “pro” dubbed machines. The truth is, however, that there is nothing “pro” about Apple products anymore. Anything outside of the most mundane happy-path is often broken in strange ways. And the typical Apple drone response is: “Who does that anyway? I certainly don’t! So manage your expectations.”
Apple software is buggy and broken beyond belief. Hardware might be good, but that is increasingly beside the point, given how broken Apple software and UX are.
> The truth is, however, that there is nothing “pro” about Apple products anymore.
Anymore? I feel like a lot of people romanticize old Apple. For example, the first few versions of OS X were pretty bad. They did finally settle down, but then entered a yin/yang good/bad release schedule. There was time when installing a beta or even .0 version was almost guaranteed to lose data. Now, I tend to run betas pretty early on. There are some bugs, but they mostly work.
I don't see Apple products as any more or less 'pro' than they have always been. Once the software catches up, my m1 MBA doesn't seem anymore or less capable than my G4 Powerbook. TBH, the M1 MBA is already more capable because at the time software had already moved on to Intel while I still had my G4.
IMO, the biggest missteps that I would consider less pro on the portable side would be the 1 port MB, and the keyboard. Both of which Apple has moved on from, though the keyboard took way too long to address. I know people complain about ports, but I find usbc (unless there's only 1), the most pro thing about apple laptops because they can morph into anything a pro may need.
>The truth is, however, that there is nothing “pro” about Apple products anymore.
The truth this is a tired of chessnut that we've been hearing for 15+ years...
And yet, judging from their laptop choices at conferences, developer videos, and so on, over half of industry leading devs across many communities (from Node to Java, and from Go to Rust and whatever else) use those laptops.
>Apple software is buggy and broken beyond belief.
Not my experience. That said, you're free to use the non-broken Windows and/or Linux OSes!
I have not been saying that in the past. I am saying it now. I also use Windows increasingly more and more, and experience significantly less bugs, even in old Win32 APIs which still “just work”, unlike Apple’s many many frameworks (new and old) which are broken in different ways. Or its system software, which keeps getting rewritten in half-assed ways with missing feature, and then never revisited or fixed.
How many people buy a brand new machine and the first thing they do is wipe it and re-install?
Well, everyone who would sell their machine could go through that process. And if this (https://news.ycombinator.com/item?id=26204426) is correct then I would not call it an 'edge' case.
Besides It's a safety feature. You need it to work when nothing else works. It should be extremely reliable. And if it's not then the only 'reliable' feature of the Mac you have is DFU mode and booting from another Mac.
I don't personally know anyone who does that.
I personally don't know anyone who dealt with failing computers who doesn't do that. Besides you need partitions for multi boot configuration.
All your examples are not of people buying a brand new m1 and immediately re-installing. They are examples why it needs to get fixed, but right now likely has a low impact on the user base.
> How many people buy a brand new machine and the first thing they do is wipe it and re-install?
I do it on every brand new machine I buy; they come with Windows or Ubuntu, and I use another Linux distribution. And for Windows users, AFAIK for a long time the recommendation has also been to immediately wipe it and reinstall, to remove any shovelware which came with the machine.
>The reverse is also true. People think if they experience a bug, then everyone experiences the bug and it's the most important bug ever.
The reverse is not always true, because some people have certain knowledge and if they experience some bug they could analyse it and make an educated guess about the quality of the product and importance of the bug.
I personally have seen many bugs in Disk Utility. On my previous Mac I couldn't make volume to be named as I wanted in certain conditions no matter how hard I tried with diff scenarios. Disk Utility was simply renaming it to something else. Something unrelated. It was obviously a bug and it was reproducible. And this was not the only one. Another one I've seen that Disk Utility simply cannot format the drive. I've seen more and it's getting worse and worse. After seeng those I was wondering do they test it at all? Does one who tests it knew how to do it?
>People seem to think that if they don’t experience a bug, it’s an “extreme edge case”
And they're usually right. From millions of units sold, a small number usually has a bug, unless it's an OS bug independent of the hardware (and also not dependent on the use of a specific software setup, e.g. some ACE plugin or some app messing with some system).
1000 people having the problem and 100 reporting it and some news sites picking it up can make for a big fuss in blogs and posts, but it's still around 1 / 2000 people having the issue in a 2M production run.
I am mostly speaking about OS bugs, not hardware. The hardware side of Apple is still of high quality. It’s the software that is becoming increasing terrible.
1/2000 is still a big number, considering the Apple numbers. I doubt their hardware has so many issues.
> "Very fast" some say? I still wish to see it, because so far it feels just like the one I had before or slower! I am not talking about benchmarks, I am talking about day to day experience.
Really sounds like you had some hardware problem. For example, running a java app test suite runs 2x as fast on my M1 MBA than the 2017 MBP it replaced. Instant on from sleep, instant new Safari tabs, and terminal tabs/windows, are all examples of things that have greatly sped up my day to day experience. Add that all this in a machine with no fans, and not much bigger than my iPad Pro.
I do run into bugs here and there, but it's hard to tell if they are Big Sur bugs or M1 bugs. I wouldn't consider any of them showstoppers or occurring enough to be more than a minor annoyance.
It's like an iPad or phone now where it doesn't seem like it ever actually shut off. So when you open the lid it's running. Small difference, but incrementally better.
From my experience, an instant is much longer under OS X. My older Macs under OS 9 did indeed wake up instantly, but since then I’ve been happy if it takes a couple of seconds.
When Apple ran diagnostics, did they also say nothing was wrong?
It's completely possible to have dud hardware that is not caught by software. I'm about to send an old MBP to get its battery replaced because it's bulging, yet the battery health tool says it's fine.
As far as your m1, something is definitely wrong. It's not just benchmarks have shown that it's fast, but in day to day usage of many common programs. Your m1 experience is the exception, which likely means a problem with your machine.
I have noticed my M1 Mac Mini behaves a bit strangely when I have my external USB3 HDD plugged in. The mouse motion is stuttery, but unplug the drive and it is smooth as silk. There's evidently something low-level going on, probably USB error recovery.
Might be worth checking if you have any external hardware causing issues?
My main bug bear is my HP Z27 DisplayPort monitor disappears when it goes to sleep. Have to unplug/replug the cable to get it to come back.
That’s an old issue (also had it with a 2012 Mac mini). It’s interference from your USB 3.0 device/cable on the 2.4 GHz ISM band.
Potential ways to fix the issue:
• use a short USB 2.0 extension cable (like this¹ 6″/15cm one) to force the USB connection down to USB 2.0 (not ideal, but guaranteed to solve the problem)
• move the USB 3.0 HDD further away from the Mac mini and mouse
• try swapping out the USB 3.0 cable for one with better shielding
For more information about the issue, see Intel’s white paper titled USB 3.0 Radio Frequency Interference Impact on 2.4GHz Wireless Devices here².
> I have noticed my M1 Mac Mini behaves a bit strangely when I have my external USB3 HDD plugged in. The mouse motion is stuttery, but unplug the drive and it is smooth as silk. There's evidently something low-level going on, probably USB error recovery.
I've had similar sounding issues with a keyboard, mouse, a couple of USB audio devices plugged in via a hub to USB 3 (nothing high bitrate like a USB drive). I need to unplug one of the USB sound devices to stop this if it happens. I think all these devices are USB 2.0 at best.
I've had that problem with the bluetooth mouse and an Intel Mac Mini. I've also had similar problems with a DisplayPort monitor. Perhaps it isn't M1 specific.
>Might be worth checking if you have any external hardware causing issues?
I have no external drives at all.
Well I do but , because they are usb and turns out for this brand new Mac that should be 'everything I ever need' apparently I need to buy a huge adapter sizing like 'half of this Mac' to make it work with peripherals.
It doesn't even have sd card reader, not even micro sd card reader. To 'save space' and for better comfort I believe.
Somehow raspberry Pi manages to boot from such device. I wish the Apple engineers knew about such amazing feature of micro sd card.
If they just could figure out how this hi-tech technology of booting from micro-sd cards works they could do amazing things like booting without a need for 'Another Mac'! ...
If you open Garage Band, then record sound track then while your track is playing you would go to settings and
you connect a bluetooth speaker. Then in Settings->Sound section you would click on tab: Input. It works?
Try to stop playing track and start to play it again. All Works?
Try to select bluetooth for input and play again.
Works?
The simplest scenario is when BTSpeaker is already connected, you just run Garage Band. Open settins->sound>input and then press play in Garage Band.
Works?
It sounds like you're frustrated, and it sucks being in that spot. I'm not going to spend my day verifying your bug.
You've got an issue with bootability, with screenshots, and with audio all in the same system. Send it back. Get a replacement Mac, or get a PC for all anyone cares.
Whatever you decide to do, your experience so far has been anamolous compared with lots and lots of other folks.
>I'm not going to spend my day verifying your bug.
It's not my bug. It's their bug.
>your experience so far has been anamolous compared with lots and lots of other folks
Or perhaps those "other folks" just didn't bother to check it properly? Just like you didn't bother? Looks like rarely some one cares too much about the quality these days. I do care. There should be some love in making product and some respect to those who use it.
If I thought this problem is specific to this machine only I would not bother to write about it here too much. I have some experience and knowledge for goodness sake . I've seen Disk Utility not functioning too many times already.
I expected intelligent people here to be able to listen and take note and realise that it's not some stupid machine that doesn't work. I've seen too many bugs already and this machine is just another example of it.
You say it's a "hardware bug"? Does it really makes any difference? Even if it is? I think it's about quality, in that case software quality and another machine would come if not with the same problems then with similar problems.
>And what I and a lot of others are saying is that no, they are not experiencing similar problems.
Didn't you say before that you didn't try and not even intended to? How you can say after that that you do not experience it? I mean how would you know? It's pure guess.
How would you know about others? You simply can't.
I didn't state anywhere here that _all_ are experiencing what I have , I only presented facts of my personal experience and suggested that this is worth checking because I sense tendency. Too many problems for one computer.
Instead of listening to the facts people are trying to downvote and avoid seeing the fact or inventing some stories about some mysterious hardware problem that I didn't observe. Apple diagnostics didn't observe it too.
I do listen, there is always a possibility of hardware problem but even if it's a hardware issue it's still an issue speaking about quality. One thing is to have such possibility in mind and another immediately jump into conclusion that there are no problems.
Modern software is full of bugs, many of them people usually do not notice, but they are still bugs and they are still there. It's such a surprise for anyone around here? I mean any developer with some experience aware that barely something works if it works at all.
I state observed facts and get downvoted even for that on the site that supposed to be respectful to the facts and reason. I do not understand should we listen to BS or Facts in 21th century?
No one have pointed out to one mistake or were I was incorrect in my statements. They just downvote and trying to ignore what have happened in reality. I would not suggest a hardware problem in such case, I would suggest a 'mind' problem of those who have managed to downvote perfectly checked facts and reasonable arguments.
Apple diagnostics isn't perfect. My first Mac ever (17" powerbook) also passed diagnostics, but it still had a lot of "weird" issues that came down to a faulty mainboard. ¯\_(ツ)_/¯
I've had similar issues with Windows and had to resort to install media, so multiple redundancy on Apple is something I'd love to see make it to me as a non-Apple user.
> Goodness, my feeling is: Can Apple make at least something these days that 'just works' ?
This is absolutely true. And what do the "thought leaders" on Hacker News do when a person relates an earnest, first-hand experience? They down-vote it until it's barely visible.
Thank you. I it is really disappointing. I mean if this is happening on this site what we can expect from others.
I see downvotes for even stating facts.
Dang! Can we stop bullying by downvoting for facts and leave just upvotes?
I am really getting sick of it. Some topics became untouchable. You say something about free software and you get downvoted. Ok somebody doesn't like those ideas ok. But shutting up those who do is not the way. Or it shouldn't be the way isn't it?
I like to read opinions I do not like. At least I can consider that point of view. Make them gray ok, make them down ok, but let me read them at least! I'll go down to see the smartest people, it's fine, but can we stop deleting messages because of bullying?
I literally have to copy thread before reading, because I know some good points could be deleted. Or so hard to read because they are almost invisible.
Macbook Air M1 doesn't have sd card reader. To 'save space' and for better comfort I believe.
After a deep research I have found that Raspberry Pi manages to boot from such device.
If Apple engineers only knew about such amazing feature of micro sd card reader.
If they could just figure out how booting from micro-sd cards works they could do amazing things like booting Mac without a need for 'Another Mac' when recovery partition doesn't work!.
Somebody, please tell Apple about this possibility. It's amazing technology !
Apple has traditionally had really good boot options.
I've been able to put the OS on a usb stick and boot/run from that. Just hold down option at boot time, and select the drive you like from the list.
You could also boot from network boot volumes if they were properly set up.
I think you could probably boot from an SD card if you put it in a usb reader.
But the best - by far - is target disk mode. You boot mac #1 hold down the T key and it turns into a disk.
Then you can cable it to mac #2 and it will show up as a drive. Additionally you can hold down option on mac #2 and choose the disk from mac #1 as the boot drive, and then boot from it. This is really good when say #1 has no display or has a display problem.
I don't know how well this works with the new M1 macs, and also with T2 and/or encrypted partitions.
> Then you can cable it to mac #2 and it will show up as a drive. Additionally you can hold down option on mac #2 and choose the disk from mac #1 as the boot drive, and then boot from it. This is really good when say #1 has no display or has a display problem.
Does this work when #1's disk is encrypted (as it should be) on a Mac with the T2 security chip?
> Does this work when #1's disk is encrypted (as it should be) on a Mac with the T2 security chip?
That should work.
I'm pretty certain that when I attached an encrypted macOS drive in Target Disk mode, the Recovery System prompted me to provide credentials in order to decrypt the drive.
I have a MBP with a broken display that I've been reluctant to fix/replace due to some data I don't want to go missing on it ("repair" with these seems to always involve a new logic board..), I'll have to try this thanks! Though I'm not sure I even remember the password anymore...
> Just hold down option at boot time, and select the drive you like from the list.
This would not work on M1 Macs.
First: because they've changed keys .It's no longer 'option' key. It's 'power key' holding long enough.
Second: because when M1 Mac fails to load Recovery Options it would also fail to load this 'boot options choice screen'.
It's not like with previous Macs.
So you stuck there until you connect to another Mac and use Apple Configurator and pray it works. Iv'e spent 3 hours making it work and I Cann't reproduce the steps. Those listed in official site did not work for me. And Apple Configurator seems to have own bugs I've encountered.
So the traditionally the option-key menu is a firmware menu, while the recovery partition is loaded from the disk or ssd.
Are you saying the new macs don't have firmware separate from the ssd? With everything more integrated and optimized I would believe it, but it would be a big change with respect to admin and diagnostics.
Most PC bioses have options to control all of the boot options, e.g. disable booting from any external media, network, etc. The Mac has always been "special", but I think you can configure those things too.
No you can't, when Apple silicon Mac with M1 cannot load Recovery options. It shows exclamation mark and url to go for support, where you learn that it would boot only with the help of 'another Mac'
The comment I was replying to was about booting from an SD card. There’s no reason to think that would work differently from the USB booting option that already exists.
>here’s no reason to think that would work differently from the USB booting
Inability to boot my previous Mac from sd-card could be that reason? Or am I missing something? Because it was not possible to boot from sd-card last time I've tried.
What I’m saying is that there’s already the ability to boot from removable media (USB). If your problem with the existing solution is that it has some limitations, then it doesn’t make sense to demand SD boot, because it would probably have exactly the same limitations.
That procedure will create a boot partition on the built in drive and use it to jump to the partition on the TB3 drive - because the built-in bootloader doesn't have support for loading or even enumerating drives other than the built-in one.
That's just bonkers.