I love Linux but after 5 years of using Apple products exclusively, I have learned to hate everything except Apple hardware.
Linux on macbooks would be an amazing dream.
Linux on macbooks would be an amazing dream.
M1 performance in laptops is great but that's only been true for a year or so, for the last 5 years Apple laptops have been hot garbage.
As an owner of fully loaded 2018 mbp (i9/upgraded GPU, 32 GB ram) I can without a doubt say it's the worst premium device I've ever owned. The battery runs out on me after 1 hour meeting - and I have >100 cycles on it. The amount of heat and noise it produces is surreal, and that's after I opened it and cleaned up the vents (which clog easily) and added thermal pads to connect VRM to chassis (which helped significantly, prior to that the CPU would downclock so bad I couldn't use anything on my device after 15 mins in a Google meet connected to a 5k monitor).
Not to mention all the bugs I had to go through with it - wasn't untill 6 months after I purchased the device that I could actually use my 5k on my USBC monitor on full resolution - only started working after an OS upgrade.
And the keyboards failing being a known problem they replace out of warranty because they recognise their design sucks (luckily I use external keyboard 95% of the time).
Apple hardware was mediocre at best for the price they charge, up until M1.
In the mobile space, they are faster but as someone who switched from an iPhone to Samsung - I can't really say it matters. Phones are good and well rounded but nothing spectacular, hardware is on par with Samsung.
Again using the Mac/iOS combo is really nice so the software integration is next level, but considering their business practices I refuse to get locked in to the ecosystem, it's just too limiting.
And Linux on M1 would likely be subpar to any premium x86 device, Linux support sucked even on x86 Macs.
There were some good things, the displays were excellent, the touchpad is still the best in class, and the size/weight were excellent.
But usb-c only was step too far, the Touch Bar wasn't the right move for a Pro audience (i), there wasn't enough room for cooling the intel chips and the keyboard situation was farcical - it's the primary interface to a Mac (can you imagine the outcry if iPhones had touch displays randomly not working, doing extra touches etc?!).
I think Apple gets a lot of leeway because the 2008-2015 MacBook Pros were probably the best laptops you could buy.
Having owned a 2009 MacBook Pro which in my opinion was the best laptop I'd ever owned and never made me question the amount of money I spent on it. The 2016 MacBook Pro was the exact opposite (mainly due to the keyboard being so bad).
I'm glad Apple have come to their senses and course corrected. I do wonder though for people that have only seen the 2016-2019 era if they will bother to try Apple again...
(i) I understand it probably would've made it too expensive to produce but I think the Touch Bar would've gone down well on a MacBook Air where I would imagine there's a lot more hunt-and-peck typists that'd appreciate and notice what's being displayed on the Touch Bar. As a touch typist I never looked down to see the Touch Bar so it was a mostly wasted on me.
I'm one of those angry bastards, and I even own an M1 Macbook Air. The hardware is impressive, no doubt, but MacOS frustrates me so much these days that I cannot daily-drive it for my workflow. Plus, once you make the switch to Linux it's really hard to see Apple products as an upgrade anymore. You're giving away your freedom, and condemning yourself to paying $8 to manage your windows properly or $10 just to hide some statusbar icons. And when all is said-and-done, I can't move that statusbar to the left side of my screen... God it frustrates me endlessly. When I saw how Big Sur redesigned everything, I just gave up on the OS altogether. The thousand papercuts I feel on MacOS are reduced to a couple hundred on Linux, the majority of which I can automate away without worrying about some bigger company pulling the rug out from under me.
I really wanted Apple Silicon to be a barnburner for me, and I was hoping beyond hope that they would take the extra space savings to add an M.2 drive or an easier to repair chassis. At this point though, I think I'm contented to just stop caring. Apple courageously headed in a direction I'm not ready to follow in, so I cut them loose in exchange for all my sweet creature comforts. And how comforting it is.
They're currently only Intel based, but there's a marketplace where you can buy or sell just the mainboards once the AMD, RISC V, or ARM64 models become available.
I wonder if it will ever happen outside of very limited use cases.
Apple will just throw some crap random docs at them, best case scenario. They won't discourage it actively but they won't help them either.
The time budget for high quality engineers working on this will also be limited as there's no market that I think of, only itch scratching. Which generally takes you to "good enough/mediocre" not to "awesome for the general public".
It's exactly because they don't want thousands of Devs wasting time at all. Why reimplement something that already works well as a computing platform?
There is no good reason why they would want to do this. But it's not like they're blocking everything and locking it down. They just haven't published docs etc - which can carry a significant maintenance cost.
Apple provides the complete package. If you're homebrewing on top of their PCs, that's your problem and you're actually more of a nuisance to them. They don't actually get any benefit from you and you, being an advanced user, are very likely to have demands that are not representative of the general population and that also risk being difficult to accommodate. What could you possibly do that the All Mighty Apple™ can't do better?
Actually they don’t get any penny from me because I can’t run Linux on M1. I like MacBooks. I like OSX.
I’d love to buy one of those M1 and even keep OSX on it but I’m never going to buy a computer whose OS updates could be stopped at any time with no replacement possibility.
In the last decade or so, we've seen Apple flip-flop their designs and priorities between content creators and content consumers. If you can remember back to the 2000s, with the first-gen Mac(book) Pros, those were actually pretty remarkable computers, albeit very expensive and arguably overpriced from a performance per dollar standpoint. They were proper workstations -- if you could afford them. They also ran Windows via Boot Camp, with first-party drivers. Apple recognized the incremental value of supporting "advanced users" then, even if it was just a small niche.
Then the iPhone and iPad came on the scene, and quickly overtook not only the Mac market but the computing market in general. Content consumption, and App Store profits, thus became most of Apple's revenue, and their priorities shifted accordingly. Once Jobs died and Jony Ive took over, they tried to refactor their computing devices to be more like their consumption devices, changing the Macs from workstations to, I don't know what, designer furniture with screens? The Mac Pro (trash can) was obsolete within a year, the Macbook Pros went from fantastic workstations to experimental art featuring terrible keyboards and no ports and broken-by-design cooling.
Enter the present day. Jony Ive left (good riddance!) and they are again free to pursue the content creator market and cater to user demands, not the holier-than-thou vision of their annoying designer. The Macbook Pro is back with a vengeance (hopefully the desktop Pro soon), now featuring a vastly superior chipset, with a powerful & efficient CPU/GPU, ML chips, hardware encoders, etc. If only it had a USB-A port. Anyway, those are all signals Apple is starting to take the workstation market more seriously again.
The market has also changed in the intervening years. Any ol' laptop will suffice for most users. For some, iPads with a keyboard or Chromebooks are more than enough. For the remaining power users, what are their options? Surface, ThinkPad, Latitude, and now Macbooks again, finally. But that requires accommodations for "pro" workflows and tooling, and these early signs are encouraging, or they wouldn't have bothered much past the Macbook Air.
If Apple was able to support Boot Camp for Windows, there's no reason they can't provide similar support for other ARM operating systems, whether that's Linux or Windows for ARM. Parallels can already run Windows on a M1. Adobe has ported much of their stuff over to M1, Unity is halfway there, Unreal is thinking about it, Docker works... which is to say, the industry is excited about the possibility that Apple is going to start caring again. Hopefully. Maybe.
Personally, I'm also intrigued by the possibility of M1 in data centers, in rack mounts, with amazing performance/watt metrics. If that happens, surely Linux development on M1 will pick up alongside, and that may trickle back down to desktop users? One can hope. I don't think Apple has ever been in this situation before: where its own chips (as opposed to integration and design) risk disrupting the status quo. It's an exciting time. We'll see if it lasts...
If you're not running their software, you're not in their ecosystem and they can't upsell you.
Not anymore - it's now about selling services to get recurring income. E.g. App Store (to extort money from developers), Apple Care, Apple Pay and iCloud (from users) etc.
That is why they are locking down the Mac systems too, and making it a closed system like the iPhone / iPad platform. They have been doing this for a while now. The only thing that now differentiates an M1 mac to an iPad or iPhone is that the bootloader is not locked on the M1, and so you can still theoretically install another OS. (Ofcourse, the practical reality is that due to the closed box nature of the M1 Apple Silicons SoC, non-macOS OS can't exploit the full benefit of the hardware and will always offer subpar performance).
It is quite evident now that Apple has been locking down the macs:
1. The first few Intel Mac Minis allowed you some level of customisation of both the hardware (change RAM or HDD / SSD) and software (install other full featured OS).
2. Then came the Mac Minis with soldered RAM and SSD. You could no longer customise the hardware. Software was still customisable and you could still install other OSes. (Recall that Apple even offered free drivers for another OS, i.e. Windows).
3. The current generation of M1 Minis now doesn't allow you to customise both the hardware (everything is soldered) and the software (no other viable OS except macOS). Technically you can install other OSes, but the reality is that currently only crippled versions of Linux and xBSD is available and practically the only full-featured OS available for it is macOS.
The only difference between the M1 mac platform and the iPhone / iPad platform is the unlocked bootloader on the macs. And the only reason they haven't locked it yet is because they know that as soon as they do it, many will avoid it. They are just waiting for the M1 Macs to reach a criticial level of acceptance before they do so. (And till then, the lack of an alternative full featured OS to the macOS, currently works to their advantage).
> Asahi Linux is a good example of that
Asahi Linux is a distro built on existing ARM based Linux. Apple Silicon is built with ARM chips, so it is not surprising that ARM Linux could be ported to it. But it will be never be as full featured as macOS on the M1 because no device drivers exist for it on the M1 platform. While their aim to reverse engineer the hardware is laudable, I see it as a pointless endeavor as it will take years, and Apple doesn't want any other OS to run on it. From their perspective, any other full features OS hinders with the built-in plannned obsolescence built into the M1, and also cuts into the recurring revenue model built into the services in macOS (iCloud, App Store, Apple Care etc).
Asahi Linux contains device drivers for the M1 Macs, quite a bit of their work has been accepted to the Linux kernel already. GPU drivers are pretty close too.
Apple does explicitly allow other OSes to run on their hardware, they explicitly enabled that.
That's why I started this project backed by a Patreon. I knew this couldn't be accomplished via itch scratching; it has to be a job (not quite full time yet, but a good chunk thereof).
FWIW: the userspace GPU driver is now passing 90% or so of the GLES tests running on macOS; only the kernel side is not started yet. The ML cores are a niche use case and broadly not useful for anyone not doing machine learning on Linux (but some people are interested in them and have started taking a look). And power management already works to a significant extent; in the next week or so I'll be submitting v3 of the PMGR driver and v2 of the cpufreq driver for upstreaming, and starting work on SMC (needed for battery/charging info). And the security chip (SEP) has already been poked at; the mailbox interface to it is shared with other chips and has already been upstreamed, and the protocol on top isn't terribly complicated. It's mostly a matter of building a little driver to give userspace access to it, and writing some tooling to use it.
Unfortunately I'm a late pragmatist so I'm probably half a decade away from using your work :-)
https://tbkconsult.com/wp-content/uploads/2016/05/Crossing-t...