33
votes
Can Windows make the jump to ARM like Apple did?
I'm seeing a lot of news in my feed about Qualcomm chips approaching laptop performance, such as
https://www.theregister.com/2023/10/24/qualcomm_x_elite/
Will this turn out any better than the last few times Microsoft tried to break away from Intel? Would you want such a laptop? Will it wake Intel out of its complacency?
I'm gonna go with "probably not".
Apple can pull these maneuvers because they're a niche player in the hardware market, having a tight control over an ecosystem with users that are highly conditioned to re-buy their software periodically.
In the case of Microsoft tablets, they tend to be much less useful if you can't run an arbitrary old Windows program.
And if Lenovo puts out a 2in1 that is 80% as good as a Microsoft 2in1, but the Lenovo one does native x86 and the Microsoft one does not, the Lenovo one will devour the marketshare.
Less than Mac users re-buying software, third-party Mac devs (especially small/indie devs) tend to stay in near-lockstep with platform developments, architecture changes included. Many Mac apps had free updates adding ARM compatibility within the first 6-12 months, and by the two year mark they’d nearly all moved over.
There’s a few components to this.
First, Mac devs have traditionally been decent in this aspect, dating all the way back through the 90s where they went along with both the 68k → PPC transition and later the PPC → x86 transition.
Second, modern macOS and its predecessor NeXTSTEP were built to be largely architecture agnostic. This means for a lot of apps built with AppKit/UIKit, adding compatibility with a new arch is nothing more than ticking a checkbox and recompiling.
Third, because Apple has pulled off transitions like this repeatedly without issue, devs trust them to prioritize the new arch and make it work instead of starting it out as a side project that eventually withers on the vine and feel comfortable investing in the new arch.
While Windows isn’t too arch-dependent, MS has yet to pull off an arch transition, refuses to commit fully to new archs (with silliness like WinRT and restricting ARM to weak mobile devices), and has cultivated a conservative third party dev culture that leans on backward compatibility in perpetuity. A lot of things are going to have to change to make Windows for ARM work out.
Basically all modern operating systems are, including Windows NT and derivatives.
I think the main difference here is that Apple is a hardware company, and their software doesn't (officially) run on other vendors' hardware. If they want to change something, they can do it, and third party developers have to adapt to stay afloat whether they like it or not. The developers are "decent" because if the platform they target loses support, they're out of business.
That there are numerous manufacturers of the same basic platform consumers typically use for Windows, and numerous driver vendors out there means that there is a natural resistance to radical change that just doesn't exist for Apple. Windows has to play nice with the manufacturers and consumers in ways that simply aren't applicable when one entity sets the rules.
I personally think that this is the absolute most important factor. Apple commits to things. MS just doesn't do that, and that's not necessarily a bad thing. The flip side of never committing to something new is that you never have to totally give up anything old, and that's why a lot of IT folks LOVE Microsoft: stability.
I would actually assume that because Windows has been on x86 since before it was called x86, it probably is pretty architecture dependent. But they probably cleaned some things up in their previous, failed, attempt to move to ARM.
Alternatively, the entire architecture of Windows post-NT is designed to avoid such a lock-in. Who knows how many shortcuts they took in the last generation, though...
WIndows NT (the base for all modern Windows, including 10 and 11) was originally developed on the i860 architecture explicitly to avoid making it dependent on x86 specific quirks or features. There's been various versions of release Windows for MIPS, Alpha, Itanium, and PowerPC, in addition to the x86, AMD64, and ARM versions that we see today.
Windows NT was designed to be architecture independent from the beginning. They had it running on DEC Alpha quite early
https://virtuallyfun.com/2023/10/02/windows-nt-3-1-on-dec-alpha-axp/
and they’ve gotten it to run on quite a few different architectures since but haven’t had much of a serious commitment to any of them. It is nothing like Linux in that respect.
Microsoft has a translation layer for both 32 bit and 64 bit programs. You can already run any legacy windows programs on an arm windows install.
So I heard, microsoft did not properly implement the x86 concurrency model (unlike apple), so many applications do not work or spuriously crash. Is my knowledge outdated or incorrect?
Yea the existence of the translation layer is basically no better than me going
"Oh but I can run Office on Wine, I don't need Windows"
It might be true....but there's a lot of hidden "if everything works perfect" in that statement.
If there's a mission-critical piece of software that isn't going to get an ARM port anytime soon, that laptop needs to be leaps and bounds better than any other competitor's x86 product. "Oh your one app you use every day will crash twice an hour like it's 1998 but you get 3 hours more battery life" ain't gonna cut it.
I mean, you bring up something interesting. I use Linux for gaming, and after Valve investing in Proton, ~85% just work™, 5% works with some tweaking (because they barely work on windows anyway) and most others are deliberately broken by devs because of anti-cheat. In other words: most if not practically all games would work if it weren't for anti-cheat. You would think this would mean that Office or Visual Studio would work well too then, but this is not true at all. This indicates that its code is not very platform agnostic and breaks as soon as anything is not exactly 64-bit Windows.
That is pretty likely. Games are in some ways easier to to deal with as they generally don't do very complex system specific API calls outside DirectX calls.
I honestly don’t know. I do use windows 11 arm on my m1 Mac through parallels. I have yet to see an issue running 32 or 64 bit programs on it. Although I avoid windows software if at all possible, and if not I look for an arm compiled version of it.
You can, but the performance is atrocious compared to Rosetta 2. A big reason why are custom instructions Apple integrated into their ARM cores.
I wouldn't even say it's that.
Macs aren't really used for games, and that's about the only software the average person wants to continue using 10+ years after release. Everything else (browsers, productivity apps, creativity apps) is either web-based, evergreen, or open-source. There aren't really people still using Photoshop 8, they are all either on creative cloud or have moved on to open source alternatives that have long since surpassed Photoshop 8 in functionality. This is true regardless of what operating system people are on.
Audio software is a strong counter-example. People very much want to keep using decade-old VSTs, whether they're effects or instruments, and many of them are no longer supported.
The current situation with Apple Silicon is you have to run your entire DAW under Rosetta to use x86 VSTs. I assume this means you then can't use Apple Silicon-compiled VSTs. (I still use an Intel Mac at home; my work Mac is M1 though. It's definitely a much more solid machine than the TouchBar lemons, but the architecture change is an annoyance. It creates issues with Docker at work too.)
Some currently-supported plugins aren't even Apple Silicon. Yamaha VOCALOID comes to mind. V6 came out in the past year, and I believe the advisory is still "use Rosetta." And, of course, V4 is still the most widely used version, due to the most popular voice banks being for that version.
The exception is Apple's Audio Unit format (AU). x86 AUs can apparently run when the DAW is in Apple Silicon mode, as the OS has a translation process for those. AUs are fairly common, but certainly not universal.
There’s at least one community I know of that’s centered around keeping old PowerPC G4/G5 Macs functioning for the purpose of using DAWs, VSTs, etc from that era.
But the average person doesn't even know what a VST is. I was very careful to make that the scope of my comment. I'm well aware there are businesses and whole industries that are dependent on ancient Windows software.
Where I work we still use Access ‘98 occasionally since future versions of Office won’t open those databases.
It is amazing, you can install Office ‘98 on Win 11 and it works great. Clippy works right and all the lame ways it tries to take over your desktop still work.
Mind you, we still occasionally get a box of IBM cards, sometimes punched in unusual formats, where I work.
https://en.wikipedia.org/wiki/Punched_card
Sounds like you've never taken a glance at the screen at the dentist's office. Lots of medical software is write-once.
Even as recently as a few years ago I've seen old XP machines sitting around disconnected from the internet for running the X-Ray machines.
Not to say that these are going to be targets for new ARM Windows machines, but they're a small microcosm of the wider Windows ecosystem where lots of stuff like that exists.
I was very careful to qualify my statement with "average person".
The reason backwards compatibility is so important to Windows is precisely because there are tons of businesses and even whole industries that are dependent on ancient Windows software.
But to a typical consumer shopping for a new laptop? Unless they play PC games, they most likely do not care if their new computer can run 25 year old programs.
When this is your target, changing things up becomes a lot easier. You can realisitcally get by with only ~5 years of backwards compatibility, as Apple did by ending support for 32-bit software before rolling out Apple Silicon Macs. By deprecating crusty older APIs and 32-bit support beforehand, they limited the scope of what Rosetta 2 needed to support heading into the switch. I think that is a big part of why the transition was pulled off so much more seamlessly than we've seen with Windows and even Linux.
You're right. But I also think the "average person" market is also much, much smaller than the "medium-large business market" where this has some serious potential problems.
If Microsoft wants to make serious inroads on the hardware front, that's where they have the strongest advantage (central AD policy administration).
Which is why the average person continues to be saddled with an operating system that is a bizarre incoherent mess of designs from three different decades.
You can't get rid of the old control panel, because x device still installs a custom module for it and the company that makes the device disappeared 10 years ago. You can't replace the old and very limited explorer context menus for the same reason, so now we have two of them...
There should be better solutions to many of these problems, but big companies are averse to change.
For whatever it’s worth, if using PS 7, CS1, or CS2 on modern hardware without the added encumberances of something like WINE or a VM were an option, I absolutely would do so. Newer releases of Photoshop have added almost nothing of value while increasing bloat exponentially, and while alternatives (Affinity, Pixelmator, Krita, etc) are getting better all the time I don’t think I’ll ever be able to work as quickly in those as I can in 2000s era Photoshop.
I'd be very, very surprised if PS 7 hasn't been fully worked out in Wine/Proton right now. There are many people in your boat. It'd just be a matter of finding the needle in the haystack.
Oh I’m sure PS 7 works in WINE very well by this point, but as noted in the earlier post that still adds extra layers I’d rather not deal with. Really what I want is a PS CS1 that’s feature-locked, but evergreen with development being focused entirely on fixing bugs, efficiency, and compatibility with new OS releases.
And is there even value in it, when the majority of software is on x86 and Microsoft could simply expect hardware to improve? The M1/M2 machines aren't fast with surprisingly low power consumption because they are ARM. They are so at the same time as being ARM. Apple designed efficient chips, which they've been doing for awhile for their phones, and just so happens to use an ARM architecture.
There isn't really a reason that Intel or AMD couldn't also develop such a chip using x86. (Modern x86 processors already even use what is basically a RISC instruction set under the hood, which is abstracted by the x86 interface.) And they're definitely trying. The Alder Lake CPUs are a major step down that path, and use a similar setup of performance and efficiency cores.
It would help a lot of Intel and AMD managed to shift the greater public’s focus away from raw performance towards performance per watt. At this point computers have been way beyond fast enough for the vast majority of users for many years, and so for mainstream CPU lines it no longer makes sense to push for maximum performance at the cost of all else.
There are still of course some users who do need maximum performance, and for them Intel and AMD can have separate lines focused on that (similar to the HEDT product line that Intel used to have).
There is a possibility, but signs point to no. Apple can get away with it largely because they aren’t afraid of abandoning legacy software. But that’s practically Microsoft’s claim to fame. They’ve already spent a lot of time and money trying to get Windows on ARM off the ground and have had a giant flop with it not too long ago with Windows RT. Developers who make Windows programs are also tepid at best when it comes to getting their stuff running on ARM natively, and the performance of their x86 comparability layer is noted to be very slow so Qualcomm’s chips have to be considerably faster than apple’s.
In short, there are a lot of factors that have to come into play in order for this to take off, and it’s not likely that all of them will.
Qualcomm needs to offer support for their chips as well. From the article:
Qualcomm is the main reason why Android phones have such terrible support for updates and why Google has had to jump through so many hoops over the years to move as much as they can to Play Store updates.
A PC with only 3 years of support is dead in the water compared to an Apple offering.
Qualcomm is the reason why I am convinced ARM is not really taking over the Windows/Linux laptop space any time soon.
And if Intel is able to actually stick to their promised roadmap, then the benefits of switching to ARM are going to start evaporating in the next few years.
It's a big contributor. Not offering driver updates for their chips makes it basically impossible for device manufacturers to update their older phones beyond the support window that Qualcomm offers. Of course, the super cheap phones didn't even get updates for the support window anyway, but that falls on the manufacturers.
Why doesn’t Microsoft launch a totally new thing not called Windows? Drop all legacy concerns for it and build a forward facing platform without the baggage.
Microsoft puts a lot of value in existing brands. A recent example of this are all the things that were carrying the “Visual Studio” brand, which included:
They cling to brands to the point of causing confusion and only drop them if their reputation has been trashed, like how it was with Internet Explorer.
Like Teams (for home or small business) and Teams (for work and school). And OneDrive and OneDrive Business (aka Sharepoint). And Skype and Skype for Business. And Edge and [New] Edge.
Even Office and Office (for Mac). Outlook is/was so different from its Windows counterpart.
They are literally the worst at naming.
From personal experience, the "Visual Studio Code" thing works very effectively - basically everyone assumes VS Code is the Visual Studio, the first time around. I can't say whether it's a good idea, but it definitely does something.
Internal fiefdoms would cannibalize it. There is a reason why you have cmd.exe and powershell simultaneously.
I believe cmd is kept around for compatibility, I don't think it is actively worked on. All that effort goes into powershell.
But I guess that is the other problem Microsoft has. They come up with some cool new thing, but they have to keep the old thing around too so they don't break some ancient software someone's business relies on, and now you have two context menus in the shell.
There was evidentally a beef between the people who came up with powershell and the maintainers of cmd.exe, thusly powershell is its own entity and not integrated within cmd.
There is also terminal now too. But that can use the others, I think.
That's almost more of a wrapper around cmd, powershell and any wsl distro than another separate command processor (like more of an xterm than ksh).
That might be a good idea, if MS could be trusted to not put it out pre-maturely and actually support it long enough to gain traction. I have a feeling it would come out half-baked and die on the vine like the Zune and Windows Phone 7.
It still boggles me that the Zune came out only months before the iPhone. That is how far behind Apple they were in this space.
There have also been times when Microsoft has been ahead of the curve in some ways, but was unable to identify what needed to be done to make a new product category work.
For example, remember XP/Vista tablets/convertibles? In concept not all that different from iPads and today’s 2-in-1’s, but because they messed up on key aspects (e.g. needing a dedicated touch UI, styluses being optional, and battery life and fanless operation being the focus instead of performance) they lost that market entirely to iOS and Android tablets.
Because then they are practically starting from scratch competing with their old OS (they already have problems with that with people being stubbern to upgrade), Mac OS, and even Linux. Basically all OS's that already have established a foothold.
From a business point of view, that makes no sense when they have an already established OS that dominates the market already.
Nokia had the same business strategy. Having something that you already have a foothold in and becoming stubbornly entrenched on it will eventually lead to a new product overtaking your business.
Nokia hasn't had a 20-30 year old entrenchment that is almost a stranglehold on the business. I mean MS as one point had to bail out Apple just to prove they weren't a monopoly and get regulations put on them, and that was like 20 years ago (I remember in college people mocking me that I'd have to give up my Mac cause Apple was going to go under). And while the competition is a little healthier now, it still has an uphill battle to become the standard go to for most (hell, at one point in the past 10 years or so you'd see apple fans rumor that apple will abandon the mac cause it's not the moneymaker as iphones and ipods is what got apple popular. I'm glad they've been wrong so far). Windows has firmly entrenched itself and made it an uphill battle for anything else to even come close to dominating. It doesn't make sense to throw that away and give your competition a leg up by now no longer having to fight against that.
I mean, as some one who prefers Mac, I'd love to see MS do something like that cause it would give Mac more chances to get more used (and therefore more software made for it) but I really don't think it would be a good move for MS.
Nokia had a huge market share in mobile phones since the 80s. They were dominated by smart phones in in 2007 and rapidly lost market share after that. They relied on an entrenched product strategy and failed to back innovation. The company is worth a fraction today of what it used to be worth.
I don't think the threat is necessarily from Apple, it could be from anywhere.
They also don't have to abandon Windows to pursue a new strategy. Fragmenting their market share across multiple products hasn't really been an issue for them in the past - they have multiple product lines in competition with each other already.
I think a good but ultimately disappointing example of the innovation I'd love to see from Microsoft is what Mozilla was doing with Servo. I wish Mozilla had been able to keep that project running.
Exactly! In some sense they obviously cannot make the jump that Apple did because they have sort of tried and failed in the past. But they're also in a very different situation... The (kind of incredible) backwards compatibility is pretty much the entire point of Windows and the whole x86 platform. Apple has far more control over their ecosystem and can just break things at a certain point, but Microsoft kind of can't...
I do really want good non-Apple ARM hardware. The Apple Silicon Macs seem really killer, and I'd like something similar for the Linux world... But I do worry a bit about the transition. There's so much history and legacy software in the x86 world, and it'd be a shame if eventually the compatibility layers just stopped being supported, or if they weren't viable for running certain applications fast enough. It's probably not the most important thing, but I also just think about the huge catalog of games that we have on the PC platforms... You can (relatively) easily run games from 30 years ago on a modern day machine, and that's kind of incredible. I wouldn't worry so much about super old games because they'll have more limited requirements, but more modern games might be harder to run with an ARM translation layer in the mix... And maybe eventually they'll drop the special hardware considerations that make emulating x86 more effective on Apple Silicon processors too? Why take up silicon with it if all current software has moved to ARM?
The other reason is that I don't think there is anything inherent to the x86 ISA that requires it to be more power-hungry than ARM. A big reason Intel has fallen behind is because they got stuck on 14nm for so long. If they can catch back up to TSMC, I think they can put out chips with similar efficiency and the ARM/x86 distinction stops meaning all that much to consumers.
Nobody actually cares what ISA their laptp uses, they care about app compatibility, performance, and battery life. ARM needs to consistently deliver on both of the latter to convince people that the former is not as important. This is where Apple's new MacBooks have succeeded, though it helps that their x86 emulation is also way ahead of everyone else making the transition a lot smoother.
Other companies trying to make ARM laptop SoCs are still going to run into the problem of Apple being a node ahead of them because they tend to buy out TSMCs production cpacity for a good while when they're new.
The real advantage of ARM is competition. Intel and AMD have an enforced licensing duopoly over x86-64. It's quite literally impossible for anyone else to make an x86-64 chip. Their existing CPU architectures were iterated on from desktop computers, and that is an environment where power was, for all intents and purposes, free - the thermal characteristics were more of a limitation than the raw power draw.
I wouldn't be surprised if a new company, that made a new architecture that used x86, but iterated on it from a mobile or IoT perspective, would also get a much more efficient CPU. But that's impossible, because no other company will ever make x86 CPUs, and Intel and AMD aren't going to throw away decades of work.
I don't think it's only process - Qualcomm's latest chips are quite good as well, certainly more efficient than x86 processors regardless. But it's mobile heritage, that Apple also has, that was allowed to exist because you could get licenses for ARM.
I think this is what Intel is already doing starting with Alder Lake. Their big-little core design is cribbed straight from mobile SoCs, and it has already netted them some big gains in laptop performance and efficiency. AMD meanwhile has been riding on TSMCs smaller process to keep efficiency up on their CPUs while retaining a more desktop/server oriented design.
Competition is great. Despite Moore's Law being basically dead, this renewed competition has made the last 4 years way more exciting in this space than the preceding decade.
Given the legacy hardware deprecating moves they have been making with 11, I can imagine them making a similar move with Windows 13, probably as a special edition.
Note that I said Windows 13. First they have to see Windows 11 succeed in the market, which it may not, which means they will do something conservative with Windows 12 to keep their enterprise customers happy. This is why you should only ever use every other version of windows. I expect to stay on 10 as long as it is practical to do so.
I have the same fear, but I bet there will always be a pro version with a fixed license cost. Business customers will never go for it, because it will convert a capital asset into an ongoing liability.
It's not just about Intel, you have to count AMD too. It's about architecture of CPU.
Can Windows jump away from x86 to ARM? Well, until there are lots of machines using ARM, I don't see Microsoft doing that. If MS announced today that Windows will be only ARM since tomorrow, what desktop or laptop would you buy to run such Windows on it?
But the movement will probably be in that direction. Problem is that MS doesn't want to invest money where they can't make any (why make ARM Windows if there are no ARM machines?) and part/laptop makers won't incest money where the majority of users is not (excluding Mac, Windows competes only with Linux and there is very low coint of Linux desktop users = why make laptops or motherboards with CPU based on ARM if there is almost nobody to buy them?).
Maybe we should ask How fast is RISC-V developement going to be? It can outrun ARM transition in desktops and laptops.
AMD and NVIDIA are working on ARM. Actually the only player that doesn't appear to be at this point is Intel. Microsoft is definitely working on it already. We know this because it would be stupid to wait for hardware and because there's already a Windows for ARM build available. The only question is how many resources they are giving the ARM team.
I know about Windown on ARM. But that is still the smaller portion of the team. It would be stupid if they ignored ARM, I agree. But the market is still more or less non-existent at this point.
I really want to see more fanless design laptops from the Windows side of things. If this chip allows that and provides the performance they're claiming it to have, I'd totally check them out!
I'm more of a Mac or Linux guy, but ya. The Arm based Macbook Airs are incredible; no fan, no heat, super thin, still fast. Its a game changer.
If I could get the macbook air but more linux friendly that would be amazing.
That Thinkpad looks pretty cool.
I did try Asahi but like you said, it has a long way to go before its daily driver ready. I don't know the status at the moment but when I tried it there was no way to adjust screen brightness so you just got 100% brightness all the time. Its a cool project though.
I think so. Microsoft has seemingly put a lot of effort into x86 > ARM translation and that's obviously one of the key elements to making this transition work. Clearly on some level it's technically possible because Apple did it, but obviously Apple has certain advantages that Microsoft doesn't so it might mean it was easier for them or their ceiling was better in translation performance, but overall I think Microsoft understands the importance of this translation piece.
Additionally, I speculate that Microsoft actually has an angle back into the mobile phone space if they can successfully transition laptops/tablets to ARM. It's Microsoft's opportunity to get a redo on what many of their past CEOs have said is their greatest mistake, losing out on the mobile phone market.
And here's the thing:
They're not good enough to pull it off. They might as well just go full Edge and embrace their mobile OS as another worse-than-stock Android experience...like Samsung.
They don't really have to be that good. Google just sucks so bad, and Apple is such a closed ecosystem that someone with enough firepower who can hit the right notes still has a fighting chance. As far as western companies go, Microsoft is probably the only one that fits the bill. It's been clear for over a decade that they've been attempting to leverage their desktop dominance into more mobile computing, of course with laptops already but more specifically tablets. Now I know you could say their lack of success in doing this for a decade is reason to believe that it won't happen, but it's a slow and arduous process. That was their problem with Windows Phone before, which they have acknowledged since, they gave up too early. They're not going to come in guns blazing and take everything over night at this point, they were far too late to the show, but have to take steps and build it up. Even the Windows 11 taskbar, when I read about why they started over from scratch on it, it became clear it was yet another part of the process in positioning Windows to be more capable for tablets than before.
Of course it's all speculation since that's exactly what the topic was asking for, but I think Google sucks so bad and Apple is too restrictive that there's possibly room for someone else to work their way in at some point.
Virtually nobody really cares about privacy...not if doing so will be inconvienient.
Thus the only way a real competitor has a chance of emerging is with a compatibility layer which allows execution of already-purchased Android or IOS apps. And Apple will certianly sue the bejesus out of anybody who would dare reverse engineer that.
And as others have mentioned, Microsoft has left a giant wake of dropped and poorly supported software and hardware that almost rivals Google. I'd give Valve better odds of successfully entering the mobile market at this point than Microsoft (though I still think that is virtually 0 as well).
If you’re looking for Apple Silicon-like battery life on a Windows laptop I think it’s much more likely we see super efficient x86 CPUs than we get a good ARM version of Windows.
I don't think it's likely to happen. There are a confluence of factors that allowed Apple to make the jump that are significantly more difficult barriers for the PC industry, which is lacking the vertical integration Apple can leverage.
It's been very frustrating to watch the rest of the industry essentially watch this happen and not bother to compete. I personally don't like Apple software, but I've already switched my primary work machine to an Apple device anyway, because PC simply cannot and seemingly does not want to compete.
Long-term, I think desktop computing's market share is likely to continue to shrink over time, and the lack of investment in decent hardware is one of the biggest driving factors. In the glory days of the '00s, a desktop or laptop computer at home was a necessity. Nowadays, I think most people can get by with a tablet, or even a phone. Mobile devices are eating away at the PC sector the same way they ate away at landlines (by making them obsolete), although the process has been slower.
I’ve seen some people suggest in online discussions that Apple did this intentionally to create a market for ARM-based machines, but that seems unlikely to me.
I believe that what happened is that Apple believed Intel when they said they’d have smaller process nodes ready to go by 2015, when the first of the infamous touchbar MBPs appeared. Had Intel been able to deliver on those promises, the thermal capabilities of the touchbar MBPs probably would’ve made a lot more sense, since die shrinks usually come with significant improvements in heat output and power consumption. That didn’t happen though, and instead Intel got stuck in their 14nm rut, but by that point Apple had sunk huge amounts of money into engineering and tooling for the new chassis and was effectively stuck with that design until it paid for itself.
There's a lack of ambition in the PC space. People aren't that aware of it, but Intel is much more interested in eating the lunch of every other player in the PC space than they are in enlarging or even maintaining the capabilities of the PC platform.
One sign of it was the introduction of decidedly inferior "integrated graphics" around the time Windows Vista came out which had a graphics model that requires a GPU and set the stage for GPU computing. They are still at it with decidedly inferior discrete graphics. I think Synaptics much have patents that give it a hold on the industry because Intel has never tried to take a bite out of the touchpad market but Intel has integrated second and third rate solutions for almost all of the "chipsets" that go into a laptop for a long time. They've starved PC platforms for IO (on paper there are a lot of PCI lanes but in practice all but the most expensive CPUs and motherboards have limitations that don't really let you use them) and held back development of standards like PCIe.
They've been jealous of smartphone vendors however...
With regards to integrated graphics, about the only nice thing I can say about Intel’s is that they’re generally painless to use under Linux. But then again, so are those on AMD’s Ryzen APUs, which are considerably more powerful…