>I hope they realize that Windows is in significant danger, the majority market share for Desktop OS is not guaranteed anymore.
i agree with most of what you said, but this is borderline fantasy.
the majority of home market share is not guaranteed, sure. with how good gaming is on non-windows machines now, there isnt much for a home user to get locked-in with (except games that require windows-only malware i.e. anticheat)
but government, institution (hospitals, universities, etc.) and large non-tech enterprise? that will be windows for at least 20 more years even if they started to change everything now (which they arent). and the number of machines in those places absolutely dwarfs the number of home installs.
Decline often happens slowly, gradually and then suddenly. Could anybody imagine Intel where it is now ? This could happen to Microsoft and is probably already happening as we speak.
And I'd imagine that this decline accelerates as _developers_ begin migrating to other platforms, since the applications they created are what made that platform appealing to non-developers. That's why Steve Ballmer was jumping up and down, shouting, in a sweaty fervor. Say what you want about pre-Nadella Microsoft, but they definitely recognized the importance of having lots of developers writing software for Windows. And they treated developers like VIPs.
> decline accelerates as _developers_ begin migrating to other platforms
developers don't control what platforms an enterprise would use. Vendors don't dictate the platform either - vendors sell to a customer, and so it makes sense that the customer dictates the platform.
when migration to different platform happens, it's because there's something disruptive that enterprises need to move to, or a new class of enterprise without existing/legacy baggage sees competitive advantage in moving. This happened to IBM when their mainframes no longer offered competitive advantage over the newly minted PC platforms.
If/when windows become lackluster in terms of a required feature, or did not keep up with a needed feature that an alternative platform provides, then the switch will happen fast. What that feature might be i dont know - if i knew, i'd be making it.
> developers don't control what platforms an enterprise would use
They might not control it directly, but they absolutely influence it. Linux was on the losing end of this for many years, as common end-user enterprise software was native and only available for Windows (or in the case of Microsoft Office, nominally available for Mac OS but with fewer features and lots more bugs). That was Microsoft's moat and it started leaking when web applications became ubiquitous. That leak later accelerated when those web applications had to work on mobile operating systems (namely iOS and Android) that Microsoft did not own and could not control.
> Vendors don't dictate the platform either - vendors sell to a customer,
> and so it makes sense that the customer dictates the platform.
There are plenty of counterexamples here. I used to have two legacy SGI machines in my cubicle at work precisely because a vendor dictated the platform to that Very Big Enterprise company many years earlier.
Similarly, many people buy Macs solely to run Logic Pro or Final Cut Pro, because the vendor (Apple) dictated the platform by discontinuing the Windows versions. Apple doesn't have the market share Microsoft has, but unlike Microsoft, they can maintain strong control because of their breadth (OS and hardware for desktop, tablet, and phone, plus high-end creative software) and because a lot of their customers are all-in on Apple's ecosystem.
This comment is multiple years out of date at this point…Lunar lake was considered an exceptionally successful Mobile processor platform that was extremely competitive with Apple M series chips, Panther lake is supposedly improving on that by an even greater degree, and Nova lake is supposedly an extremely promising upcoming desktop platform, and their Arc graphics are excellent for the price point. This is like writing off MacBooks during their thin and hot touchbar/butterfly keyboard era and then assuming they’re that way still in 2026 when they’re now extremely well received laptops lol
I do believe the OP's comparison was with AMD, not Apple. Intel has fallen into financial difficulties following product failures and AMD's strong competition, a far cry from Intel's heydays.
Apple barely comes into the picture, with less than 10% of the worldwide PC market, and was not a factor in Intel's decline.
governments, institutions, and large enterprises (like, thousands of people) do not have the power to do anything "suddenly". they have contracts, and cash flow concerns. you cannot suddenly replaces tens to hundreds of thousands of machines.
20-50 years down the road? maybe! they (microsoft) surely arent doing themselves many favors. but they are certainly not in "significant danger" today.
It means that, today, a lot of enterprises begin pondering the question, and then about a year from now, they start seriously studying and prototyping it, and then "suddenly" in 2029 Microsoft starts seeing a deluge of defections. It means a whole bunch of peopling finishing the conversion all at once, relatively speaking, even if that "all at once" is 3-4 years away.
To put it another way, the thresholds where people get annoyed enough to quit are highly correlated to each other. If individuals on HN are posting "I don't want to switch, I've been working this way for decades now, but Windows has crossed the line for me, I've switched to Linux, and it was easier than I thought it would be", then corporations and governments are having very similar deliberations internally.
This is probably a more accurate model for how "influencers" seem to work than the idea that some crazy guy in your organization falls in love with Product X and evangelizes it internally. I'm sure that happens and is a real force, but this correlation-of-experience effect is probably bigger on the whole. If Product X was good enough to make an evangelist internally, or more germane to this conversation, to make some a mortal enemy of it internally, it's usually because it was a good enough or bad enough product to be able to do that in the first place, and eventually everyone will figure it out in exactly the same way, just later.
20 years is way too large a minimum estimate. If Microsoft responds correctly that might be good, but if they just decide to rest on their laurels and extract whatever value they can out of Windows while they can, Windows would never last 20 years of that. Even the slowest organizations can move faster than that. After all, to cut Microsoft's revenues off at the knees, they don't need to remove every last Windows 2000 server in their backoffice they can't upgrade, they need to cut out just the majority of desktop licenses.
> lot of enterprises begin pondering the question, and then about a year from now, they start seriously studying and prototyping it
Not sure about big enterprises, but I already see this happening in the mid-size, non tech company market.
I'm an IT manager and has been a sysadmin/ops for my entire career, and the past ~4 years I've been seeing a pretty consistent shift toward companies my company does business with deploying more and more macs. Windows is still dominant in my industry, but the cracks in the wall are widening. It's gotten to the point that I'm genuinely surprised now when I see Windows when someone screen shares.
Apple silicon is just too good and the generations coming into the workforce now don't have a "default" windows familiarity that we used to have. They're coming in needing to be trained on how to use a PC in general, windows or not, having used nothing but chromebooks and mobile OSes.
Now, Office OTOH is more entrenched than windows. Even the macshops I interact with are all on M365. Macs are managed with Intune, users & SSO with Entra, Defender for EDR, and of course the office apps. And that's why Microsoft probably isn't as afraid as it seems when it comes to Windows. Even without Windows lock-in, there is very real M365 lockin that is far more entrenched than the endpoint OS.
i disagree. unless intuit is also rewriting quickbooks, dassault systèmes is rewriting solidworks, every bank is rewriting their custom windows-only software, every government branch is rewriting their custom windows-only software, etc. and every company is willing to retrain 95% of their employees on a new operating system, have increased support requirements for a few years at least, etc.
not even touching the capital required for such a transition that in many cases has questionable benefits (from a business perspective).
time will tell! i have first-hand experience with how fast banks move, so i will stick by my 20 year minimum. happy to see otherwise, though.
in any case. what i replied to was a claim that windows is in "significant danger" today. it is not.
They already have. You can't buy QuickBooks for desktop anymore unless you want Enterprise, the expensive $4k+/year subscription. They dumped the Pro/Pro Plus and moved all those users to QuickBooks online.
And now they've launched Intuit Enterprise Suite in an effort to move the QBE customers into Online. The writing is on the wall there, desktop is going away.
It's also happening in more specialized areas too. I work in waste management/recycling, and this industry was traditionally windows heavy with thick clients on desktops. Even the truck scale software is moving to web interfaces, as are the dispatching and asset management.
OS increasingly doesn't matter for most knowledge work.
Yeah, there are going to be industries that will probably never move, certainly not within a 20 year timeline, but there are a ton that are moving or have moved entirely to SaaS and web apps.
In 20 years I expect basically all of these to move to web-based interfaces and away from thick clients. You're already seeing graphics heavy use cases like CAD do this (Onshape has been hugely popular and is cloud native on Linux). Even behemoths like SAP are increasingly web enabled through fiori.
It's an interesting case to me. The company I work for has been shipping systems on windows since the 90's despite pretty consistent requests from customers to ship hardware on Linux. 2 years ago we started creating our own Linux distribution and this year started shipping products on it. We still ship a lot of stuff on Windows 11, but that market share is starting to shift now. 10 years from now I could see us completely moved to our Linux distro. Now, what's actually interesting is that it wasn't customer requests or efficient capital allocation that drove this. Microsoft effectively forced us to do this against our will by a combination discontinued products and handling of Windows 11 and now that we've spent the capital we won't be going back.
You can't abandon Windows because of software X, Y, Z. Over the years vendors move to multiplatform as more and more customers ask for it. These changes are slow but steady. And one day you find out that the last "must have" software is not limited to Windows anymore. That's when the dam breaks.
> i disagree. unless intuit is also rewriting quickbooks, dassault systèmes is rewriting solidworks, every bank is rewriting their custom windows-only software, every government branch is rewriting their custom windows-only software
Up front they won't need to do a full rewrite. They'll only need to make it work well enough under Wine.
At a source level, tools like Avalonia's xpf make porting WPF apps to other platforms easier:
of the stupid enterprise-y software like quickbooks, solidworks and other proprietary stuff that i have used, they barely work well enough under native windows. not to mention, even sticking them in a windows VM voids any support contracts.
You can't suddenly replace them, but in a lot of cases you can find that over an extended period more and more people choose the MacBook option from IT rather than the Windows one.
most people are not willing to learn an entire new operating system for no reason, though. this might happen in tech-based companies, sure, but banks? accounting firms? ive never seen them offer macbooks.
this is also ignoring all of the critical software that is windows-only (e.g. quickbooks, solidworks, bespoke programs in banks and government).
point is: microsoft is not in "significant danger" today.
An increasing number of people are coming into the workplace never having used a Windows machine, or only occasionally having done so when it was absolutely necessary.
>An increasing number of people are coming into the workplace never having used a Windows machine
i would love to see your numbers for this. what does "increasing percentage" mean? 1% -> 2%? 10% -> 20%?
i teach at a college level, in tech, and would estimate ~5% of incoming students have any experience with something other than windows on a pc, at best. outside of tech, i would estimate ~2%.
surprising! what i know of canadian banks is admittedly little, so they might be moving faster than the banks i am familiar with. may i ask what department? do you know if it is managed by intune?
Most people dont even use the operating system. They look for the apps menu, then click what they want to run. Most people can switch between OSes easier than you think because there really isn't that much difference in how they work on the surface.
users are one component, but you are still ignoring/forgetting the rest.
user management, file management, security, windows-specific software, auditing requirements, required capital investment, lack of competent linux sysadmins compared to windows sysadmins, and so on.
expanding your vpn to support more employees working from home is much easier than replacing hundreds of thousands of machines, all of the windows-only software that runs on those machines, training all of your employees on a new operating system, cancelling all of your existing contracts... you get the point.
well, it happened with Teams meetings replacing fancy CISCO equipments.
It happened with all the vpn+shared drives buried to just use SPO.
different experience,I guess.
Did your employees got trained? or just sent the link to 3 'online trainings'?
Managers manage to switch to Mac seamlessly. I am sure the rest will follow with cheaper macs now available. And now, with 'office on the web', you can use basic office everywhere. (even on Debian)
>And now, with 'office on the web', you can use basic office everywhere. (even on Debian)
office is a tiny, almost negligible, piece of the puzzle. quickbooks, solidworks and other cad software, bespoke software, security software, user management, permissions management (replacing active directory), contractual obligations, the millions of dollars required in implementation, the millions more dollars required for increased user support, and so on...
but, again, just to reiterate: i am disputing that windows is in "significant danger" today.
I resemble that remark. I've got to wonder how many people are starting to cut over to Linux/Mac or just stopped caring about being patched on Win10.
A couple weekends ago, I made the overdue call to kill my dual booting with Windows 10 and go full Linux. I'd considered finding a copy of the embedded Win10 long term version or paying for the patching. The local account was one of the things holding me back from doing the update. I knew I could muck with things to still have it, but figured it would be yoinked away later. Similar thoughts to updating the old threadripper that no longer qualifies for Win11. The reinforcement came from all the blasted copilot integration -- notepad, paint... just looking like evasiveness was going to be everywhere.
For a long time my 'good' box was Linux, my old box was Windows. So much just works. I still have an M2 with Windows 10 on it, but it is not in any machine right now. Will see if I run out of space and need it before they actually provide something I'd want to even have on my desk.
Perhaps on the consumer front. But if you have some 30yr old factory with machines, those machines probably use windows software, and you can be reasonably sure that that software will still run in 20 years.
It's not like apple just deciding that we don't do 32bit applications anymore all of a sudden. However many ads MS will shove in windows, as long as they can run software that they depend on, companies and factories won't care.
The home market is interesting, because they do need to address that as well. I'm not sure how many people are switching to macs, and even fewer are switching to Linux, That's not Microsofts problem, not on the large scale.
If you have a PC at home right now, and you're not technically inclined, and Windows is driving you nuts, you're just not getting a new PC again. More and more people are managing without PCs at home, using their phone or a tablet.
To many of us, the idea of your phone as your primary computing device is complete bonkers, but more and more people are choosing that option. Microsoft isn't really giving them a reason to stay, because every time they fired up their laptop Windows updates starts rolling in, taking forever, the UI keeps bugging them about things they don't care about and now there's ads in the start menu. So will Windows attempts to boot, the average person already did the thing they needed to do on their phone.
Windows Home Edition, or whatever it's call now, isn't competing with Linux and Mac, it's competing with Android and iOS, and it's losing.
That's correct. Furthermore
if RAM prices keep going up and staying up, many people won't buy a new PC and they will switch everything on their phones. So the current market could be the undoing of Windows.
This exact same thing (literally another german state i think) almost happened about 20 years ago and Microsoft freaked the fuck out. Thats where all the TCO nonsense came from - just one german state trying to de-microsoft.
I think Microsoft won, too.
I think theyre terrified of positive examples. Especially ones with FAR lower TCO and lower geopolitical risks.
out of curiosity, "large parts" of "one german state" is how many machines roughly?
i am suspecting that it is probably nowhere near enough to put windows in "significant danger". however, i am rooting for their success and hope that they thoroughly document (and publish) the process. i have never seen a transition like that go smoothly, let alone when it is in government.
Note that projects like these often fail not for technical reasons, not even cost, but political pressure from other parties, pressure from people that worked for ages in the administration and, well, have some problems to adjust to new software.
There is also a push from the German state to switch to open source or at least European solutions. There is the Deutschland-Stack, for which the IT planning council made open source mandatory: https://www.heise.de/en/news/Deutschland-Stack-IT-Planning-C...
And so on. At my day job more and more customers are reconsidering cloud adoption, especially M365 and such.
thanks for the link! it is unfortunate that they do not provide numbers, just percentages. i would love to know exactly (or roughly) how many machines "80%" is.
and the "80%" seems slightly misleading, because it is 80%, not including the tax administration. i have no idea what overall % of machines are inside or outside of the tax administration.
it also appears like this is mostly about software like office, rather than operating systems?
>"outside the tax administration, almost 80 percent of workplaces in the state administration have already been switched to the open-source office software LibreOffice."
switching away from office is significantly more realistic than migrating away from windows altogether, and something that every business can and should absolutely consider doing soon.
anyways, seriously, good for them. as i mentioned elsewhere, i hope that they are thoroughly documenting their experiences and are willing to share them after completion.
This alone obviously doesn’t put Windows in danger, but if it does go over well then it’ll mark a turning point; A large non-techy institution getting away from Microsoft’s castle and being better off for it would signal to the world that it’s not only doable, but could even be worth it. It’ll take a while, but this could be the start of the end for Windows.
That "significant danger" was a bit of dramatization on my part. I don't expect anything to significantly change in the short term. I was more referring to long-term tidal-like change, which would be very hard to stop once momentum builds up.
> with how good gaming is on non-windows machines now, there isn't much for a home user to get locked-in with
The options for the average user are not linux or windows, but only macOS or Windows. Gaming is abysmal on macOS on any of the current hardware.
That said, I agree with you that there's less-and-less gaming lock-in on windows, but that's because the majority of gamers are gaming on iOS and android.
>That said, I agree with you that there's less-and-less gaming lock-in on windows, but that's because the majority of gamers are gaming on iOS and android.
I don't think you are aware of how much the landscape has changed regarding gaming and Linux.
2% (linux, really 1% steamOS and 1% other linux) and 1% (macOS) makes it sound much less impressive than "2x".
The options for an average user, who does not use steam and is not in the steam hw survey, are just macOS and windows.
The options for a serious gamer who uses steam (a tiny fraction of PC users) is clearly just Windows or SteamOS at this point, or more likely Windows + a steam deck (which is half of the 2% there, SteamOS).
Or just gaming on iOS / android, like most gamers do these days. The steam hw survey isn't really representative of gamers since the vast majority of them game on consoles and phones.
i agree with most of what you said, but this is borderline fantasy.
the majority of home market share is not guaranteed, sure. with how good gaming is on non-windows machines now, there isnt much for a home user to get locked-in with (except games that require windows-only malware i.e. anticheat)
but government, institution (hospitals, universities, etc.) and large non-tech enterprise? that will be windows for at least 20 more years even if they started to change everything now (which they arent). and the number of machines in those places absolutely dwarfs the number of home installs.