Found an interesting comment, which is also funny in a way.
It is a reply to this conversation.
The linked article is The Downfall and Enshittification of Microsoft in 2026
2025 wasn’t bad either. Retiring Windows 10, while requiring new hardware to upgrade during an economic downturn, and aggressively pushing unpopular features nobody asked for.
Whatever led to this sequence of floundering and cartoonishly stepping on rakes feels like it should be a case study.
Just pure greed. Never been in short supply at MS HQ, but this time short-term-ism has finally defeated the “long con” faction.
So, the Person 3 writes a truly great comment:
I don’t agree (having worked with the Windows team for a few years). There were several factors:
TPM 2 for Windows 11 was an important baseline for a load of security features that they were being hammered for not having. Apple and most Android phones had security coprocessors that did more than TPM 2 and so it was viewed as table stakes for security by a lot of government and corporate customers. The Microsoft sales org is focussed entirely on these kind of customers because of the company!s scale. Getting a $1M contract is barely worth a sales person’s time. The ones that move the dial in their org are at least tens, if not hundreds of millions of dollars. This means that they simply do not talk to SMEs, let alone individuals. They have a partner ecosystem to sell to these people but the feedback channels are not good.
They have an ingrained culture of lying to management. Partly this is incentives being misaligned, but there’s also fear of things leaking to journalists. No one wants a slide deck about how far behind a competitor they are to leak, so mid-level managers delete those slides when they present to senior leadership. This means that senior leadership has no idea how bad things are. They get a rose-tinted picture and reduce headcount in the org (and the mid-level manager gets promoted for making things good).
Windows is full of technical debt. The virtual memory subsystem alone is more code than a minimal Linux configuration. It has a load of copied and pasted code that made SQL Server marginally faster on one specific thing. Only three people in the company understand it (and one is semi-retired). Win32k is a massive attack surface that has vast numbers of things in the kernel that have no business being there. All of this happened for good reasons in the ‘90s but modern computers look nothing like the computers for which all of these optimisations were done and they hurt both maintainability and performance. Worse, most ‘anti-malware’ systems hook into these things and they can’t change them significantly without other vendors crying antitrust and claiming they did it to make Defender the only usable option. Apple isn’t a convicted monopolist, so was able to kick all third-party code out of the kernel. Windows still lets third parties run code in interrupt handlers (ever wondered why interactive performance is so bad on Windows? Take a look at how many things run in ISRs: every other OS moved to a policy of ‘ISRs wake threads, they don’t do anything directly’ twenty years ago).
A lot of the senior people on both the product and engineering sides don’t use competitors’ products at all, so they have no idea of what the competition looks like. They’ve been using Windows and Office for decades and believe it’s better. And they’re so used to Windows and Office that, if they do briefly try something else, the fact that it’s different makes it seem worse.
There are a bunch of cultural issues that reward shipping new features without any reference to user demand or how they integrate with other things. It’s not quite at the level of Google’s ‘get promoted for shipping a new chat system’ phase, but it’s quite bad.
But, first and foremost, senior leadership has no vision. I asked a bunch of senior folks what they thought computing should look like in 10-20 years. None of them had an answer. You can’t define a strategy unless you have an idea of what the strategy is aiming towards. And, because this comes from the top, no one with any kind of vision stays long enough to get promoted, with very few exceptions. It was noticeable that about half of the people I respected and enjoyed working with quit in the year before I did, and none of the ones I thought were a waste of space did (some got promoted).
There are more issues, and now that the stock price is tanking the board should be asking Nadella some hard questions.
While I truly appreciate some mindful pieces of wisdom, all I want to say is ‘Microsoft is full of idiots, lol!’ when I see any discussion about them. They’re like cartoonish villans, but also very stupid, like it’s some stupid movie.
Interesting thing is, the Person 3 says ‘hey, they aren’t all idiots!’ and then they tell how idiotic M$ is :)
Also, an interesting series of articles: How Microsoft Vaporized a Trillion Dollars. Inside the complacency and decisions that eroded trust in Azure—from a former Azure Core engineer.
Linux is just superior on Desktops now. (I mention desktop only because servers… only incompetent idiots run their servers on Windows. Plus a couple of very niche usecases where its necessary, I believe.) And now we have games. This user nicely summarised the situation:
And we probably shouldn’t discount the actual areas of improvement either. It used to be that for a large segment of the population, Windows could get almost infinitely bad before they would consider switching, because they play games and you just couldn’t really get games to work on Linux. The fact that the trade-off is no longer “I will not be able to play any games”, but instead “some games may be a bit more buggy or a few have anti-cheat which makes them not work but most just work out of the box”, is an immense achievement.
Ah, and the blog post itself, it’s good.