I am fifty years old. I have watched the evolution of software updates across my entire adult life, and the trajectory is something nobody talks about.
When I was a teenager, an operating system update was a Christmas gift. You would celebrate. You would read the changelog with genuine excitement. Updates came once a year, maybe twice, and they made things better.
In my thirties, I learned to wait. Let the brave ones go first. Be a few months behind. Let the valiant beta tasters discover the landmines so you do not have to.
Now? I actively refuse to update. Even security patches. Because at some point in the last decade, the equation flipped: updates became more likely to break things than fix them.
Microsoft’s January 2026 Patch Tuesday is the perfect illustration of why.
What happened
January 13: Microsoft ships its monthly update. Standard procedure. Billions of PCs get it automatically.
January 14: PCs will not shut down. Some will not boot. Blue screens. Outlook crashes. Remote Desktop login fails.
January 17: Emergency patch #1.
January 25: Emergency patch #2.
Two emergency patches in ten days. For an update that was supposed to make things more secure.
Why this keeps happening
This is not a one-time screwup. This happens regularly. We just only hear about it when it is catastrophic enough to affect everyone. When it affects a subset of users, they do not make enough noise for it to become news. The failures that do not trend on Twitter are still failures.
The reason is structural, not human. Windows is a 40-year-old codebase that was never rewritten from scratch. Layer upon layer upon layer. Kernel drivers interacting with user-space applications interacting with hardware abstraction layers interacting with security modules that were bolted on years after the foundation was poured.
Nobody fully understands this system. Not even Microsoft. Whoever writes a patch must be scared shitless, because there is no way to predict every interaction across every configuration. Too many moving pieces, all connected, all capable of bringing the whole house of cards down.
The real problem: marketing runs the product
Here is the thing that nobody inside Microsoft will say publicly: the people pushing Windows forward are not the engineers. They are the marketing team.
Engineers want stability. They want to test more, ship less, fix the fundamentals. Marketing wants features. New UI. AI integration. Copilot everywhere. Ship it now, fix it later.
When marketing runs your operating system, you get a product that looks impressive in a press release and blue-screens in a hospital.
This is not unique to Microsoft. This is the inevitable result of any company where the revenue team has more influence than the engineering team. But when your product runs on billions of machines in critical infrastructure, the consequences are not “low engagement metrics”. The consequences are banks that cannot process transactions and nurses who cannot access patient records.
Before you say “just use Linux”
I spent years deep in the Linux kernel. I know it intimately. And I have to be honest: Linux has the same fundamental problem.
It is a monolithic kernel. One bad line of code in a driver can bring the whole system down. The Linux kernel is absolutely infested with bugs. They have their own version of this problem, just with a different community structure managing it.
The difference is not that Linux is better engineered. It is that Linux does not auto-update billions of machines simultaneously. The blast radius of a Linux kernel bug is limited by the fact that sysadmins choose when to update. Windows users do not get that choice.
What I keep thinking about
We are living on the bleeding edge and most people do not realize it. Everything is interconnected. Everything is just-in-time. One small mistake in a monthly patch can be a business-ending event for someone.
The testing pipeline cannot possibly cover every hardware configuration, every driver combination, every enterprise deployment scenario. So they ship, and the first hour of deployment IS the test. Billions of machines are the QA team.
When it works, nobody notices. When it fails, hospitals have PCs that will not boot.
This is not going to get better. It is a fundamental structural problem with how operating systems have evolved over decades. No amount of process improvement fixes a 40-year-old monolith. The only real solution would be to rethink what an operating system is from the ground up.
But nobody at Microsoft is going to propose that. It would take years, it would cost billions, and the stock price lives on quarterly results.
So we get Patch Tuesday. And sometimes, we get Blue Screen Wednesday.
Somewhere, an engineer at Microsoft is reading this and nodding. They know. They have always known. They just cannot say it out loud.
Sources: BleepingComputer, The Register