How do I use Windows?

On my laptop, Windows is not the main character. My real daily system is EndeavourOS GNU/Linux, where I live, work, and keep the things that actually matter to me. Windows sits beside it as a dual-boot guest: a tool I reach for only when I need a very specific piece of software, a particular game, or some school-related application that insists on running there. It is useful, but it is never in charge.

Before anything else: this page is not an invitation to argue about operating systems or to get mad at me for having Windows installed at all. I am describing how I personally manage a system I sometimes have to use. It is not advice that everyone must follow, and it is definitely not a statement that Windows is “better” or that I am abandoning GNU/Linux. This is just the compromise that works for my hardware, my school life, and my gaming needs.

Think of this as a threat model written in plain language: I accept that I sometimes have to touch nonfree software, but I do not hand it my whole life. I decide what data is allowed to be near Windows, which accounts are allowed to touch it, and how much network visibility it gets. Everything else stays safely on the GNU/Linux side, where I have far more control.

Free software first

My heart is firmly with free software. The GNU Project exists to give users the freedom to run, study, share, and modify the programs they rely on|the classic four freedoms that make software free in the sense of liberty, not just price. In the usual GNU wording, that means: the freedom to run the program for any purpose, to study and change how it works, to redistribute exact copies, and to distribute modified versions so that the whole community benefits from improvements. Those are not marketing slogans; they are concrete powers that a user either has or does not have.

GNU itself is a complete operating system built from these principles, and adding the kernel Linux to it gave us the GNU/Linux systems we use today. When I talk about “my system,” I mean that world: code that respects my freedom and a community that believes users should be in control of their own machines. When I boot into EndeavourOS GNU/Linux, I feel like I am walking into my own house. When I boot Windows, I feel like I am borrowing a key to a locked office.

I try to fight for that free software world in small, practical ways. Whenever I can, I choose programs under the GNU General Public License (GPL) or other strong copyleft licenses that guarantee that the code|and the freedoms around it|stay with every user down the line. The GPL exists specifically to protect the right to share and modify software and to make sure those rights cannot be taken away later. When I write or share my own software, I aim for that same model: if it helps people, it should help them in a way that keeps them free.

I also care about the layers below the operating system. Projects like Libreboot, which replace proprietary BIOS/UEFI firmware with free firmware, are steps toward laptops that are free all the way down to power-on. The same goes for GNU Linux-libre, which maintains versions of the kernel Linux with non-free blobs removed so that fully-free GNU/Linux distributions can exist. Even if my current hardware is not completely libre-bootable yet, I still think these efforts matter a lot and I try to support and amplify them where I can.

In day-to-day life, “free software first” means that my password manager, my note-taking tools, my development environment, my music player, and my browser setup all live in the free software world. That is the default. Proprietary tools only show up when there is no realistic alternative for a specific task, and even then I try to keep them boxed in and temporary.

Why I still touch Windows at all

So why touch Windows at all? Because, realistically, some parts of my life still expect it. There are school tools that only support Microsoft’s platform, games that either rely on proprietary launchers and anti-cheat systems or simply misbehave on GNU/Linux, and a few corner-case applications that have no good free replacement yet. Instead of pretending those needs do not exist, I isolate them: they live on a separate Windows install whose purpose is very clearly defined.

Using Windows this way changes how I think about it. I am not trying to “embrace” the platform as it is shipped; I am trying to contain it. I treat it as a compatibility layer that just happens to be a full proprietary operating system. That means I am careful about accounts, telemetry, and which online services get to see that machine. It means avoiding the trap of logging into everything, syncing everything, and letting the system turn into yet another data-gathering endpoint.

Concretely, this also means that certain categories of data simply never cross the line into Windows. My long-term archives, my main password store, my private notes, and my “this is my life” photos live on the GNU/Linux side. Windows sees only what it needs for that session: a game library, a few documents for school, maybe a temporary copy of something I am testing. When I am done, that data either moves back to GNU/Linux storage or gets deleted.

I could try to pretend that I live in a world where all institutions respect free software and everything runs smoothly on GNU/Linux. I would love that world. But the one I actually live in has exam portals, proprietary launchers, and locked-down school tools. My answer is not to pretend they do not exist; it is to put them all in one fenced-off place and refuse to let them creep further into the rest of my life.

What I dislike about Windows 11

When I complain about Windows 11, I am not doing it for drama. A lot of the design goes in exactly the direction free software people have been warning about for years: stricter hardware requirements like TPM 2.0 and “approved” CPU lists that quietly push people toward new machines, setup flows that try very hard to make you use a Microsoft account, default telemetry that is simply on unless you go hunting through settings, and a desktop full of suggestions, store entries, and “experiences” nobody asked for. That is not neutral design; it is a constant pressure to accept less control over a computer you supposedly own.

On top of that, the system keeps drifting toward “your computer as a service terminal.” Features are pushed through store-style mechanisms; advertising and recommendations creep into more places; and more parts of the system assume that you are always online, signed in, and happy to let remote services sit between you and your own files. For someone who cares about user freedom, this feels less like an operating system and more like a managed kiosk that happens to live on your desk.

Some of the newer “smart” features are just the same pattern in a louder form. Screen-scraping assistants, timeline-style activity logs, and features that record what you have on your display so they can search it later all have the same fundamental problem: they are watching you so that they can help you. Maybe some pieces run locally, some encrypted blob sits on disk, some rules exist about when data gets sent out|but the basic power structure is still upside down. I want tools that I command, not tools that observe me and then offer to be helpful.

If you are coming from the Free Software Foundation side of things, none of this will sound new. I agree with the basic criticism: Windows 11 is nonfree software that takes away user freedom and digital autonomy. It is not something I celebrate. For me, running it in this fenced-off way is damage control, not endorsement. I do not recommend it as a first choice to anyone; I just refuse to pretend that school requirements and game compatibility magically disappear because I wish everything already ran on fully libre systems and libre hardware.

How I fence Windows in

The first decision, then, is about purpose. My Windows install is clearly scoped: it exists for school work that genuinely needs it, for gaming that does not behave well on GNU/Linux yet, and for the occasional piece of general-use software that has no good free alternative elsewhere. If something does not fit into those roles, it stays on my GNU/Linux side|or it does not get installed at all. That boundary alone keeps the system lighter, cleaner, and easier to reason about.

The second decision is about trust. I do not pretend Windows is privacy-friendly; instead, I surround it with extra layers. On the network side, I use DNS filtering (for example, with services like NextDNS) to cut off obvious telemetry and tracking domains before they even reach the system. Inside the OS, I disable what I can, turn off unnecessary “experiences,” keep a close eye on permissions, and ignore any feature that exists mostly to serve the vendor rather than me. The goal is not perfection|it is to reduce the surface area to something I am willing to accept for this one specific machine.

Storage is another line I draw on purpose. My GNU/Linux partitions hold my real home directories and long-term data. Windows sees, at most, a separate shared data partition or a few specific folders I decide to expose. I do not mount my entire GNU/Linux home directory inside Windows; I do not hand the proprietary side full read–write access to my whole life. If a file lives only on the Windows partition, I treat it as disposable or backed up elsewhere.

Visuals and workflow matter too, not just because I like things to look a certain way, but because interface friction is exhausting. Stock Windows is built around habits and patterns that do not match how I work on GNU/Linux. So I nudge it until it feels closer to home: a cleaner desktop, a taskbar that behaves the way I expect, a file manager that does not get in my way, and shortcuts that match what my fingers already know. The more consistency I have between systems, the easier it is to drop into Windows for a task and then leave again.

At the command line, I refuse to give up the tools I use on GNU/Linux. I want the same basic commands, the same pipelines, the same muscle memory. On Windows, that means installing full GNU and Unix-style utilities and wiring them into a custom PowerShell profile. Over time, that profile has grown into something like 15 000 lines of configuration: aliases, functions, prompts, wrappers around GNU coreutils, and a lot of glue code to make PowerShell behave like a Unix-style shell as much as possible. When I open a terminal there, I can type ls, grep, or tail without thinking about it, and my scripts behave in a way that matches my GNU/Linux side instead of fighting it.

Software installation is another place where I pull Windows toward the way I use GNU/Linux. Random installers from random websites are a last resort. I treat the system as if it had a package manager from the beginning: software comes from catalogs I trust, gets updated in one place, and leaves cleanly when I am done with it. I do not want the system to turn into a museum of old versions and half-forgotten tools just because they once solved a temporary problem.

Media is one of the simpler parts. I prefer a single, solid free player that can open almost anything, rather than a pile of store-based apps and streaming-driven interfaces. One good player, locally installed, is enough for my use. It keeps things predictable and small: the machine plays what I give it and does not try to nudge me toward stores, libraries, or content I did not ask for.

Finally, there is the “what I simply will not do on Windows” list. I will not generate or store long-term encryption keys there. I will not treat it as my main email or messaging workstation. I will not rely on it for the only copy of something important. When I boot that OS, I know I am stepping into a space where my freedom is limited, so I act accordingly.

My complicated soft spot for Microsoft

With all of that said, I do still like Microsoft in some ways, and I do not think that is a crime. I grew up in a world where a lot of people’s first contact with computers was through one version of Windows or another. Some of the engineering inside those systems is genuinely impressive, even if the software around it is nonfree and the business model is one I disagree with. I can respect the work of individual engineers, designers, and writers without signing up for every decision their employer makes.

I also like some of the tools that have come out of Microsoft’s ecosystem, especially where they are released under free software licenses. Projects like PowerShell (as a cross-platform shell under the MIT License), winget (the Windows Package Manager), and parts of the .NET stack show that the company is capable of publishing real, usable free software and not just talking about it. I still read the licenses, and I still keep my distance where I need to, but I do not pretend these projects do not exist or that they are automatically bad just because of the logo on the repository.

I also enjoy some of the aesthetics: certain fonts, parts of the design language, bits of the terminal experience, and the way some tools feel in daily use. None of that cancels out my criticism of the nonfree parts, but it does mean my feelings are more complicated than “everything here is terrible.” Human beings built these systems. Some of them clearly care about craft and usability, even when the larger platform is heading in a direction I dislike.

To me, there is nothing wrong with liking some things Microsoft makes|fonts, UI details, the feel of certain developer tools|while also being honest that the overall platform is nonfree and not aligned with the goals of the free software movement. You are allowed to appreciate craft wherever you find it. What matters is what you normalize and what you build your own life around. In my case, that center of gravity is still firmly GNU/Linux and the wider free software world.

If you think in Free Software Foundation terms, you can picture this as separating feelings from strategy. Emotionally, I can enjoy a slick terminal font or a thoughtful UI detail from a proprietary system. Strategically, I try to put my time, code, and recommendations into software and communities that respect user freedom. Liking some of the art and engineering that comes out of a big nonfree vendor does not require me to stop criticizing the parts that take away control from users.

Why macOS still looks interesting (from a distance)

I do not run macOS myself, but I still find it interesting for one big reason: under the glossy desktop, it is basically a Unix system. Apple’s Darwin base combines the XNU kernel with a lot of familiar Unix-style components, and macOS has gone through formal Unix certification. In practice, that means a real POSIX-style environment under the hood, a shell, standard tools, and a layout that feels much closer to traditional Unix than many people realize.

Apple also publishes a significant amount of system code in publicly available source releases. If you are curious, you can browse through Darwin-related components and see how some parts of the system are structured. That does not magically make macOS a free operating system, and a lot of what matters most to users is still locked down and proprietary. But it does mean there is a concrete trail for people who want to study how a modern Unix-style desktop OS is assembled.

From a free software point of view, macOS is still nonfree. You cannot treat it as a replacement for a fully-free GNU/Linux system or a completely libre stack. At the same time, I think it is okay to admit that a mainstream, polished desktop that is also a certified Unix has some appeal|for example, in how easy it can be for people to move from macOS terminals and Unix command-line habits into the GNU/Linux world later. The technical closeness can be a bridge, even if the license terms are still a wall.

So when I say I “like” macOS, what I really mean is that I like the idea of a widely used Unix-style desktop, I like some of the design language, and I appreciate the technical choices that make it feel cohesive. I do not treat it as a free system, and I do not hold it up as a model for what freedom should look like. It is another example of how you can appreciate parts of a thing without pretending the whole package fits your ethics.

Putting it all together

All of this is held together by one rule: Windows is not my main operating system. My real work, personal documents, and daily life live on EndeavourOS GNU/Linux. Windows is there when I need it and out of the way when I do not. I keep it lean, update it through controlled channels, and shut it down when I am finished. It is a sidekick, not a home.

From the outside, it might just look like “a Windows install.” Under the surface, it is a deliberately constrained environment: filtered at the network level, tweaked in the interface, bridged with full Unix-style tools, and kept on a tight leash through habits and tooling. You do not have to like my setup, but please do not be angry that I run this OS at all|this is just the practical balance I have chosen for myself in a non-free world.

For the other half of this story|the system that actually is my main OS|you can read more on my About GNU/Linux page, where I talk about GNU, the kernel Linux, and why that combination is where I really prefer to live.

“But Windows 11 has Copilot and AI!”

I hear this a lot, usually as if it ends the conversation: Windows 11 has Copilot, “AI features,” smart this and smart that|so why not just give in and enjoy it? The short answer is that most of these features work by looking over your shoulder. Some of them watch what is on your screen, some index what you open, some talk to online services whenever you ask for help. Even when parts run locally, the design assumes an assistant that sits between you and your own computer, with code I cannot study or change. That is the opposite of what I want out of a personal machine.

I am not against the idea of smarter tools. I am not even against “AI” as such. I am very interested in open-weight models, free implementations, and on-device tools that you can actually study, fork, and run under your own terms. If I can have a small, local model on GNU/Linux that helps me search my own notes or files without phoning home, that is exciting. What I do not like is the idea that the price of having a helpful assistant must be letting a proprietary system watch everything I do and then trusting a company’s promises about where that data goes.

So when I disable Copilot things, it is not because I hate the underlying technology. It is because I want that technology on my side, in my hands, under free licenses, and running as close to my own hardware as possible|ideally on GNU/Linux, not embedded deep inside a proprietary desktop that I only use as a backup.

If you are an FSF member or just someone very strict about software freedom, you can think of this as a kind of quarantine. I do not pretend that Windows 11 respects my freedom. I do what I can to keep it boxed in, and I spend my real energy on GNU/Linux, GPL-licensed software, and, when I can, hardware and firmware that move me closer to a completely free stack. Until the world catches up, this is how I live with one foot in the free system I actually want and one foot in the proprietary mess I sometimes cannot avoid.

Questions & Answers

So what do I actually use on Windows?

• NextDNS as a DNS-level filter to cut down on telemetry, trackers, and junk domains: nextdns.io
• Windhawk to patch and customize the Windows shell and desktop behavior: windhawk.net
• Microsoft PowerToys for window tiling, shortcuts, and power-user utilities: PowerToys
• Scoop as my main command-line package manager: scoop.sh
• winget, the built-in Windows package manager, when it fits a particular app: winget
• UniGetUI as a graphical frontend to manage Scoop, winget, and more from one place: UniGetUI
• KDE applications for Windows so I can use familiar tools like Konsole, Dolphin, and Kate on Windows: KDE for Windows
• GNU coreutils and other Unix-style utilities, installed via Scoop and integrated into my custom PowerShell profile so I have a full Unix-like toolbox on Windows: GNU coreutils
• VLC media player as my main audio/video player for local media: videolan.org/vlc

All of this runs alongside EndeavourOS GNU/Linux in a dual-boot setup on the same machine. Windows stays as a careful, privacy-conscious side OS for school work, gaming, and a bit of general use|kept lean, watched closely, and never treated as my primary environment.