“I went back to Linux and it was a mistake” — what this kind of story gets right (and what it misses)

/ General / By

“I went back to Linux and it was a mistake” — what this kind of story gets right (and what it misses)

Every few months, someone publishes a version of the same confession:

  • “I tried desktop Linux again.”
  • “It’s better than it used to be.”
  • “And I still bounced off it.”

If you’ve used Linux for years, these pieces can feel like déjà vu. If you’ve never used Linux, they can read like a warning label.

But they’re valuable — not because they “prove” Linux is good or bad, but because they surface the gap between Linux as a capability and Linux as a consumer product.

A recent Verge essay framed that gap bluntly: the author revived an older Dell XPS 15, installed Ubuntu, enjoyed the speed and quiet, and then ran into the familiar frictions — firmware edge cases, odd device support, confusing installs, and the feeling that the “better option” is always a different OS.

Let’s unpack what’s actually going on here.

Desktop Linux in 2026: the paradox

Desktop Linux has never been more capable:

  • Gaming is genuinely viable thanks to Proton and Vulkan.
  • Creative tools exist (and some are excellent).
  • Distros are polished.
  • Hardware vendors certify machines.

And yet, many users still experience it as “fragile,” “fiddly,” or “a house of cards.”

The paradox comes from the fact that Linux is an ecosystem, not a single product. Capability is distributed. Responsibility is distributed. And when something breaks, the question is often: whose problem is it?

On macOS, Apple owns the whole stack.

On Windows, Microsoft owns most of the stack and hardware vendors target it aggressively.

On Linux, the stack is shared across:

  • kernel developers
  • distro maintainers
  • desktop environment teams
  • package format ecosystems
  • hardware vendors
  • app developers

That sharing is a strength — it’s also why “it works” isn’t guaranteed.

Where these Linux disappointment stories are usually accurate

1) Hardware support is better, but edge cases still hurt

In the Verge story, two early pain points were typical:

  • A fingerprint reader that doesn’t work.
  • A firmware / EFI partition issue that complicates updates.

Linux hardware support has improved massively, but it is still uneven in areas where vendors treat Linux as an afterthought:

  • biometric readers
  • some Wi‑Fi chipsets
  • laptop sleep / suspend
  • GPU driver quirks (especially on mixed vendor systems)

Even when Linux supports the hardware, it can be less “integrated.” A Windows driver might include a management UI, power profiles, and vendor tuning. Linux may have the core support but not the same user-facing polish.

2) Multi-boot and partitions are still a footgun

Dual-booting has always been a classic trap.

It’s not that it can’t work — it’s that the cost of a mistake is high, and the error messages are rarely written for normal humans.

When something goes wrong with EFI partitions, it can feel like you’re working inside a box of unlabeled cables.

3) Installing apps is, bizarrely, more confusing than it used to be

This is one of the most important points.

In 2006–2012, the mental model was fairly straightforward:

  • use your distro’s package manager
  • maybe add a PPA
  • compile from source if you’re brave

In 2026, you can install:

  • distro packages (apt/dnf/pacman)
  • Snap
  • Flatpak
  • AppImage
  • vendor .deb/.rpm installers

That’s great for flexibility — but it’s awful for predictability.

Two users on “Linux” can be on the same desktop environment and still have wildly different app environments.

When an app “quietly fails to install,” it’s not a Linux philosophy problem. It’s often a UX failure in the packaging layer.

4) Sleep, external devices, and audio routing are still not boring enough

The story also mentioned:

  • sleep behavior that is inconsistent
  • external drive / SD reader weirdness after sleep
  • Bluetooth wake limitations
  • Steam games refusing to use an external audio interface

These are exactly the kinds of problems that make people feel like the system is unstable — because they show up in normal life.

You can tolerate a missing fingerprint reader.

But when your docked setup breaks after sleep and you have to reboot, the OS stops feeling like a tool and starts feeling like a project.

Where these stories often miss the bigger picture

1) “Linux” is not one experience

Ubuntu is not CachyOS is not Fedora is not Bazzite.

A user who installs Ubuntu on a random older laptop is not testing “Linux.” They’re testing:

  • Ubuntu’s defaults
  • the GNOME stack
  • the kernel version Ubuntu ships
  • the device firmware
  • the vendor’s BIOS quirks

Sometimes a small change fixes the whole experience:

  • different kernel
  • different power management defaults
  • updated Mesa stack
  • different audio configuration

That’s not an excuse — it’s the reality of the ecosystem.

2) The best consumer Linux products are often appliances

The Steam Deck is successful because it’s not “Linux as a hobby.” It’s Linux as a controlled appliance:

  • known hardware
  • curated OS updates
  • predictable graphics stack
  • clear support boundaries

That model can be copied.

For general desktops, it’s harder because users bring random hardware.

But the lesson is clear: Linux succeeds when someone owns the integration layer.

3) The “better option exists” feeling is partly about mainstream app gravity

A Linux desktop can do most things. But it often loses to:

  • macOS for audio production with commercial plugin ecosystems
  • Windows for certain games, peripherals, and enterprise apps
  • iPadOS for certain single-purpose consumption tasks

That’s not because Linux is inferior. It’s because ecosystems have gravity.

A creator who lives in Ableton and depends on specific VSTs is not choosing an OS — they’re choosing an industry supply chain.

Linux can fight that with:

  • better compatibility layers
  • better commercial tooling support
  • stable packaging standards

But it’s a long game.

Practical advice: when Linux is the right answer

If you’re thinking about trying Linux in 2026, the key is to match the tool to the job.

Linux is often a great fit for:

  • a “browser machine”
  • coding and dev tools
  • privacy-focused everyday computing
  • older hardware revival
  • gaming on Steam Deck / Bazzite-style setups

Linux is often a frustrating fit for:

  • complex audio production workflows with commercial plugins
  • “everything must work perfectly with this specific dock and device chain”
  • enterprise apps tied to Windows-only tools

And the single biggest success factor is boring: hardware choice.

If you pick a laptop with good Linux support, Linux feels like a normal OS.

If you pick a laptop that barely works on Windows because of firmware quirks, Linux will inherit those problems.

Bottom line

Stories like “Linux was a mistake” aren’t anti-Linux propaganda — they’re honest reports of what happens when a powerful ecosystem collides with everyday expectations.

Desktop Linux is good enough to be someone’s primary OS, and for many people it’s the best choice. But “good enough” isn’t the same as “boringly reliable across random hardware,” and the modern packaging landscape can make basic app installation feel more confusing than it should.

The path forward isn’t more ideology. It’s more integration: better defaults, clearer packaging standards, and hardware that treats Linux as a first-class target.

Sources

n English