Very recently I responded to a post about X11 going away, and the comments filled up with the usual suspects claiming Wayland isn’t stable, that it breaks everything, and my personal favorite, a link to that infamous GitHub gist as if it were holy scripture.
The person who posted the gist was simultaneously running KDE Plasma on Wayland, and the irony was lost on them.
After 40 years in software engineering I’ve seen plenty of technology transitions, and some deserve the pushback they get, but Wayland’s reputation at this point does not. The arguments against it are frozen in 2020 while the technology moved on without them.
That Gist Everybody Cites
If you’ve spent any time in Linux forums you’ve probably seen it, “Think twice about Wayland. It breaks everything!” It gets passed around like a technical white paper, but it isn’t one.
The author is probonopd, creator of the AppImage packaging format, and he’s not a Wayland developer or a compositor maintainer. He’s a packager whose format doesn’t bundle the Wayland Qt plugin by default, which means AppImages built his way break on Wayland sessions. That’s not Wayland being broken, that’s a packaging decision.
His gist started around 2020 and gets periodic updates, mostly adding complaints and rarely noting fixes. The conflict of interest is obvious since his project has compatibility issues with Wayland, so he’s motivated to discourage adoption rather than fix his tooling.
Not a reliable source, nothing timely or relevant to contribute.
The Major Claims: Then vs. Now
Let’s walk through the greatest hits and check them against reality.
NVIDIA Support
The 2020 complaint: Flickering, graphical glitches, unusable without workarounds because NVIDIA doesn’t support the implicit sync that Wayland compositors expected.
The 2026 reality: NVIDIA driver 555 landed in June 2024 with explicit sync support via the linux-drm-syncobj-v1 protocol, and KDE Plasma 6.1 and GNOME 46.1 both support it. Ubuntu 24.10 defaults to Wayland for NVIDIA users, the flickering is gone, and the architectural dispute that caused years of pain got resolved.
If you’re running a supported NVIDIA card with drivers from the last year, Wayland works, and I’m typing this on an RTX 4070 Super.
Screen Recording
The 2020 complaint: Wayland provides no capture APIs, OBS doesn’t work, and SimpleScreenRecorder abandoned Wayland support.
The 2026 reality: OBS Studio has had native PipeWire screen capture since version 27, released in 2021, which was five years ago. It works on GNOME, KDE, and wlroots-based compositors, you add a “Screen Capture (PipeWire)” source, select your display, and you’re done.
The gist still lists this as broken, but it isn’t.
Screen Sharing
The 2020 complaint: Zoom only works on GNOME, Jitsi is broken, and “there is nothing we can do from the Jitsi Meet side.”
The 2026 reality: xdg-desktop-portal handles screen sharing across compositors, and Zoom, Microsoft Teams, Discord, and Google Meet all work. The portal architecture that people dismissed as “Red Hat bloat” turned out to be the solution.
Global Hotkeys
The 2020 complaint: No equivalent to X11’s XGrabKey, and applications can’t register system-wide keyboard shortcuts.
The 2026 reality: The GlobalShortcuts portal exists and is implemented in major compositors. Applications need to adopt it, which takes time, but the mechanism is there, making this an “in progress” situation rather than a “broken forever” situation.
Window Positioning
The complaint: Applications can’t set their own window position on screen.
The reality: This is true and it’s intentional. Wayland’s security model doesn’t let arbitrary applications reposition windows or grab input from other windows, which prevents a whole category of attacks that X11 happily enabled. You can argue about whether the tradeoff is worth it, but calling a security feature “broken” misses the point.
What “Stable” Actually Means
The goalposts on “Wayland isn’t stable” keep moving, so let’s look at when major projects decided it was stable enough to ship as default:
- Fedora (GNOME): Wayland default since 2016, ten years ago
- Ubuntu: Wayland default since 22.04 in 2022, including NVIDIA systems in 24.10
- GNOME: Wayland default since GNOME 41 in 2021, five years ago
- KDE Plasma: Wayland default since Plasma 6 in February 2024
- COSMIC: Shipped stable as Wayland-only with no X11 session offered
These projects have tens of millions of users, their developers aren’t idiots, they have bug trackers, QA processes, and reputations on the line. When they flip the default it’s because they’ve determined the experience is ready.
The notion that all these teams simultaneously made the wrong call and random Facebook commenters know better is not a serious position.
What I Actually Run
I’m not a Wayland evangelist, I’m an engineer who evaluates tools based on whether they work for my use case.
My daily driver is a System76 Thelio Mira R4 running COSMIC desktop on Arch Linux, with an AMD Ryzen 7 9800X3D, 64GB of RAM, an RTX 4070 Super, and a 3440x1440 ultrawide at 165Hz. This is not a simple setup.
What works: gaming through Proton with Elite Dangerous, Monster Hunter Wilds, and World of Warships all running with MangoHud without issues, and development with RustRover and terminal emulators for building COSMIC applets.
What’s still rough: some older applications that hardcode X11 assumptions run through XWayland which mostly works but occasionally has quirks, and a handful of niche tools haven’t been updated and probably never will be.
That’s it, and the list of problems is short and getting shorter.
Why the Myth Persists
Old information has momentum on the internet, forum posts from 2019 still rank in search results, people repeat what they heard years ago without checking if it’s still true, and confirmation bias kicks in where one bad experience becomes “Wayland is broken” even if ten other things worked fine.
The probonopd gist is well-SEO’d and it looks authoritative if you don’t check who wrote it or when the claims were last verified. It gets shared which boosts its ranking which gets it shared more.
There’s also a contingent of X11 users who simply don’t want to learn new workflows, and that’s a valid personal choice, but it’s not a technical argument. XWayland exists specifically so those users can keep running their X11 applications on a Wayland compositor, and the transition path is there.
What’s Actually Still Rough
I’m not going to pretend everything is perfect, and some things are genuinely still in progress:
- Color management is being actively developed, the protocols exist but adoption is incomplete
- HDR support is coming but not fully baked in all compositors
- Legacy software that will never be updated will eventually break as X11 fades, an inevitable consequence of progress
- Some edge cases in XWayland still cause minor issues with specific applications
The difference between these real issues and the mythological “Wayland breaks everything” narrative is that these are bounded problems being actively worked on, not fundamental architectural failures.
The Bottom Line
The “Wayland isn’t stable” argument made sense in 2018 or 2019, by 2021 it was getting stale, and in 2025 it’s misinformation.
If you’re running GNOME, KDE, or COSMIC you’re probably already on Wayland and didn’t notice. If you’re gaming on Linux with recent NVIDIA drivers you’re likely on Wayland. If you’re using a laptop with a recent distro, Wayland.
The ecosystem moved on, but the discourse hasn’t caught up.
Next time someone drops that gist in a comment thread, ask them two questions: when was the last time you actually tried Wayland, and have you checked whether the issues you’re citing have been fixed in the last five years?
Usually the answer to both is no.
