There is no question that Linux is way more responsive general use, and Windows 11 isn't responsive no matter how powerful your hardware may be, even on a Raptor Lake the whole OS is jerky, especially keyboard input in all applications like Visual Studio, Notepad etc. Having used every major version since 3.0, Windows is clearly worse than ever. At home you can certainly mitigate some of these issues by disabling spyware etc., but you probably can't do that stuff on your work computer.
But as for gaming, we should keep our eyes on the target - to get gaming companies to use Vulkan and compile games for Linux natively. Wine/Proton is certainly impressive work, but from what I've seen it's still a hit and miss when it comes to individual games. While I haven't tried any top games launched from the last four years in Wine, but from what I've seen having a "top rating" in Steam with Proton doesn't guarantee it will work at all.
From the user perspective, pretty much any major (regular) distro and derived distros should be fine.
From the developer's perspective shipping software to be deployed on all kinds of distros requires some effort that a Windows dev wouldn't be used to; like do you ship just a tar-ball, a shell-script that's distro independent, or do you want like deb, rpm packages, or perhaps snap(with its own issues)?
I would argue the options for Windows aren't that great either since they moved on from win32, and MS has never been good with API versions, so having a complex GUI application for Windows means a lot of workarounds or rewrites for different Windows versions. While I don't know nearly all major client applications out there, I rarely see any heavy application use any of MS' GUIs(especially not .NET), it's usually one of two: 1) custom widget API built on top of OpenGL etc., or a well established widget toolkit like Qt (or sometimes GTK). But when it comes to enterprise stuff, it's usually much worse.
For sure, the ingame GUI is obviously rendered in the graphics API of choice. But games still usually needs a way to show error messages and sometimes some type of launcher. Doing this with pure X11 isn't that hard, ~200 lines will give you a styled message box, the probably only issue there is that fonts doesn't look too nice.
Those that use libraries like SDL will get such features built-in, along with lots of others including audio, input handling and special controllers such as joysticks. This is comparable to the feature set DirectX offers for Windows, and is appealing to many indie developers. But SDL is more like an engine framework, as it controls some the main control flow of the engine for you and offloads it to a library, so you wouldn't get the precise timing and threading control you want for a high-performance game engine. I consider it more of a learning tool than anything else. Taking an already built game engine and porting it would be cumbersome. Using something more barebone like glfw is probably the smarter choice, or better yet build your own, which is what I've done for 20+ years with OpenGL. The others will still remain though; like IO from the OS and audio. The latter of which is an area where I think we need an upgrade/replacement of the very outdated OpenAL. Linux does have the advantage of providing a very bare metal access to the physical devices in addition to events from X11 or Wayland. One last thing outside my expertise is networking; I haven't touched network programming since the early 2000s, so I don't the state for game development there.
My point being, if your game is already in Vulkan or OpenGL, porting the graphics is very simple as you say, the other parts probably needs some effort.
).