Personally my arch install is almost boring me with how stable it’s been - and if anything goes wrong, it backs itself up before and after every single update plus on every boot just cuz, so I can roll back to wherever I want. I’ve put a lotta work into building out all these redundancies I’m happy with, and arch has been so goddamn stable I haven’t even had an excuse to use them. The process of getting a complete install was absolutely not “it works” - but now that I’m there, yeah, it really does just work. My only complaint is that I don’t have any reason to tinker with it more.

Honestly I’ve found the opposite of what you said, where on Debian based distros I commonly had to go to a project’s git repo and follow readme instructions to build when it wasn’t in an apt repository. Meanwhile on arch, the only thing you have to install manually is yay and then afterwards everything is in the AUR. Not saying that makes arch more user friendly than Debian (obviously), but that one aspect I do actually find easier on arch at least if you’re willing to use an AUR helper.

I don’t think I’d call it anything wrong, but the subvolumes definitely do make it different for installation purposes so that following ext4 instructions for bootloader configs or kernel arguments could put you on the wrong path

Some desktop environments (plasma 6 at least) support it when running in Wayland - my daily driver for months has been an HDR setup (with a nvidia gpu, even) and it’s been great. It’s not quite ready for non-technical users, but imo it’s not far off and I can’t wait for it to be more common.

Yea as opposed to the windows method of “just open regedit and navigate 8 folders below HKEY_CURRENT_USER to change some ambiguous system variable in hex” lmao

I’ll take editing a text file in /etc/ for my configuration any day

I remember being able to figure out a solution with the kwin docs and dfeet for introspection, but it was a while ago and my memory is less than stellar lol. I ended up running all my window-related logic in the kwin script (js) and just using python+dbus to see if the script had been injected and do it if not. If you go the same route (though the python is unnecessary this way if you aren’t using it for something else, like running one of the windows you wanna manipulate) the workspace global variable stores all windows in stackingOrder, so looping through that list (for (window of workspace.stackingOrder) {…}) is an easy way to check each window. I definitely remember docs on the workspace/windows part, but tbh I think I just introspected with the jsconsole.log and the log kwin prints it to (journalctl maybe?). Sorry I don’t remember more about the process, I got into the kwin scripting for all of an hour cuz of the Wayland regression of not being able to control your own window’s size/position in qt so it was a bit of a hack fix I haven’t had to think about since.

I’ve had good luck interfacing with KDE windows using dbus in python, although if I remember right the more granular control also required javascript (still passed thru dbus with python). The only python library I needed was Qt (for their dbus implementation because I was already using their event loop, but any other lightweight dbus package should work fine).

I hadn’t, and it was definitely worth reading so thanks for the link. I’m still not sure exactly where I stand on the big AI companies relentlessly scraping everything for training data, but that was very convincing that copyright laws aren’t the solution (and I already believed better labour laws were needed for artists, though the details of exactly how music artists are getting shafted were new to me). Thanks for the link!

Whether or not you agree with AI image generation, the authors of this study have pulled off something impressive. This particular study isn’t going to be the single most important thing to humanity this year, sure, but they made a pretty clever stride in pushing a developing field forward and you don’t need to be excited about the field itself to appreciate that.
I’m assuming your dislike for AI image generation is based on the plagiarism issue, which is absolutely valid, but model architecture is separate from training data and the concepts here are perfectly usable with a more moral training set. The companies scraping all the data - OpenAI, google, and to a much lesser extent stability AI - are the ones to blame for that problem, not researchers working on model architecture.