The COPR package didn’t work for me on Nobara, so I had to build from source, but it works great. There are a couple of things I don’t like, but overall seems pretty neat.
If I can get Xwayland to work nicely for steam with high refresh rates, then it seems like this might be the WM for me until COSMIC-DE comes out.
Lately I’ve been using chatGPT to create a bunch of small custom python programs to do stuff like this (if I can’t easily find an existing program to do what I want).
For example I would tell it something like:
<span style="color:#323232;">Create a python program that does the following:
</span><span style="color:#323232;">
</span><span style="color:#323232;">-asks the user for a directory to process
</span><span style="color:#323232;">
</span><span style="color:#323232;">-sorts the files in that folder according to file type, placing them into appropriately named sub-folders, eg all image files into a folder named "images", all music files into "music" and so on.
</span><span style="color:#323232;">
</span><span style="color:#323232;">-creates any new sub folders before moving the files
</span><span style="color:#323232;">
</span><span style="color:#323232;">-moves the files verbosely
</span><span style="color:#323232;">
</span><span style="color:#323232;">-gives the user a notification upon finishing
</span>
You can customize it to do exactly what you want, and it takes only seconds for it to give you the code.
I can’t even begin to tell you how much time it’s saved me over the last few weeks, automating simple stuff that would normally take ages.
I think that’s a good start, but the baseline of what AI can do. These scripts are around since filesystems have been invented. And you can do this with one (lengthy) shell command. Or one of the already existing file sorting utils. (something like this [Edit: see next comment] or Hazel or DropIt) With those you can even configure if it should recusively visit subdirectories and do individual subdirectories for the filetypes or mangle everything together for example in one big unsorted mp3 directory.
What I’m waiting for (I’m not OP) is something that looks at the content of the files. Do a directory for all the manuals I downloaded for the household appliances, find out on which event I took a photo and make a correctly named album for that, find the project files for my diverse electronics projects and file them into seperate directories together with related info. And find the mp3 files and TV recordings with a mismatch of metadata and folder structure.
I think what you’re describing is definitely the way things are heading. I would love a teachable, automated AI personal assistant. But I think I’ll wait for an open source, hardware agnostic version that I can self-host.
Does this “obviously” have to use AI? I can see a tool that sorts files into folders based on file extensions, modification dates and/or metadata could get the work done.
And if organising files by content (e.g. “my zoo trip”, “meetings with Xenia”) is that important, doing it manually seems like a better idea because accuracy is presumably important.
I don’t really see the distro hopping argument either. Even if you don’t share your home directory between installs, presumably you copy over your files as directories rather than individually pouring them into one super folder?
I think a lot of folks with very limited IT knowledge think AI will solve things that have been solved for decades.
The issue is availability and elitism. A noob user doesnt know how to find this stuff, google is so rotten that its not help anymore and pros often just shit on them instead of making a comprehensive wiki.
If you want just boot your system and not have to worry about setting up keybindings, my best suggest is ArcoLinuxB i3 Edition and Garuda Linux i3 flavor, you really don’t have to worry at all for that, and you can use the i3 reference card to learn the most common keybindigs.
I have had it installed for a while and I check it after every update. I can’t use it yet as my daily driver because of scaling issues. The desktop scales properly but windows do not. Fonts are too small and the cursor is tiny. I figured out how to scale the cursor manually but I couldnt scale the windows.
(begging forgiveness, I haven’t read the comments yet).
Regarding backups - I started with using Ubuntu and its Backup application. This application is a front end for a command line package called Duplicity. One of the things that annoyed me about the backup app was that I couldn’t work out how to reschedule the scheduled backup.
Taking control of my own backup setup was the answer. Learn about bash scripting so you can create a short bit of code to handle your backups. Read up a little on duplicity, read up a little on mounting remote file shares, read up a little on setting up an ssh key for encrypting your backup.
This may be an heretical thing to say but I found ChatGPT quite useful in answering these questions (as always with anything you get from an LLM, double check it’s answers against reliable sources).
Thanks for mentioning this. I‘m actually scripting quite a couple of things in bash and some in python already. I had the exact same idea.
But one reason I wrote the post was because I wanted to share my experience with debian (and ubuntu) for users that are less experienced than I am.
I even have a custom made backup script for the 50 services I run on my two ubuntu servers. It is even self cleaning.
Also tried chatgpt but so far I didnt have any luck. The code it spat out (was for screen brightness control) didnt work. But I did get it to work in the end.
there’s the “add tiling features to a DE” path – Pop Shell / Cosmic DE is the best known, but KDE has some pretty decent options and there’s a couple Python scripts (at various stages of readiness) for Xfce
or the “add a DE to a tiling window manager” – Regolith is the best known here (basically swapping i3 for Mutter), but along those lines it’s “relatively” easy to swap out window managers in the desktop of your choice (i3 + Xfce being an easy choice)
If there was a regolith but based on river or dwl I would definitely do the switch, cause i do like a more dynamic tilling workflow compared to the manual tilling
linux
Hot
This magazine is from a federated server and may be incomplete. Browse more on the original instance.