I think a lot of engineers underestimate the amount of non-technical effort that goes into running platforms with users.
It’s a shame I think there are some really interesting problems that a less profit driven dating app but currently anything without a big marketing budget isn’t going to have the userbase to support it.
Modular computing always seems to fall victim to Moore’s law/similar, everything core get so much better every 5 years, that by the time you want to refresh anything, it’s usually time to upgrade the whole thing.
Periferals are nice but USB is already a multivendor connector standard and if the choice is 3d printing cases or trusting a vendor to exist in 10 years I’m betting on 3d printers for now.
Hope I’m wrong but this isn’t the first time I’ve seen a project like this.
At the end of the day the n900 was the ultimate portable Linux machine, but it died because within 10 years you could do all of that on an android device AND have a decent phone too.
Start on any standard distro do a net install or equivalent and pick a different DE, I’d recommend Ubuntu/Fedora/Debian/Suse, but YMMV
It supports sharing your clipboard from mobile to desktop you just need to select the click click share & then click on your desktop.
Krita is what I use but I also find text handling difficult so I always do text last.
Can you label rules, that would be a better approach IMO.
Not familiar enough with UFW but could you parse the output and store the rules number as a variable if this is all one long running script?
Why the move from SDDM to PLM?
Do you have examples?
Because most of what you are listing is stuff that has been using ML for years (possibly decades when it comes to meteorology) and just slapped “AI” on as a buzzword.
What advances?
LLMs don’t have anything to do with abstract ideas, they quite literally produce derivative content based on their training data & prompt.
The same can be said of the approach described in the article, the “GPLv4” would be useless unless the resulting weights are considered a derivative product.
A paint manufacturer can’t claim copyright on paintings made using that paint.
Seems like the easiest fix is to consider the produce of LLMs to be derivative products of the training data.
No need for a new license, if you’re training code on GPL code the code produced by LLMs is GPL.
flatpaks are all updated at once, just like distro packages, so yeah you might need to commands, but that’s still very different to having each application update itself (and the security hell implied by that)
Also I think pkcon can manage your updates across various backends (unless you are on Arch, where I think there are both technical & ideological objections to having a simple tool that just works)
What language? What does it do?
I’m no fan of GitHub but if you set up dependabot (or equivalent elsewhere), you can ramp maintaince down to running PRs through tests every few months (assuming it’s not software with a large attack surface)