Im on mobile, but I’ll try to circle back later and provide more details and links. Also I use NixOS, Niri, and DMS, but also was on KDE before switching over, so a similar setup but not the same.

Problems 1 and 3 are related. Basically there should be some configuration file for polkit/pam, and in there you have to configure to allow fprintd (or try and see if theres a setting for howdy, I know theres extra config you have to do for fido2 and it should be similar if possible) to unlock greetd. Should should also be something about unlocking kwallet on login, but with kwallet you have to do some weird settings on top of it to enable to login password to work, or setup a blank password wallet and set thag to default. I switched to gnome-keyring and seahorse (gui) when making the change to niri for the reason you mentioned, its better supported and needed less tinkering.

Problem 2. Kde keeps like a cache of its “.desktop” apps somewhere, and you probably have to blow it way and update xdg on where to find the “app launchers”.

For extensions or extras. Honestly I would try to stick with anything noctalia has in its plugins store, and let it handle any “enhancements”. The only other things id recommend is when youre looking for replacements for KDE apps, like dolphin, Kate, etc, is to look for all the lightweight wayland native ones that people typically install when they are building their own “DE” on top of niri (instead of installing noctalia or DMS), aka look for post where people are showing off their waybar setups. I switched to whatever the popular image view is, among other tools, and even though they seem like they’d lack features or be more geared towards terminal users, they’ve been great. Replacing dolphin has been the only hurdle, i cant find anything that’s as clean and powerful yet, but have high hopes for Cosmic to mature so i can start stealing from thag instead.

Again, apologies for the vagueness, I dont have my config in front of me, hopefully its enough to get you on the right track.

Its most likely illegal, but depends on where you live (some country might have more flexible copyright laws that have that situation as an exception).

But I would consider that ethical and moral. When HBO deleted a bunch of their cartoons for the tax ride off, one of the creators of the cartoons tweeted that they themselves had to pirate their own show just so their kid could still watch it.

I pre-ordered so I read about it in their updates/blog, but they have been actively working on getting everything mainlined.

Search for “Suspend-and-Wakeup on i.MX8MP” and they mention it briefly, they went more in detail about it in previous blogs though: https://mecha.so/blog/mecha-monsoon-update

On top of working with the chip manufacturer to get the cpu side mainlined, they’ve also been working with the contractor firm “Igalia” (who’ve been hired to work on mainlining and upstreaming a bunch of open source projects the past few years) on getting the gpu drivers cleaned up and “mainlined” as well.

Honestly a lot of the work they’ve been doing will benefit other linux on arm projects, especially mobile. I agree with you though, if a project isn’t putting the work in to get everything mainlined, then its not worth considering (atleast to me). I wish Ubuntu touch and sailfish all the luck in the world, and respect the work their doing, but for me its not a option.

No matter what I said it wouldn’t stop people rationalizing that I actually double super secret meant that its OK to genocide and starve children within (most likely) a few years, so id probably just have fun with it and sing the opening to “backstreet back”, including the bass line in accoplelo.

So right now im kind of doing option 1, but .local (I’ll switch to .internal) and not using * ( which I will change and see if it let’s me use the subdomains I configured in caddy that werent working before)

Option 2 it what I considered switching too, I have adguard running on my router, but was testing option 1 when I decided that I might want to skip ahead to something more advanced. I really dont want to have to install certs on each device to get past the warning for https.

Option 3 is on my backlog, I have a vps I planned on setting up pangolin on later down the road after I got some of my services spun up, but maybe I should take care of that now instead.

Option 4 is something I briefly looked into before realizing its way over my head, and im also pretty risk adverse.

Sumarry: so maybe option 3, and possible option 2 if its still needed when on my local network without having a client running.

Donnie Darko and Final Fantasy 14. The Shadowbringers expansion would kind of fit the mood.

I didn’t pirate that movie, I was training my AI. I didn’t hack that company, I was demonstrating value to my shareholders. I didn’t run that red light, the self-driving AI I vibe-coded hallucinated that the light was green, only happens 20% of the time. I did my homework, but the AI broke the guardrails and deleted it. I didn’t fart, those are just the fumes from the data center that my city council were bribed to force through even though 99% of their constituents showed up to voice their objection against it.

“What we own to each other”, a talk based entirely on me using misrembering philosophy quotes from “The good place” to justify my personal beliefs.

Definitely not what I should pick, but the first thing I came up with within 30 seconds.

Follow up talk would be “platos lair”, explaining how we’re all delusional because we’re so upset that “vimms lair” isn’t what it use to be due to Nintendo being assholes.

A percentage of how much our core values/morals/beliefs align, weighted by how important they are to us.

It would make it super easy to know upfront who I should include in my life and who I shouldnt.

Right, by process I mean people have to do that extraction manually, or setup some process or workflow to handle it, but the point is that it doesnt retain anything outside of the scope of the PR. So the bosses whole plan is fundamentally flawed, which would be a useful thing for OP to bring up. Trying a different approach and instead of objecting to it on principle, which the boss showed he doesn’t care about, instead express that even theoretically its impossible for it to “learn” without doing additional steps.

LLMs are non-deterministic, so of course its not consistent, but you can increase the chances of it trending towards more useful output by weighting the dice aka with skills, instructions, etc. And just to be clear, I agree with what you said, its a fact. I tried to express that with the quotes around “learn”, but maybe it didn’t come off that way.

Github copilot doesn’t retain anything you say in a comment, outside of the PR. It wont improve over time unless you have some process to extract comments and “learnings” out of PRs into either the org wide “copilot-instructions.md” file or into repo specific “Agent.md”, “copilot-instructions.md”, and “skill” files.

I did see that post previously, but thank you for linking it!

The info on the kernel is exactly what I had in mind!

At some point I would love if you wrote a guide on how you got started, things you learn along the way, pitfalls to watch out for. Not only around getting everything working, but how to upstream it to the kernel.

I know in a previous post you asked for help on contributing to the kernel, and now you already have PRs! That step has always intimidated me so seeing how you approached it would be super helpful

Link to repo its forked from - https://github.com/Maxteabag/sqlit

Learning one language makes it easier to learn another, so dont worry about which one you start with. Basically the “language” or vocabulary changes, but the underlying “grammer” (as in noun, pronoun, subject, adjective, etc) is for the most part universal. Like knowing what a library, functuon, method, class, loop, variable, etc will be transferable.

For a project, I dont know what you do for work, but you mentioned vba macros in spreadsheet, so id recommend building off of that. As someone else with adhd, the more interested I am in something, the more motivated I am in pushing through roadblocks when its not instantly easy. So id recommend trying to automate some task you do manually for work, something you know a lot about and know what the inputs and outputs should look like, and write something for that.

Yeah I usually have the opposite problem op has.

I dont write as much code as people thing engineers do, most of what I do is translating what the user says they want to what they actually want

But the code i do write is so novel and tailor made to our implementation of standards, that AI isn’t useful at all.

What’s weird is the few times im doing the kind of work OP is talking about, is the only time its actually useful because it can generate boilerplate code that interacts with common libraries and tools, which it has a lot of training data on.

Thanks for going through the effort!

Just a heads up, archive.today might not be useful ever again, your issues with it could be related to this: https://en.wikipedia.org/wiki/Wikipedia:Archive.today_guidance

So a lot of documentation in general typically only covers beginner or expert levels, with almost no intermediate.

It’s always “here’s how to say hello world” or “here is how to extend our implementation at a low level”.

So it’s nice to have examples of intermediate use cases for power users who need to do more than the most basic examples, but don’t plan on submitting PRs to add or change functionality.

More info, guides, and examples are always better than less (as long as it’s accurate).

I’ll definitely have to take advantage of the hardware acceleration.

In regards to gaming, what kinds of games have you typically used this for? I’m not thinking about using this for competitive shooters or something like that where every ms counts, but what about games like a souls like where timing a parry or something matters, have you tried anything like that?

Lacking those features is perfect fine for me, I’m looking more for a remote display and input implementation anyway.

One last question, have you experimented with any compression or “enhanced” encryption for the ssh connection? If so, has that impacted the responsiveness at all?

You have no idea how grateful I am to you right now.

This has been on my personal project backlog for awhile now. I have run into so many issues and headaches with possible rdp or vnc solutions for my desired use case, especially with wayland being a must have.

I have recently fell in love with ssh and was planning on looking more into waypipe as a possible route to take as I kept seeing it recommended, but the examples and documentation i found was always “generic” or surface level and didn’t have enough of the pieces I needed to scrape together.

I’ve been putting it off due to only having surface level knowledge about all the pieces, and you merged like 10 of them in one go.

My biggest question to you is what is performance like? Like picture quality, audio delay, and latency/responsiveness around mouse and keyboard inputs? How does it compare to using something like an ipkvm, rdp, rust desktop, etc?