I’ve started doing podman quadlets recently, and the ini config style is ugly as hell compared to yaml (even lol) in docker compose. The benefits outweigh that though imho.

USB GOOD

Sodium is basically ubiquitous and sodium batteries are much easier to produce, which is not to be neglected amid rising geopolitical tension. They are also a lot safer to handle. The technology has improved to be ready for the market in almost no time and is still improving rapidly.

They won’t be in your smartphone or wireless earbuds where energy density is key, but for large storage grids or maybe even cars, they might replace LiPo or LiFePO4.

It’s more than that honestly, or less, depending on the perspective. Most people share their data with “only our 900 bestliest partners” or more everyday. Every tweet, login, whatever metafata can be more valuable and easier to compute than voice or video, depending on what you are looking for.

I’m not familiar with Nova launcher, it seems to have some USPs to it.

But these days I always recommend Lawnchair. It’s built on top of the Pixel launcher, which in itself is a very good launcher IMHO, and refines it to be more customizable.

https://github.com/yt-dlp/yt-dlp

Although I absolutely understand that the common folk will have a hard time downloading porn in the CLI.

You will own nothing and be happy.

Yes and no. I have enjoyed reading through this approach, but it seems like a slippery slope from this to “vibe knowledge” where LLMs are used for actually trying to add / infer information.

At least Orca Slicer works fine for me in Wayland.

Screen sharing is a great example. I used to have issues with it, but since about a year I’m able to share my screen in the MS teams PWA in Firefox and even the Discord flatpak without a hassle.

Do you have some examples? Most things I (and others) do are in the category “showing a desktop”, multiple desktops with different resolution / scaling / refresh rate, maybe opening a virtual monitor using krfb.

Wayland has been a complete game changer for me regarding performance and reliability (as soon as it hit a certain stability lol).

You’re absolutely right! Here is a recommendation for companies to invest in that produce fake eyelashes for car headlights:

In Signal, push notifications simply act as a ping that tells the app to wake up.

Google only pings your phone via FCM, the Signal App then polls the message itself.

Edit: oops, you meant the pop-up. You can disable them from showing the message.

It’s a tale as old as software development itself: Fuck around with your license, find out by getting replaced by a fork.

Although Redis came back and reversed the relicensing one year later lol:

https://dirkriehle.com/2025/05/03/re-relicensing-to-open-source-explained/

Those are rookie numbers

mstsc.exe was exactly the one I used. I vaguely remember that I had used that successfully some years ago (in a much bigger environment with proper certs) and it worked.

Regarding NLA: I believe that I would have to disable that on purpose no? It was on a very vanilla Windows 11 install. I just looked and regarding any other settings than the ones listed in mstsc.exe, I get told that Windows 11 Home edition doesn’t support RDP lol.

I recently used the old Windows RDP client they refer to. I tried to connect to a Windows VM and it didn’t work. Had to pull out some old log utility tool and filter a while to recognize the server didn’t use a valid TLS certificate (lives inside a VPN) and the handshake failed. Tried disabling cert validation (although I’m not sure if that one obfuscated option did exactly that), still no luck.

I then tried KRDP in KDE. It asked me if I’m sure I wanted to connect since the cert is self signed. It even showed me the cert info and I know the issuer in person. I accepted and got in. Easy as that.

This is a great summary. Also, see the GrapheneOS teams list of hardware prequisites. They have said themselves before that this is not a very high bar, but that there are just no other manufacturers focusing on security. Apple claim they do, and get recommended as second best option in the forums sometimes, but given the walled garden approach, it must be next to impossible to develop against, even if they unexpectedly completely open up their ecosystem.

They would probably also have to take a lot of criticism for their implementation approaches and their brand integrity would suffer immensely. Maybe there would even be some new undisclosed vulnerabilities to fix, like the goto fail bug breaking SSL encryption ten years ago.

Compare it to the new open source Nvidia drivers for Linux, they have taken quite a while to develop since Nvidia announced the release, and I don’t know if they have yet reached the performance levels of the proprietary ones. Doing this for a whole Phone, given they even fulfil the requirements hardwarewise, will probably take a decade. And in this decade, ten new iPhones will be released…

I’ve heard maintainers getting spammed with AI generated Issues and stuff. This is just a stupid meme from some weeks ago, trying to merge some Anime ASCII Art into the Kernal Panic log message, because the kernel should be more weeb-friendly.

You’re right, Linux Kernel development happens on an own cgit instance. Given how much Linus loves distraction, he’d be better off by not looking into Github at all.

Just called it the other day, it’s a reality TV show. That’s why they acquire all the media outlets, the viewer rates are insane.