That actually sounds great then. It probably also depends on how the phone is used, but that likely applies just the same to play services. I mean comparing if you’re always in WiFi, mostly on mobile or often on bad mobile.

But if it barely shows up for you, that’s good to know. I just remember when first setting up notifications that the options without play services had semi-scary warnings associated with them, which stopped me from using them (and having play services running anyway, it didn’t seem to make sense to not just use them).

Home assistant by default uses and relies on play services for notifications. It’s what allows instant and reliable notifications without also causing significant battery drain.

This can of course be turned off, but it comes with significant warning labels that battery and reliability will suffer. It has to hold a web socket connection, or frequently poll for updates/status.

From my understanding system level support is needed to make this efficient and low battery, but I don’t know if graphene or similar roms provide something like this (or if they even can)? I would also be very interested to know.

This isn’t an important choice. If you don’t like the distro or you are having issues, it’s like 10 minutes to switch to a different one.

That being said, I would not recommend Ubuntu for a variety of reasons. CachyOS is a great option, but there are others. Just try it, if you have issues or don’t like it, try something else.

Check the commit history. Claude is part of basically every commit that I saw for multiple pages, I didn’t check further.

Long-term cause it has at this point been around for actual decades.

If you’re gonna use an overlay network like Tailscale I would highly recommend also taking a look at netbird. Similar concept, but fully self hostable if the need arises (but not recommended unless you have a specific need).

There isn’t one. That’s exactly the point. The original article had issues with vendor-specific passkeys (Google, Apple) that essentially are only accessible through their ecosystem.

But that’s exactly the problem. Those that were the target audience for the article do not have well managed passwords. For them the passkeys arguably do increase security, while having a bit more “lock-in”.

For those that have well managed passwords (like probably most in this community), they are perfectly fine to use because they don’t have lock-in either. I have full control over my used passkeys.

Basically you’re not getting the “normies” to use proper password management, and for them they do provide additional security and convenience. They can always add more passkeys on different vendors, you’re not limited to just 1 per service. Yes, adding them is a hassle, but so it’s managing passwords and following higher security standards in general.

My various servers, firewall and nas are relatively power optimized. Everything including switches (most with some 10g ports) and access points is 160 W or so combined. I can’t give you monthly usage, as I’ve got solar+battery so it’s frequently just 0 kWh for some days. It’s also rather hard to separate from the other power usage in the house when solar doesn’t fit cover it.

Generally not possible either. Unlocking the bootloader always wipes the phone. So unless it’s already unlocked, or it can be unlocked with some sort of exploit instead of using fastboot, it’ll be worse than a (bad) normal backup.

Or another TV of the same manufacturer that has Internet access.

Stable vs Unstable does not mean anything to do with bugginess or prone to breakage

It would be great if that’s how the words were actually used, but in most discussions I’ve seen it is certainly not the case. Also, when talking to new users, they are supposed to magically understand and know the word is used in a niche way in this context? It certainly won’t be what they take from it when hearing or reading it.

Unstable means packages update frequently including major versions

That’s what we have rolling or bleeding edge for. Sure, unstable can mean that when used within the same project/context. So having different releases be stable and unstable for the same thing, but across projects/distros that’s just confusing, especially for new people. Yes I’m aware of the historic reasons and the roots in software development, but I’d argue it isn’t helpful to convey the practical reality of the differences.

I didn’t claim anything is bad, anywhere.

Also I didn’t “try to use Debian in a way it wasn’t inteded to be used”. I am using debian on 90% of my servers/vms/containers to this day. Have been for literally decades. To give a specific example: I was running a nextcloud, which required PHP, and wanted to update that. The newer version reuqired a higher PHP version, because the currently installed PHP version was EoL and not supported anymore (by PHP maintainers or Nextcloud). So updating required removing the debian-provided PHP version and installing one from other sources (external repo or compile from source), which honestly defeats the point of a “stable” distribution if the software provided is EoL, doesn’t it?

And to be clear this is just one example I can specifically remember. The “issues” I encouter when using debian are usually of that general type. And to be clear, that’s mostly the point of using it on a server, where I use it, but even there it sometimes goes a bit too far.

And finally: I was originally arguing the semantics of calling debian stable and arch UNstable, which it is not. It is rolling, it might be bleeding edge. It’s also isn’t the same as having a stable and unstable release branch (within the same distro) when applying the label to the most stable release of different distros.

I didn’t say “it suddenly breaks”. I said I have more issues with it, which is decided different from “it just works”, which has been my experience on CachyOS. Specifically I mean issues with installing (up to date) software, or using software that has requirements that I just can’t install the required version cause the repo has a version from 3+ years ago. Not everything has to be “bleeding edge”, but being so stable that everything is utterly outdated is also a kind of unstable.

Sometimes this goes so far that the “stable” version shipped by debian is already unsupported upstream with all the security implications that entails (I had that once with PHP, I think).

Compare the other extreme Debian, which is very stable.

In the sense that is unchanging, sure. But that’s one hell of a bad word to describe that in this context. In my experience the incompatibilities caused by only ancient versions of things available in the repo of even a newly released Debian are not helpful. Stable wouldn’t be the word I choose to describe the situation, or the fiddling and tinkering that results from that. Often having to go outside of the repo to solve, defeating the whole point of said repo.

Arch in my limited experience is much more stable, in the sense of “just works” and always having up to date software. If I want to use something, I can just install it (from the repo), the chance of it not being the current release are slim.

Similar to NetGuard there is Tracker Control. It also uses the VPN “slot” but it can additionally run a wireguard tunnel itself now. So you can combine access control, adblocking on DNS level and actual vpn.

Edit: added link

I have no idea what people keep talking about when they say Arch/CachyOS is “unstable”. I have had so many more issues with running Debian (usually on servers) over the years. Since I swapped my desktop to CachyOS over a year ago, I have basically no issues at all. Just one very minor thing that fixed itself after a few days (bug in kde on new version released). For like 15 months of use, that’s spectacular compared to any other system I’ve ever used (especially including Debian and of course Windows, obviously).

What exactly is the goal? How do you make “no issues” and “it just works” more stable?

I’m remoting into my work machine with RDP in a similar way, also using multiple (local) screens. I happen to be on CachyOS (Arch), but I would assume this works on any distro.

I just call xfreerdp3 directly with the appropriate command line options to span my local monitors which I want to use. I only use 2 of the 3 I have, as the 3rd remains for my local stuff.

Zero issues. Just works once setup.

A lot of people keep suggesting Remmina, but I couldn’t get that to work even slightly. Let alone with fine grained control over monitors.

Many solutions exist, rdp clients aren’t exactly rare. I’m personally using xfreerdp for the extensive options: I can connect to a Windows system (at work) and have it be full screen on 2 of my 3 local monitors. It’s fast, reliable and just works honestly.

But what you seem to be looking for is something more like RustDesk or vnc. You want to look at and help with an existing session. If you connect with remote desktop, the local monitor gets locked, so you can only take over, but not show or help. VNC over the open Internet is generally not a great idea for security, but wire guard exists, as does TailScale and/or NetBird which means you don’t need to expose it. So that also works.

Having your own collection is great. But it doesn’t provide the service Spotify does (or any streaming service). 80% of the time I listen to discovery-type generated playlists. I want to find new music. This is fundamentally impossible with the music I own. This is something you can’t self host. Even if you have a vast collection of music you don’t know (by whatever means your get it), you still need the algorithms to pick the music that you’re likely to like.

I really wish I could. I self host basically everything else. Even tried some local music similarity training for “smart playlists”. It’s kinda neat at best, but no where remotely close to the music discovery of Spotify and other online services. You need the massive amounts of users to derive that data.