@Marvelicious @Epzillon Oh heavens no. I'm using it for the wife-acceptance-factor. I would have used a discrete Jellyfin box of some kind if it was up to me. My main point is that the smarts in smart TVs are trash, and is best turned off.

Yah, I'm not going to do that.

I think this can happen if the feed itself (the rss/atom file) has more articles than you specify. As far as I know this setting really only applies to articles it cached that are no longer visible on the feed.

All I do is look into the open issues, the community, docs etc. I don't remember auditing the code.

audience already agrees that complicity in genocide is an acceptable tradeoff to software freedoms I talked about that to show one possible counterbalance between liberty and usages which are probably not explicitly wanted by the authors. Another common example of freedom/restrictions is someone wanting to have their software permissively licensed while also not allowing cloud vendors to resell access to it. That's how you end up with licenses like Elastic's. Or, if you want another example of "free", look at the distinction between the GPL and the BSD license as it applies to Sony and the Playstation. One of the reason Sony chose BSD for the basis of its gaming system is because the BSD license allows for commercial usage. In that sense it is MORE free than the GPL, which would not allow the type of usage Sony did with the Playstation without conferring more responsibility to Sony, for instance, releasing their source. Under BSD they have no obligation to do so, hence it is more free in that respect. My whole point is a lot of people say "I want my software to be freely licensed" but they do not realize that they may be unintentionally opening the door to usages of the software that they do not want to see. One should not pick a license that allows for unfettered usage of the software if you have certain ways you don't want to see it used. As a final parting example, look at Prusa and their printers. They release the firmware and designs as open source. They they later get angry when companies clone their designs. This is permissible under the license. This is making Prusa want to lock down their future designs to avoid that usage. Anyone considering licensing of their own software should think very carefully about what usages they support or object to and license the software accordingly. If you release your software as BSD licensed and some company comes along and makes a billion dollars with it, you aren't owned a cent under that agreement. If this makes you angry, don't pick BSD.

Even on something like tor browser or tails?

I'm not spotting it. "AI" is only mentioned once. The key and secret in the docker compose don't seem to be API keys, but keys for directus itself (which upon a careful reread of the article, I realize is not FOSS, which might be anpther reason people don't like it"). Directus does seem to have some integration with openai, but it requires at least an api key and this blog post doesn't mention any of that. The current setup they are using doesn't seem to actually connect to openai at all.

You have a few questions here, which ones do you want answered? To configure the camera, you should have defined it in the config. That you don't know this means you should go back to the docs and read the setup section start to end. I write my frigate clips to an NFS share. I mount it on the host and bind Mount the path in my container. You can also mount NFS directly in a container, but it comes with extra steps. LXC is not a proxmox-specific thing. You can run lxc containers on almost any Linux and you can manage multiple containers with other software (lxd, incus, etc). At one time, docker was based on lxc, but both docker and lxc have evolved significantly since then. LXC and docker are indeed similar, but one aims to provide an OS-level environment and the other simply a software environment.

I want to believe you have given this some thought, but for someone with as long a sea log as yours, you seem to have forgotten what happened when we "gave it time to sort itself out" for other services that are now completely entrenched in our lives and have made them worse for it. apps for everything not raising more complaint about the erosion of our privacy by private corporations not defending open standards like PDF and now PDFs are a security and compatibility nightmare "hey, maybe subscription models can be applied to printer ink" etc, ad nauseum AI itself is fine, and its been used for good (solving protein folding). But AI in just about everything else is awful. It wastes energy and water. It is actively making people dumber. I'm fighting a losing battle at work with fools who wholesale believe AI answers on any question and others who literally vibe code. If you truly believe ai is going to be better in the long run, you have not been paying attention to the last 30 years of technology becoming trash.

Yeah that might end up being necessary, I haven't found any easy solution to provide the 5A/5V required by a Pi5. Still pondering the power supply issue.

I also like Stalwart. It's easy to setup and does its job very well. I'm just a bit nervous that the development team consists of one (!) person. Btw, can anyone recommend an e-mail client that speaks JMAP?

Okay then I will recommend the new Beelink ME mini https://www.bee-link.com/products/beelink-me-mini-n150 And also the "pocket NAS" from CWWK.net currently looking at a refurbished optiplex 3060 with an Intel I3-8100T Those are fine but will be less power efficient than the mobile CPUs. Do you mean building a multibay NAS myself? I looked into it and it didn't seem really cheaper It definitely is. Also most NAS have a very weak CPU. Because a "NAS" really doesn't need one.

I use this to help my grandma remotely! The two steps needed were to join her into my Tailscale network and set up SSH with key authentication only. Now I am able to SSH into her computer and enable VNC (remote control) and connect to the VNC-server over an SSH-tunnel like this.

Not sure, I use it containerised though

ArchiveBox strikes me as being a rickety pile of hacks, but it does mostly do its job out of the box. The built-in search is abysmal however and must be replaced with one of the other options to be useful.

I went through this about 6 months ago. Just build playbooks from basic to specific. I did so in three parts: Container creation Basic settings common to all my hosts Specific service config & software Ansible assumes you have a hierarchy of roles to apply for each service, so layering playbooks this way should help

However, the Small Town Murder podcast is excellent.

None so far. And I am using pretty much all the features. It is also great for my current migration from docker compose to kubernetes. A newt client on both and I can just switch on the pangolin side.