Oof… it actually happened to me and it’s not 1 problem but 2 namely :

• you ran out of disk space while updating
• AND one of the messed packages is one that is required for the upgrade process, e.g. curl or wget (sorry can’t recall which it actually is)

so that leaves you in a terrible stable. You can still clean up this mess BUT that’s tricky. Basically you have to

• actually find out what’s taking up space (often old kernels) or “just” give up on data temporarily (basically you move your /home, or part of it, to a USB stick) via rescue mode (you need to be familiar with the CLI) or remount the disk on another working system
• get the actually missing packages via another working system then install the minimum you need then finish the update

For me it was on a small temporary system (e.g. RPi for HomeAssistant) so it was basically easier to recover from a recent backup after formatting.

It’s annoying but it’s actually not that bad.

You’d have settings for when to stop seeding, e.g. 1:1 ratio minimum, duration of the track xN, etc with a reasonable default. Suggestions welcomed.

I would recommend against a new player when existing scriptable ones like vlc and mpv already exist.

Instead what I would do is a plugin for either, eventually repackaged as its own player (if somehow installing the script itself is too much for some) for which the script would

• include a very small torrent client
• point that client to the torrent (which AFAICT is still not public, so for now a reconfigurable URL)
• include a search function that when it fails, proposes to search within the trimmed cleaned torrent metadata then does the torrent download then plays.

Because they literally wrote the book on lock-in https://fabien.benetou.fr/ReadingNotes/InformationRules and they tried with all their might to stop free software https://en.wikipedia.org/wiki/An_Open_Letter_to_Hobbyists so beside the money and power they have been strategically at it for decades. Dependency is deep in the product.

Please feel free to help right now. You can still move to the EU if you want to but if you take for example NLNet they fund open source work for anybody anywhere in the World, you “just” have to propose something that is new and needed with a focus on the Internet.

Send an email to people you care about with your new email, they reply or not it’s OK you still have their email address anyway, no big deal.

Source : I’ve migrated away from GMail.

It’s a VERY specific tool that needs

• a lot, like World scale, amount of data and that has repetitively been done WITHOUT permissions from authors of that data
• huge amount of data must be processed and this is done in enormous datacenters that consume radically MORE than traditional ones without GPUs
• energy and cooling for those very specific new datacenters that then becomes unavailable to the local community, energy produced that is often rushed and typically more polluting

So I think it is fundamental to distinguish

• “AI” as a theoretical researcher field, public research focusing on processing CERN data, weather forecast, genomics, medicine, etc that is indeed a tool that might produce results that helps us all

versus

• commercialized for-profit “AI” with GenAI and LLMs as blackboxes mostly used for spam, scan, low quality code, etc.

When one amalgamates one with the other, knowingly or not, they do the marketing for the later.

Because dockers record with regards to security is questionable

Works with Podman

• stop the container (safer)
• scp only its mounted directory once to the target, via ssh key (no password)
• next time rsync instead to add only the new files, ideally via a crontab script

I can recommend Storage Box by Hetzner starting at 1TB for ~4EUR/month. Been using it for a year or so.

On the RPi microSD, on a USB stick, on a SSD,… entirely up what you have available and what you need.

Immich fits on a RPi so not sure one needs a lot of memory for that.

Funny I have the opposite experience.

I use KDE Plasma, Firefox, konsole, etc and sometimes, no idea when and why, I just pick a file then drop it somewhere else, including ON the terminal… and it works?! Like it brings the full path for that file and then I can compose with CLI tools, amazing!

I’m quite used to the terminal so I rarely use drag&drop (mv, cp, scp, rsync, etc just work) but when I do I’m actually often positively surprise that totally different software made with different interaction paradigms (e.g. GUI vs CLI) do work well together. Overall I think https://specifications.freedesktop.org/ is quite impressive.

Gosh… wish I could upvote twice. Feels like we just gave a low cost (for now) chainsaw to anybody who wish they had a pocket knife then say “There, you can cut anything with that!” and somehow they forgot they can just buy some OK stuff from Ikea or a nice artisan. The need to “build” anything without taking a minute to know, not even the state of the art, whatever already exist out there and “fix” it by “personalizing” it is nuts.

Let’s not “vibe code” anything when reliable solutions already exist!

AFAICT it relies on https://git.ffmpeg.org/gitweb/ffmpeg.git/blob/HEAD:/libavfilter/vf_scdet.c

Neat, made me curious, seems to rely on https://ffmpeg.org/ffmpeg-filters.html#scdet-1

There are actual radios, like FIP.fr or plenty of other ones that just give you the m3u8 and their playlist. It’s not federated but it’s less centralized than most platforms like SoundCloud or Apple Music.

If you do want a specific song and album you can pay for it via BandCamp and get the actual file, DRM free, to play on any device.

If you do want a song or album it BandCamp does not have, or you already have a copy of, e.g. physical CD, and you want something less centralize SoulSeek still works.

PS: I have been running my own PeerTube instance for years now but I don’t use it for music, just videos.

Haven’t actually tried it myself but check https://github.com/yt-dlp/yt-dlp#sponsorblock-options

I’m listening to podcast on a daily basis. I’m using Escapepod on a deGoogled Android phone.

Basically I pay for premium feeds, this way the author of the podcast get money and I don’t ads. One example is “Tech Won’t Save Us” via Patreon https://www.patreon.com/c/techwontsaveus/posts?vanity=techwontsaveus and another is 404 Media.

How it works is that premium podcast feed have a uniquely generated URL that I then use on my podcast client. It then downloads the episodes from the feeds, no ad. It’s very straightforward.

No need for YouTube (terrible, part of Google/Alphabet) nor Apple (terrible as it’s doing everything possible to grow it’s closed ecosystem) to have quality content.

It’s all just speculations, both what you suggested and what others said.

You are on the right path with your screenshots but you might not be measuring the right thing.

So, you need a (paper) notebook to record objectively (not your biased feeling assuming a pattern that might not exist) when it happens and for how long. Only from then can you backtrack to WHAT causes it. Sure you can have some hypothesis (update related, screen attach/detach, BIOS, RAM, etc) but that should NOT lead to your data acquisition.

So you htop is nice but AFAICT it’s just about CPU and memory, it’s not about e.g. IO so consider instead iotop, in particular if one process is some indexing (e.g. locatedb). Theoretically if it’s not CPU/memory (which you are saying it’s not the case) then it basically just leaves IO, that can be again indexing, some heavy process that is bottlenecked on disk access, but can also be a bug, e.g. BT pairing/unpairing that happens faster than you can notice.

Think of this as a fun investigation that leads you to better understanding of your setup, good luck.