That solution still introduces lots of friction. At the volume and rate that these bots want to be traversing the internet, they probably don’t want to be fully graphically rendering pages and spawning extra browser processes then doing text recognition to then pass on to the LLM training sets. Maybe I’m wrong there, but I don’t think it’s that simple and actually just shifts solving the math challenge horizontally (i.e., in both cases, the scraper or the network the scraper is running on still has to solve the challenge)

I’m sure you meant to sound more analytical than anything… but this really comes off as arrogant.

You make the claim that Anubis is negligent and come and go, and then admit ton only spending minutes at a time thinking of solutions yourself, which you then just sorta spout. It’s fun to think about solutions to this problem collectively, but can you honestly believe that Anubis is negligent when it’s so clearly working and when the author has been so extremely clear about their own perception of its pitfalls and hasty development (go read their blog, it’s a fun time).

ucore (soon to be cayo) on my home server!

yeah this is the way. as part of my borgmatic script I bring down all the stacks that have databases, let the backup run, then bring those stacks back up. as long as the containers aren’t running (and as long as the container properly closes itself down, not usually something to be worried about though), any method of data back up should be fine.

I do this with quadlets and systemd targets now but before I was doing it with a bunch of docker compose down commands.

It is quite convenient for restoration, as you say

This is awesome. Bookmarking.

Am I totally off-base in thinking that MagicDNS and pluggable DNS nameserver overrides are a huge feature of tailscale?

I love that I can refer to my tailnet devices just via their machine name. I use it everywhere. And also that I can just slot in my NextDNS ID so that any device running tailscale now automatically uses that, and I don’t have to mess with my shared router settings or per device settings. Is all that actually really easy to set up outside of tailscale? Cuz if it is and I just somehow missed that when doing all my research, I’ll happily give plain wireguard or other mesh orchestrators like NetBird a go.

And I already know that mDNS is not the answer. That protocol is simply not reliable enough.

I think you may have shifted the argument a bit.

We’re not “back to where we were 100 years ago”. Bandcamp exists and pays artists for song purchases. It’s not perfect, and the selection of Bandcamp and the few other services like it are sometimes limited, but there ARE ways to buy digital music and have a non-negligible amount of the money go directly to the artist.

I think you’re trying to make an argument for just pirating digital format music. I would say, don’t just throw up your hands and go straight there by default, try to buy the music first, and then if you can’t or really can’t afford it, then by all means download the music in other ways.

Every spotify downloader I’ve seen just matches to youtube and uses yt-dlp in the backend. I may have missed a new one though.

Here’s my “low complexity, medium effort, full legal, full quality” solution:

Start actually buying your music. I go down the list in descending order of convenience:

• Bandcamp
• Qobuz
• Apple iTunes (not Apple Music)
• Physical CDs (for ripping)

Tag all your music with Picard (or wrtag if you only buy full releases, there’s a GH issue for other cases) or beets. Picard is the simplest and most feature complete right now and has a nice GUI. Then upload your tagged music to your Navidrome.

Then use a tool like

• https://github.com/WilliamNT/tunesynctool
• https://github.com/blastbeng/spotisub (check my fork for a better functioning version) These will match songs from your spotify playlists to songs in your subsonic-compatible server (which Navidrome is) and recreate your spotify playlists using the music it finds in your Navidrome. These syncing tooks can have misses and you may need to do some log-digging or issue-opening to find out why, but I’ve gotten them working fairly decent and plan on doing some work to improve them some day.

It’s a nice, fully legal, fully self-hosted stack. Not NEARLY as convenient as having them auto-ripped for you from youtube, but like you said, there are quality and metadata concerns when ripping from youtube.

I agree with everyone saying you can run what you want on most any hardware. Only thing I’ll throw in is that with older/more used parts, you’re at a slightly higher risk of hardware failure, so if you wanna store data on there that you really don’t wanna lose, consider looking into online backup storage services. I’m not sure of their international availability but some good ones:

• borgbase
• backblaze

if it’s not a lot of files that you actually want backed up you might be able to get away with free/cheap tiers of google drive/one drive using rclone

Good luck!

Yes! This is the developer of gonic, if anyone didn’t know, which is a delightfully simple, fast, and stable opensubconic compatible music server. Been using it for years, I recommend. I’m very excited for wrtag to grow, as I really want something simpler and more automated than picard and beets

I use picard, which works great, but is by no means automatic. Keep an eye on https://github.com/sentriz/wrtag, I think it’ll be really clean and highly functional in the near future.

It’s still worth using, IMO. They present themselves as very privacy respecting and technologically savvy. I can say after years of using it, it feels that way. They never get in my way and consistently release quality, stable updates.

I didn’t think I’d get as much of a kick out of knowing that my random shuffle is truly random, but I do.

Self hosting music that I purchased is a really liberating feeling

95% of my homelab lives on a single server, and everything I do is within containers. So, my documentation is just keeping all my compose files in a git repo and writing in comments when necessary. It’s fairly self-documenting, and I haven’t found the need to break out of just using containers for everything, besides a couple things like setting up mergerfs or cockpit, but that’s all plug and play nowadays with stuff like https://projectucore.io/

Of course, I don’t have any other things set up in my physical layout or network stack… but all that stuff would probably just go into an entry in my notes (obsidian/wiki.vim).