Sure, but my script only gets rid of the second and later parameters, i.e. ones with & not ?. Personally I don’t think I’ve ever seen a single site where an & param is critical. These days there few where the ? matters either, but yes YT is a holdout.

automatically modify a YT link to a designated invidious instance or whatever

Surely that’s the link that should be posted, then?

using your GitHub

It’s not yours, it’s Microsoft’s.

This is a really good question. I’ve also been wondering why there seems to be no obvious go-to service for blogging, i.e. full-form authored text, in the same way there are for photos (Pixelfed), video (PeerTube), and of course microblogging and discussion forums like this one. Seems like an oversight.

Yes, there’s WordPress. But IMO WordPress is just overkill for most use cases, with its massive database backend. Text is text, the web was designed for text and it worked before databases existed. A static site generator will generate a flat text site just fine (I’ve used them) but you need to host it.

Someday I’ll try self hosting but for now, I’ll pay for decent services.

Maybe you shouldn’t even need to try?

I’ve changed my mind on this one. I used to believe in the utopian internet dream of everyone hosting their own stuff on their own domains. But managing domains and hosting are both a PITA. They require money, technical expertise (because security), and commitment (or else your site goes away). The URL of a blog article posted, say, right here is probably going to be more permanent that it would be on the average private site. And Archive.org is recording the content either way. I’ve come to the view that sites should be left to organisations, and individuals should do themselves a favor and just affiliate themselves to one of those sites. Against payment if appropriate.

Which leaves the question of which site?

You never define “clean”.

To strip excess URL parameters (i.e. beginning “&”, almost certainly junk) if the clipboard buffer contains a URL and only a URL (Wayland only):

if url=$(printf '%s' "$(wl-paste --no-newline | awk '$1=$1' ORS=' ')" | egrep -o 'https?://[^ ]+') ; then
  wl-copy "${url%%\&*}"
fi

There’s a lot of debate around the Bluesky relay and whether or not it’s posible to “scale down” ATProto.

Luckily, with all the attention that Bluesky and ATProto have gotten lately, there’s a lot of interest in running alternative relays just in case things go bad, and I think it may be possible that funding for a public good relay might materialize to protect from the potential failure or compromise of the Bluesky relay.

But I also want to draw attention to another non-obvious thing about ATProto: the relay is not required.

Let’s do a thought experiment. […]

Etc. But I’m certain everyone here has actually read the article so they’ll already be familiar with that thought experiment.

Fair point.

So. Nobody gonna push back? It’s quite a convincing article. Especially this bit:

This is something I want to put a bit more focus on: how important the PDS is.

Giving people their own PDS is soooo crucial to having a free ( as in freedom ) internet. This needs to be something that I can:

• Self host or have somebody else host for me
• Download all of my data from, whenever I want to, and
• Grant other apps read and write access to

That last piece is crucial to the existence of all of these different ATProto apps. We are giving people their own data store and then letting them connect that to all kinds of different tools and experiences.

It does at least seem like the protocol is more sophisticated, and so perhaps carries more potential, than (say) the one powering this site.

Good job. But don’t worry if you have to look up answers. I’ve been at this for 20 years and I still have to look up and double-check basic syntax like the classic find -exec one. No big deal if only takes a couple of seconds.

This is definitely the sort of thing that LLM AI tools can help with, in theory.