Talon is closed source, What’s available on GitHub is the community command set - effectively just configuration

Definitely! The author is super friendly in IRC too (and Xe, the anubis author, is there as well lol)

They make one request per IP. Rate limit per IP does nothing.

I finally dealt with the AI scrapers hammering my Forgejo instance - https://jade.ellis.link/blog/2025/05/18/actually-stopping-forgejo-ai-scraping Hopefully next week I’ll be able to get back to actually programming Continuwuity rather than fighting fires.

The problem (as matrix people found out the hard way) is some media & content is very illegal. Most individuals really don’t want even the chance of being exposed to CSAM or gore, and neither do server operators want the chance of that being shared from their server or written to their disk when that can result in police at the door. You need default-on moderation that is very powerful, and end users should never be distributing media they don’t want to. This pushes towards centralisation of nodes run by experts, and heavily punishes true P2P models.

Windows licenses ain’t cheap. Never have been.

Here are some interesting feeds I follow, mostly tech-focused and quite Rust heavy:

• Alexis King’s Blog https://lexi-lambda.github.io/feeds/all.atom.xml
• Blog on Asahi Linux https://asahilinux.org/blog/index.xml
• brson https://brson.github.io/feed.xml
• dystroy https://dystroy.org/blog/atom.xml
• ecton https://ecton.dev/rss.xml
• fasterthanli.me https://fasterthanli.me/index.xml
• Faultlore https://faultlore.com/blah/rss.xml
• Graphite - Blog https://graphite.rs/blog/rss.xml
• Ink & Switch https://www.inkandswitch.com/index.xml
• Jade’s Website https://jade.ellis.link/blog/rss.xml
• Lord.io https://lord.io/feed.xml
• Mara’s Blog https://blog.m-ou.se/index.xml
• matklad https://matklad.github.io/feed.xml
• Raph Levien’s blog https://raphlinus.github.io/feed.xml
• Tulir Asokan https://mau.fi/blog/index.rss
• Xe Iaso’s blog https://xeiaso.net/blog.rss

Generated by opening an OPML export in firefox, running the following script and deleting a bunch of feeds:

 "- " + [...document.querySelectorAll("body > outline > outline")].map((f) => `[${f.getAttribute("text")}](${f.getAttribute("htmlUrl")}) \`${f.getAttribute("xmlUrl")}\``).join("\n- ") 

Not right now.

You should! You don’t have to be qualified to have something interesting to say - and you’re doing it, so you know more than most people!

That sounds fun! Have you written about it?

It matters a little bit - Google measures performance on real devices through CrUX, and that feeds into their rankings - but not much. There’s no real incentive to go for a Lighthouse performance score above 80 or so.

I use https://fedoraproject.org/coreos/ for my server/website. My host doesn’t offer it as an image so I have to upload it myself, but I use an ISO I made with the CLI to automatically set up everything anyway. It works pretty well, I configured auto updates and I can just forget about it.

Yeah, or even the inbox in lemmy. It’s a surprisingly common thing.

Regarding your first paragraph, this results limit is per page. To get the next page, you take your timestamp of the last item and use it in from_time, or whatever you’ve called it. It’s still a pagination technique.

Regarding custom sorting, some of the techniques in the article can do this, some of them can’t. Obviously timestamp based pagination can’t, however the ID-based pagination that I mentioned can.

This whole article was sprung from a discussion of exactly that case, because users often simply don’t delete notifications. It’s very common for users to have years of undismissed notifications stacked up under the notification bell, and it’s not a good experience to load them all at once.