fucking finally. though at this point I'm pretty happy clanking out qml quickshell plugins.

nostr doesn't need to be the origin of all content, but provide a handle for reckoning with it all. I am still bullish on one of the many many routes to more content being RSS to nostr and nostr to RSS syndication. I use https://drss.io as my podcast listening app across all platforms now, though it still has sharp edges. you can export an opml, upload it, and it will sync all your feeds, one npub per feed, and pull them every 5-10 minutes. you can then read the content in any nostr client, they become articles. podcast episodes get an imeta mp3 tag.

we need to lean into the lists/sets, wot curation, decentralized nature and not worry about eyeballs. just keep pushing the right direction to allow signal to surface, identity to be verifiable, and infrastructure to be decentralized.

in the long run I do believe that content publishing on nostr first makes the most sense, with all the distribution channels being downstream. I think if we build the bridges, content will start by flowing in. Then as the properties of nostr that make it beneficial become more salient due to bad behavior by the countries and corporations of the world, content will start being published more natively and flow out.

it's an optimistic view, but that's the point. what we should be moving towards.

I wish we could assume Mermaid, and less urgently LaTeX support, for anything rendering markdown. like this little guy I made for reading clanker-ese: https://stigmergic.dev

for a while now, yes. if you tell it not to in your `CLAUDE.md` it will respect your rule.

I see weird behavior when the context is past 200k. the further past, the weirder.