brine.dev

Platform-Brain is brutal. It is nearly impossible to escape. The Fediverse itself is an example: let's remake all the platforms... but more ethical? This was a bad move or at the very least, a missed opportunity.

Twitter is both horrible and perfect. Posts the size of a punchline and the ability to @ anyone is a perfect cocktail for hot takes and dunking. But an "ethical" version of this is just kinda... boring.

But boring is good. Boring is right, boring works. Boring clarifies, cuts through, and captures the essence of the evolutionary spirit. It's platform-brain that convinces us that we need the ethical colosseum, and its lobby is powerful. Only to be defeated by giving up reach, the dopamine of "likes," and ultimately, attention itself.

Good luck with that.

While writing discover I was constantly challenging my own platform-brain. No ethical retreads, just look at a problem and find the best way(s) to solve it. And let me tell ya... it was fucking hard.

As its "clever" name reveals: discover is about discovery. How do I find blogs I want to read? And given those blogs, can I build a non-algorithmic feed that takes the place of platform algorithms? The premise itself had me skating perilously close to the open ice of platform-brain.

"Likes" for example. People LOVE likes. It is trivial to add a script that can track likes on blog posts in the directory. But likes are a keystone of the colosseum. Dopamine that rewards performance. Not evil in and of itself... until they become a measurement target. Now you’re pitting the sword of a post about AI against the wood block of a post about 19th-century typography.

So I traded "Likes" for Mentions. If you REALLY like a post, link to it and respond from your own blog; your platform. If this seems "boring" it's because it requires effort. Friction. It says: I read this, it lives alongside my own thoughts, and I’m willing to host the connection. It's the web's slow drip of dopamine delivered by the enemy of the platform: RSS.

And my respect for RSS led somewhere completely unexpected...

During testing, I made a quiet rule: if a feed doesn't deliver the full-text posts—no snippets, no "read more," no clickthroughs—it doesn't get into discover. Controversial? Maybe. I understand why people do this. All good. It just doesn't fit with how I built discover.

But the really interesting bit: writers who DO publish full-text feeds are the ones I gravitate towards. These are the people writing for themselves; that are over themselves. These are my people and I'm curious what they have to say. This is the antithesis of Twitter's punchline economy and proof that the boring and generous include my favorite voices.

Is this... unreasonable?

Let me quote Mike Watt: "start your own band". I welcome anyone to fork discover and curate to their taste; their rules. If you do, hit me up. I'll add the ability to share playlists between instances (and also mentions).

Federation by fork!

#openweb #indieweb #discover #rss

So... I wrote discover. I've been dreading writing this post because I have SO much shit spinning around in my head that I'd rather talk about. But it's adjacent to discover... so I need to start there. Fine.

discover is a curated list of RSS feeds worth following. Feeds from people writing, making and thinking. High quality signal.

See, this is why I didn't wanna write this. I just re-wrote the about page blurb. But here's the deal, it actually addresses the two biggest issues I see with the #OpenWeb: discovery and algorithm.

Another directory? Yes. Kinda. Directories are fine but they often include EVERYBODY. This is an altruistic ideal but drudging through a sea of "my first blog post" feeds is a subpar way to find feeds with high signal.

The epiphany was staring me in the face, literally, daily: playlists. Our platform overlords know this. Playlists are a great way to curate by vibe. It works for music, why not blogs? Slap an evocative title onto feeds that share connective tissue and you've got a compelling story. Throw in search & tags; photos and excerpts, and you're cooking!

But that's the fun tech part. The less fun realization is you have to curate to signal. You have to leave low-signal blogs behind. This was hard. Easier were leaving behind people who just want clickthroughs; viciously truncated feeds with little to no meat on the bone. Gone. Blogs that are social media engagement funnels or link farms? Nope. But if you're over yourself (that's honestly a big one) and doing and curious and sharing and writing and give a shit? I wanna read you.

And no judgements. To each their blog. But let's be clear: if you're thirsty for reach, you belong on a platform.

The algorithm. So the feed section started off as a shopping cart. Add the feeds you like to your "cart" and "checkout" via OPML. But when I was testing it... I started digging it as a legit feed. I did NOT want to build a reader. There's a zillion of 'em and people are very particular about readers. No thanks. But I did build a basic reader into discover to help me curate. And I just enjoyed reading that way. So out went the cart and in came a human-curated algorithm. The feedback I'm getting on it is wildly positive. And I love it too. I love it when a plan comes together.

Now, the other bit that was really bugging me: trackbacks. I originally had Webmentions wired up but it was just too much. Too much fuss; too little adoption. Not it. I struggled with this... but the answer was again, staring me figuratively in the face: I already had a trust layer. ~200 hand-curated feeds, fetched every hour. I bet they were linking to each other? They were! 108 cross-links. Real writers citing each other's real work.

So just surface those cross-links into a mentions feed that an author can subscribe to and see the conversation. Dope. But it's discover as a gatekeeper (which I typically loathe) that makes this possible. Otherwise it's back to the Pingback days but with today's AI bots. We all know where that goes.

So the TL;DR is: I wrote trackbacks with pure RSS. Thank you RSS!

#discover #openweb #indieweb #rss