RSS as personal context

I’ve been thinking about RSS feeds as a way to give AI software better context about me. Not context in the technical sense of stuffing more tokens into a model (though I guess it is also that), but in the human sense: a better starting point for understanding what I care about.

This started yesterday when I asked Notion AI to use my Feedster MCP server to pull the latest articles from the RSS feeds I subscribe to. Today I took it a step further and asked it to review the latest articles in my “Technology blogs” channel and suggest topics that seem to be bouncing around the collective zeitgeist, e.g. things I could write about here.

Feedster gave me 30 recent posts from that channel, including content from The Pragmatic Engineer, SemiAnalysis, MacStories, Pixel Envy, Ftrain, Daring Fireball, Michael Tsai, Asymco, and a handful of other independent-ish technology publications. Notion AI read through them, picked out common threads, and created a database with ten potential blog ideas, each tied back to its sources.

The initial list was pretty good. It included ideas about how AI-generated code is changing code review (this one was actually not very good… way too obvious), physical buttons making a comeback as interfaces for agents, age verification turning into identity infrastructure, and a post about why boring software is often the best software.

But the ideas still felt like they could have been written by just about anyone interested in technology, so I pointed it at this website’s feed and asked it to try again. It read my recent posts and colophon, figured out that I care about personal computing, designing and building developer tools, RSS, and the open web, and then reworked the list.

The first idea became this post.

Personal investment

When we talk about giving AI better context, we usually mean handing it a doc, pointing it at a folder of markdown, or writing a very long prompt. This works fine for specific tasks, but it doesn’t say much about the person doing the prompting. If I ask a tool to find interesting technology writing on the web, it still has to decide what interesting means and will probably fall back to some combination of what is popular, highly ranked, well-linked, or easiest to retrieve.

An RSS subscription list works differently because a lot of those decisions have already been made. Every time I subscribe, I’m making a small choice about whose work is worth investing my time in. One feed doesn’t reveal too much, but a few dozen of them start to describe a point of view. Organizing those feeds into channels makes that point of view even more explicit.

In Feedster, channels are basically playlists for websites. I have channels for technology blogs, software builders, tech news, the indie web, science, art, sports, and other things I consistently read. I chose the sources, but I don’t know what they’ll publish next.

This is very different from the kind of personalization we’ve gotten used to on social media. Those systems watch what we click, how long we linger, and what makes us come back, then build a profile we can’t really inspect or edit. A feed list is literal. I can tell when a channel stops being interesting to me, because reading it feels like work. I can then choose to remove it or double down on my investment and spend time tailoring it further.

I like reading blogs

There is an uncomfortable contradiction in using AI this way. One reason I built Feedster is that search engines and AI answer engines are already giving people fewer reasons to visit individual websites. If an agent reads all of my feeds and returns a tidy summary, it could make that problem worse.

I don’t want to build a feed reader that eliminates reading. The version I find compelling is closer to a reading guide: show me what’s new, connect ideas across sources, give me enough context to decide what to open, and always point me back to the original writer.

RSS is good raw material for this because each item comes with a title, source, publication date, and canonical URL. The content has a home. The software can use it, but it shouldn’t erase where it came from.

That’s what worked in the blog-idea exercise. Notion AI connected pieces about AI code review, developer-focused hardware, age gates, and identity infrastructure to things I’ve written about before. Those connections helped me decide where to go deeper, and the database kept the links so I could return to the source material when I started writing.

The point isn’t to give me fewer things to read. It’s to help me spend more time with the things most likely to change how I think, while preserving some of the older web’s useful wandering. AI is good at compressing everything into a clean answer, but the best version of this should still leave room for detours.

Personal context infrastructure

RSS feels a bit like Markdown to me right now. Markdown never went away, but LLMs made everyone remember why plain text is useful: people can read it, software can parse it, and it moves around easily. RSS has a lot of the same boring strengths. It’s open, portable, and structured around links back to the original source.

That turns out to be pretty useful here. Instead of saying “go look at the web,” I can say “start with these sites.” These are the writers and publications I’ve already decided are worth my time. Look there first, then branch out if you need to.

I’m not sure what this becomes in Feedster yet. Better briefings, related posts, read/unread memory, resurfacing old links — all of that seems possible. But I care more about getting the relationship right. I want the tool to make my feeds easier to use, not quietly turn them into another inbox I stop actually reading.

Maybe RSS isn’t just a leftover from the old web… It might be one of the better ways to tell new software where to begin.