How I Built a 161-Feed RSS System to Survive Ancient Near East News Overload

The Morning Flood

Every morning, same scene: phone buzzing, inbox piling up, and that nagging sense that I've read a hundred headlines but remember nothing. It's not just tech news—Ancient Near East scholarship does the same. One day it's a new cuneiform tablet, the next a recycled excavation report dressed as 'breaking.'

Once, we worried about too little information. Now we're drowning. The issue isn't access; it's judgment. I had to build a system to take back control.

Step One: Ditch the Scattershot, Build a Ladder of Trust

My first attempt was embarrassingly naive: a simple agent that scoured the web and handed me five items. It worked—sort of. But it kept serving the same story three times in different outfits. A new Assyrian palace find would appear as an official announcement, a news report, and a blog post, each claiming to be 'new.'

The fix wasn't better algorithms; it was better sources. I spent a weekend curating 161 RSS feeds, each chosen for its proximity to the original fact. I organized them into a trust ladder:

• Primary sources: official excavation reports, museum announcements, academic journals—the clay tablets of our time, unaltered by interpretation.
• Quality press: outlets like Biblical Archaeology Review and Science News, which cross-check and add context.
• Specialist blogs and aggregators: sites like Ancient Origins or Rogue Classicism, which turn dry data into readable stories.
• Independent scholars and archaeologists on social media, adding on-the-ground flavor and personal notes.

Step Two: Tame the Feed with a Tree-Structured Reader

Managing 161 feeds is a chore unless you have the right tool. I use Folo, an RSS reader that lets me organize everything into a tree of folders: 'Mesopotamia,' 'Egypt,' 'Anatolia,' 'Levant,' 'Methodology,' and 'Museums.' It's like curating my own journal, updated daily.

The interface is old-school—no algorithm deciding what I see. I decide. That's the point. I'm the editor-in-chief, not the algorithm.

Step Three: Give Your Agent a Backbone with Folo CLI

The real magic happens when you connect Folo to an AI agent. Folo's CLI lets the agent read my unread items, already filtered by my trust ladder. No more random web scraping. The agent gets a curated stream.

I set up a morning task: at 8 AM, the agent reads the last 24 hours of unread content and applies 'editorial judgment.' It merges duplicate stories, cross-references details, and prioritizes official sources. The output is a clean brief with links back to the originals—no hallucinated URLs.

Teaching the Assistant to Read Your Mood

At first, the agent kept pushing stories about the same topic—say, the latest theory on Hittite linguistics—even after I'd stopped caring. I had to tell it, 'Less linguistics, more on new excavations and artifacts.' It saved that preference to a memory file and adjusted future picks.

That's the beauty of a good assistant: it doesn't read your mind, but it remembers your complaints. A little scolding goes a long way.

From Feed to 'Cyber Newspaper'

Text dumps get tedious. So I asked the agent to render the top five stories as a clean HTML page: card layout, minimal design, with 'Why it matters' sections. The result looks like a personalized newspaper, not a chat log.

I even used the same method for a long-running saga: the rumored 'foldable iPhone' of the Ancient Near East—wait, no. I mean the ongoing debate about the location of the Garden of Eden. Conflicting reports, reinterpretations, and outright fabrications swirl endlessly. I had the agent build a 'dynamic wiki' that tracks every claim, its source, and its credibility.

The page includes a status summary, a timeline of key claims, a frequency chart of independent sources, and filterable cards tagged 'Confirmed,' 'Corroborated,' 'Single Source,' or 'Unverifiable.' It's like a detective's board for archaeology.

The Payoff: Anxiety Fades, Clarity Emerges

With this system, I no longer panic about missing a 'big story.' I see patterns—which claims have consensus, which are just noise. The noise is filtered upstream, so what reaches me is already meaningful.

In a world of AI-generated slop, the antidote is curation. Build your own dam against the flood. Subscribe to sources you trust, let the agent do the sifting, and you'll find that separating signal from noise becomes second nature.

It's not about knowing everything. It's about knowing what matters.