Why Your News Feed Feels Like an Information Black Hole—and How to Build a Care Practice

The Morning Scroll That Leaves You Empty

Every morning, I do the same thing: pick up my phone and see what happened overnight. Dozens of unread messages in WeChat, a new round of trending topics on Weibo, a long list of "important updates" from news apps, and then a few tech communities, each shouting "just published" or "major update." After ten minutes of scrolling, a strange feeling settles in: I know a lot of things, but I can't name one thing that actually matters today.

A phone launches. An AI model updates. A company announces a new plan. Every item looks worth tapping, but after reading them all, almost nothing sticks. We used to worry about not having enough information. Now the problem is the opposite—there's so much that we can't finish reading it, and it feels like being swallowed by an ever-expanding black hole. It eats not just time, but also our judgment about what deserves attention.

Taking Back Control: A Care Practice for Information

So I decided to run an experiment: build a personal news-filtering system with AI agents, and take back control of what I read each day. This isn't about being productive or staying ahead—it's about caring for your own mind, protecting it from the sludge of low-quality content. It's a care practice, really, one that treats your attention as a finite resource worth guarding.

Step One: Stop Letting Algorithms Pick Your News

My first idea was simple: set a scheduled task in an agent to run every morning at 8 a.m., gather five tech news items, and send them to me. That gave birth to my first briefing assistant. The results were decent. Before, I'd hop between news sites, apps, and Weibo to piece together what happened in tech overnight. Now, while making coffee, a briefing is waiting on my phone.

But the novelty wore off in days. The same model launch, written three different ways, would take up three slots in the briefing. A story I read yesterday would be retold by another outlet today and pop back as "latest." I can't blame the agent—it wasn't slacking. Searching the whole web gives you a bag of loose bits: official announcements, media reports, secondary interpretations, reposts, clickbait—all looking like "new info." Hand that mess to an AI, and it can summarize fast, but it can't tell you what's genuinely worth your time.

Building a Trusted Source Pool

To cut down the echo, I started with the sources themselves. Over the years, I've accumulated 161 RSS subscriptions. But quantity isn't quality. To stay clear in this information black hole, you need a clear filtering and management system.

• Primary sources: Official blogs from companies like OpenAI, Google DeepMind, and Anthropic. These give you raw, unaltered info first.
• Authoritative media: Bloomberg, The Information, Business Insider, WSJ, Reuters, Caixin—they have strong reporting networks and cross-checking, offering a solid macro view.
• Good secondary sources and aggregators: The Verge, Techmeme, TechCrunch, MacRumors—they turn dry raw info into readable, deep content.
• Bloggers and KOLs: Big tech influencers on Weibo, Bilibili, etc. They're for peripheral stuff—hands-on experiences and unique takes.

If you don't mind, you can also subscribe to ifanr and APPSO—they're excellent sources too.

Manage Your Feeds Like a Tree

When you have over a hundred sources, managing them gets messy. Sticking them all in one list is a recipe for confusion. So you need a unified tool. I use Folo, which is basically an RSS reader. RSS might sound old-school in the age of algorithmic feeds, but for actively aggregating info and staying in control, nothing beats it.

With Folo, you can organize your subscriptions into categories, like flipping through a magazine you've made yourself. In my Folo, I have six core sections. From Bloomberg, Ars Technica, Wired, TechCrunch, to DIGITIMES, MacRumors—all sorted into lists for tech, gaming, culture, AI, and cars. Opening Folo feels like reading a tech magazine I've slowly built.

Make Your Feeds Agent-Friendly

More important than organization, Folo has a CLI that turns those 161 feeds into a library your agent can call directly. Once configured, the agent reads unread items from my subscriptions. It's no longer grabbing random headlines from the web; it's eating from a pool I've already curated. Plus, each item has a direct link, so no more hallucinated URLs from the LLM.

Training Your Assistant: Learn to Say "Not That"

Once the info pool was solid, a new problem appeared: industry hot topics aren't the same as your personal interests. For a while, open-source model news filled the tech pages. Day one, I'd click. Day two, another breakdown of the same model's parameters. Day three, I knew it had nothing new for my day. What I actually cared about were specific hardware changes—a new laptop's specs leaking, a phone's release date.

A human would scroll past automatically. An agent won't. It only knows the topic is hot and sources are still writing about it; it doesn't know I've had enough. So I just told it: "Too much AI news today. I want more consumer electronics and hardware. Remember that for future pushes." The agent automatically created a MEMORY.md file and wrote that preference into its long-term rules. And it worked—the next delivery was full of consumer electronics, with the AI model news gone.

That's what I love about agents: they don't magically understand you, but they remember what you dislike and what you care about. A good assistant is often trained by being scolded.

Printing Your Own Cyber Newspaper

Even with the agent doing the sorting, the chat format felt lacking—text squeezed together, not great to read. Since AI can write code, why not turn all those scattered clues into my own "cyber newspaper"? I upgraded the prompt: take the final five deep briefs and turn them into an HTML page with a minimal style, card layout, with title, core facts (multi-source), and why it matters, plus buttons for source links. The result is a clean, white-card view with multi-source verified facts and commentary, and a button to jump to the original article.

The same method works for tracking a long-running topic. Take the foldable iPhone: leaks, denials, more leaks—each one looks huge, but together they're just the same question rephrased. I asked the agent to do a harder task: stitch the foldable iPhone clues into a live-tracking "dynamic encyclopedia." It built a single HTML page with a status summary, a tree diagram of specs, a timeline, a keyword frequency chart, and clue cards each with credibility level, status, earliest date, and source links. The agent sorted everything into "confirmed," "multi-source," "single rumor," and "unverifiable," and each claim traces back to a source.

The Calm After Filtering

After seeing that foldable iPhone tracker, I felt less anxious about when it would actually launch. Not because I lost interest—no rumor summary can replace holding the real thing. But when you can see which clues are consensus from the supply chain and which are just clickbait, the fear of missing out evaporates.

That calm comes from cutting out low-quality info. In 2025, Merriam-Webster chose "slop" as the word of the year—originally meaning something like swill, now used for AI-generated, low-quality digital content. There's more info than ever, wrapped in prettier packaging, which makes judging harder.

In this age of digital slop, building your own reliable sources is the best defense against AI garbage. That's the whole point of this system: AI can collect, dedupe, and organize, but it shouldn't outsource your judgment. The flood of outside info isn't going to stop. Instead of trying to swim faster, build yourself a small dam: subscribe to sources you truly trust, keep diverse voices, and when you see a claim, go back and read the original.

Whether it's curating sources or having an agent follow your preferences, all we're doing is putting a gate upstream of our information flow. What reaches you is already filtered. Then it's easy to see what deserves a close read and what's worth a glance. Being able to extract something truly useful from the noise—that's meaningful enough.