How I Built an AI Editor to Save My Mornings from Information Overload

The Morning Scroll That Left Me Empty

Every morning, I do the same thing: grab my phone and try to catch up on the night. WeChat pings with dozens of unread messages. Weibo trends have already rotated a few times. News apps push a long list of "important updates." I hop between tech communities and see more "just released" and "major update" posts.

After ten or fifteen minutes of scrolling, a strange feeling settles in. I know a lot of things happened, but I can't name a single one that truly matters today. A new phone launches. An AI model updates. Another company announces a fresh plan. Each headline looks worth a click, but after reading them all, almost nothing sticks.

We used to worry about not having enough information. Now the problem is the opposite: too much information, way more than we can process. It feels like being swallowed by an expanding black hole that consumes not just time but also our ability to judge what's worth our attention.

So I ran an experiment: I built an AI agent to filter the news for me, to take back control of what I see each day.

Step One: Ditch the Random Scraper, Plug into a Real News Network

My first idea was simple: set up a scheduled task in my agent that runs at 8 a.m., collects five tech news items, and sends them to me. That gave me my first briefing assistant.

The results were decent at first. Before, I had to jump between news sites, apps, and Weibo to piece together what happened overnight in tech. Now, while I make coffee, a summary is already waiting on my phone.

But the novelty wore off in a few days. The same model launch, written three different ways, could occupy three slots in the briefing. A story I'd already seen yesterday would resurface today, reworded by another outlet, wearing the label "latest" again.

Honestly, I can't blame the agent. When it searches the whole web, it gets a bag of loose, unverified info: official announcements, media coverage, second-hand interpretations, reposts, clickbait—all looking equally "new." Hand that pile to an AI, and it can summarize fast, but it can't judge what's genuinely worth your time.

To cut through the echo chamber, you have to start with the sources. Over the years, I've collected 161 RSS feeds. But quantity isn't quality. To stay sane in this information black hole, you need a clear system for filtering and managing them.

Action One: Build a Tiered Source Pool Based on Trust

In an age where content gets reshuffled and AI-generated junk is everywhere, the core principle for getting useful info is "trace to the source." Every retelling strips away context or twists meaning. The closer you get to the original point, the clearer the picture.

So when picking sources, I focus on the ones that actually create content, ranked by reliability:

• Primary sources (official announcements & earnings reports): OpenAI News, Google DeepMind News, Anthropic News—company blogs that give you unaltered, first-hand info.
• Authoritative media (serious journalism & investigations): Bloomberg, The Information, Business Insider, WSJ, Reuters, Caixin—they have strong reporting networks and rigorous cross-checking, offering objective macro views.
• Quality secondary sources & aggregators (vertical analysis): The Verge, Techmeme, TechCrunch, MacRumors—they turn raw info into readable, deep content.
• Bloggers & KOLs (tips & expert opinions): Weibo tech influencers, Bilibili tech YouTubers—they fill in gaps with hands-on experience and unique viewpoints.

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

Action Two: Organize with a Tree Structure

When you have over a hundred feeds, managing them gets messy. If you stuff them all in one list, you'll lose your mind trying to find anything. You need a single tool to hold them.

That's where Folo comes in. It's essentially an RSS reader. RSS is a protocol that pushes the latest content from websites to you. In this age of algorithmic feeds, RSS might sound old-school, but for actively aggregating info and staying in control, there's still nothing better.

With Folo, you can treat your subscriptions like a custom magazine, neatly organized by category. In my Folo, I have six clear sections. From Bloomberg, Ars Technica, Wired, and TechCrunch to DIGITIMES and MacRumors, all sorts of media are sorted into lists for tech, gaming, culture, AI, and autos. Opening Folo feels like flipping through a tech magazine I built myself.

Action Three: Open an Interface for the Agent

More important than all-in-one organization is that Folo is agent-friendly. It offers a CLI tool that turns those 161 feeds from a simple RSS reader into a database my agent can call directly.

Once configured, the agent can read unread items from my Folo subscriptions. It's not grabbing random headlines from the whole web; it's reading a batch of sources that already passed my first filter. Plus, each item has a direct link, so the AI never fabricates URLs.

I upgraded my morning briefing with this prompt:

Use Folo Read https://api.folo.is/skill.md and follow the instructions. Read the unread items from my Folo subscriptions from the past 24 hours. Then perform these "editor-in-chief" judgments:

• Merge duplicates: identify multiple sources covering the same event, don't repeat them.
• Cross-verify: compare details across sources, synthesize into one in-depth brief.
• Prioritize originals: if there's an official announcement or first-hand foreign report, use that and mark "multi-source verified."

Step Two: A Good Assistant Gets Scolded into Shape

Once the info pool was solid, a new problem appeared: industry trends aren't the same as my interests.

For a while, open-source AI models dominated the headlines. Day one, I click. Day two, another breakdown of the same model's parameters. Day three, I already know it doesn't add anything new to my life. What I really cared about were specific hardware changes—like the specs of a new laptop or the launch date of a new phone.

Humans naturally scroll past what doesn't interest them. Agents don't. They only know a topic is still hot and sources are still writing about it. So I told my agent directly:

"There's too much AI news today. I want more consumer electronics and hardware news. Remember that for future briefings."

The agent automatically created a MEMORY.md file and wrote that preference into its long-term rules. And it worked. In the actual runs, it carefully compared and filtered according to my preference—you can see it in the model's reasoning process.

Looking at the delivered results, the AI model news was gone, replaced by consumer electronics and hardware. This is what I love about agents: they don't magically understand you, but they can remember what you dislike and care about. A good assistant is often shaped by being scolded.

Step Three: Stitch the Fragments into a Cyber Newspaper

Even with my agent organizing and summarizing, the chat interface felt lacking. The text was cramped, and the reading experience wasn't great.

Since AI can write code, why not have it turn scattered clues into a personalized "cyber newspaper"?

I upgraded again: "Take the final five deep briefs and format them into an HTML page with a minimalist UI, card layout. Include a title, core facts (multi-source summary), and why it matters. Put source links as buttons at the bottom for easy tracing."

Now I get a clean view of the news. No more dense paragraphs. Clean white cards lay out the verified facts and commentary, and if something interests me, I click the button to jump to the original.

The same method works for long-running topics. The foldable iPhone is a perfect example: leaks, denials, and more leaks cycle endlessly. Each one looks like a big deal, but together they're often just the same question rephrased.

I challenged my agent with a tougher task: track the foldable iPhone rumors and stitch them into an ongoing "dynamic encyclopedia."

I asked for a self-contained HTML page titled "Foldable iPhone Rumor Tracker." It had to search for keywords like foldable iPhone, iPhone Fold, Apple foldable iPhone, and include:

• A 100-word status summary at the top.
• A tree diagram with "foldable iPhone" as the root, categorized by release time, form factor, screen & hinge, specs, price, and unconfirmed.
• A timeline of key leaks with first appearance, latest changes, and current status.
• A keyword frequency chart counting independent valid sources, not reposts.
• Clue cards with specific claims, credibility, status, earliest date, and original source links.
• Filtering by category, credibility, and status.
• A source list.

I set rules: prioritize Apple official, supply chain announcements, analyst original reports, and original media; merge duplicate reposts; clearly distinguish "confirmed," "multi-source corroborated," "single rumor," and "unverifiable"; don't present unconfirmed info as fact; every credibility rating must be traceable.

The result was a beautifully structured page. At the top, a 100-word summary that wasted no time, directly stating production progress and core mysteries. Then a parameter tree diagram showing everything from screen ratio and liquid metal hinge to final price. The timeline and keyword frequency chart revealed how rumors evolved and converged. Each clue card carried a label: "Confirmed," "Multi-source Corroborated," or "Single Rumor." For hot topics like "no crease," I could see exactly how many independent sources backed it, and click through to Ming-Chi Kuo or Bloomberg's original report.

This multi-dimensional cross-referencing saved me from piecing together truth across dozens of pages. It handed me a logically organized "investigation report."

After seeing that page, I'm actually less anxious about when the foldable iPhone will launch. Not because I've lost interest—rumors can't replace actually holding it. But when you can clearly see which clues are supply chain consensus and which are just clickbait, the fear of missing out fades away.

Why This Matters for Your Time

That anxiety disappears because you've cut off the low-quality noise. In 2025, Merriam-Webster chose "slop" as its Word of the Year. Originally meaning "swill," it now describes the flood of AI-generated, low-quality digital content.

There's more information than ever, packaged to look better, yet judging it gets harder. In this era of digital slop, building your own reliable sources is the best defense against AI junk. That's the whole point of this system:

AI can collect, deduplicate, and organize, but it shouldn't outsource your judgment. The flood of information will keep coming. Rather than swimming faster, build a small dam upstream: subscribe to sources you trust, keep diverse voices, and when you see a conclusion, go back and read the original.

Whether it's manually curating feeds or having an agent work to your preferences, what we're doing is putting a gate upstream. What reaches you has already been filtered. Then it's easier to see what deserves a deep read and what's worth just a glance.

Being able to extract something genuinely useful from the noise—that's already meaningful enough.

Practical Takeaways

If you're drowning in notifications, here's what I'd suggest:

• Start small: pick 10–20 sources you truly trust, not 100.
• Use an RSS reader to organize them, so you control the flow.
• If you're technical, look for tools with APIs or CLIs so you can automate the filtering.
• Train your AI assistant like a human: tell it what you like and dislike, and it'll remember.
• Don't let the tool become another source of noise. Set limits on how often it updates.

Your mornings should be for coffee, not for drowning in headlines.