---
contentId: 4d118945-8aa0-4608-ae75-c62b17fa3ffb
title: Building a Curated Technical Monitoring Feed Is Hard
description: Being overwhelmed is our new normal. To keep up with the news, we need to use both AI and collective intelligence. That's the only way.
date: 2026-11-09
---

We're living in a strange world. Every time there's a technological breakthrough, the world accelerates. And the more it accelerates, the more technological breakthroughs we have. It's a vicious circle.

Until now, that acceleration was manageable. We had time to adapt, to learn, and to understand.

We had time.

But AI changed everything. Not because it breaks the circle, but because it amplifies it. Since the release of ChatGPT in November 2022, anyone can create content, and services are adding more features than ever. Things are changing faster than we can adapt, and the amount of news is overwhelming. _It's a real problem, and it's only the beginning._

Keeping up with the news is freaking hard. It's not just about finding good articles. It's about reading them, understanding them, absorbing them, and remembering them for later reference, all while keeping your critical thinking intact and not falling into the trap of clickbait, sensationalism, and misinformation. That's exhausting.

To deal with this news overload, we need both AI and collective intelligence. That's the only way.

That may sound counterintuitive at first, but our brains aren't designed to process so much information or remember everything we read and find. Compared to AI, we're far behind on many tasks. This is one of them, so our only chance is to use AI effectively while collectively fighting for good content.

_I don't want to start a debate about whether the world really needs AI or whether this is the world we want to live in. The company I work for pushes me to use AI as much as possible to stay competitive, so I adapt to it. But I'm happy to discuss this topic in the comments if you want._

## A personal journey

I've spent the last six years finding an efficient way to keep up with the news. I've tried many things. Most of them failed, but I've finally found a way that nearly works for me, so I built a tool around it to make it more efficient and available to everyone.

At first, I was simply searching Google for specific topics. Sometimes, I found interesting articles or content, read them, and then moved on.

However, this technique has four major drawbacks.

1. You have to be active. Articles won't come to you. You have to search for them.
2. You have to know what you're looking for. If you don't know that a technology exists, you won't search for it, and you won't find it. This is terrible for open-mindedness and serendipity.
3. It takes a lot of time. Most of the time, the highest-ranked articles come from Medium or Dev.to, and they are always bad. No intro, no conclusion, and sometimes, not even any real content. I honestly stopped visiting these platforms, but that makes it even harder to find good content. Sometimes, it's like a treasure hunt, and you have to craft the perfect query to find a good article.
4. You have to remember everything you read and find. If you read an article on a really interesting blog, you have to remember the name of the blog and the name of the article. And sometimes the article you read is really good, but the timing is wrong. Timing really matters. _It's like meeting the right person during the wrong phase of life._

So I tried subscribing to some newsletters. It wasn't any better. There was too much noise and not enough value. I quickly gave up on this technique.

Then I tried using RSS feeds. There was still a lot of noise, not enough discovery, and way too much work to filter the good content from the bad. Also, a lot of websites don't even provide RSS feeds. _This is my personal experience. RSS readers might suit your needs, but using them takes a lot of effort._

While trying all of these techniques, I also decided to use X, formerly Twitter, as a source of news. At first, it wasn't very useful because figuring out who to follow takes time. So whenever I read an interesting article, watched a good video, or met someone I found interesting, I'd follow them on X.

For years, it didn't really work. I didn't find a lot of good content, but I felt I was heading in the right direction, so I kept at it. And then, step by step, things improved. Now, I know that I can open X and instantly find the latest news and the best articles or videos, with multiple points of view in the comments. It works really well for me.

But all of this comes at a cost. It took me years to build it, and not everyone has the time or wants to do the same. And I spend a lot of time on X. My network continues to grow, and I get a range of perspectives, which is essential for me to keep my critical thinking intact. _Everything comes at a cost, and the quality of my X feed is no exception. It's sometimes really bad for my mental health._

And even today, timing is important. My X feed shows me an article, but if it's not the right moment for me, I don't know what to do with it. And when I later realize that I need it, I don't remember where I saw it, and I can't find it anymore. It's really frustrating.

Until now, I haven't been able to address the fourth point. I can find good content, but I can't remember it for later reference. I don't have a good way to save it, find it later, and share it with others for discussion.

## A pragmatic solution

That's why I've built [soubiran.dev/news](https://beta.soubiran.dev/news), a curated technical monitoring feed open to everyone, with summaries, critical analysis, and a place to discuss the news. The goal is to address all the points I mentioned above.

The project is designed around the idea that not everyone has the same needs, the same preferences, or the same way of consuming content. So I've built multiple entry points to access the content:

1. The website. You won't be able to read the full articles on the website, but there will be a title, a summary, and links to the original article or the Discord discussion.
2. The Discord server. Each article will have a dedicated thread in the relevant category, with a summary, key points, and critical analysis. The discussion will happen in the thread, and the first message will link to the original article.
3. The RSS feed. For those who want to use it in their own feed reader, there will be an RSS feed available with the same content as the website.
4. The MCP server. Connect your AI assistant to the feed by adding `https://connect.soubiran.dev` as a remote MCP server. It's built into [api.soubiran.dev](/infra/services/api-soubiran-dev). After signing in, you can ask it to browse the latest publications, search in natural language, or retrieve a summary and critical analysis without leaving your assistant.

And more to come...

The feed will be based on my personal reading, but everyone will be able to submit articles to the feed. Each article will be reviewed and analyzed before publication to check its quality and relevance to the feed. Opening submissions to everyone is really important if I want to include a range of perspectives. _Social media algorithms are really good at creating echo chambers. By allowing everyone to submit articles, I want to fight against that as much as possible._

AI will generate the summary and critical analysis and assign a category to each article. The goal is to save us time while still providing valuable content. Before spending time reading an article, just read the summary to see whether it's worth it. After reading the article, use the key points as takeaways and the critical analysis as a second opinion, with sources to explore the topic further.

Finally, use the website or Discord to easily retrieve the article later, read it when needed, or quickly revisit the summary to refresh your memory before sharing it with someone or using it as a reference.

I've been using it for months now, and it works really well for me. I hope it can work for you too, and I can't wait to see and read the articles you'll submit.

I'll share the technical details later.
