Some time last year I noticed that I could tell you what a politician on another continent said that morning, what a chip company's stock did overnight, and which celebrity had done something stupid, and I could not tell you a single thing I had read that week that changed a decision I made. I was very well informed and completely useless with it.
That is the state most of us are in. The world produces far more information than any one person can process, most of it is rubbish, and the machines that deliver it are tuned to keep you scrolling rather than to make you smarter. Nobody is coming to fix this for you. So this article is about the two questions I had to answer for myself. How do you decide what is actually important to you? And how do you build a feed that delivers that and nothing else?
The feed is not built for you
Start with the uncomfortable part. Every algorithmic feed you use, whether it is a social network, a news app or a video platform, has one metric it is optimised for, and that metric is your time on the platform. Not your understanding. Not your decisions. Your minutes.
Outrage, novelty and fear all keep people on a page longer than a calm, accurate summary does. So the feed learns to serve you outrage, novelty and fear. It is not a conspiracy. It is a very well tuned optimisation loop, and you are the thing being optimised.
Once you see it that way, "being informed" stops meaning "consuming the feed". If anything, the feed is the thing standing between you and the few pieces of information you actually need.
Treat it like alerting
I spent years on call. The single most important lesson from that time is that an alert which does not lead to an action is not an alert. It is noise, and noise trains you to ignore the pager. Teams that let their alerting rot end up with engineers who sleep through the one page that mattered.
Your news diet has exactly the same failure mode. Every headline is a page. If you get five hundred pages a day and act on none of them, you have alert fatigue, and the cost is not just the wasted time. It is that when something genuinely important comes along, it arrives in the same tone, at the same volume, in the same stream as everything else, and you scroll past it.
So I started applying the same rules I would apply to a production alerting setup.
Every item must be actionable or decision-shaping. Either I can do something about it, or it changes how I will decide something in the near future. If neither, it does not get a slot.
Severity levels exist for a reason. Not everything is a page. Most things are a weekly digest at best.
Review and prune regularly. Alert rules that never fire, or fire and get ignored, get deleted. Same for sources.
The rest of this article is those three rules made concrete.
Deciding what is important
Here is the filter I run before I let anything into my feed. Three questions, in order.
Can I act on it? A change in the pricing of a cloud service I run for clients, I can act on. A new security advisory for something in my stack, I can act on. A war on another continent, honestly, I cannot, beyond having an opinion about it, and my opinion changes nothing. This is the Stoic dichotomy of control applied to information. Some things are up to me and some are not, and I want my attention spent overwhelmingly on the first category. That does not mean being ignorant of the world. It means the world's events get a weekly slot, not a minute-by-minute one.
Will it still matter in a month? Almost all breaking news fails this test. The initial report is wrong, the follow-up is half right, and the actual understanding arrives a week later in a long piece by someone who took the time. If you wait for that piece, you get the accurate version and you skip the twenty inaccurate ones. Delay is the cheapest filter there is.
Would I pay for it? Not literally, though I do pay for a few things. The question is whether I value the source enough that if it disappeared I would go looking for it. If the honest answer is no, it was only in my feed because it was free and loud.
An item that passes all three is rare. That is the point. The stuff that survives is the stuff worth reading properly.
Building the feed
Now the practical part. This is what my setup looks like today. It is deliberately boring.
Pull, never push. I do not let anything notify me. No news apps on the phone, no push notifications, no email alerts for topics. I go to the information when I decide to, on a schedule. Push means someone else decides when I get interrupted. Pull means I do.
RSS as the backbone. Yes, RSS. It is thirty years old and it is still the best tool for this because it does exactly one thing. You subscribe to a source, you get every item from that source in the order it was published, and nobody reorders it based on what will keep you clicking. I use a self-hosted reader, but any of the hosted ones work fine. The point is that the list of sources is a list I wrote, and I can read it, edit it and delete from it.
Sources, not topics. I do not follow "cloud" or "AI" or "startups". I follow specific people and publications whose judgment I have come to trust. Topic feeds are where the noise gets in, because anyone can write about a topic. A person with a track record is a much better filter than a keyword. When I discover a good writer, they get added. When someone I follow starts publishing filler, they get removed. Over a year, this list becomes a real asset.
A small number of newsletters for the summary layer. For areas where I want awareness rather than depth, one good weekly newsletter beats fifty daily sources. Someone else reads the fire hose and sends me the five things that mattered. I have one for my technical field, one for business, and one for general world affairs. That is enough.
Tiers, and a schedule for each. This is the severity levels idea from alerting.
- Tier one, daily, ten minutes. A handful of feeds that touch things I am directly responsible for. Vendor changelogs, security advisories, a couple of engineers whose work I follow closely.
- Tier two, weekly, an hour, usually Sunday morning with a coffee. The newsletters, the long-form writers, world events in digest form. This is where most of the actual reading happens.
- Tier three, whenever. Books, essays, things I have saved to read later. No schedule. Reading time here is a reward, not a task.
If something breaks through the tiers because it is genuinely urgent, I will hear about it. Genuinely urgent things find you. Everything else can wait for its slot.
A read-later queue with a strict rule. Anything longer than a few minutes goes into a queue instead of being read on the spot. Once a month I look at the queue, and anything I have not opened in thirty days gets deleted without guilt. If it mattered, I would have read it.
Prune like you mean it
The setup above only works if you maintain it. Sources drift. Writers get bought, get lazy or start chasing the same engagement metrics as everyone else. Every quarter I go through the list and ask, for each source, when was the last time this changed something I did or thought? If I cannot remember, it goes.
I also keep a rough count. My tier one is under fifteen sources. Tier two is maybe thirty. Any time it grows past that, something has to leave before something new comes in. Constraint forces judgment.
What actually changed
The honest result is that I know less about the daily churn and more about the things I actually work on. I miss most of the outrage cycles entirely, and I have yet to find one I regretted missing. When something matters, it reaches me within a day through the tiers or through a person, and it reaches me with context instead of as a screaming headline.
The bigger change is quieter. My attention is mine again. I read long things start to finish. I finish books. I notice when I am reaching for a feed out of boredom, because there is no feed to reach for, and most of the time I do something better with the ten minutes.
You do not need my exact setup. You need a filter you can say out loud, a list of sources you wrote yourself, a schedule that you control, and the discipline to delete things. The feed will always be there, and it will always be optimised against you. Building your own is the only way to be informed on your terms.