We are not built to process the whole world at once.

The problem is volume, not bias

For most of human history, the news that reached you was news you could do something about. A flood upriver. A change in who collected the tax. A shortage at the market. Information traveled about as far as your ability to respond to it, and the two stayed in rough proportion.

That proportion is gone. A feed can deliver every catastrophe on the planet to you within the hour, and nearly all of them are things you can feel but not affect. Your body does not make that distinction. It answers a distant emergency much the way it answers a near one — with a small readiness to act that has nowhere to go. Repeat that a few dozen times a day and the result is not a better-informed citizen. It is a tired one.

Most attempts to fix the news treat it as a quality problem: rate the outlets, flag the falsehoods, balance the perspectives. Those are worth doing, and we do some of them. But a perfectly accurate, perfectly balanced feed of two hundred things you cannot act on would still leave you exhausted. Accuracy was never the binding constraint. Volume was.

What we mean by “useful”

Newsful is a portmanteau of news and useful, and we mean something specific by the second half. A story is useful to you when you could plausibly do something about it, or when it will plausibly do something to you. Everything else is, at best, a matter of interest — and interest is not the same as need.

So two filters run ahead of all the others: proximity and actionability. A zoning hearing eight blocks away, a school board vote, a road closure on your commute, a bill moving through your statehouse — these are small stories by any national measure, and they are the ones where your attention converts into something. A leadership crisis on another continent may be far more consequential in absolute terms and still be, for you, weather.

This is why we do not rank by recency, and why we do not rank by engagement. Recency assumes the newest thing is the most important thing, which is how a news cycle is built but not how a life is. Engagement assumes that whatever holds your attention deserves it, which is the assumption that produced the problem in the first place. We rank by whether a story is relevant to where you actually are, and whether there is a door in it.

Quality signals still matter, but they arrive late. Once we know which stories are yours, we use assessments of bias, sensationalism, and depth to decide which account of them to show you and which to leave out. They filter the shortlist. They do not build it.

There is one deliberate exception. Emergencies — severe weather, public safety events, environmental hazards near you — skip the ranking entirely. When something is genuinely urgent and genuinely close, calm is not the priority.

Why your profile never leaves your device

To do any of this we need to know some fairly personal things about you: where you live, what you care about, what stage of life you are in. That is a precise portrait of a person, and it is exactly the kind of portrait the rest of this industry collects, stores, and eventually sells or loses.

We decided not to hold it. The details that shape your feed stay on your device and are converted there into a mathematical representation — something we can match against stories but cannot read backward into the facts that produced it. That representation travels with your request, gets used, and is discarded. What remains on our servers is an account and a few benign settings: nothing that would tell a curious employee, an acquirer, or an intruder what you care about.

This costs us things. We cannot show you a history we do not keep. We cannot tune the product by studying what any individual person asked for. If you clear your browser storage or move to a new device, you will have to tell us about yourself again, because we genuinely cannot recover it. We think that is the right trade, and we would rather explain the friction than design it away. A calmer feed should not cost you a dossier.

What we are not

We are not a bias-rating service. We assess stories for slant and sensationalism because it helps us choose between accounts of the same event, not because we think those scores are a verdict. Reasonable people read the same article differently, and a number on a page can imply a precision that no such judgment actually has.

We are not a replacement for local journalism. Everything Newsful surfaces was reported by someone else, usually by a newsroom working with less than it had a decade ago. We are a way of finding that work, not a substitute for it. A version of this product that ended up starving its own sources would have failed at its purpose.

We are not infinite, on purpose. There is no endless scroll here, and no reason to check back every twenty minutes. A feed you can finish is the point. If Newsful is working, you will read it, close it, and go do the thing you read about.