Most people who follow tech read too much and learn too little. The fix is not a better feed. It is a three-layer diet: one aggregator for breadth, a few niche newsletters for depth, and primary sources — filings, announcements, actual documents — for the claims that matter. Set up once, it takes roughly twenty minutes a day.
The word matters here. To build, as Merriam-Webster defines it, is "to form by ordering and uniting materials by gradual means into a composite whole." A news diet is the same. You are not collecting articles; you are assembling a system where each layer does one job and skips the rest.
Why does more reading leave you less informed?
Because feeds optimize for volume, not understanding. An aggregator shows you what everyone is covering right now. That is useful exactly once — as a signal of what the day considers important. Read the same story five times across five outlets and you have spent an hour learning one fact.
The deeper problem is that aggregation flattens sourcing. A press release, a leaked memo and an audited filing can all arrive as identical headlines. If you cannot tell which is which, you are not reading news. You are reading packaging.
Our analysis: the noise is not accidental. Volume is the business model of most news surfaces. A diet that fixes this has to be built deliberately, layer by layer, with each layer allowed to fail quietly if it stops earning its slot.
What belongs in each layer of the diet?
Three layers, three jobs:
- Layer 1 — one aggregator, skimmed. Pick a single headline surface or feed reader. Scan titles only. The goal is a map of the day, not comprehension. Five minutes, hard stop.
- Layer 2 — two or three niche newsletters. Choose writers who cover one beat deeply — venture, AI policy, semiconductors, whatever your work touches. A good beat writer reads the primary documents so you do not have to, and says when a claim is company-claimed versus filed. Ten minutes.
- Layer 3 — primary sources, on demand. When a story actually matters to a decision you face, go to the document: the filing, the terms page, the company's own statement. This is where the twenty-minute budget gets spent.
Notice what is missing: an infinite scroll of general tech headlines. That layer produces the feeling of being informed without the substance, and it is the layer to cut first.
How do you choose sources worth keeping?
Judge each source on three questions, the way you would read a pitch deck — looking for the fracture.
- Does it show its sourcing? A story that names its documents can be checked. A story that says "reports suggest" cannot. Keep the first kind.
- Does it separate claims from facts? "The company says revenue doubled" and "revenue doubled" are different sentences. Sources that keep them distinct are rare and worth their weight.
- Does it admit what it does not know? Confidence about open questions is the loudest tell of a weak source.
Run every source through that filter for a month, then cut anything that fails two of the three. A smaller, sharper list beats a long one every time.
What this means for your daily routine
Practical steps, in order:
- Pick one aggregator and give it five minutes at a fixed time — morning coffee, not every notification.
- Subscribe to two or three beat-specific newsletters. Unsubscribe ruthlessly after four weeks; if a newsletter has not changed how you see a single issue in a month, it is decoration.
- For anything that affects a real decision, find the primary document before you act or repeat the claim.
- Once a quarter, prune the whole stack. Sources drift. A newsletter that was sharp last year can coast this year.
The fixed time matters more than the sources. Doomscrolling is not a content problem; it is a boundary problem. A diet with edges cannot spiral.
How do you read a tech story skeptically?
Three habits cover most of it. First, check who is claiming: a company's own announcement is a company statement, not an independent fact. Second, look for the number behind the adjective — "fast-growing" means nothing; "growing" with a figure and a source means something. Third, ask what has to be true for the story's premise to hold. If a funding story never mentions the market the money is chasing, the interesting part is missing.
For an example of what deep primary-source reading looks like in practice, see our breakdown of What Is a SAFE? How Startup Convertible Equity Actually Works — the instrument only makes sense once you read the actual terms rather than the coverage of them.
Context also compounds. A story about GPU shortages lands differently if you have read GPU Supply and Cloud Costs: The Constraint on AI Startups. A regulatory headline makes sense against EU AI Act GPAI Obligations: What Startups Must Track. Background reading is the highest-return part of any news diet, because it turns each new story from a claim into a testable one.
Where should a founder or operator start?
Start with the layer closest to money. If you run a company, that means funding coverage and filings — our funding section tracks rounds with terms in the lede, and the broader tech news hub covers the market conditions venture behavior follows. If you are watching public markets, IPO coverage shows how private-company claims get tested against public disclosure.
One closing note on expectations. A good diet will not make you the first to know everything. It will make you among the first to know what is actually true — which, for anyone making decisions rather than making conversation, is the only version of informed that counts.
