The brief is the proof. The standard is the product.

An operator's morning reading problem is a flood with no filter, never a shortage: five newsletters, two markets apps, a feed that mixes a Fed decision with a founder's vacation photos, and somewhere in there the one item that will actually move a quote you honor this week.

I know what the forty tabs actually contain because I kept them open for years. Two markets apps that disagree with each other before breakfast. A newsletter that spends four paragraphs on a company I will never touch.

A feed where a genuinely load-bearing regulatory change scrolls past between a hiring announcement and someone's conference selfie. The reading time is the cheap part of that. The real cost: the one item with your name on it arrives dressed identically to the thirty-nine without, and the sorting falls to you, at six in the morning, before the first real decision of the day.

I have spent twenty-plus years building systems and businesses, and I started in audit, which ruined me in a good way: I do not trust a numberI cannot trace. What I kept meeting in the business-news aisle was the opposite discipline. Confident prose, vibes-grade sourcing, and a $100-a-year subscription that summarized what a wire service said yesterday.

The item that matters to a person running payroll is rarely the headline. It is the second-order line: the surcharge that follows the strait, the covenant that follows the rate.

Somebody had to build the thing that reads for that line every morning. Nobody had.

Your actual morning question is brutally practical: of everything that happened yesterday, what touches my costs, my people, my customers, or my loan, and what do I do about it before Friday. Nobody was answering it.

The gap has a shape, and it is worth naming precisely. The paid tier answers questions an operator is not asking: allocation questions, exit questions, which-fund-is-buying questions. The free tier answers no questions at all; it manufactures a feeling of being informed, which is a different product from information.

Real analysis exists. It costs several hundred dollars a seat and it is written for people who allocate capital, not people who make payroll. The free tier underneath it is engagement media: it optimizes for opens, which means it optimizes for feelings.

Look at who the well-built briefs serve. Investors have them. Tech workers have them. Policy people have them.

The person running a $1M-$50M business, the one who signs the truck lease and eats the diesel surcharge personally, gets either the investor's brief with the wrong lens or the vibes brief with no lens. (I keep a candid map of that field at Filtered vs the field.)

I ran operations for businesses like these as a fractional COO, and my clients lived in that gap. They did not need more news. They needed someone to read the flood against their P&L and hand back the four things that touch it.

The morning question has a property the news industry hates: most days, the honest answer is short.

A brief built for it must be willing to say "three things today, and the third is minor."

No engagement-funded publication can afford that sentence. The economics select against honesty about thin days. So the fix was never going to be a better newsletter in the same business model; it had to be a different machine with a different reason to exist.

And no, an aggregator app does not close the gap either; I tried several for years. Aggregation solves collection and leaves the expensive part untouched, which is judgment: knowing that the third paragraph of a boring procurement story is the one that moves a freight quote, and that the loud market story moves nothing an operator controls.

Judgment is what the paid tier charges hundreds a year for. Building it into a machine, and holding the machine to an auditor's standard, was the actual work.

What you get is a two-minute morning brief and a full elaboration on the blog, five signal categories deep. Behind it, Filtered reads hundreds of sources overnight: wire services, regulators, markets data, trade press, the places operator-relevant news actually breaks.

The part I will put my name on is what happens between reading and shipping. Every edition passes more than 130 automated checks before it reaches an inbox.

Every figure has to trace to a source we can show you; a number that cannot prove where it came fromdoes not ship, and neither does arithmetic we performed ourselves and dressed as fact. When a section has nothing worth your time, it ships empty rather than padded.

A morning with the machine, at the altitude I can describe: overnight it reads, deduplicates, and ranks. Then the verification passes run, and they are rude on purpose.

A figure that cannot show its source gets the draft flagged. A story whose headline and body disagree gets held. A section with nothing above the bar ships empty, with no apology and no filler, because an operator's trust is spent in cents and refunded in years.

Multiple independent judgment passes read the result before I do, and I read it before you do.

When the checks and I disagree, the edition is late rather than wrong.

It has shipped late. It has not, to my knowledge, shipped a figure it could not trace, and when it ships an error anyway, the corrections page exists precisely because "to my knowledge" is not a standard.

I will not describe the internals here, deliberately: the checks, the rankings, and the judgment passes are the product of years of tuning and they stay ours. What you get to hold me to is the output contract: traced figures, named sources, corrections in public, and a brief that respects the two minutes it asks for.

The two minutes are a design constraint, not a marketing line. The email carries the day's five beats compact enough to finish over one coffee; the daily blog elaboration carries the full mechanism for anyone who wants the depth; and the standing guides (the five linked below) hold the evergreen teaching the daily items assume.

Three layers, one standard, and you choose your depth each morning instead of me choosing it for you.

The reader I build for runs a real business: $1M to $50M, a team, debt with covenants, vendors with terms, customers with moods. Freight brokers re-quoting lanes. Contractors watching an SBA close. Agency owners whose clients read the same scary headline they did.

Filtered's test for every story is "whose Tuesday does this change," never just "is this important." If the answer is nobody who runs payroll, it does not ship.

For twenty years my clients have been founders in this band, and the pattern never changes: they are drowning in information that was priced, written, and framed for someone else's job.

I also know this reader from the inside of the inventory room, not just the advisory seat. I co-founded and ran operations for a leather goods e-commerce brand, manufacturing overseas, and physical inventory does not forgive the way a slide deck does: a freight delay is a stockout with a date on it, never a narrative.

That experience is why the brief obsesses over the unglamorous lines, freight and fuel and payment terms and covenants, and treats them as front-page material when they move. For the person signing things, they are the front page.

A filter that selects everyone selects no one.

So, just as important, who this brief is not for. It is not for traders; there are no picks, ever. It is not for people who want macro commentary as a spectator sport.

And it is not for the operator who wants comfort: on a day when the news is genuinely bad for your cost lines, the brief will say so plainly and hand you the move, which is the only kind of comfort that survives contact with a P&L.

Free buys you something the paid class structurally cannot offer: the right to ship a thin day thin. A subscription that costs $100 a year owes its readers the feeling of volume, and the feeling of volume is where padding is born.

This brief owes its readers only accuracy and brevity, in that order. On the day that incentive ever changes, this page changes first, in plain language, before anything else does.

Filtered costs nothing because the daily brief is how we earn the right to be in your morning. It is the demonstration, live every weekday, of the verification standard described above. Charging for it would put the proof behind the same paywall problem it exists to fix.

"Free" deserves the same scrutiny I ask you to apply to everything else, so here is the model stated plainly. Advertising-funded briefs sell your attention; their incentive is opens. Subscription briefs sell you depth; their incentive is renewal, which is healthier but gates the product.

Neither model is Filtered's. The brief is the public proof of work for a company that builds analytical systems, and it wins only when it is accurate enough, for long enough, that operators trust it.

Accuracy as the growth engine is a slower business than outrage. It is also the only one I know how to be proud of.

The candid business answer, because operators deserve one: Filtered is built by CoreiBytes, and readers who outgrow a daily brief will find our analyst product, Meraj, when they go looking for more. No brief-to-invoice funnel tricks, no "free chapter" games. The brief stands whole on its own.

Corrections are public. A material error gets fixed in place with a dated note, and the record lives at /corrections. The email is immutable once sent; a correction to a shipped edition lands in the next one.

Some things never ship. Figures we cannot trace. Derived arithmetic presented as sourced fact. Stock picks and buy/sell calls, ever (the full line is on the disclaimer page). Culture-war enforcement stories dressed as business news. Padding, when a section is honestly thin.

The standard is checkable without trusting me. Pick any edition from the archive and audit it the way I would: every figure should name its outlet or its data series in the sentence that uses it, every source link should resolve, and every claim of consequence should show its chain rather than assert its conclusion.

If you find one that fails the audit, the corrections page is the mailbox, and the fix will be public. A verification standard that requires you to take the verifier's word is a slogan; this one is designed to be spot-checked by strangers.

Empty beats padding, every time. When a section has nothing that clears the bar, it ships short or ships empty, and the brief says so. This is the rule I am proudest of and the one the engagement business model makes impossible: the machine is allowed to tell you today was quiet. A reader who has watched it do that is a reader who believes it on the loud days, and that belief is the entire asset this project runs on.

Sources are named where the claim lives, not in a footer. Outlets are attributed in the sentence; data carries its series and its observation date; the standing guides link FRED and EIA at the point of every figure so you can check the number without trusting the prose. Audit ruined me in a good way. The habit transferred.

The five standing beats each keep a living guide: the operator economy, AI for operators, growth signals, people and ops, and capital for operators. That is the whole surface.

Judge it daily: pick an edition, follow three source links, check three figures against their series, and decide for yourself whether the standard holds.

That test takes five minutes and requires believing nothing on this page.

The brief carries my name because a standard needs a person attached to lose something when it slips.

The corrections page is mine. The thin days are mine. So is tomorrow's edition.

I am a three-time founder, a fractional COO, and a systems architect. I started in accounting and audit, hold a Master's in Accounting, and have built systems for teams at companies you know and founders you don't.

I have built and sold three companies, run operations for a physical e-commerce brand with overseas manufacturing, stood up a twenty-person delivery team in about a month when a project demanded it, and have been writing about AI eating business workflows since 2018, back when that was a niche opinion. Filtered is the system I wanted every one of those rooms to have at 6 a.m.

If you have read this far, you are probably the reader this whole page describes, and the fastest way to test that is not another paragraph from me. Read tomorrow's edition, or open the archive and read yesterday's.

Follow one source link. Check one figure. Then decide whether this belongs in your morning, on the evidence, which is the only basis I want the decision made on.

Every daily post and guide on this site carries the same byline: drafted by the Filtered engine, approved by Habib Ferdous. The wording is literal.

The engine handles the volume. It reads and ranks the day's sources, drafts under written editorial rules, and runs 130+ mechanical checks before publication. Those checks are designed to surface sourcing, figure, and quality issues for review. The full scoring and release logic stays inside the system. Readers can judge the part that matters: named sources, figures tied to those sources, and a public record of corrections.

I wrote the rules, review the finished draft, and make the final editorial call on what ships, what stays in draft, and what needs correction.

The engine can make mistakes, as any machine can. Human review reduces that risk; it does not make the process infallible. The corrections page exists for what gets through: a public record of material errors and the changes made.

You will not find the phrase "AI-assisted" here. It is the wrong description in both directions: it understates how much of the drafting is automated, and it understates how much a named human owns the result. The honest version is the byline itself.

That is the division of work behind the byline. Filtered drafts. I approve.

Filtered provides information, not advice. The Disclaimer and Terms apply to everything published here.