The cheap article is the expensive one

Someone sends me a link to an AI writing tool every couple of weeks, usually with the same note attached: it does the same thing as yours for a fraction of the price. Sometimes it's forty cents an article. Sometimes there's a free tier. The question underneath is fair, so I want to answer it plainly: why does Mastheads cost more than that?

Here is the honest version. A cheap article is not cheap. It is a cost you have deferred, and the interest is worse than the sticker. The forty cents you save at generation is a down payment on a bill that arrives later, in a currency that's harder to pay: your time, your rankings, and your readers' trust. Let me show you the three invoices.

Invoice one: you finish it yourself

The pitch for a cheap generator is that it writes the article for you. Read the fine print and it writes a draft for you. Those are not the same job.

A first draft from a volume tool is confident, fluent, and completely unverified. Before it can go live under your name, someone has to do the work the tool skipped: read every factual claim and check it against a real source, cut the three paragraphs of throat-clearing that say nothing, fix the statistic that's off by a decimal, rewrite the opening so it doesn't sound like the last four hundred openings, add the internal links, size the image, and decide whether the piece was worth publishing at all.

That work didn't vanish because a model produced the first pass. It moved - off the tool's balance sheet and onto yours, priced at your hourly rate instead of the tool's per-article one. A "forty-cent article" that needs forty-five minutes of a competent editor is not a forty-cent article. It's a forty-five-minute article with a small head start. And if you skip the forty-five minutes to keep the price real, you haven't saved the money - you've just published something nobody checked.

Invoice two: it drags the rest of your site down

Here's the invoice most people don't see coming, because it doesn't arrive per article. It arrives all at once.

Thin, near-identical pages produced at volume are the exact shape Google's scaled content abuse policy was written to demote: many pages, little added value, whatever produced them. And search engines increasingly judge quality at the level of the site, not just the page. A pile of weak articles doesn't sit quietly in a corner earning nothing - it drags on the credibility of everything around it, including the handful of pages you actually worked hard on.

So the cheap-content bet carries a nasty asymmetry. The upside is a few dollars saved per post. The downside is a sitewide quality signal that can take months to climb out of, long after you've forgotten which sixty articles caused it. You didn't buy cheap content. You bought a lien on your whole domain.

Invoice three: one bad claim costs the trust in all of them

Publish enough unchecked articles and, statistically, the pipeline will eventually ship an invented statistic, a wrong date, a quote nobody said, or a scraped image with someone else's rights attached. Not because the model is malicious - because at volume, "rare" becomes "scheduled." One hallucination in fifty articles is four hundred of them across twenty thousand.

And trust doesn't degrade politely, one reader at a time. A reader who catches a single fabrication doesn't file it as a one-off; they re-evaluate everything else you've published, and they're right to. The whole reason a real publication is worth more than a content farm is that a reader can extend it credit - assume the next article is true because the last hundred were. Break that once, visibly, and you've devalued the entire archive. That's the most expensive invoice of the three, and it isn't paid in dollars. It's paid with your name.

The number that actually matters

All three invoices come from measuring the wrong thing. The figure a cheap tool advertises is cost per article generated. The figure that runs your business is cost per article you'd actually publish under your own name.

Run the math honestly. A tool that produces fifty pieces for five dollars, of which maybe six survive a real read, didn't cost you ten cents each. It cost you about eighty-three cents per usable article - plus the hours you spent finding the six, plus whatever the other forty-four did to your rankings if you got tired and let some slip through. "Cheap per article" is a precise measurement of a quantity that doesn't matter.

What the extra actually buys

So where does the money go at our end? Not into fancier sentences. Models have written competent prose on demand for years; that part is a commodity. It goes into the machinery a cheap tool leaves out, because leaving it out is what makes the cheap tool cheap.

• A source gate that reads each candidate source against the specific story - right event, right dates, real substance - and drops a topic it can't properly support instead of padding it to length.
• Independent quality checks that can hold a weak draft rather than ship it on schedule.
• Review by default, so nothing goes live without your word unless you deliberately turn that off.
• A named editor of record on the byline, which only means something because a real human is accountable for what runs.
• An authenticity trace on every article - the sources it used, the checks it passed - so "we verified this" is something you can open and see, not something we ask you to take on faith.

All of that makes Mastheads produce fewer articles than a volume tool would, and it's slower on purpose. The slowness is the product. Every piece that clears the gates is one you can stand behind; the ones that don't clear them are precisely the ones that would have mailed you an invoice later.

Slop is a loan

Cheap content is a loan against your reputation. It's taken out in your name, the terms are buried, and it comes due on a date you don't get to choose - a core update, a reader who fact-checks you, a rights holder who finds their photo on your page. The forty cents was never the price. It was the minimum payment, and the balance compounds.

We priced Mastheads so you would never have to take that loan. It costs more at the register. It is, by a wide margin, the cheaper way to run a publication that's still standing a year from now.