LinkedIn Built the Slop Machine. Now It Hands Out Smoke Detectors.

Because think about what a report button admits. Nobody ships a button for things people can handle on their own. There is no “Seems like a bad take” button, because you can tell a bad take when you see one, and you scroll. The slop button exists because a million professional readers, people who evaluate writing for a living, hit something in their feed and could no longer say what they were looking at. The click does not mean I found a machine. It means I am not sure anymore, and I would like someone to check.

That uncertainty did not come from nowhere. It was manufactured, on this platform, over about a decade, and the story of how is better than the button.

The scale of what got made

Start with what the flaggers are up against. AI detection firm Pangram ran an analysis of the major text platforms in July and found LinkedIn to be the most AI saturated of them all. More than 40 percent of its long posts were flagged as machine written from the first word to the last. Not assisted. Written. And of all the AI content Pangram flagged across every platform it measured, close to two thirds came from LinkedIn alone.

The wider web is heading in the same direction. A Pew Research Center study published in August found 10 percent of pages sampled in July carried significant signs of AI authorship, up from about 2 percent five years ago. Paciva reported this spring that automated traffic passed human traffic for the first time. The dead internet theory spent years as a fringe joke. The joke now has footnotes.

So a million clicks in two weeks is not surprising. What deserves your attention is who handed out the button.

The arsonist has a tip line now

The slop did not blow in through an open window over the course of the last week. It was commissioned, and the receipts are public.

Originality.ai measured engagement on LinkedIn posts alongside AI detection and found machine written posts underperform human ones by 45 percent on average, except in one category. In leadership and inspiration content, the machine written posts beat the humans by 75 percent. The platform’s own reward system discounts synthetic writing everywhere except the genre that relies on sincerity, which means the manufactured epiphany is not a plague on the feed. It is the feed’s best paying product. People noticed the price and responded the way people naturally respond to prices.

Then there is the pen itself. Until this push, LinkedIn offered a feature called enhance your post, a native tool for turning your rough draft into the very kind of writing the button now exists to report. Per the announcement from LinkedIn’s own product chief, that feature is being removed. Hold that sequence in your head. The company shipped the machine that writes the slop, ran it while the feed filled, and is now collecting it back while handing you a button to report the output. Not ironic at all… at all.

One more detail from the same announcement, and it is the quiet one. Since the crackdown began, views on content the platform classifies as slop are down 40 percent. Which means the dial existed the whole time. The distribution of machine written content was never whether it was a setting, and somebody just turned it down, which tells you it could have been turned down at any point in the last three years.

The button is not the fix. The button is the receipt.

Why a million professionals cannot tell

Here is the part nobody at the platform will say, and it is the actual reason the button struggles.

The house style of LinkedIn was engineered before the machines arrived. In 2017 a growth strategist named Josh Fechter worked out that the platform showed only a post’s first line before a see more link, and that the click on that link counted as a quality signal. So he wrote in single sentences with a hook on top and a personal story underneath, BuzzFeed named the style broetry, and the format outlived the exploit by nine years. A generation of professionals learned to write in a cadence built for a click counter. Then the language models trained on a decade of it.

So the machines write LinkedIn the way they do because LinkedIn created the roadmap on how to write like a machine first. The convergence ran both directions, humans compressing toward the format that paid, models learning the compressed format as what professional writing is. The button cannot tell the difference because, on this platform, there stopped being one.

Which is why the button’s real victims are human. Ask anyone who writes their own material what they fear in that feature, and it is not slop getting caught. It is getting flagged for writing that came out of their own head, in the style the platform spent a decade teaching them.

The tax honest writers pay now

That fear has already changed how people write, and this is the part you will recognize from your own drafts.

The notorious em dash is now read as a machine tell, so people strip it from sentences they wrote themselves. Words like delve get retired without a farewell. Clean parallel structure gets broken on purpose. Some people leave a typo in, the way you might leave a scratch on something to prove it is not a rendering. Writers are editing their work to look worse, on purpose, because polish became evidence against them.

Call it the proof of human tax. Every post now carries a second job on top of saying something, which is demonstrating that a person said it. And the tax lands hardest on the people the platform claims to want, the ones who write their own material, because they are the only ones with something at stake in the verdict. The slop factories do not care about being flagged. They will make forty more posts today.

I pay a version of it myself. I cut em dashes from my writing years ago as a style rule. That choice now doubles as an alibi, and If I’m being honest, resent needing one.

Before anyone feels too righteous

The million clickers and the slop posters are not two different populations. At 40 percent of long posts, they cannot be. Some meaningful share of the people flagging machine content in the morning scheduled their own machine drafted thought leadership that afternoon, and the platform’s numbers say so even if nobody’s memory does. The feed does not divide into victims and offenders. It is one population, sending and receiving its own flood, one cheap post at a time.

Regular readers will recognize the shape, because it is the same one that runs your inbox and your calendar. Cheap to produce, expensive to receive, and every coping tool handed to the receiving side one complaint form at a time. Block, report, mute, leave, and now a fifth button that files a report and changes nothing about what reaches you tomorrow. The receiving side of the feed still does not own a single tool that works for it. It just got another way to tell the sender’s platform how the flood feels.

Run the test

Open your last post and read it as a stranger holding the button.

Count the sentences only you could have written. A number from your own work. A mistake with your name attached. A position that could cost you a client. A machine can imitate your cadence for free, but it cannot pay your prices, which makes cost the last authorship signal that survives.

If the count is zero, the problem is not that you might get flagged. It is that there is nothing in the post a flag would be wrong about.

What ends this

Not the button. Flags do not fix supply. Prices do. The slop stops when synthetic sincerity stops paying, and that payout is a setting inside the platform, the same one that just proved it can throttle distribution 40 percent on command. Until the reward changes, the report button is a suggestion box bolted to a firehose.

For the person writing, the way through is the same as the way it has always been, just with higher stakes. Write what a machine would not dare. The real figure, the named loss, the dated call someone can check, the position half your market will refuse to sign. Style stopped proving authorship the day the models finished learning yours. Cost still proves it, because cost is the one thing the slop cannot afford.

And for the person receiving, the honest conclusion is the one this newsletter keeps arriving at from different directions. The feed, the inbox, and the calendar run on the same asymmetry, and none of them will hand you a real tool for your side of it, because you are not the customer of the flood. You are the destination.