You get no notification. Nobody laughs at you. Forty people see your post, you assume the algorithm was in a mood that day, and you try again on Thursday.

That is the shape of the thing. The penalty for bad content has stopped being embarrassment and started being silence.

Nobody tells you. You just stop travelling.

The slop button

LinkedIn users can now report a post as AI slop. The platform is killing off its "enhance your post" AI rewriting feature, and it plans a nudge that tells you when other people have flagged your writing as machine-made. It says it blocks more than 200,000 AI-generated spam comments every day before they appear.

The scale explains the panic.

LinkedIn's editor in chief listed the tells. Bland takeaways. Strange formatting. Recycling-bin emoji paired with a call to repost. Broetry. And the "not this, but that" sentence pattern that chatbots adore. The commonest red flag of all is an articulate comment that does nothing except summarise the post above it.

The rush to Reddit

Reddit is the most-cited website in AI answers, according to the visibility firm Parse. If a chatbot recommends a jeweller, it read Reddit first. So a platform that sat at the edge of digital plans for a decade is now at the centre of them.

Reddit users feel exactly as they always did about brands. When Brilliant Earth advertised for a Reddit strategy manager, tasked with shaping how the brand appears in conversation in a way that would carry through to search engines and language models, a moderator of the jewellery subreddit warned the community with a siren emoji, describing the place as possibly the last corner of the internet still holding honest company reviews.

An industry has grown up to bridge the gap. One agency, nine months old, tells clients to lurk for weeks and post only when they have something genuinely useful to say. Others sell the shortcut. Thirty-five to a hundred brand mentions a month from $2,500. Reputation packages including positive posts and the removal of negative content from $4,500.

Two brands are held up as the ones that cracked it, and they cracked it the same way. Sonos sent an employee into its subreddit under the name KeithFromSonos to answer technical support questions and translate corporate announcements into English. Fidelity built popular subreddits by hosting finance discussion and publishing guides.

Customer service and education, running the marketing.

Then the cautionary one. HelloFresh put a community team member called Katie into its customer subreddit late last year. She handled feedback, explained delivery delays and recalls, and stayed cheerful through rude complaints. In April she went silent. The company says it is evaluating insights.

Everyone in that subreddit noticed.

Why it is one story

On Reddit it is a moderator with automated tools, plus whatever the AI models decide to cite. On LinkedIn it is a classifier deciding whether your words leave the building. In both cases a machine now sits between what you write and anybody reading it, and in both cases the machine is tuned to detect the same thing: writing produced by somebody with nothing to say.

The old bargain was simple. You owned a page. You posted whatever you liked. The platform showed it to whoever followed you, and paid you in reach for the trouble of turning up. That arrangement has quietly been withdrawn on both platforms inside the same month.

What to do about it

Check whether you are already invisible. If your LinkedIn reach quietly collapsed this year and your posts read like they were assembled rather than written, you have your answer. The platform has stopped handing your work to strangers, and it has no obligation to tell you.

Read your own last ten posts for the tells. Bland takeaway. Repost call to action. The "not this, but that" pattern. If three of the ten trip the list, that is your reach problem sitting in plain sight.

Name a person and give them time. That is the entire Reddit strategy that works, and it works on LinkedIn too. A human being, present over months, answering questions somebody actually asked. Sonos did it. Fidelity did it.

Every platform has priced attention the same way for fifteen years. Post volume, post frequency, post consistency. That pricing is being replaced by something closer to a credibility check, run by software, applied silently, with the result visible only in a reach number nobody thinks to interrogate.

The gate is still there. The gatekeeper has changed.