Most of what you are reading on LinkedIn right now was never written by a person, and the person posting it is lying to your face about it.
A recent breakdown of long form LinkedIn posts found the majority reads as fully AI generated. Not assisted. Generated start to finish and posted like it came from a real brain. A few weeks ago LinkedIn finally did something about it.
Here is what changed. LinkedIn shipped a button that lets any member flag a post as “seems like AI slop.” That flag feeds straight into their algorithm. This sits on top of a reach penalty rolled out back in March, which already cut visibility on generic AI content by close to half. They also pulled their own “enhance your post” rewrite tool and replaced it with a basic proofreading tool instead. LinkedIn built an entire detection system around killing generic AI content.
If you are a founder leaning on AI to pump out posts, this is the moment to think, not panic.
Most people are going to panic and swear off AI completely. Back to typing every post by hand, back to inconsistent posting because life gets in the way. That is the wrong move. Here’s the thing. The problem was never AI. The problem was AI with no voice behind it.
There are two ways to use AI for content right now. Way one, you type a prompt, copy the output, post it, and pray nobody clicks the button. That is the version getting flagged and buried. Way two, you feed the model your actual voice, your sentence rhythm, your way of making a point, until what comes out could not have come from anyone else’s account. That is the version that survives this.
I did not swear off AI. I built an AI agent that writes my LinkedIn posts for me, and I spent the last few weeks teaching it to sound exactly like me.
Here is how. I pulled years of my own YouTube transcripts, raw and unedited, and used them to map exactly how I talk. Not what I say, how I say it. The way I open with the point instead of a warm up. The way I compress an idea into two or three things instead of ten. The way I never hedge a claim. The way I reach for a real comparison, a Ferrari versus a Toyota, instead of explaining something in the abstract. All of that got built into a voice profile the agent now follows.
Guess what happens when you do this right. The agent is not replacing your voice. It is scaling it. You are not hiding that AI touched the post. You are making sure nobody can tell, because the thing they are reading actually is you, just written faster.
This matters more than most founders realize. You are not just avoiding a flag button. You are protecting your reach and your credibility with the exact audience you are trying to reach. A founder scrolling LinkedIn can smell a generic AI post from three lines in. The flat tone. The advice that could apply to any business because it was never actually about your business. That is what gets ignored, and now flagged.
If you are still copy pasting AI straight into your post box, that button has your name on it. Sooner or later someone clicks it, and LinkedIn quietly starts burying your content.
If you are ready to build something that writes like you instead of writing like everyone else’s AI, that is exactly what I am doing right now, and it is exactly what I help founders set up for theirs.
Got a question about any of this. Reach out and let’s talk about it.