I want to tell you what actually happened with the last article I published here.
The one about diminishing returns on imagining outcomes. The one that ended with “I wrote every word of this. I also had help.” That line wasn’t an accident and it wasn’t a throwaway. It was a door left deliberately open for whoever was paying close enough attention to walk through it.
Here’s the full story.
The Conversation That Started It
I was in a working session with Claude — not using it as a search engine, not copy-pasting prompts into a box, but actually working with it the way you’d work with a sharp collaborator who knows your business, your voice, and what you’re trying to build.
We’d just finished something technically interesting. A content directory for a restoration company website where every single cell in every table was a live anchor link to a real published page. 512 posts audited, matched, verified, and built into five tables with 97 anchor links. Published and verified green in about twelve minutes.
And I had this thought: the interesting thing isn’t the output. The interesting thing is the mental state that made the ask possible. The fact that I described it like a game and not a project because I was standing far enough back to see the whole thing clearly.
So I said to Claude: I want to write about that. About imagining outcomes. About the 60% threshold. About where planning stops paying off. And I want it published on Will’s Take.
And then I said something else. I said: write it as my ghost writer. Write it in my voice. But leave a reward in it for the readers who figure out it wasn’t entirely mine.
Why Ghost Writing and Why the Wink
Ghost writing is as old as publishing. Every biography written “with” someone, every CEO thought leadership piece, every speech delivered by a politician — the named author and the actual writer are often not the same person. That’s not a secret. It’s just how a lot of content gets made.
What interested me was doing it honestly. Not hiding it, not pretending, but also not putting a disclaimer at the top that kills the experience before it starts. Instead, burying a tell inside the piece that rewards the reader who catches it — and gives them something real when they do.
The line “I wrote every word of this. I also had help” is technically true in both directions depending on how you read it. I shaped the idea. I gave the direction. I approved every word. Claude wrote the draft. We both own that piece in different ways.
The readers who caught the wink immediately understood something about how I work now that I couldn’t have told them directly without sounding like I was bragging about my tools. They discovered it. And discovered things land differently than told things.
What It Actually Took
The reason this worked — the reason Claude could write in my voice convincingly enough that the wink was subtle rather than obvious — is because the setup was already done.
Claude has been working in my environment long enough to know how I think. It knows I use plain language even for complex ideas. It knows I’m self-deprecating but not self-doubting. It knows I care about the insight being real, not just sounding smart. It knows I end things on a specific kind of note — not a summary, not a call to action, but a thought that keeps going after the article stops.
That’s not prompt engineering. That’s relationship. It’s the result of actually using a system enough that it understands you, the same way a good assistant or a longtime collaborator eventually learns how you work without being told every time.
I gave Claude the raw idea — the 60% threshold, the diminishing returns curve, the way AI compresses it — and said write this as me. And it did. And I read it and it sounded like me on a day when I’m thinking clearly and not rushing.
That’s a strange feeling, honestly. A good kind of strange.
The Treasure for Those Who Find It
If you read the diminishing returns article and caught the wink, here is what I want you to take away from finding it.
The fact that you caught it means you were reading closely enough to notice when something shifted. That’s a skill. Most people skim. Most people extract the headline idea and move on. You stayed with it long enough to feel the seam.
That same quality — the willingness to stay close to a thing and feel where the edges are — is exactly what makes the difference between someone who uses AI as a shortcut and someone who uses it as a lever. Shortcuts get you to the same place faster. Levers get you somewhere you couldn’t reach before.
The readers who found the wink in that article are the ones who will understand what I’m actually building here. Not content. Not a blog. A way of working that takes the best of what human thinking does and pairs it honestly with what AI does, and doesn’t pretend either one is sufficient on its own.
Why I’m Telling You This
Because transparency is more interesting than mystery, once the mystery has served its purpose.
The wink worked. People found it. The conversation it started was better than the one a straightforward disclosure at the top would have started. The experiment landed.
Now the behind-the-scenes is its own thing. And this one — unlike the last one — I wrote.
Every word.
Mostly.
Frequently Asked Questions
What is AI ghost writing and how does it work?
AI ghost writing is the practice of using an AI system to draft content in a specific person’s voice, with that person shaping the ideas, approving the output, and taking authorship credit. It works best when the AI has deep context about the author’s voice, values, and way of thinking — producing drafts that require minimal revision because the system already understands how the person communicates.
How can you tell if an article was written by AI?
AI-written content that is well-calibrated to its author’s voice is often indistinguishable from human-written content. The tells, when they exist, are usually subtle: an evenness of tone that lacks the irregularities of natural human writing, or a completeness that feels more considered than conversational. The best AI ghost writing is designed to pass a close reading — and sometimes contains deliberate signals for readers paying close enough attention.
Is using AI to write articles ethical?
Ghost writing has been a legitimate and widely used practice throughout the history of publishing. The ethical question turns on transparency: whether the relationship between the named author and the actual writer is disclosed, and whether the ideas genuinely belong to the person taking credit. AI ghost writing where the human shapes the concept, direction, and standards — and the AI executes the draft — is a modern form of a practice as old as professional communication.
What makes AI ghost writing work well?
AI ghost writing works well when the system has accumulated genuine context about the author over time — their vocabulary, their values, their aesthetic, the specific ways they think about problems in their field. A well-calibrated system can produce drafts that sound like the author on their best days: clear, direct, and specific in ways that reflect real expertise rather than generic expertise.
What should readers take away when they discover a piece was AI ghost written?
The discovery that a piece was AI ghost written is most interesting when it changes how you read the piece rather than dismissing it. The ideas were real, the direction was human, the standards were set by a person with genuine expertise. What AI contributed was execution — the sentence-level craft of turning a clear idea into readable prose. That contribution is real and worth understanding, not as a gotcha but as a signal about how professional thinking is changing.
Leave a Reply