So I scanned the Manifesto, the essay this whole publication stands on. The result: 100% AI.

I read the number, sat with it for a minute, and decided the right move was to show it to you myself.

What the number measures, and what it cannot

The scanner answers one question: how were these sentences typed. It is probably right about mine. The prose in this publication is written by AI, working from my ideas, my structure, and my stories. My “How I make this” statement has said so since the day the feature launched. The scan does not reveal anything. It confirms.

Here is what the number cannot measure. It cannot tell you who lay awake resenting the people he loves most, and then built a system to fix it. It cannot tell you whose 360 review triggered the leadership change, whose father went through heart surgery in the middle of everything, who actually takes the supplements, runs the routines, and books the couple trips this publication describes. It cannot tell you who decides, line by line, what goes out under this name.

That part is not assisted. That part is my life.

Why I delegate the typing

I run four lives: Work, Personal, Father, Marital. A senior role at a Swiss bank, my own health and energy, two young children, and a marriage I treat as a life of its own. This entire method exists because I refuse to drop any of them. Which means my energy has to go where it is irreplaceable: living the method, deciding what it says, and being on camera saying it.

The writing itself is done with AI for the same reason a professional editor cuts my videos. Nobody watches a film and asks whether the director personally spliced the reels. The craft I sign is the judgment: what is true, what is useful, what is mine to say. So I read, challenge, and approve every line, word by word. Nothing goes out that I have not signed off.

If you have read this far, you may notice something: this is the method applied to itself. A system that tells you to architect your energy deliberately would be a strange thing to produce by ignoring its own advice.

The part the scanner cannot see: the training

The 100% also hides what it took to get there. Before this publication existed, I spent hours upon hours, across months, rewriting my website and my first essays with the AI. Line by line. Rejecting drafts that sounded polished but empty. Correcting every passage that made my life sound tidier than it actually is. Pushing back until the writing came out the way I actually talk about these things: direct, personal, and honest: about what works, what fails, and what it costs.

In a real sense, I trained it to write the way I would write if writing were where my energy went. The voice you read here took months of judgment to build, and takes judgment every single week to keep. This very essay went through four rounds of my corrections before you read it: an image replaced, a framing softened, a defensive paragraph cut entirely. None of that shows up in a percentage.

What I will not do

Substack lets publishers disable the scan on any post. I will not, on any of them. It also would not be hard to rewrite text until detectors go quiet; an entire cottage industry sells exactly that. I will not do that either, because it is the only move in this whole topic that actually looks like hiding.

The platform put it well when they launched the feature: the use of AI is not necessarily the problem, but a lack of transparency around it definitely is.

What I can promise

Some of the best writers on this platform type every word by hand, and you can feel it in their sentences. That craft is real, and nothing here is an argument against it.

My promise is a different one: everything here has been lived by me, decided by me, and verified by me. Every claim carries my name, my face is on every video, and the responsibility for all of it is mine alone.

The number will not change, and it does not need to. Everything behind it will keep being lived first, decided by me, and signed by me.

I am Philippe. I will see you in the next one.