TRANSPARENCY

How I use AI.

Real connection does not happen with AI. AI can help make it possible.

I use AI to bring this hour to you, without sending your private words to a model.

The hour itself

This is not a product built from a prompt in a weekend.

I wish it had been that simple.

I designed and built this hour myself. But most of the work was not in what I added. It was in what I kept taking away.

Bringing it back to its essence took most of those ten years.

AI helped me edit the writing. It never got a vote.

It started as a live seminar, with real coffins. Then I rebuilt it, version after version.

Testing, failing, starting again. Full time for the past five years. Collecting valuable feedback and developing it further.

Every now and then, a voice asks why on earth I am doing all this.

I have ignored that voice a hundred times. Otherwise there would be nothing to launch.

It cost me time. A lot of it. Vacations and trips I passed up, countless weekends and evenings.

At a desk, building a product about conscious living.

I know that paradox from the inside. This hour grew out of it.

Where no model goes

No model reads what you write in this hour.

That includes your fifteen answers, your farewell letters, a message you keep to yourself, and a sentence you leave half finished. None of it is interpreted, summarized, judged, categorized or answered.

A model comes near this hour in exactly three places. Here they are.

fifteen questions
lettersmessage
  • Synchronizing my recorded narration with the words on your screen
  • Speaking your name in your letters
  • Moderating a message you choose to make public

Nothing beyond that. Your private words are never sent to a model, not even in shortened form or with your name removed.

You can read exactly how your words are processed on Your Words.

The voice is mine

I recorded everything you hear me say, in a studio, in both languages.

There is one exception: your name.

It is spoken aloud during the hour, and there are too many names to record them all. For that, I use a voice model trained on my own voice. The rest of the sentence is my real voice.

One thing still does not sit right with me. At the most personal moment of the hour, your name is spoken by a voice model, not by me.

I have not found a better solution. I would rather tell you here than hope you do not notice.

The music

Nine pieces carry the hour. Eight of them were composed in code, layer by layer, around a root note I measured.

AI supplied the raw material: five rough layers, including a voice, a bowed string, a pad and a low drone.

The ninth is a licensed piece by musician Quincas Moreira, created by a human long before this project existed.

A prompt produced a single stereo mix. It could not loop seamlessly or hold the same tonal world for an hour.

That is why the pitch, the breathing and the loop are in the code.

Dozens of versions were discarded before one survived. I know what that says about me.

The rest of it

Writing

The ideas and sentences are mine. AI is my copy editor and, on larger pieces, my assistant director. I remain the director.

Email

The thinking is mine. AI drafts a first version. Then I rewrite it in my own words. I press send myself. Always.

Video

I stand in front of the camera. No avatar has ever spoken on my behalf. I do use AI for editing and post-production.

The site

Built and refined with AI.

Why AI remains a tool

For me, this hour is about connection.

First with yourself: feeling what is there, acknowledging it, and giving yourself what you actually need underneath the noise.

Then with the people around you, from that place.

And then with what goes beyond you.

Technology carries that hour to the other side of the world at three in the morning, in your own language. Its role goes no further.

Missing something here? Email me.