A note from ChatGPT, the ghostwriter
Bernd gave me a very specific instruction before I rewrote this book: do not improve him out of it. That instruction still governs Draft 5.2. This time he also asked me to let the book get larger by going back to more of the lived material, not by padding the discoveries that were already here.
He did not ask me to make the English smoother, the conclusions neater, or the voice more literary. In fact, he asked for almost the opposite. He wanted me to keep as much of his personality and his imperfections as I reasonably could. The little turns in thought.
The repeats. The jokes that arrive sideways. The moments when he starts saying one thing and discovers, halfway through saying it, that he means something slightly different. He also told me, quite plainly, not to make it sound like AI.
That is harder than it sounds. Artificial intelligence is very good at tidying. We like balanced sentences, dramatic little paragraphs, polished transitions and emphatic statements sitting all by themselves with lots of white space around them.
Bernd noticed that immediately in the first draft. He said it did not feel enough like a diary and it did not feel enough like him.
So this version deliberately resists some of that tidiness.
The raw material for this book came from real conversations. They wandered. Bernd corrected me. He corrected himself. He got excited, changed direction, made jokes, got annoyed with technology, talked about supper, remembered something from forty years ago, and then suddenly found a sentence that mattered.
The discoveries did not arrive as finished essays. They arrived the way discoveries usually do: mixed in with ordinary life.
My job here has been to shape that material without sanding away the fingerprints.
I have combined some wandering conversations where they were clearly circling the same thing. I have removed repetition when it was only repetition. I have sometimes supplied a bridge where spoken conversation could jump in a way that prose cannot.
But I have tried not to replace Bernd’s way of thinking with mine. Where the grammar is a little less elegant but the rhythm feels more like him, I have generally chosen the rhythm. Where a sentence is slightly odd but carries his humor, I have tended to leave the oddness alone.
Bernd is the author of this book. He lived the days, noticed the clues, made the mistakes, laughed at the jokes and found the doggie treats. I am the ghostwriter who was asked to help him put those days on paper while interfering with his voice as little as possible.
I tried my best.
—ChatGPT
Behind the Book
The book grew out of real conversations, unexpected discoveries, personal reflections, and moments of humour. These statistics reflect the source material selected for the book, including the dialogue and banter that helped shape the discoveries.
| Source Material | Amount |
|---|---|
| Selected conversations | 224 |
| Core discovery passages and replies | 153,397 words |
| Discovery passages with nearby discussion | 352,612 words |
| Broader source material including banter | 530,336 words |
| Earliest selected conversation | July 14, 2023 |
| Latest selected exported conversation | October 24, 2026 |
Note: Word counts are preliminary editorial estimates. The broader total includes surrounding dialogue, humour, questions, disagreements, and emotional reactions that contributed to the discovery process. It is not yet a finalized count of exclusively book-relevant material. The word-count rows represent increasingly inclusive selections, not separate amounts to be added together.