A Book Is Not Written in a Straight Line

What self-publishing taught me about discovery, authorship, and using AI without handing over the wheel

I thought finishing the manuscript would be the hard part.

That is the romantic version of publishing, isn’t it? You wrestle an idea into language, survive the doubt, type the final sentence, and emerge holding a book. Maybe there is coffee. Possibly rain against a window. The mythology is strong.

What I actually had was a manuscript.

Publishing Don’t Think Straight taught me that writing a book and making a book are two different creative acts. The first asks whether the idea deserves to exist. The second asks whether you can keep hundreds of decisions aligned long enough for another person to hold it in their hands.

My process was not write, format, upload, sell. It was write, rearrange, discover a structural problem, repair it, proof it, question it, research a government form, upload it, panic at a platform status, revise it, and then learn that distribution was an entirely separate job.

In other words, the publishing journey was very on-brand. It did not think straight.

Discover before you decide

I now use a five-part philosophy in much of my creative and strategic work: Discover, Think, Plan, Create, Grow. I did not sit down at the beginning of the book with those five words taped above my desk. The framework became clearer later. But looking back, the first useful work was discovery.

Discover is not the polished research phase where you collect enough information to prove your first idea was right. It is the messier act of noticing what the project is trying to become.

Don’t Think Straight grew around people who had built, performed, invented, led, or endured in ways that resisted the conventional route. The material eventually became 49 Journey profiles and four broader Reflections: Bold Originality, the Reality of Abundance, Giving as Growth, and the Obsession Effect. Each profile needed more than biography. I wanted a reader to leave with a Takeaway, a Call to Action, and a Reflective Challenge, something that turned another person’s story into movement in their own life.

That created a combinational problem. I was not simply gathering facts. I was comparing patterns across industries, personalities, timelines, setbacks, and behaviours, then asking what remained true when the surface details changed.

This is one place where AI became genuinely useful. It helped me generate angles quickly, compare large clusters of information, challenge categories, and see unexpected combinations. It could take a pile of notes that looked unrelated and help me ask sharper questions of it. That rapid ideation shortened the distance between ‘there may be something here’ and ‘here are five possible structures worth testing.’

But speed is not authorship. AI could suggest a connection; it could not decide whether that connection was honest, useful, or mine to make. It could organize ingredients. It could not taste the meal. Every meaningful choice still came back to judgment: what belonged, what was merely clever, what sounded like me, and what I was willing to stand behind in print.

Thinking is not decorating the first idea

After discovery came the less glamorous work of thinking. That meant testing the premise rather than immediately packaging it.

Was this a collection of famous success stories? A motivational book? A creative workbook? A challenge to realism? The answer changed as the manuscript changed. Later, the clearest tension became less about celebrity and more about self-authorship: I know what I want to build, but I keep waiting for permission.

That distinction mattered. If I misunderstood the book, every later decision, subtitle, cover, website copy, video ideas, bookstore description, would be polished in the wrong direction.

This is why I have become suspicious of creation that begins too quickly. Making things feels productive. It gives you visible evidence of effort. Thinking can look like delay, even when it is preventing six months of beautifully executed confusion.

AI was useful here too, particularly as an argumentative partner. I could ask it to attack the premise, compare alternate framings, find repetition, or show me where an idea sounded inspiring but vague. Sometimes it produced nonsense with great confidence, which is also useful training. You learn to separate a fluent answer from a sound one.

The goal was never to have a machine tell me what the book meant. The goal was to create more pressure against my own assumptions.

The manuscript looked finished before it was finished

By the summer of 2025, the book had moved from a 117-page stage toward roughly 180 pages. I printed the whole thing and marked it by hand. All 180 pages.

The physical proof was valuable. A book behaves differently when you can flip through it. You notice pacing, page turns, visual repetition, empty space, and the odd fatigue that never appears when you scroll a PDF. It was also a ridiculous amount of paper to consume every time the manuscript changed.

That full print became the exception rather than the process. Later passes moved to PDF annotation and selective printing. One physical review for the object; digital passes for copy and layout; individual spreads for problems that genuinely needed paper.

Even then, the book kept finding new editors. During one especially unplanned final pass, my sister Melissa edited the manuscript while we were stranded at Pearson Airport for more than ten hours amid cancellations and delays. Publishing advice rarely includes ‘secure an airport floor and recruit a family member,’ but reality does not respect the clean case study.

The deeper lesson was that visual polish can disguise editorial incompleteness. Once text is laid out like a real book, your brain starts treating it as finished. A handsome page can make a placeholder feel intentional. It can make a typo look typeset rather than wrong.

I learned to separate the passes: structure, copy, placeholders, layout, technical preflight, and physical object. They are different forms of attention. Trying to perform all of them at once is how obvious problems become invisible.

The first real milestone

The milestone was not the moment I felt certain. Certainty remained extremely punctual in its absence.

The milestone was deciding the book was coherent enough to move from an evolving manuscript into a production system. That is where Plan and Create began, and where the trouble became much more technical.

I had discovered what I was making. I had thought hard enough about what it meant. Now I had to build a repeatable way to turn the idea into 147 final pages without breaking the book every time I touched it.

That is the subject of Part Two: the invisible work between a manuscript and a book, including one tiny red square in InDesign that managed to become a surprisingly effective villain.

Next in the series: The Invisible Work Between a Manuscript and a Book