Created At: Tue Jun 09 2026
Good Writing = Good Thinking
When Form, Content and Expert Thought align.
If you are following the constant debate about AI-generated writing, you may have already formed an opinion.
Whether you immediately rule out content that is formed in a specific, AI-like way. Or whether you enjoy the perks of high-volume writing and publishing.
For me, most of the debate misses the most crucial point entirely.
The question isn't whether AI can write. It is, what the writing inspired in the first place. Where the content is actually coming from.
The Assumption Worth Challenging
Most people treat writing as the final step. You think, then you write. The writing is transcription — the formalization of thought that already exists.
But this assumption is wrong and the error is consequential.
Writing isn't the output of thinking. It is thinking. The act of forcing an idea into a sentence reveals whether the idea actually holds. Vague thinking produces vague sentences not because the writer lacks vocabulary, but because there is nothing precise enough to say. The sentence is the test.
This matters enormously once you introduce a tool that can produce fluent sentences on demand.
What AI Actually Does to the Equation
AI removes friction from production. It can take a brief — an argument, a position, a set of points — and give it form at a speed no human writer matches. This is genuine leverage, and there is no honest reason to dismiss it.
But leverage amplifies what's already there. That is the entire definition of leverage.
Give AI a sharp brief built on hard thinking and it produces work that reaches further and faster than the thinker could alone. Every idea gets the form it deserves. Volume becomes possible without dilution. The constraint that used to be "how much can one person write" becomes "how much does one person have to say."
Give AI an empty brief and it produces something that looks identical from the outside — structured, fluent, confident — and means nothing. The reader feels it without being able to name it. The sentences are grammatically correct. They are also, functionally, noise.
This is not a technology problem. The tool performed exactly as intended. The input was empty. So is the output.
Form Follows Function
Louis Sullivan wrote it for architecture in 1896. It applies to every sentence.
Form is how something is written. Function is what it needs to do. When those two are aligned, the writing is invisible — it simply works. When form becomes the goal rather than the vehicle, the writing collapses into performance.
Here is what this means in practice: AI can execute form. Given a clear functional requirement — inform, persuade, reassure, move to action — a capable model will find a structure that serves it. The short declarative sentence. The concrete example after the abstract claim. The contrast that signals hierarchy.
These are not mysteries. They are rules, and rules can be followed without understanding why they exist.
What AI cannot do is determine the function. That requires knowing what the writing is for, who it's for, and what specifically they need to understand or feel at this moment. That determination is not a writing task. It is a thinking task. And it has to happen before the first word is written — by a human, in full.
Content is King
The phrase is older than the internet and has been ignored just as long.
Content is not volume. It is not originality for its own sake. It is the specific, accurate, relevant thing a reader needs at this point in their decision — nothing more, nothing less.
The failure mode is not writing too little. It is writing around the point. Filling the space where an idea should be with language that approximates having one. AI makes this failure faster and more scalable. A person with nothing to say now produces more of nothing, at higher velocity, with better formatting.
The person who has actually done the thinking faces a different situation. The ideas are already there. Every argument has been stress-tested, every claim traced back to something real. The AI doesn't generate the content — it gives it room. Each idea can be developed properly, given the structure it needs, extended into territory that would have taken weeks at the keyboard alone.
The bottleneck was never the writing. It was always the thinking. AI just makes that visible.
The Gap Widens
The popular argument is that AI levels the playing field. More people can produce more content. The barriers to publishing have collapsed.
This is technically true and strategically backwards.
What AI levels is access to form. Anyone can now produce writing that looks like writing — structured, competent, superficially coherent. The floor has risen. The ceiling has not moved.
The ceiling is determined by thought quality. By the precision of the observation, the honesty of the argument, the specificity of the insight. None of that is generated. All of it has to be brought.
The person who was already producing interchangeable content — safe, hedged, derivative — now produces it faster. The person with a genuinely distinct point of view can now give every idea the space it deserves. The distance between them grows.
AI doesn't democratise good writing. It makes clear that good writing was never about production.
What This Means in Practice
Use the tool. Use it freely, without apology. Delegate the form to a system that executes form well.
But do the thinking first. Hard enough, long enough, until the argument is specific and the claim is real and the reader's situation is understood precisely. Then write the brief. Then produce.
The discipline hasn't changed. It has concentrated. Less time at the keyboard. Harder work before you open it.
The writers who will be indistinguishable from each other in three years are the ones who reached for the tool before they finished the thought. The ones who won't be — did the thinking first.
Form follows function. Content is king. Both are true. Both have always been true. And both are downstream of something no tool has ever supplied - the capabilities of an expert thinker.