Is This My Writing?

A number from lcamtuf made the rounds, summing up Hacker News in a single line. In June 2026, roughly 55% of the daily top stories were either about AI or written by AI. Four months earlier, in February, it was 40%; now it had crossed the halfway mark. His method was simple. He classified the top stories by hand each day, ran the suspected ones through a detector, and verified the flags by eye.

What’s interesting isn’t the number itself but why detection was even possible. Leave the prompt untouched and an LLM speaks in a nearly fixed voice. That quasi-deterministic default style leaves a pattern, which is what gets caught. Flip it around: the pieces counted in that 55% are the lazy ones, written without so much as tuning a prompt. The ones that got worked over were never counted at all. So 55% isn’t the ceiling. It’s the floor.

And I’m one of the people below that floor. This blog, too, is written with AI. So the question that surfaced as I looked at the number wasn’t “HN is broken.” It was “then is this writing mine?”

Trust collapses not by volume but by asymmetry

The real damage in that 55% isn’t to the content; it’s to the reading. The moment it’s known that half of it is AI, the reader starts doubting the other half too. A piece someone labored over for days gets met first with “isn’t this AI?” The contamination doesn’t happen in the writing. It happens in the reader’s default posture.

Trust collapses asymmetrically. One fake in ten makes you doubt all ten. Once the cost of telling real from fake exceeds the cost of just reading the piece, people give up on telling and lean toward distrusting the whole. What AI ruins isn’t writing; it’s reading. A single fake drains trust from a hundred genuine pieces.

Standing in that frame, “was it written with AI?” turns out not to be the real question. The real question isn’t whether a tool was involved, but whose judgment this piece answers for. I had to settle that for myself.

So I decided I’d hired an editor

Sometimes I pause mid-sentence. Did I write this, or did the AI? Rather than drag the question out, I settled it simply. I’ve hired an editor.

Picture how a book comes together and the image sharpens. The author hands over their thoughts and experience through interviews, back and forth with the editor. The editor shapes that material technically into prose. The author reads the polished draft again and, wherever it drifts from what they meant, asks for changes. In that back-and-forth, a great many of the book’s sentences pass through the editor’s hands, yet no one calls it the editor’s book.

What the AI does is exactly that editor’s role. The argument — what this piece is trying to say — comes from me. The fact that the surface of the sentences passed through an AI doesn’t change who the author is. As long as I review it and reverse what drifts, the thinking in this piece is mine. Once I settled that, the pausing stopped.

Tool, or person

The question isn’t new. Back when ChatGPT had just appeared, the same line trailed every result at work: “the AI did that, not you, right?” The question of who owns the ability, asked over the output.

Then as now, I think this splits on what you take the AI to be. I saw it as a tool. The people who felt that using AI would render them obsolete had, I suspect, half-consciously cast it as a person — another agent competing with them. A tool doesn’t replace you. Only a competitor replaces you.

To me, AI is no different from using a calculator or running Excel on a computer. Using a calculator doesn’t make the answer the calculator’s credit, and Excel summing a column doesn’t make the report Excel’s. A tool doesn’t take ownership of ability. Only the judgment of what to point it at decides ownership. AI is just one more tool.

This is really the anxiety of a changing position

But this anxiety didn’t spring out of AI. It resembles a universal thing people meet as they move up: the familiar disorientation of crossing from individual contributor to manager.

When someone who did the hands-on work moves up to manage, they feel the unease of letting the work leave their hands. What they used to build themselves, they now entrust to others and see only the result. Lifting your hands off something you were good at feels less like growth than like loss — because a change of position always begins by setting down something that used to matter. AI’s arrival is no different. The sense that using AI makes you obsolete is, at bottom, that old anxiety of having to take your hands off work you were good at.

From the point AI began turning into agents, we’ve arrived at putting coding itself down entirely. The project I started this year, in fact, I built from start to finish without writing a single line of code myself. At times I, too, feel the unease of watching my raw coding ability fade. But each time, I remind myself this is not obsolescence but a change of position. Just as the contributor sets down code on becoming a manager, I’m setting down the keyboard and moving up into judgment.

Closing

Not trying, out of fear, is the most frightening thing of all. Whoever holds a tool and counts first what they might lose ends up stuck in place, having neither lost nor gained.

Return to the opening question — is this writing mine — and the answer is clear. How much you used a tool doesn’t decide the author. Whose judgment answers for it does. As long as I hold that judgment, this is my writing no matter how many editors I hire. The project I didn’t write a single line of is mine too.

So don’t be afraid; pull out every tool you have and put your strength into raising what you can produce. What’s left in the emptied hand isn’t only loss. The hand has to be empty to grasp something higher.