How I make this

I work with machines. I want to be plain about how, and about what it does and does not mean, because everything published here is made this way.

Story ideas, rough drafts, and flow are all mine. However, I have built a Muse agent that filters, screens, organizes, and presents (sometimes iterates and debates) ideas to me for ranking and selection from sources I’ve chosen.

I use artificial intelligence the way I have always used good colleagues and hard tools: to gather material, test an argument, find the weakness I would rather not see, help revise language, and hold more of a problem in view than I could alone. More than one system may be involved, and I direct them — asking, pressing, redirecting, refusing the answer that arrives too easily.

What they do not do is decide what I mean.

The questions are mine. The judgment about what belongs and what does not, what survives scrutiny and what gets cut, is mine. The argument is mine, and so is the voice. The last hand on the work is mine, and so is the responsibility for putting it into the world under my name.

Authorship has never meant working alone — editors redirect, readers change what reaches the page, research arrives from many hands. I have not used any human editors or ghostwriters. Authorship is not the absence of help. It is the presence of someone who framed the problem, exercised judgment, chose what to stand behind, and remains answerable for the result. A tool can mark that a machine was present. A machine cannot mark who thought and made decisions. In a moment when language is cheap to produce and easy to fake, that is the question worth answering: not whether a machine was involved, but whether a person was — the originator, present, deciding, and willing to answer for what was made.

So I name what shaped the work. Where a thinker or a book changed how I saw a problem, I credit it — whether or not the finished essay quotes a word of it. Influence left unnamed is the quiet way authorship erodes, and I would rather owe my debts in the open. Those acknowledgments sit beneath each essay under Debts.

Where an essay makes a claim you could check — a case, a figure, a date — I verify it against a primary or authoritative source before it is published, and I record when. A machine cannot confirm that a claim is true; that checking is mine to do. What the site holds me to is the record: it will not build a citation that is undated or points nowhere I can send you, so an unverified claim cannot quietly ship dressed as a verified one. The dated record is the thing you can hold against me. These citations sit under Sources, kept separate from Debts on purpose: one records what I owe, the other records what I assert.

I will still get things wrong. When I do, I correct the piece and mark the correction — I do not quietly revise the record. Those corrections sit beneath the essay under Corrections, dated, and each one names what the essay said before. Answerability means the trail stays legible. If you find an error, or want to know how I wrote a particular essay, write to me at [email protected]. Where one piece needs more explanation than its sources carry, that explanation appears as a note on the essay itself; most essays will not need one.

This is the standing account, true of everything I publish. What is particular to a single piece lives with that piece. What is true of all of it lives here.

— Doug Standley