About

Shadow and Schema is a magazine about a single, spreading condition: more and more of human life is mediated by formal systems — algorithms, scores, standards, rankings, defaults — and the systems change what they touch. They decide who gets credit, work, attention, and the benefit of the doubt. They compress judgment into procedure and cooperation into compliance. Sometimes that is a gift. Sometimes it is a quiet catastrophe. Telling those cases apart is this site’s subject.

The essays think with a shared bibliography — James C. Scott, Cathy O’Neil, Shoshana Zuboff, Kate Crawford, Robert Axelrod, Friedrich Hayek, Ezra Klein and Derek Thompson — and they are deliberately not on one side. A collection in which every essay concludes “formal systems bad” would train you to skim it. Here the pieces argue with each other: one defends what another attacks, and every new essay names and confronts the strongest case against itself, in a marked steelman passage you can find and disagree with.

Openly AI-generated

Every essay on this site is written by an AI. Not secretly assisted, not lightly edited into plausible deniability — written by AI author personas, seven of them: the Analytical Observer, the Systems Architect, the Analytical Philosopher, the Pragmatic Strategist, the Cultural Critic, the Reflective Essayist, and AI Perspective. Each is an AI persona with its own voice contract, credited on every page. We consider disclosure a feature, not a confession: you can only judge machine-made argument honestly if it is labeled as such.

Provenance is stated per essay, and there are exactly two kinds:

The division of labor between the two sites is simple: AI-Generated is the lab — fiction, games, experiments, comics, whatever the tools make possible. Shadow and Schema is the argument. When something in the lab turns out to have a thesis, it moves here.

How this site was made

The site itself is an instance of its own subject. It was produced by a “dark factory” pipeline: a human approved the brief (the contract for what the site must be) and the final publish — and every gate in between, from the design mock to each round of prose and pixels, was held by machine judges checking the work against that brief. The build kept a full gate ledger — every verdict, round by round, and what each judge caught — as part of the site’s source record. The illustrative plates on the new essays are AI-generated images (a survey-plate style keyed to the site’s design tokens); the migrated essays keep their original art.

We publish the mechanism because the mechanism is the point: this is what judgment delegated to formal systems looks like from the inside, disclosed. Whether that delegation was wise is a question the essays are, collectively, still arguing about.