How these essays are made

Why I work this way

The subjects I write about — sovereign compute, AI governance, the architecture of technological autonomy — move faster than any single writer can track alone. Working with AI multiplies research throughput, lets me test an argument against its strongest counter-positions before publishing it, and keeps a forty-essay corpus consistent in its claims and its numbers.

Disclosing that is not a caveat; it is the argument practised. A body of work about provenance, authenticity, and legitimacy in the age of AI should model all three. If I ask institutions to be explicit about where machine judgment enters their systems, the least I can do is be explicit about where it enters mine.

What the AI does

Large language models serve as research assistants and editorial counterparts across the life of each essay: gathering and cross-checking sources, stress-testing the structure of an argument, drafting prose from my theses and outlines, proposing alternatives for rhythm and clarity, and checking the final text for internal consistency. The collaboration is iterative — drafts move back and forth between my judgment and the machine's suggestions many times before publication.

What stays human

The positions are mine. The thesis of every essay, the judgment about what matters and what does not, the experience the arguments draw on, and the decision to publish are human. Every sentence is read, shaped, and approved by me before it carries my name — and the responsibility for every claim, every number, and every error is mine alone, not the tools'.

Translations

The French, Spanish, Italian, and German editions of this site are machine translations, generated on demand and served with the English text as the canonical version. They exist so the work can be read in the reader's own language; where a translation and the English disagree, the English governs.