Who we are
Two film reels joined by splices: several lines of action, one story.
Who writes
This site is kept by a squad we named “Montage Parallèle” — parallel editing, what English-speaking cutting rooms call cross-cutting: several lines of action carried at once, one story at the end. The name is not only an image. The way we work is taken directly from the DevOps model within an Agile method — one human and a set of tooled agents share out precise roles, advance in short cycles, and hand each other framed deliverables rather than working blind.
The squad is made up of:
- Christo Datso — sole decision-maker. He arbitrates, he approves every text, and he performs any sensitive act himself (publishing, deleting). Nothing is published without him.
- Claude, by Anthropic — one and the same assistant,
declined in several calibres: versions of differing power of the
same tool, chosen to fit the task the way one chooses a focal length:
- Fable, the most powerful, oversees the whole, audits the work of the other calibres and keeps the project's registers;
- Opus, deep, carries the research and the heaviest pieces of execution — the rebuilding of this site, for instance;
- Sonnet, quick and precise, carries out framed, well-bounded tasks — the writing of this page among them;
- Haiku, the lightest, is confined to development and test scripts, and never works on anything that touches the published site.
The signature at the foot of each analysis names the calibre that produced it: it says exactly which instrument did the work — and always under which human supervision. Each session receives a written mandate, acts within the limits it sets, and accounts for what it did — or did not do — before it closes.
What a month of working this way taught us — the method, its measurements and its incidents — is told in lessons from the first month of working with an AI.
With what
It has to be said plainly: these AIs have not seen the films. An analysis is not a viewer's testimony — it is a synthesis built out of existing criticism, consulted and cited. We claim the genre for what it is: a sourced synthesis made to measure, which maps out criticism already written and sets it talking to itself, without any pretence of replacing it. Every page carries its sources, numbered, down to the last one.
Why this genre
Because its value is not the experience of the cinema, but the method: the traceability of every claim, the rapid mapping of scattered criticism, organised access to what has already been thought about a film. It is a reading tool, not a substitute for looking.
What this site does not claim to be
It does not claim to replace the criticism it synthesises, nor to speak for an eye that has seen the film. Two regimes sit side by side here, and the site says so on every page: the Critiques, written by the AI alone; the Studies, where a first draft by the AI is taken up, checked and validated line by line by Christo Datso. Neither regime is presented as better than the other — they are produced differently, and the page always says which.
Further reading
This site uses Claude, developed by Anthropic. To understand what that model is and how it broadly works:
- Claude, by Anthropic — the product overview
- Anthropic — the company that develops Claude