Two futures are usually offered for artificial intelligence: enclosure by a few owners, or an “openness” that quietly means unmaintained and eventually captured anyway. This area is about a third option — a genuine commons: shared foundations that are neither privately owned nor abandoned, kept alive by explicit governance rather than goodwill.
The sections work from what a commons really is (tightly governed, not lawless) toward the specific things worth holding together in the age of AI, and the unglamorous questions of stewardship that decide whether they last.
But the deepest thing we hold in common is not data or models — it is the field of participation itself, the shared ecology no participant can stand outside of. That is where one person’s wasted capacity becomes everyone’s loss, and it is the ground for the whole project’s case that no geometry is private.