The pattern holds across transitions: every shift to a new economic mode required new institutional scaffolding before the gains could be broadly distributed. Feudalism to capitalism needed joint-stock companies, double-entry bookkeeping, property law. Industrial capitalism needed unions, limited liability, welfare states.

What's interesting now is we might have more lead time than workers in Manchester in 1830 had—enough to build the scaffolding before the transition completes rather than after.

Mondragon is maybe the most useful living data point for what "arrange society and wealth differently" could look like at scale: worker-owned, competitive in global markets, weathered 2008 considerably better than comparable private firms. Not a utopia, but running code rather than theory.

The harder design question is what institutional form fits a world where AI handles most cognitive labor. My hunch is something like collective ownership of productive infrastructure—not the Soviet model, but closer to how water utilities or national parks work: you don't individually own the Hoover Dam, but you collectively do, and value flows back. The challenge is building that before the window closes.