What does governance need to become?

## The civic hacker and the senator

Taiwan's Sunflower Movement occupied the legislature for twenty-three days in 2014. Out of that wreckage, a thirty-five-year-old civic hacker named Audrey Tang built vTaiwan — a platform letting citizens deliberate on national policy through structured digital conversation. By 2019, it had shaped binding regulation on everything from Uber to telemedicine. Twenty-three million people, doing something the rest of the world was still theorizing.

That same year, the United States Congress was asking Mark Zuckerberg how Facebook made money if it did not charge users.

Governance, as most of the world practices it, runs on architecture designed for populations that communicated by letter and died before sixty. The U.S. Constitution was drafted by fifty-five men in 1787. The Westminster model dates to the 1200s. Brilliant for their moment.

## The immune system

The bottleneck: the people who would need to vote for structural reform are the people reform would displace. Term limits require a congressional vote. Electoral college abolition requires ratification by legislatures whose members benefit from the current allocation. Every serious reform proposal runs into the same lock. Structural self-preservation, operating with the quiet efficiency of an immune system attacking the surgeon.

Tang's team built a workaround — digital deliberation at scale, routing around legislative bandwidth. In Washington, Jason Brennan published *Against Democracy* and argued the bandwidth limitation might not be technical. Catholic philosophers pointed out that both miss the point: the principle is altitude — decisions at the lowest competent level, neither king nor mob. And in Bloomington, Indiana, Elinor Ostrom spent decades studying governance that actually worked — irrigation commons in Nepal, fishing villages in Japan — and concluded the Western fixation on the nation-state was historically provincial.

## The machine on fumes

Trust in government across the OECD has dropped below 40 percent. The machine still runs. It produces less and less of what anyone asked for. Whether the fix is a software upgrade, a personnel filter, a structural principle, or a paradigm replacement is where the camps diverge — and the bottleneck that blocks all of them is the same lock with four different keys, none of which the locksmith has any incentive to hand over.