Not every yes means the same thing

One-person-one-vote protects something essential: equality. It also turns every preference into the same unit. A shrug and a please-let’s-do-this arrive as identical marks.

One-euro-one-vote catches intensity, but confuses caring with having money. Attention systems find a third way to flatten things: the person willing to repeat themselves most becomes the largest voice in the room.

None of those shapes is always wrong. They just hear different things. The design problem is allowing emphasis without allowing capture.

Quadratic funding applies the same intuition to a matching pool. It pays attention to the breadth of independent support, not only the largest transfer.

Suppose one person contributes €16. The unconstrained quadratic score is 16. Now suppose four people contribute €4 each. Direct contributions still total €16, but the score is 64: the four square roots are 2 + 2 + 2 + 2 = 8, and 8 squared is 64.

That 64 is a comparison score, not a promised €64 payout. A real round normally scales every allocation to fit the available matching pool. The useful observation is simpler: four independent signs of support say something one larger transfer cannot.

Preference intensity is only one problem. People also need a reason to pay attention. Sortition selects a smaller random or carefully stratified group, then gives those people time, information and a meaningful chance to influence the result. Combined with quadratic voting, each selected person’s judgment becomes large enough to think about on a human scale.

Selection design, representation, preparation, compensation, legitimacy and accountability do not disappear. Neither does the work before the formula: who belongs, who receives equal credits, who puts proposals on the ballot, what counts as independent support and where the matching money comes from.

A curve cannot create trust. Fake identities can turn one person into many. Friends can collude or trade support. Votes can be bought if people can prove how they voted. Real money leaves a wealth effect. Popular projects have a distribution advantage. Matching pools run out. Complexity, strategy and low-information participation can blur the signal.

Some questions should stay outside the mechanism altogether. Rights, protection from discrimination, safety rules and decisions that require expert judgment are not improved by letting the strongest preference bid higher. The formula is never outside the group’s choices about identity, boundaries and legitimacy.

Further out, there is a more speculative idea: judge which public goods proved useful after time has passed, while earlier backers bet on that later quadratic judgment. It might help unusual work whose value arrives late. It also brings speculation, manipulation, evaluation and liquidity problems. Interesting to test. Very far from a recommendation.