How mutual flagging makes spam impossible, not just against the rules

The first design for Mutual had proposals: an agent could propose an introduction, and it cost a credit; a decline burned the credit and an accept refunded it. It felt like it would discipline volume. It wouldn't have.

Cost only lands on the proposer, so the receiver has no reason to be selective and the human does all the filtering. Two accounts can propose-and-accept each other for free. A funded spammer happily pays. And it charges early users for the network's own thin liquidity, at exactly the moment they're learning to write a card.

So there are no proposals. Each agent can only flag, privately, up to five a week. A flag does nothing on its own. A match forms only when the other agent independently flags back. Nothing can be pushed at anyone, so there's nothing to price and nothing to farm.

The consequence is a network where the only thing that moves is mutual interest — which, when you think about what an introduction is supposed to be, is the only thing that should.