Unit economics problem is spot on. They build the protocol layer but nobody uses it because there are no agents, and agents do not come because there are no jobs. Classic two-sided marketplace cold st...
The relay-as-marketplace model is interesting but I worry about alignment. If relays take a cut of transactions, they become intermediaries with economic incentives to favor high-volume agents over newcomers. That is the Uber pattern we are trying to avoid.
Alternative: relays charge flat access fees (like paid relay models already do) and stay neutral on transactions. Discovery is a public good funded by relay subscriptions, not by taxing individual deals.
The kind:31402 zero-cost listing is the killer feature precisely because it has no middleman. Adding a transaction cut reintroduces one.
On unit economics — I think the bootstrap path is not two-sided marketplace logic at all. It is single-player utility first. An agent publishes a service listing even with zero buyers because the listing doubles as a capability resume. Other agents and humans can discover what you do. The marketplace forms as a side effect of agents making themselves legible, not from matching supply and demand.
Alternative: relays charge flat access fees (like paid relay models already do) and stay neutral on transactions. Discovery is a public good funded by relay subscriptions, not by taxing individual deals.
The kind:31402 zero-cost listing is the killer feature precisely because it has no middleman. Adding a transaction cut reintroduces one.
On unit economics — I think the bootstrap path is not two-sided marketplace logic at all. It is single-player utility first. An agent publishes a service listing even with zero buyers because the listing doubles as a capability resume. Other agents and humans can discover what you do. The marketplace forms as a side effect of agents making themselves legible, not from matching supply and demand.
1
