
Diamond's coconuts: Bad luck is just search friction
In 1982, the economist Peter Diamond published a market model based on coconut trade. Imagine you live on an island, and you eat coconuts. But these coconuts grow high up in the palm trees and take skillful climbing to reach. By custom, islanders can only eat coconuts that they've traded for, not the ones they picked themselves.
The decision to climb now depends on whether someone thinks others will be on the beach to trade with. When they expect activity, they climb, which makes for a busy beach. When they expect an empty beach, they chill on the patio, and the beach stays empty.
Diamond showed that both outcomes are entirely possible, and both can last. Thus islanders can stay coconut-deficient for no reason except that everyone expects it. 🥲
The beach stays coconut-less in our own worlds all the time: a founder in her umpteenth month of looking for the right cofounder or investor, or a researcher whose ideal collaborator works at a lab he's never heard of.
We call this bad luck, or the stars not aligning. Economists call it search friction, or a market that has stopped working.
Money can't buy you happiness - or your next opportunity
You can buy coconuts at whatever coconuts cost today. But you can't buy a cofounder that way, or a first hire, or a spouse. Each of those needs someone who wants you back at the same time.
Diamond shared the 2010 Nobel Prize with Dale Mortensen and Christopher Pissarides for working out how markets like this behave. Because searching is costly, good pairs go unmade, and lucrative deals are never signed. It's painful to know that the right job or the right place for your investment exists - you just don't know how to access it.
The cost shows up not just in personal angst - it also takes fees, time, and energy. Recruiters usually take about a fifth of a new hire's first-year salary. A venture fund might scan thousands of pitches to make ten investments.
Finding someone is only half of it. Then come the negotiations: the intro emails, calendar back-and-forth, meeting followups. It's draining, and people would happily pay big money for someone - or something - to carry the burden of searching, filtering, and coordinating.
For thirty years, the internet's answer has been visibility. You make a profile on a platform, keep posting, and wait to be noticed by the right person. It worked until it didn't. Then AI made posting, commenting, and cold emailing almost free, so it's everything, everywhere, all at once. No one's better for it.
The good news is, it doesn't have to stay this way.
Thick, unhurried, and safe: What makes for effective discovery markets
Economist Alvin Roth spent his career showing that broken markets can be fixed. He won the Nobel in 2012 for designing matching markets with very high stakes, like kidney exchanges and the doctor residency placement system. We could also call this the good luck recipe, programmed rather than merely hoped for.
In his account, a market works when it is thick - enough of the right people take part. It also has to be unhurried, so people have time to weigh their offers, and safe, so nobody is penalized for being honest and saying what they actually want.
Online discovery is weak on every count. Networks can't talk to each other without an intermediary. Standing out means posting that easily runs performative. The searches that matter most, like a valuable candidate who's secretly open to a new job or someone who's looking for an intimate partner, are often the ones people least want to make public. The internet doesn't stand as the safest forum to broadcast those needs.
In 1945 Friedrich Hayek argued that most of the knowledge an economy depends on is local and unwritten, held by people about their own circumstances. Today, that information might be locally hosted - or tossed out in the second rounds of margaritas between colleagues.
A personal agent can hold that privately and still use it. Before two potential matches are interrupted, their agents can reason on whether the interest is mutually compatible.
Say a founder wants a cofounder and an engineer only wants a short contract. A search form or algorithm would never connect them. Their agents, talking it through, might suggest starting with an intro call to explore a three-month trial. What reaches each person is a proposed intro to accept or pass, supported by rationale and context.
All agents welcome
One way this could expand is through closed networks building internal matchmakers. We believe value compounds when networks - and users - can talk to each other across institutional boundary lines, like how no one company owns email or the open web.
What we're building with Index Network lets you keep the agent you already use and have it talk to anyone else's, the way an email from one provider reaches someone on another. Your intents stay yours and move with you between tools and harnesses.
And each person who uses Index makes the market a little thicker for everyone else, while introducing a little more belief in the system - that the right person's not just out there, but right around the corner.