Trusting the Next Stranger

“Trusted third parties are security holes,” Nick Szabo wrote in 2001, years before anyone could buy a bitcoin or transact with a stablecoin, and it became one of the founding aphorisms of everything that followed in crypto. While usually read as a critique of banks and clearinghouses. I read it as something bigger: a summation of one of the oldest unsolved problems in economics. Every time the economy needs to grow, it runs into the same wall: someone has to transact with a stranger they have no reason to trust — and every leap forward in economic history has been, underneath the technology of the moment, an answer to that problem.

Take the Tang Dynasty merchant. Bronze coin was heavy and risky to move in bulk, so from around 804 CE merchants began depositing coins with a trusted agent in one city and carrying home a paper certificate redeemable for the same amount in another: feiqian, or “flying money.” The paper carried no value of its own. It stood for coin kept somewhere else, engineered so an agent hundreds of miles away would honor it as if it were the metal itself.

Four centuries later, the same wall showed up in thirteenth-century Italy. Trade was expanding across Europe, but moving gold and silver over land and sea to pay a supplier in another city was slow, expensive, and an open invitation to theft. Italian merchant-bankers solved it with the bill of exchange: an instrument that let a merchant in Florence instruct an agent in Bruges to pay a merchant neither of them fully knew, on the promise that the paper itself carried the credit. The real invention was trust engineered into a financial instrument, because the old kind – knowing your counterparty personally – didn’t scale.

Two civilizations, four hundred years apart, arrived at the same answer, because the problem was never really about the gold. It was about trust that could travel faster than the person vouching for it.

By the time Marco Polo reached Kublai Khan’s court, nearly five centuries after feiqian, the Chinese state had taken the moneychangers’ trick and formalized it into imperial policy — paper currency, backed by nothing but the emperor’s seal and the threat of execution for forgers. Polo could barely believe what he was watching and said so in his Travels:

“All these pieces of paper are issued with as much solemnity and authority as if they were of pure gold or silver… the money is then authentic.”

A Venetian merchant, standing in front of the first paper money an empire had made its only legal tender, trying to explain to the folks back home why paper could be trusted. Sound familiar?

Two centuries after our Florentine merchant, in Venice, the constraint had moved. Merchants were pooling capital into joint ventures with partners who couldn’t watch each other work in real time. Luca Pacioli’s 1494 Summa de Arithmetica printed, for the first time, a systematic method for double-entry bookkeeping, what he called “the Method of Venice.” It gave every transaction two sides and every partner a shared, auditable record of what had actually happened. The ledger became the thing partners trusted instead of each other.

By 1602, wouldn’t you know it, the constraint had moved again. The Dutch East India Company needed capital and continuity that no individual merchant’s lifetime or personal credit could supply, so the Dutch invented a new kind of counterparty altogether: a chartered company that could own property, sign contracts, and outlive its founders — the corporation as a legal person, answerable to the state that chartered it and, through tradable shares, to the public that funded it. That single invention did more to shape the next four hundred years of capitalism than any trade route or product.

The pattern holds every time. Before an economy can extend itself to a new kind of counterparty — a foreign merchant, a passive investor, an anonymous cardholder, a server on the other side of the internet — someone has to build the infrastructure that makes trusting that counterparty safe: an identity, a ledger, and a way to know a transaction is actually, finally, done. We generally tend to remember the commerce that followed. We, as often, forget how much slower and more painful the infrastructure was to build, usually assembled after a crisis rather than before one — a currency debasement, a bank run, a stock bubble — because trust, historically, has been expensive and institutions are what we built to carry the cost.

So what changes when the stranger isn’t a person at all, but a piece of software? Everything, and also nothing.

We’re at that wall again, except this time the new counterparty isn’t a person or a company. It’s an agent: software that can hold a balance, decide how to spend it, and transact with another piece of software it has never encountered, with no human approving the individual step. This is a new kind of economic actor in its own right, arriving faster than any before it, and it needs exactly what every counterparty before it needed: identity, a shared ledger, and finality, except now on a timescale of milliseconds instead of months.

What’s actually new, and what makes this moment different from the ones before it, is that for the first time we don’t have to wait for the crisis to build the trust infrastructure afterward. Identity, settlement, and finality can be written into the protocol before the agent ever shows up. That is what a decade of building open, programmable money and settlement rails was for, even when it didn’t look like it at the time. Crypto spent ten years arguing about what it was building. It turns out it was building the ledger the next stranger would need — and the next stranger is here early.

This is the intersection: the agentic economy and the onchain economy aren’t two trends that happened to arrive at once. One built the instrument. The other is the counterparty that instrument was always going to need. Circle’s mission has always been to build the Internet Financial System — trusted, programmable money and the infrastructure to move it — since long before “agent” meant anything other than a person representing your interests. USDC is the money side of that: a dollar built to move at the speed and finality software actually requires, not the speed a bank teller can process by hand. Arc is the venue side: an economic operating system built for exactly the kind of activity a bill of exchange or a joint-stock charter used to make possible, except native to software from day one instead of retrofitted onto it centuries later.

None of this is finished. Identity for agents, reputation that can’t be gamed, credit for a counterparty with no employment history — these are being built now, not solved. But the direction is the one thing history is consistent about: every economy that has ever grown did it by learning, at real cost, how to trust its next stranger. We’re the first generation that gets to write that trust down in advance instead of cleaning up after the fact. The next stranger is already here. For the first time, so is the ledger.