For more than a thousand years, the intellectual core of the Western world was held captive by a single institution: the medieval Catholic Church. It became enormously powerful as it controlled the dissemination and interpretation of the Bible. In other words, the Church controlled who could read the Bible, and what it was permitted to mean, which the laity was forced to accept. Priests read from the Bible in Latin, a language common people could not read; the Church supplied the literate administrative class that every ruler across Latin Christendom depended on to govern. Most of our knowledge of medieval Europe is downstream of the monasteries tasked with creating the small literate class, copying nearly every extant book and keeping the records of statecraft that royals depended on. Through its control over virtually all information, the Church’s control over belief was perhaps the most concrete monopoly the world had ever known; only sanctioned beliefs were permitted.

What undid the Church’s monopoly was not a competing faith, but instead a rival technology. The mechanism was Gutenberg’s printing press. While his own Bible was in Latin, his technology made copies cheap to distribute across languages. Within decades, printers across Europe were producing Bibles in vernacular German, Italian, Dutch and French. In fact, Luther’s German New Testament sold between 3,000 and 5,000 copies in three months. The press did not make books cheap to produce but cheap to copy.

The Catholic Church lost its monopoly over information. That loss, not any single heresy, was the Church’s final blow: Once a farmer could read the Bible in his own language, the priest no longer gatekept the Bible’s meaning as printing presses, dotted around Europe, printed the Bible in vernacular languages. The people now had the ability to interpret holy scripture for themselves.

The Thirty Years War, one of the deadliest in Europe’s history, culminated in the Peace of Westphalia in 1648, which splintered the continent into a patchwork of separate sovereigns. Fragmentation created what a universal church never offered: exit. A heretic or merchant could cross a border into a different regime — most, after the Reformation, with their own national church. Westphalia made rulers compete for mobile capital: this competition between sovereigns created a constraint on any one of them. A prince who expelled his Huguenot weavers or Jewish bankers lost their wealth and tax base to a neighboring state. Exit was expensive, so rulers learned to tolerate what they could not afford to lose.

As the medieval world gave way to the early-modern age of capitalism, a new axis of control emerged: money. Capitalism made prices a new sort of information. Silver prices on European trading floors reflected what was happening an ocean away in Potosi. Information had escaped; money, itself just information about markets, would take three more centuries. Between Westphalia and World War II, money was tethered to metal, independent of any single throne. A merchant who disliked or distrusted any king’s coin could hold gold, or move. Then, in the twentieth century, the door slammed shut. The 1944 Bretton Woods conference made the US dollar convertible to gold and pegged other currencies to the dollar. Indeed, the US dollar’s reach extended to the farthest ends of the earth through SWIFT, the Belgian messaging network that nearly every international bank transfer must pass through; and through correspondent banking, the system that enables smaller banks worldwide to hold accounts at large American banks to access dollars. Cut off from either, a bank anywhere on earth can barely move money across borders. Sanctions made these rules enforceable. The might of the US dollar would follow you across every border; exit by geography was made impossible.

The resistance to dollarization was born of the same war that built this financial order. World War II had made cryptography an important weapon: the Allies broke Nazi codes at Bletchley Park; everyone understood that codebreaking helped win the war. In 1976, the US government therefore classified strong cryptography as a munition, legally equivalent to missile technology. That same year, two mathematicians, Whitfield Diffie and Martin Hellman, invented public-key encryption: anyone could publish a “public” key that scrambled messages, and only their “private” key could unscramble them. For the first time, two strangers could communicate in secret without ever meeting to exchange a code. Diffie and Hellman had created the cryptography to guarantee privacy, but the US government still considered it to be a threat to national security. Their research had been funded by the military itself. RAND Corporation, the Air Force’s think tank, had sponsored work on public-key cryptography when it was researching how to communicate confidentially.

David Chaum was the first to see what this math could mean for money. In 1982, the year he earned a PhD in computer science from Berkeley, he invented the blind signature, a way for a bank to certify that a digital coin was real without seeing who held it. It gave the digital coin a key property of physical cash — no one tracks where you spend a dollar bill. By the end of the decade, Chaum had turned the blind signature into DigiCash: money stored on your own computer, untraceable and impossible to freeze. But DigiCash was doomed. Being a company, everything ran through the centralized DigiCash apparatus, thus the currency could never be truly independent. Its libertarian customer base had no appetite for centralization. When the company went bankrupt in 1998, DigiCash died with it.

But Chaum’s most important legacy might lie in the people he inspired. They came to be known as the cypherpunks, a play on the cyberpunk genre of science fiction: programmers who believed privacy would be guaranteed in a battle won by code rather than through the law. The cypherpunks arrived in full force in 1992, when three men — Tim May, a young, retired physicist from Intel; John Gilmore, Sun Microsystems’ fifth employee; and Eric Hughes, a Berkeley mathematician who had worked with Chaum himself — invited 20 of their favorite programmers and cryptographers to a house in Oakland. Their task was to invent the tools that would let individuals escape the state’s surveillance. May, Gilmore, and Hughes were all radical libertarians. They believed the defining political question of the era was whether governments would use the internet to surveil individuals, or whether those individuals could use cryptography to escape governments entirely. May’s argument was simple: if the state cannot monitor you, it cannot control you. “The State will of course try to slow or halt the spread of this technology,” May had warned in 1988, “citing national security concerns, use of the technology by drug dealers and tax evaders, and fears of societal disintegration.” He was right. However, math offered a way out: it is permanent and immune to power.

The Cypherpunks mailing list, run off Gilmore’s personal server toad.com, grew to hundreds of programmers. It became the coding sandbox for remailers that erased the sender from emails, cash protocols that bypassed banks, and the proof-of-work concept that would later undergird Bitcoin. Teams of these anarchist programmers teamed up to build; their ethos was Hughes’ catchphrase: cypherpunks write code. “Privacy is necessary for an open society in the electronic age,” Hughes stated in his A Cypherpunk’s Manifesto. “Privacy is not secrecy. A private matter is something one doesn’t want the whole world to know, but a secret matter is something one doesn’t want anybody to know. Privacy is the power to selectively reveal oneself to the world.” Given the NSA’s warrantless wiretaps, the FBI’s infiltration of antiwar groups, and a president who had resigned over bugging his rivals within living memory, the cypherpunks couldn’t trust the government to protect their privacy. Rather, the plan was to write software that would make surveillance impossible.

The state was listening and took the threat of unbreakable civilian encryption seriously. Phil Zimmermann, an antinuclear activist and programmer, had spent years wanting to put military-grade encryption in the hands of ordinary people. When a 1991 Senate bill proposed forcing communications providers to hand over personal data, he acted on the moment. His encryption protocol, Pretty Good Privacy, or PGP, was released online for free. “It’s personal,” Zimmermann declared in the user guide. “It’s private. And it’s no one’s business but yours.” PGP allowed its users to scramble their email so thoroughly that even the National Security Agency couldn’t read it. Within weeks it had users across the world, and it remains the most widely used email encryption to this day. The US government opened a criminal investigation against Zimmermann for exporting munitions soon after, treating the release of encryption software as though he had shipped missiles to foreign adversaries. Two years later, in 1993, the US government tried to mandate the Clipper Chip: standard encryption for the internet, with the NSA in control. The battle was waged in three separate ways: technical, political and economic. Matt Blaze, a Bell Labs researcher, published a paper showing the chip could be bypassed, rendering it useless; the Electronic Frontier Foundation organized opposition, calling it unconstitutional surveillance; and tech companies lobbied against it, warning that foreign customers would never buy products with backdoors. Thus, Clipper was dead by 1996.

The Zimmermann case was dropped for a simple reason: once the source code was on the internet, it was out of any one person’s control. The code could be replicated, copied, studied. The US government unwittingly discovered what the Church had four centuries earlier: that you cannot prosecute a copy out of existence. No court ever guaranteed a right to encryption, but the government retreated. The cypherpunks won.

With the government in retreat in the fight over encryption, cypherpunks turned back to the foundational project: fully sovereign money. The attempts came quickly mostly from the cypherpunk mailing list itself, each solving one piece of the puzzle. Adam Back’s Hashcash in 1997 was not even designed as money, but as a scheme to stop email spam through forcing every sender’s computer to solve a small mathematical puzzle before an email could be sent — trivial for one email, catastrophic for a million. With this scheme, Back had invented something more consequential than email spam filtering: proof-of-work, a system that requires computers to burn electricity on elegant math puzzles before they can participate in the network. This made computer power itself the price of entry.

Wei Dai’s b-money came the next year. It was the first design for online money with no issuer, no company, no bank, and no owner. How it worked was that every participant kept their own digital copy of the ledger, and money was created through Back’s proof-of-work mechanism. Nick Szabo’s bit gold, also in 1998, came closer still. A computer scientist and lawyer, Szabo had spent years studying why gold worked as money. Its essential property, he decided, was scarcity that no one could control; his bit gold recreated this scarcity digitally, chaining Back’s proof-of-work into a permanent, timestamped record that looked like money without a mint. But he could never solve the final problem. The missing piece was a solution to the double-spending problem. Digital money has a flaw physical cash does not: a digital coin is just information stored as 1s and 0s, and this information can be copied. A dollar bill cannot be copied, per se — if you give a dollar bill away, it’s not in your possession anymore. In other words, if you spend the same coin twice with two different people and both transactions look valid, who decides which one counts? Every earlier design either trusted a company to decide, like DigiCash, or assumed some authority would settle disputes. Without an authority, the system fell apart.

For nearly thirty years, no one could crack the solution to the double-spending problem. Hal Finney, a cypherpunk from the mailing list and one of the earliest PGP developers, came closest. In 2004, he built a working version of reusable proof-of-work in which Back’s computational stamps converted into tokens that could circulate from one person to another. But the conversion ran through Finney’s own server. A cypherpunk central bank is still a central bank.

In October 2008, someone who called himself Satoshi Nakamoto finally cracked it. Whoever Satoshi was — and no one knows to this day — he announced the idea in October 2008 on a cryptography mailing list descended from the original cypherpunk list, in a nine-page paper titled “Bitcoin: A Peer-to-Peer Electronic Cash System.” The design beautifully combined the pieces the movement had spent a generation assembling, like Back’s proof-of-work, Dai’s distributed ledger, and Szabo’s chained record. Bitcoin settled the question of who owns what with a simple rule: the valid chain with the greatest accumulated proof-of-work wins. Rewriting this chain would be ruinously expensive, because forging history meant outcomputing the entire honest network, a growing one. Both the structure of Bitcoin and Satoshi’s anonymity made it a movement without a leader. The Bitcoin whitepaper dazzled the cypherpunks; Finney was among the first to download the software, and in January 2009 was the first person Satoshi ever sent Bitcoin to.

The very first block declared its purpose in plain text, timestamped forever into the chain: “Chancellor on brink of second bailout for banks,” the headline from that morning’s Times of London referring to the British government’s plan to rescue its banks with public money. Bitcoin was then framed as sovereign money without a ruler, a protest against the state bailing out its own banks at the expense of its citizens. Then, in 2011, Satoshi disappeared. His departure had ensured that nobody could own Bitcoin. And in that way, he also ensured that everyone could.

In our time, crypto and AI are believed to be studies in contrasts; but how crypto achieved decentralization could inspire open-source AI builders, too. Both have the same idea of exit. Crypto lets anyone hold their own keys; open-source AI differs from its closed-source counterpart by letting anyone download model weights, the billions of numbers that underlie a model itself, and run or fine-tune it however they want. But while Bitcoin was built after the dollar’s monopoly was in place, the opportunity to build open-source AI is here before the monopolies (or oligopolies) have fully formed.

For a new generation of punks, that opportunity to set frontier intelligence free is calling.