Rough Consensus

The internet was designed for a world that no longer exists. It works in the world that does.

In 1969, a handful of graduate students wired four computers together and wrote the rules as they went. They knew each other by first name, so they built protocols that simply believed whatever they were told: that a sender was who they claimed to be, that an address was honest, that the packet arriving now came from a friend. For a network you could fit in one room, it was a reasonable bet.

Five billion strangers later, those same protocols are still running. And still trusting.

Rough Consensus takes twenty-three of them—one per chapter—and traces each from the idea that made it to the breach that broke it. A 75-cent accounting error that uncovers a KGB spy. A misconfigured router in Pakistan that erases YouTube from the internet for two hours. A forged Google certificate, signed by a quietly hacked Dutch company, turning up on the connections of hundreds of thousands of Iranians. One tampered software update that walks into the Treasury, the Pentagon, and Microsoft at once. And a plain text file the entire web honored for twenty-eight years, until the AI companies came.

The thread connecting them is the uncomfortable one: almost nothing here was broken into. The attackers used the protocols exactly as designed. The openness that let four nodes become five billion users is the same openness that lets anyone lie.

Clear enough for readers who have never touched a command line, precise enough for the engineers who run these systems, Rough Consensus is the story of how the internet actually works, why it keeps getting hacked, and why—against every reasonable expectation—it keeps working anyway.