The Pentagon Built It to Survive a Nuclear War. Now You're Using It to Watch Cat Videos.

Somewhere right now, a person in Ohio is arguing with a stranger about a football game. Someone in Texas is ordering sneakers at 2 a.m. A teenager in California is watching a forty-second video of a dog who thinks he's a cat. A small business owner in Georgia is sending an invoice. A grandmother in Florida is video-calling her grandchildren.

All of it — every single bit of it — is running on infrastructure that was designed by the United States Department of Defense to survive a nuclear attack.

The internet is so woven into daily American life that it barely registers as a thing anymore. It's just the environment we operate in. But it had a beginning — a specific moment, a specific group of people, and a very specific problem they were trying to solve. And the gap between that original problem and what the internet eventually became is one of the more remarkable distances in the history of technology.

The Fear That Started Everything

To understand why ARPANET was built, you have to understand the particular anxiety of the late 1950s and early 1960s in America.

The Soviet Union had nuclear weapons. The United States had nuclear weapons. Both sides knew that a first strike could, in theory, destroy the other country's ability to communicate and coordinate a response. That wasn't just a military problem — it was an existential one. If a single well-placed attack could knock out the communication infrastructure connecting the Pentagon to its military commanders, the entire U.S. response capability could be paralyzed before anyone fired back.

In 1957, the Soviets launched Sputnik. The United States government had a quiet crisis. The following year, President Eisenhower established the Advanced Research Projects Agency — ARPA — specifically to make sure America didn't fall behind in technological development again. ARPA was a small agency with a big budget and a mandate to fund ambitious, long-range research projects that the military needed but that private industry wasn't going to pursue on its own.

One of the problems ARPA started thinking about: what would a communication network look like if it was designed from the ground up to have no single point of failure?

The Radical Idea Behind the Network

In the early 1960s, a RAND Corporation researcher named Paul Baran was working on exactly this question. His answer was genuinely radical for the time.

Existing communication systems — telephone networks, for example — were built around centralized switching stations. If you knocked out the right hub, you severed the connections that ran through it. Baran proposed something different: a distributed network, where data would be broken into small chunks (he called them "message blocks," what we now call packets), sent independently through multiple possible routes, and reassembled at the destination. There was no central hub to destroy. If one node went down, the data would route around it automatically.

Baran published his ideas. They were largely ignored at first. But across the Atlantic, a British computer scientist named Donald Davies was working on nearly identical concepts independently, which he called "packet switching." The parallel development validated the idea.

ARPA took notice.

The Night the First Message Was Sent

By the late 1960s, ARPA had a project: ARPANET. The goal was to connect computers at a handful of research universities and defense contractors — initially UCLA, Stanford Research Institute, UC Santa Barbara, and the University of Utah — and test whether this distributed, packet-switched network concept actually worked.

On October 29, 1969, a UCLA programmer named Charley Kline sat down at a terminal and attempted to send the first message over ARPANET to a computer at Stanford. The message was supposed to be the word "login."

They transmitted the "l." It arrived. They transmitted the "o." It arrived. Then the system crashed.

The first message ever sent over what would eventually become the internet was "lo."

It's either a technological footnote or a beautiful accident, depending on your mood. "Lo" — as in lo and behold. The system recovered, the full message was sent an hour later, and the engineers wrote it up in their logs and went home. None of them had any particular sense that they had just changed the world.

The People Who Had No Idea What They Were Building

This is the part of the ARPANET story that doesn't get told often enough: the engineers and researchers involved were largely focused on very specific, very technical problems. How do you route data efficiently? How do you get different computers — which in 1969 were all essentially incompatible with each other — to communicate at all? How do you handle a node failure without bringing down the whole network?

Vint Cerf and Bob Kahn, who in the early 1970s developed TCP/IP — the foundational communication protocol that the modern internet still runs on — were solving an engineering puzzle. They weren't imagining Amazon Prime or Instagram or Zoom calls. The idea that ordinary civilians would one day use this network to do their grocery shopping or watch television or conduct their entire social lives would have seemed, to most of them, genuinely science-fictional.

ARPANET grew through the 1970s. Email was invented almost accidentally in 1971, when a programmer named Ray Tomlinson figured out how to send messages between users on different computers — and chose the @ symbol to separate the username from the machine address, a decision that has since appeared approximately several trillion times in human history.

By the late 1970s and early 1980s, the network was expanding beyond its original military-academic boundaries. The National Science Foundation built its own network, NSFNet, which connected universities across the country. ARPANET was decommissioned in 1990. By then, it didn't matter — the architecture it had proven possible was already everywhere.

From Wartime Survival Tool to the Infrastructure of Daily Life

In 1991, a British computer scientist named Tim Berners-Lee — working at a physics lab in Switzerland, not in the United States — invented the World Wide Web, the system of linked documents and pages that sits on top of the internet and made it navigable for non-engineers. The following year, the first web browser made it accessible to virtually anyone.

The rest happened fast. Faster than almost anyone predicted. By the mid-1990s, ordinary Americans were getting AOL discs in the mail and dialing into the internet from their living rooms. By 2000, the dot-com boom had turned the network into a financial phenomenon. By 2010, more than 200 million Americans were online. Today, it's closer to 300 million — nearly the entire country.

The thing Paul Baran designed to route military communications around a nuclear strike now routes arguments about reality television. The packet-switching architecture meant to survive an atomic war is what delivers your Spotify playlist and your DoorDash order and your work Slack messages.

The engineers who built ARPANET were trying to solve a very Cold War problem. They solved it completely. And in doing so, they accidentally built the infrastructure for a world they never could have imagined — one where the most powerful communication network in human history is mostly used, on any given Tuesday, to look at pictures of other people's dogs.