All Articles
Tech History

It Crashed After Two Letters — Then It Became the Internet

By The Origin Beat Tech History
It Crashed After Two Letters — Then It Became the Internet

It Crashed After Two Letters — Then It Became the Internet

If you had to pick a creation myth for the modern internet — a single moment where everything began — it would probably look something like this: a graduate student at UCLA sits in front of a hulking computer terminal on the night of October 29, 1969, types two letters, and watches the system crash.

The letters were L and O. The intended word was login. The message never arrived.

And yet, that failed transmission was the first data ever sent across ARPANET — the government-funded military network that quietly, accidentally, became the infrastructure the entire connected world now runs on.

The Problem That Started Everything

To understand why ARPANET existed, you have to go back to the particular anxiety that defined American life in the late 1950s and early 1960s: the very real possibility of nuclear war.

After the Soviet Union successfully launched Sputnik in 1957, the US government went into a kind of institutional panic about falling behind technologically. The Department of Defense created ARPA — the Advanced Research Projects Agency — specifically to fund ambitious, forward-thinking research that might give the United States a strategic edge.

One of the problems ARPA turned its attention to was military communication. The existing telephone network was understood to be catastrophically vulnerable. A nuclear strike on a handful of major switching centers could sever communication across the entire country. Military commanders needed a way to keep talking to each other even if large parts of the network were destroyed.

The proposed solution came from a concept developed independently by two researchers — Paul Baran at RAND Corporation and Donald Davies in the UK — called packet switching. Instead of sending a communication along a single dedicated line (the way a phone call worked), packet switching broke messages into small chunks of data, sent each chunk independently across whatever routes were available, and reassembled them at the destination. If one path was knocked out, the packets would simply route around it.

It was an elegant idea. ARPA decided to build a network around it.

The Network Goes Live

By 1969, ARPA had contracted with a company called Bolt Beranek and Newman to build the hardware, and had connected four nodes: UCLA, the Stanford Research Institute, UC Santa Barbara, and the University of Utah. These were research universities with powerful computers — and the goal was initially not to survive nuclear war, but to allow researchers at different institutions to share computing resources remotely.

The military survival angle was real, but the day-to-day motivation was more practical: computers were enormously expensive, and if researchers in Utah could access UCLA's machine without physically traveling there, everyone saved time and money.

On October 29, 1969, UCLA student programmer Charley Kline attempted to send the first message across the network to a computer at Stanford. He was on the phone with a Stanford researcher named Bill Duvall, who would confirm receipt. Kline typed L. Duvall confirmed it arrived. He typed O. Confirmed. He typed G — and the system at Stanford crashed.

The first message in the history of the internet was LO. As the story is often retold, it reads almost like a greeting — lo, an archaic English word for look or behold. It wasn't intentional, but it's hard not to appreciate the poetry of it.

The full login was completed about an hour later, after the Stanford team rebooted. But the crash version is the one that stuck.

From Military Tool to Something Nobody Planned

Over the following decade, ARPANET grew — slowly at first, then with gathering momentum. More universities and research institutions joined. Protocols were developed to standardize how different computers talked to each other. In 1971, a researcher named Ray Tomlinson sent the first email across the network, choosing the @ symbol to separate the user's name from the host computer. He later admitted he couldn't remember what the first message said — probably something he typed to test whether it worked.

What nobody at ARPA had fully anticipated was how quickly people would start using the network for communication rather than computation. The file-sharing and remote computing functions were what the project was built for. But researchers almost immediately started using it to send messages, share ideas informally, and have conversations across institutions. Email became the network's dominant use almost by accident.

In 1983, ARPANET split into two networks — a military side (MILNET) and a civilian research side — and the civilian portion began the transformation that would eventually produce the public internet. The introduction of TCP/IP protocols gave the network a universal language that allowed it to expand far beyond its original nodes. Tim Berners-Lee's invention of the World Wide Web in 1989 added the hyperlinked, browser-navigable layer that made it accessible to ordinary people who had no idea what a packet was.

The Accidental Architecture of Everything

What ARPANET's story keeps coming back to is the gap between intention and outcome. The engineers who built it were solving a specific, bounded problem: keep military communications alive under extreme conditions. They built something robust, decentralized, and adaptable — and those qualities turned out to be exactly what was needed for something much larger.

The internet as Americans use it today — streaming video, social media, online banking, video calls, cloud storage, the endless scroll — was built on top of infrastructure designed for a scenario that, fortunately, never happened. The nuclear war that justified ARPANET's existence never came. But the network it produced reshaped civilization anyway.

That UCLA computer lab is still there, now preserved as a kind of unofficial historic site. On the wall, there's a logbook from October 29, 1969. The entry for that evening reads, in careful handwriting: Talked to SRI host to host.

Understated, to say the least. But then, nobody in that room knew what they'd just started.