1991: The World Wide Web goes public

On 6 August 1991, British computer scientist Tim Berners-Lee posted a concise summary of the World Wide Web project to the alt.hypertext newsgroup. Working at CERN, he invited anyone interested to try the system, examine its code, and help improve it. That short message is widely accepted as the moment the Web left the laboratory and became a public resource. The first website, running on Berners-Lee’s NeXT machine at info.cern.ch, explained how hypertext could link documents across the internet.

The idea itself was older. Berners-Lee had circulated an internal proposal in 1989 and built a working prototype in 1990, complete with the first browser and the HTTP protocol. Yet until that August day the system remained an in-house tool for particle physicists. By releasing the specifications and software openly, he turned a specialised research aid into an invitation for global collaboration.

The intellectual lineage runs deeper still. The dream of a universal, linked store of knowledge echoes Vannevar Bush’s 1945 Memex, Douglas Engelbart’s 1960s vision of augmenting human intellect, and Ted Nelson’s Project Xanadu. Berners-Lee combined those aspirations with the practical packet-switched networks that had grown out of ARPANET and with the simple elegance of markup languages. The result was not merely a technical standard but a new medium for thought.

That medium became indispensable to artificial intelligence. Modern machine-learning systems depend on enormous corpora of text, images, and structured data—most of it harvested from the open Web. Search engines, knowledge graphs, and large language models all rest on the hyperlinked information space first offered to the public on 6 August 1991. Even techniques such as retrieval-augmented generation and web-scale pre-training trace their possibility to the decision to make the Web free and universal.

Today, as AI systems read, summarise, and reason over the same global library Berners-Lee unlocked, the anniversary reminds us that infrastructure can be as transformative as algorithms. A quiet newsgroup post three decades ago still shapes how machines—and humans—encounter knowledge.