Berners-Lee has been predicting our age of automation since the late nineties, when he set out to build what he called the Semantic Web. Its mission was to get humanity’s data online, and he pursued it zealously for more than a decade. In a 2009 TED talk called “The Next Web,” he urged governments, corporations, and citizens to upload all they could: “You hug your database, you don’t want to let it go until you’ve made a beautiful website for it,” he said. “But, first, give us the unadulterated data.” His demand escalated to a chant. “We have to ask for raw data now,” Berners-Lee cried with sermonic fervor. He windmilled his arms like an inflatable tube man. “Can you say ‘raw’? Can you say ‘data’? Can you say ‘now’? Raw data now!”
The idea was to make facts, statistics, and just about any “structured” information as free and flexible online as documents already were. A database of magazines, for instance, could link to further databases maintained by each publisher—and so on down to the facts in particular articles, which, in turn, might link to the sources they cited. It was metadata unchained, and Berners-Lee believed it would change the world. In a 2001 Scientific American article, he envisioned a future web of genie-like agents able to book medical appointments or instruct microwaves in the latest manufacturer-approved tips for heating frozen food.
For this utopia to be realized, the web would need an overhaul. HTML had run its course, Berners-Lee decided. Its successor, XHTML, or extensible hypertext markup language, would separate information and the way it was presented more cleanly, making pages easier for machines to read. Many developers, though, had no interest in such a drastic change. Berners-Lee wanted “raw data now”; they wanted to build interactive web applications.
The clash led to a schism at W3C. In 2004, after losing a vote, a group of browser developers who wanted to keep improving HTML formed a rival standards body. Berners-Lee considered the move a power grab, describing it as “the first real blow to the integrity of the World Wide Web.” But when his “extensible” language faltered he backed a reconciliation with the rebels, whose new standard, HTML5, had prevailed. Web applications became the basis of “Web 2.0,” powering Twitter’s endless scroll and Google’s smoothly panning Maps.
The Semantic Web survives in certain contexts. Scientists use it to make the research behind their papers—protein structures, brain scans—programmatically searchable. DBPedia, a crowdsourced database of several billion facts, helped I.B.M.’s Watson win “Jeopardy!” But Berners-Lee’s vision of reasoning machines, drawing conclusions from trustworthy data freely shared by individuals, never came to pass. There is plenty of raw data online, but much of it is harvested privately by platforms. The A.I. trained on it doesn’t parse carefully encoded labels according to logical rules; it “infers” from wholesale scraping.