added functionality generally comes along at the price of added complexity.”
This idea became an obsession of Larry’s. Years later, he called it a “seminal” book, and remembered being amazed when he first read it “that people are so focused on outside things and are not focused on the functionality of things.” (It still drives him mad to stay in a hotel and not be able to figure out how to turn the lights off “in less than three minutes.”) The book, he said, strengthened the attitude he brought to designing the Google search engine, which was the opposite approach from existing search engines like Alta Vista. If you did a search for university on Alta Vista, it heaved at you every text that contained the word university, without ranking value or assessing whether people were actually using the links. Doing the same search, Google relied on the collective intelligence of its users and returned with the top ten universities. Thinking that “your customer or users are always right, and your goal is to build systems that work for them in a natural way, is a good attitude to have,” Page said. “You can replace the system. You can’t replace the user.”
Page and Brin wanted to build an efficient search engine, one that didn’t waste users’ time. Efficient use of time was paramount for them. Neither Page nor Brin eagerly read novels or went to many movies or concerts, and they disdained games like golf that took too long to play. Once, during the early days of Google, Time magazine had arranged to photograph Sergey in a white lab coat. When the photo shoot ran over the allotted time, Sergey abruptly called out, “Red alert,” and simply walked away without explanation.
Page and Brin together, it was said, were “two swords sharpening each other.” They were not breathtakingly more brilliant than their peers, said Winograd, observing that brilliance is commonplace among top Stanford engineering students. What was unusual about them, he said, was their boldness. “Page and Brin’s breakthrough,” writes Battelle in The Search, his book on the history of search, “was to create an algorithm—dubbed PageRank after Larry—that manages to take into account the number of links into a particular site, and the number of links into each of the linking sites.” Instead of relying only on keywords as earlier search engines had, PageRank did a link analysis, counting the sites that were most frequently visited by users and jumping them to the top of the search results. They believed this “wisdom of crowds” approach was a more objective way of measuring which Web pages were most vital. The goal was to get better answers to search queries. They understood one big thing: They were establishing a formula that would harness the growth of their search engine to the growth of the Web. What they needed was massive computing power to conduct lightning-fast searches, and huge servers to store the millions of indexed Web pages.
In 1996, Larry Page’s father, a polio victim as a child, died of pneumonia. He was only fifty-eight. Bereft, Page threw himself into his project. To crawl and index the Web required enormous amounts of Stanford’s computer system, and Page and Brin were not shy about using it. Together, he and Brin harassed the computer science department to grant them extra resources. Terry Winograd, who worked on the project, recalled that “they had more of a feel of teenage kids than most graduate students—‘Don’t tell me what to do!’” Professor Rajeev Motwani, who also worked on the project, said, “They didn’t have this false respect for authority. They were challenging me all the time. They had no compunction in saying to me, ‘You’re full of crap!”’ He recalled, “The fondest memory I have of Sergey is of him walking into my office when I was sitting at my desk and he would say, ‘Bastard!’ That was the kind of thing he would do. Larry was sitting outside. It was a joke. But behind the joke was that he wanted something from me: more computer time.”
Once, Winograd said, they snuck onto the loading dock where new Stanford computers were delivered and “borrowed” them to expand their computing power. Page and Brin brought a cart to transport the crates. Some years later, Page confessed that their embryonic search engine in 1997 hogged so much computer capacity that “we caused the whole Stanford network to go down.”
The new search engine, at first called BackRub, was an object of some secrecy.