Google argues that smart employees explain success
In its antitrust confrontation with the government, the pillar of Google’s defense has been that innovation — not restrictive contracts, backed by billions in payments to industry partners — explains its success as the giant of internet search.
Its competitive advantage, it says, is brilliant people, working tirelessly to improve its products.
Pandu Nayak, Google’s first witness in the antitrust trial that began last month, is the face of that defense.
Nayak, a vice president of search, was raised in India and graduated at the top of his class at one of that nation’s elite technical schools. He came to America, earned his doctorate in computer science at Stanford University and then spent seven years as a research scientist on artificial intelligence projects at NASA’S Ames Research Center in Silicon Valley.
Nineteen years ago, Nayak joined Google and found a particularly welcoming workplace, filled with professional friends. “At the end of the day, Google is a technology company — it really values the skills that I have,” Nayak said in his testimony Wednesday.
The computer scientist’s testimony is an attempt to rebut a central argument in the case filed by the Justice Department and 38 states and territories. Their suit claims that scale is essential to competition in search. That is, the more data from user queries a search engine collects, the more it learns to improve its service, which attracts still more users, advertisers and ad revenue. That flywheel, the suit says, is fueled by ever-increasing volumes of user data.
The government and states claim that Google locks in a huge data advantage through exclusive contracts and payments of more than $10 billion a year to Apple, Samsung and others to be the default search engine on smartphones and personal computer browsers. A slender man with thinning gray hair, who spoke in clipped, slightly accented English, Nayak has a professorial air, and has taught graduate courses at Stanford. Much of his testimony was essentially a tutorial in search technology and its evolution, guided by a Google lawyer, Kenneth Smurzynski.
Nayak went through the long string of research advances at Google that have improved search quality, including developments in machine learning, deep learning, transformers and large language models — the technology behind AI- driven chatbots such as Openai’s CHATGPT and Google’s Bard.
The evolution Nayak traced was one in which innovations in language understanding have become increasingly important to gains in search quality, while the sheer volume of search queries has become less important.
At one point, Nayak cited a change where using a third less data brought “no meaningful decline in search quality.” Clever software, he suggested, matters more than more data.
In testimony earlier this week, Michael Whinston, the Justice Department’s economic expert, estimated that Google’s exclusive deals blocked rivals from between one-third and one-half of all user search queries in the United States.