As enterprising coders used MapReduce to derive insights from Google’s data, it became possible to transcribe users’ voice mails, answer their questions, autocomplete their queries, and translate among more than a hundred languages. Such systems were developed using relatively uncomplicated machine-learning algorithms. Still, “very simple techniques, when you have a lot of data, work incredibly well,” Jeff said. As “data, data, data”—stored and processed with BigTable, MapReduce, and their successors—became the company’s prime directive, Google’s globe-spanning infrastructure became more seamless and supple. The idea of distributed computation was an old one; concepts like “cloud computing” and “big data” predated Google’s rise. But, by making it intellectually manageable for ordinary coders to write distributed programs, Jeff and Sanjay had given Google a new level of mastery over such technologies. Users may have sensed that something had changed:...
learn more