#LDS15 Stefan Weitz: Machines Have Enough Data to Understand the Real World
Search has already become a critical part of our lives. But Stefan Weitz, the outgoing director of search at Microsoft, believes search will help hinge together the growing capacity of machines with the unique capabilities of humanity.
“We now have enough data to allow machines to figure out what the real world is,” Weitz said during a presentation at Street Fight’s Local Data Summit in Denver Thursday. “Technology will not replace us — it will augment our lives; and search, in particular, is about to radically enhance reality.”
Technology and devices are already breaking barriers to enhance life, says Weitz. Some technologies assess surrounding spaces and then enhance that information by providing historical data about the location. Others can analyze previous actions of a user and then push notifications to that user’s mobile device, alerting that an item of interest is on sale. Software can even analyze human body reactions to assess real responses to, for example, a television show.
“If I’m watching The Wire, you can see that I’m really into it,” Weitz says. “My heart rate is up, my eyes are focused.”
That information can be more useful for data aggregation, said Weitz, if a user rates the show. The result is that consumers are creating four zettabytes of data every year, providing more and more data for technology to use.
“We’re creating so much data that it’s almost meaningless,” Weitz says. “If you break it down to 133 billion 32-gigabyte iPads, and then stack all of them up, you’d create three great walls of china. That’s how much data we’re going to create this year. And the terrifying thing is we’re going to double that next year, and the year after that.”
But what that means is that it’s now possible – or, inevitable – to describe the real world in extremely precise detail. Search is becoming more proactive and functional than ever before.
It’s not actual artificial intelligence quite yet, Weitz said, but technology has excelled in “machine learning.” After seeing one pattern enough times, a machine can “learn” what will likely happen next.
“Machines have begun to teach themselves about the world without being explicitly programmed,” he said. “All that complex and icky data, at the other end of machine learning pipeline, the technology will show you what that data really means.”
This allows companies to predict what customers will do and what they are looking for. The future of search won’t need to listen to what you ask for in order to know exactly what you mean. Searches and results will appear before a consumer even knows he or she needs it. It will simplify everyday life by taking over the minutiae that were previously taking up time and energy.
“The quest for knowledge can be done with machine learning,” Weitz said. “It’s given search systems a brain. They can interact with us in more natural, humanistic ways.”