March 29, 2016
Silicon Valley has dubbed machine learning and artificial intelligence as the next big thing. Today’s tsunami of data has created the need to make sense of it, quickly and efficiently. Although recent focus has been on giant public clouds from Amazon, Google and Microsoft, now those companies’ abilities to use AI to parse all that data has become the latest arena of competition. All three companies are now striving to define the next gen platform, with Google in the lead and Microsoft and Amazon playing catch up.
The New York Times enumerates some of the AI resources that these three companies already offer. Google Compute Engine head Diane B. Greene reports that her company’s AI capabilities are based on “a wealth of analysis technology” gained from its autocomplete feature (used for search), which can instantaneously touch 500 computers in several locations; the billion uses of Maps and Photos; and 1.4 daily petabytes of Gmail.
Google has also made TensorFlow, the core of the machine learning technology, available as a free open source. Its idea is to make clouds available for rent with its AI resources, which Google executive Urs Hölzle believes will surpass Google advertising profits — $16.4 billion in 2015 — eventually.
“Better maps and photos is just the start,” adds computer design expert Andreas Bechtolsheim, Google’s first investor. “It’s going to be in life sciences, automobiles, everything.”
Microsoft and Amazon are debuting their own machine-based intelligence capabilities. NYT reports that, last year, “Amazon and Microsoft both added machine learning to their cloud software platforms, Amazon Web Services and Microsoft Azure,” to “help customers spot patterns and make predictions in vast amounts of data.”
Among Microsoft’s current 18 machine-learning services are face recognition, text analysis and product recommendation. Microsoft is expected to offer more AI capabilities soon.
Set up in 2014, IBM’s Watson unit, a software and a services business with 500 industry partners has had more than 80,000 developers download its software. “It’s early days, but the long-term goal is to have hundreds of millions of people use Watson as self-service AI,” said Watson division general manager David Kenny.
The stakes are big. Market research firm IDC estimates that, by 2020, the market for machine learning applications will reach $40 billion, and that 60 percent of those apps will run on Amazon, Google, IBM and Microsoft. Today, continues IDC, only about 1 percent of all apps have AI integrated, a number that will rise to 50 percent by 2018.
Machine-learning specialist Pedro Domingos, who wrote “The Master Algorithm,” a 2015 book, believes that “AI and big-data technology will remake the world. “Whoever wins this race will dominate the next stage of the information age,” he said.