Google's Software Steals Limelight From Hardware at Nexus Event

  • Company uses in-house devices to show off Android innovations
  • Software emphasis contrasts with Apple's hardware-first tech

Google Takes On iPhone With New Nexus Phones

At an event meant to showcase hardware, Google spent a lot of time on software. The Web-search company is mining and weaving together its storehouses of user data to create devices meant to woo buyers away from Apple’s smartphones and tablets.

While showing off new Nexus phones in San Francisco Tuesday, Google Inc.’s future Chief Executive Officer Sundar Pichai and a handful of other executives sought to highlight how the company’s vast machine learning systems can give consumers a better experience when navigating their own devices and the Internet. These systems analyze and predict users’ behavior and preferences from across Google’s various sites and databases.

“In this multi-screen world, it’s even more important for us as Google to do the hard work so that the user experience is simple and delightful,” Pichai said. “Increasingly, we do that by doing hard and deep computer science.”

Google’s strategy is motivated by necessity. The company does create a few marquee phone and tablet devices in-house -- sometimes with a chosen hardware partner -- to signal to the industry what components it sees as crucial to running future software, as it did with the touch and display components of the Chromebook Pixel and various features of its Nexus smartphones. But it leaves the majority of the hardware development to the third-party manufacturers that make the hundreds of other Android-based models sold in the U.S. and around the world.

Defining Feature

Google’s flagship Nexus phones account for less than 1 percent of Android-based device sales in the U.S., Brazil, China, Australia, Japan, the U.K., France, Spain, Italy, Germany, Russia and Mexico, said Carolina Milanesi, an analyst at Kantar Worldpanel ComTech.

This means Google frequently touts software as the defining feature of its devices, and the hardware additions -- such as fingerprint sensors, fast charging and low-power co-processors called sensor hubs -- are intended to signal to industry partners what hardware they might want to put in future devices to take advantage of Android software innovations.

“As we build our next version of the operating system, we build the hardware along with our ecosystem partners so they can guide the ecosystem forward,” Pichai said. “This year we’ve gone a step further. We have a more comprehensive lineup. We’ve thought about the range of what’s possible.”

AI Systems

The range of what’s possible is largely determined by how the various devices process information, then share it with Google’s artificial intelligence systems, which live on the phones and tablets and in the company’s own data centers. For example, the new Nexus phones have on-board chips that embed sophisticated machine-learning algorithms into a low-power sensor hub, which processes information from other phone components. The phones also come with free 90-day trials for Google’s streaming-music service, which uses AI to predict what songs people want to listen to.

The new Marshmallow version of Android, which will be available on the new phones, comes with numerous learning systems. One of them profiles which applications are used most frequently and selectively tunes out hungry-but-rarely-used apps to reduce power. That led to an improvement in battery life of as much as 30 percent even on older devices that were upgraded to Marshmallow, the company said, again stressing the power of its software for souping up performance.

Another tool, called Now On Tap, brings into focus the different design philosophies between Google and Apple. Now On Tap lets users draw information from individual applications and put them into others, through artificial-intelligence algorithms that figure out which bits of data are germane. That stands in contrast to Apple, which designed an entirely new hardware component named 3D Touch into its iPhones to let users accomplish broadly the same thing by varying pressure on the screen.

All of Google’s learning features are meant to “make sure the end-user experience is much easier on users,” Pichai said. The underlying message: hardware is important, but it’s the artificial intelligence embedded in Android software that will make someone really care about a device.

Before it's here, it's on the Bloomberg Terminal.