Google Minds Machine Learning
Google CEO Sundar Pichai devoted much of Thursday's second-quarter earnings call to what lies ahead, much of it focused on two "Ms" -- the mobility that is Google's present and the machine learning that is its future.
The first one, mobility, "underlines everything we do today," he told analysts, while machine learning "is the engine that will drive our future" and is already in wide use today, across Google (Nasdaq: GOOG) platforms, influencing and improving its current and future products.
Alphabet, Google's parent company, showed 21% growth in revenue for the second quarter year-over-year, hitting $21.5 billion. As always, advertising revenue drove the overwhelming majority ($21 billion) and generated the profits -- $7 million, and earnings per share of $8.42. The Thomson-Reuters earnings estimate was $8.04. The news sent Google's stock up by as much as 5% in after-hours trading.
That performance, which followed strong Facebook results, prompted this CNBC story on how Google and Facebook are basically raking in almost all the money being made in digital advertising -- though that could changed if Verizon has its way following the Yahoo acquisition this week. (See Verizon Sports Big Plans for Yahoo.)
Pichai was more interested in talking about Google Assistant and how advances in machine learning, voice and image recognition and natural language processing will enable users to get things done in a conversational way, across devices, in context. Google Assistant is being used in Allo, the new messaging app, and Google Home, its voice-activated home automation system.
But machine learning is being used across all the Google platforms, he said, pointing to the use of Google's DeepMind AI, a machine learning system now being used to drastically cut the cooling costs within the company's data centers, and the introduction of two new machine learning APIs for its cloud platform last week -- the Cloud Natural Language API and Cloud Speech API. And Pichai called out the Tensor Processing Unit, Google's custom chip that is being used to "supercharge" machine learning tasks.
"Machine learning is the engine that will drive our future, but it is already making our products better, and we use it every day," he said. "In fact, more than 100 teams are currently using machine learning, from StreetView to Gmail to voice search and more." The search process is already being improved with a ranking system based on Deep Learning, Google's native machine learning service.
On the mobile side, Google is gearing up to bring its Daydream virtual reality to the Android mobile platform, even as it continues to benefit from expanded mobile video viewing.
There wasn't too much discussion of the Other Bets, the part of Alphabet that includes Google Fiber. As a unit, Other Bets' revenue was up to $185 million, a 150% improvement, but its losses grew 30% to $859 million. A significant piece of that is capex being invested into the fiber unit, according to Google.
— Carol Wilson, Editor-at-Large, Light Reading