Google I/O

Google is muscling up its cloud platform for what it sees as an "AI first" world. 

This week at the Google I/O event in Mountain View, Calif. the company revealed the second iteration of its Tensor Processing Units (TPUs), which hold the promise of accelerated machine learning and artificial intelligence capabilities.

Going All In With Artificial Intelligence

Amidst a series of announcements about artificial intelligence, the Google Assistant, Android and many of the company’s other consumer and enterprise initiatives, Google CEO Sundar Pichai said Cloud Platform customers would be the primary beneficiary of the new hardware.

“The reason we named it cloud TPU is we're bringing it through to the Google Cloud Platform,” he said.

Google has developed a new chip that is a follow-up to the initial processor it built for its own AI services. However, this version is specifically engineered for the processing needs demanded by such high levels of machine learning. It can also be used through a dedicated cloud service for businesses and research organizations that desire such computational analysis.

The aim for Google is to both supercharge its cloud services with AI, making its Cloud Platform more appealing. Additionally, there is competition from Nvidia, which currently dominates this chip market. Intel and another batch of startups are also rumored to be getting into this space, with such cognitively aware computing considered to be the next wave.

In making the announcement, Google senior fellow Jeff Dean and senior vice president of cloud infrastructure Urs Hölzle said the machine learning model needs to be democratized for potential breakthroughs from a community of research.

“Research and engineering teams at Google and elsewhere have made great progress scaling machine learning training using readily-available hardware. However, this wasn’t enough to meet our machine learning needs, so we designed an entirely new machine learning system to eliminate bottlenecks and maximize overall performance. At the heart of this system is the second-generation TPU we're announcing today, which can both train and run machine learning models,” they said.

Cloud Jobs API: A New Job Search Tool

In many ways Google is probably thought of as a job killer, as its services and machine learning capabilities further make humans obsolete. But CEO Pichai believes the reverse can be true, with his discussion of its Cloud Jobs API, which brings some of the same precise and filtered search tools to bear on finding employment.

“We also think about how this technology can have an impact on people’s lives by creating opportunity for economic empowerment,” he said.

He said 45 percent of US employers have talent shortages and problems filling open job positions. Any complex search issue is precisely the type of problem Google is prepared to solve, with the added bonus of it now enters another market.

The job search tool was originally announced in November 2016 and has been in private alpha since. Companies that want to explore it need to sign up at the official site

It was unclear if this would grow into a full-blown competitor to equivalent services like Indeed, Monster or LinkedIn. Pichai said they were partners as part of the initial rollout, although as often happens with new Google forays this could morph into something larger over time.