Google: Unification Of Software And Hardware Will Enable ‘Always On’ AI Computing

Google says artificial intelligence (AI) will enable the era of ‘always on’ computing and that it will combine hardware and software development to take full advantage of the development.

Sundar Pichai, Google CEO, said AI would be the next stage in evolution of computing, following on from the web in the 1990s and the smartphone a decade later.

Central to this vision is the Google Assistant, the company’s machine learning powered personal assistant, which will ship with the newly announced Google Pixel smartphone.

Indeed, the Pixel is used as a prime example of this marriage of hardware and software development and is the first device to be manufactured by Google “inside out”.

A shift in IT

“We’re at a seminal moment for computing,” Pichai declared. “IT has always shifted every ten years or so. The web was arguably the most fundamental platform shift in our lifetimes. About 10 years later or so in the mid-2000s we had the mobile revolution with the smartphone.

“That brought computing to around half the world’s population and the shift continues. We are evolving from a mobile first world to an AI first world.”

This AI drive will see the creation of a “personal Google for every user”, with natural language and machine language translation helping deliver this vision.

Google Assistant will be incorporated into a number of devices and applications other than the Pixel with the idea of encompassing the phone and the home; essentially everywhere the user would be.

Always on computing

“Why should we build hardware?” asked Rick Osterloh, head of Google’s hardware unit. “The volume and complexity of information makes this the right time to work on both hardware and software.

“Building them together allows us to take advantage of the capabilities like Google Assistant. “We can deliver the simple, smart and fast experiences our users expect from us.”

The platform is in the recently launched Google Allo video chat application and is a focal point of Google Home, a new appliance that lets users use Google Assistant hands free around the home.

Google Home is designed to mimic the aesthetics of other household objects, such as candles or wine glasses and features interchangeable bases. Users simply say a command, such as ‘play music’, and the device uses machine learning to determine the correct noise levels and suggest ideal playlists for certain activities or days. It will even learn your preferred streaming service, such as Spotify.

Google Home will also return queries from the Google search engine and if it can’t find it on the Knowledge Graph, it will return a snippet from the web. Information from Google Maps, such as traffic, can also be determined, and the aforementioned natural language and text-to-speech functionality enables better translation. Needless to say it can also control smart home devices around the house.

But it’s not just Google. Assistant will be opened up to developers with ‘Actions for Google’ and an SDK will be released in the future so third party manufacturers can integrate the technology into their own smartphones and tablets.

“It’s early days, but the assistant continuously learns from, this and keeps getting better,” added Pichai. “Our vision for assistant is for it to be everywhere.”

Google is not alone in seeing the potential for AI. IBM, Facebook, Amazon and Microsoft are also investigating future applications and all five companies have formed an AI alliance. The notable exception is Apple, which is also looking into AI with apps such as Siri.

Quiz: What do you know about Android?