The giant artificial intelligence systems at Google are powered by special computer chip developed by the technicians of the company a few years ago. the chip is able to take care of large and complex processes deemed to be the key for the future of artificial intelligence.
While that was a closed thing for Google and its affiliates, the company has now opened its doors for other organizations to make use of the special chip. Companies would not be able to use the chips via the cloud computing service of the internet giant. The aim of Google is to develop a new business built around these chips which it has termed as tensor processing units, or T.P.U.s.
“We are trying to reach as many people as we can as quickly as we can,” said Zak Stone, who works alongside the small team of Google engineers that designs these chips.
Google can be considered to be not only a company that produces software butt also one that now is able to manufacture hardware as well. The company develops and builds its own hardware ranging from computer servers to networking hardware at its huge facilities. These hardware is put to use by the company to enhance efficiency of its data centers and to reduce overall costs to the company.
Google has also developed one A.I. chip which it embeds into its smartphones in addition to the T.P.U. chips.
Computer vision technology where computers are taught to recognize things using artificial intelligence if the current focus of the company. However, according to Mr Stone, the new chips would also help the company to develop a range of other services that can be monetized, in the near future.
Google’s new chips were started to be used by Lyft for enhancing its research on driverless cars during the end of last year.
Identification and saying of street signs or pedestrians by a driverless car were the ultimate aims of Lyft using the chips of Google. It is hoped that the driverless systems of cars can be trained on these aspects in hours instead of days.
“There is huge potential here,” said Anantha Kancherla, who oversees software for the Lyft driverless car project.
The chips have also helped Google to become more self-reliant and reduce dependence on chip manufacturers like Nvidia and Intel.
“Google has become so big, it makes sense to invest in chips,” said Fred Weber, who spent a decade as the chief technology officer at the chip maker AMD. “That gives them leverage. They can cut out the middleman.”
While developing special chips can alter the market dynamics of the global chip market but Google would still be using chips from the likes of Nvidia. “Who’s buying and who’s selling has changed,” Mr. Weber said.
(Adapted from NYTimes.com)