Facebook brews the Google engineer to help design his own chips

Facebook hired one of Google's chip developers, Shahriar Rabii, to help the social network in its ongoing efforts to design its own silicon, according to a report by Bloomberg . Facebook jumped on the chip development train earlier this year, when they started building a team that could design custom chips to power the server and consumer hardware. Rabii's new role on Facebook will be as vice president and head of silicon, according to an update from Linkedin bio.

This is a move that is moving with other tech giants, many of whom bring in-chip design in-house rather than relying on renowned vendors like Intel and Qualcomm. Apple has been creating its own processors for iOS devices for nearly a decade and has designed custom single use chips for artificial intelligence and other tasks in recent years. The iPhone maker also plans to use its own chips to replace Intel processors for their Mac computers by 2020. Earlier this year, Amazon reportedly launched a new initiative to design its own chips, specifically to help the AI ​​features of his Echo. line of smart speakers.

Google produces its own custom Visual Core chips for Pixel smartphones – and Rabii, recently hired by Facebook, had previously led the team that developed them. In its new role, Rabii is not likely to develop chips for Facebook brand smartphones, but the company is working on several types of hardware that could use a custom processor.

Earlier this year, Oculus VR, owned by Facebook, launched the autonomous Oculus Go virtual reality headset that is currently based on a Qualcomm brand chip. Future models can use custom Facebook chips instead. The company is also developing its own series of Echo Show smart speakers with AI features, and a custom chip can give Facebook a competitive edge at home.

Custom chips could also be used to better train the AI ​​algorithms that Facebook has patrolled on its site for hate speech, fake accounts and potentially dangerous content. At present, the company uses third-party modified GPUs from companies like Nvidia. Designing its own AI workout servers with proprietary chips, like Google with its Tensor processing units, could help with the very tricky problem of using AI instead of L & # x39; human eye to control its ever-growing platform.