Facebook is developing a machine learning engine for tasks such as recommending content to users. This was reported by The Information, citing two people familiar with the project. According to the post, the company has already developed another video transcoding chip to improve the user experience of watching videos on Facebook apps.
Facebook is acting like other big tech companies, including Apple, Amazon.com, and Google, which are increasingly ditching off-the-shelf chips in favor of in-house designs. This allows for optimal solutions in terms of cost and performance.
Back in 2019, the Facebook blog posted a message that the company was developing custom chips specifically designed for AI inference processing and video transcoding to improve the performance and energy efficiency of its infrastructure, which at the time served 2.7 billion people across all platforms. Facebook.
The company then clarified that it plans to partner with industry representatives such as Qualcomm, Intel, and Marvell Technology since existing general-purpose processors are not suitable for the workload faced by Facebook systems. However, The Information said that Facebook is currently developing chips on its own.
The Facebook spokesman said, Facebook is constantly looking for ways to improve computing performance and energy efficiency with the help of our partners and through its own internal efforts.
0 Comments