The Facebook AI has achieved some tremendous feats: it can read stories, answer questions, play games and it can be taught things. For the latest venture, the Facebook AI is getting some hardware.
Facebook AI Hardware Can Train Neural Networks
The technology is so ahead of other AIs that Facebook even open-sourced the data for the design of the hardware, so other companies can use it as a base code. Facebook teamed up with Nvidia at a large scale to work on the project.
The hardware, nicknamed Big Sur, has eight high-performance GPU boards of up to 33 watts each, making it twice as fast as their previous AI. Not only is it faster, it’s also more versatile and thermal power efficient.
Facebook Wants to Be A Leader in Artificial Intelligence
The goal of this extensive research and development is to make it easier for AI developers all over the world to improve their technology. By putting themselves way ahead of the game, Facebook is allowing other smaller ventures to take their own steps into artificial intelligence. In a rather prideful move, it’s made their blueprint available to huge corporations and small ones alike in order to make the technology accessible to everyone.
Facebook hopes that other companies will join their own R&D team for the Facebook AI, and it’s increasing their investment on GPU hardware by more than three times. They’re encouraging other teams in within Facebook to boost neural network usage onto products and services. All other AIs are underdeveloped in comparison, even though Facebook is only a social media platform. Other giants like Microsoft and Google have very solid AIs; Google’s is even capable of driving a car, but Facebook is painting a much bigger picture. Not only is Facebook involved in virtual reality, but it’s also taking huge step towards the development of artificial intelligence.