What is believed to become the next hardware revolution currently resides in one of the company’s many data centres. One of Facebook’s most promising and prized possessions has opened new possibilities in the fields of Artificial Intelligence and Machine Learning.

This new horse in the ever growing race of AI mastery has been named as Big Sur. It is a hardware system that has been put together to “train” software to improve itself over the course of its time. It has the ability to explore networks and absorb the crucial information on the road to self-improvement. Although machine learning has been there for several years, what most companies lacked was a good resource. Now with over 1.6 Billion users, Facebook has the opportunity to make a giant leap in this field thanks to the massive data set available to its new engine. According to Facebook:

We've developed software that can read stories, answer questions about scenes, play games and even learn unspecified tasks through observing some examples.

Facebook has put a lot of resources and dedication into this project making Big Sur far more versatile and efficient. Each Big Sur unit consists of a 4U chassis that houses eight NVIDIA M40 GPUs and two CPUs along with an SSD storage and changeable fan. In collaboration with NVIDIA, Facebook developed its own hardware: customised to the needs of the venture.

Ultra-fast neural networking is what is being achieved here. Neural networks allow the computers to adapt and learn tasks and methods without actually programming for them. Similar to how the human brain works. What previous projects took three months to learn, Big Sur can successfully do it within the course of a single day.

Open Sourcing and Future

Facebook decided to make the design material for its million dollar project available to the public stating that

" We’re not in the business of having secret thing"

The commitment shown by Facebook to one of its core pillars shows the importance it has laid. Initially, Bug Sur was designed to recognise pictures and mimic them but with ever growing prospects and needs that follow, it has grown to be a huge step for all researchers and people working on the same idea.

We want to make it a lot easier for AI researchers to share techniques and technologies. As with all hardware systems that are released into the open, it's our hope that others will be able to work with us to improve it.