TECHNOLOGY

Like Internet On Drugs: Meta Wants To Launch The World’s Fastest Supercomputer

Facebook’s parent company Meta plans to launch the world’s fastest supercomputer this year. It should then be used primarily for artificial intelligence applications. That’s kind of like betting the Internet on high-speed drugs.

Facebook’s parent company Meta has big plans for the future. This has become clear at least since the introduction of the Metaverse. Company boss Mark Zuckerberg envisions a virtual world of beach experiences, online tennis with stars, and endless video games in the living room.

Of course, Meta needs decent computer power for all of this. Now the company has shown what it has up its sleeve. Meta has now introduced its super research centre, the Research SuperCluster (RSC ). A supercomputer is already in use here. By July 2022, Meta wants to turn it into the world’s fastest supercomputer.

Meta-Supercomputer: A Trillion Operations Per Second

Therefore, the RSC will focus primarily on the development of applications with artificial intelligence (AI). Experts want to develop AI models that can learn from trillions of examples, operate in hundreds of languages, and create new augmented reality tools. All of this should ultimately benefit the Metaverse in particular.

Translated, this means: The meta-supercomputer is already ensuring that computer representations are extremely fast and that AI programs can be trained in the shortest possible time. This makes the meta-supercomputer one of the fastest supercomputers in the world.

This is precisely what the technology group would like to go a step further. By July 2022, the capacity is expected to almost triple to 16,000 GPUs. The computer should then carry out up to a trillion operations per second.

Why Does Meta Need A Supercomputer?

Complex computer applications, such as augmented reality or artificial intelligence, recognize hate speech online in seconds and require computer power. This not only applies to use but also to programming when you teach things to computers.

Supercomputers can provide just that. At Meta, it currently takes around nine weeks to train a model. With the expanded supercomputer, which will probably be the fastest globally, it should only take three weeks.

This would also allow for more actual and smoother experiences in the virtual world. This would be like the Internet on high-speed drugs, from streaming to video games to augmented reality.

At Meta, this performance should be used for applications in the Metaverse and protect users, says the company.

We can train models faster to determine if an action, sound, or image is harmful or harmless. This research will help protect the people who use our services today, and in the future, we will build for our Metaverse. That’s kind of like betting the Internet on high-speed drugs.

Translated, this means: The meta-supercomputer is already ensuring that computer representations are extremely fast and that AI programs can be trained in the shortest possible time. This makes the meta-supercomputer one of the fastest supercomputers in the world. This would also allow for more actual and smoother experiences in the virtual world. This would be like the Internet on high-speed drugs, from streaming to video games to augmented reality.

techbuzzfeeds

Recent Posts

Virtual Tour: What It Is And How To Do It

The virtual tour has become an exciting reality for small and medium-sized businesses. Until a…

1 day ago

Advertising Makes Its Story On Snapchat

At barely 26 years old, Evan Spiegel, the young prodigy founder of Snapchat, decided to…

4 weeks ago

Chatbots And Voice Bots: How Artificial Intelligence Is Transforming Call Centers

They answer customer calls with a voice that sounds human, giving contemplated data and not…

1 month ago

Serverless Reaches Technological Maturity

Serverless, this other way of designing and running applications in the Cloud, has gained maturity…

2 months ago

Artificial Intelligence’s Impact On Medicine And Psychology

Is Artificial Intelligence Facing An Epochal Revolution? Artificial intelligence can analyze a vast amount of…

2 months ago

Google Analytics 4: With Machine Learning, An Even More Granular And Dynamic Map Of Users’ Online Behavior

The transition from a logic of "observation" of behavior to one of evaluation of the…

2 months ago