The concept of the Metaverse gained prominence in our collective consciousness during the height of the pandemic when people yearned for better ways to connect with each other beyond video calls. The significant growth of gaming during the pandemic has also propelled it forward. However, the Metaverse has become so trendy that it now faces opposition, and people are not talking about it as much.
Nevertheless, the technologies that will power the Metaverse are advancing rapidly. One of those technologies is generative AI, which utilizes deep learning neural networks to generate creative conceptual art and other ideas based on simple text prompts. Jensen Huang, CEO of AI and graphics chip manufacturer Nvidia, believes that generative AI will be transformative and that it has only just begun. One of its biggest applications could be in the Metaverse, which places enormous demands on content as developers need to populate virtual worlds with 3D assets. Numerous companies like Stable Diffusion, Promethean AI, and Ludo AI are leveraging these technologies to automatically generate artworks and other assets for gaming and Metaverse applications. Nvidia is also conducting its own research in this field.
Many Metaverse companies are hopeful that generative AI will help provide the resources to build their worlds. Huang believes that as you input more and more prompts, such as text to flesh out a concept, the conceptual images will continue to improve. And he also believes that when it becomes reusable for various Omniverse applications, it will become evident that generative AI has reached a more mature stage.
GamesBeat: I wonder if there are people losing their enthusiasm for the Metaverse now. It was news from last year. Maybe they think it’s an idea that can die during a recession. What would you say about that?
Jensen Huang: First of all, the definition of the Metaverse, as it relates to the virtual environment in which we develop artificial intelligence software. I believe that Omniverse, the Metaverse, is essential to that journey. You can’t develop software that you can’t test. You can’t deploy software that you can’t test. With the world wanting to become software-defined, where there’s a software-defined car or a software-defined lawnmower or software-defined agricultural equipment or software-defined pick-and-place or a software-defined warehouse, we cannot transition to that new world without a virtual environment to simulate and test that software. That’s what Omniverse was built for in the first place. It seems logical to me that the next era of AI will need Omniverse!
