At CES 2025, Jensen Huang reappeared on stage to answer questions from journalists. The head of NVIDIA particularly discussed the potential role of artificial intelligence in games.
Journalists, as well as commentators from ITC.ua, are interested in how far AI assistance in games can go. Can artificial intelligence completely create them? Huang answered no.
NVIDIA introduced DLSS 4 technology, which is capable of creating three additional frames from one rendered frame. Considering also the scaling of resolution, there arises a situation where the video card generates 15 out of every 16 frames. Previously, journalists asked AMD if a video card could ever create 100% of the pixels
“Can I tell you that in the future every pixel will be created by machine learning? Yes, definitely. This will be in the future,” replied Frank Azor, the chief architect of gaming solutions at AMD.
The next day, Jensen Huang disagreed with this.
“No. Remember when ChatGPT first appeared, we said, ‘Oh, now let’s just create a book.’ But now no one expects this anymore. You need to provide something — it’s called a condition. We condition the chat with contextual prompts. Before you can answer a question, you must understand the process. The context could be a PDF, the context could be a web search. The context could have been such that you would precisely say what context is, right? The same happens in games. You have to provide context. The context for video games has to be not only relevant in terms of the narrative but also spatial, world-related, spatially relevant. The way you set conditions, how you provide this context, involves providing some initial fragments of geometry or textures — it can generate, it can enhance the rest. … In the future, 3D graphics will generate justified conditions,” — explained the head of NVIDIA.
With DLSS 4, the new NVIDIA video cards only reproduce one of the four prospective frames. The main thing, says Huang, is that they have to be reproduced, and under these conditions, others can be created.
“The same will happen in video games in the future … not just with the pixels that we visualize, but also with the geometry we output, the animation we output, the hair we will recreate in future video games.”
Huang apologized if his explanations might have seemed unsuccessful to anyone, and said that there is still and always will be a role for artists in video games.
“It took so long for everyone to realize that generative AI is indeed the future, but you need to condition, you need to create a foundation with authors, artists, intentions”.
Source: PC World
Spelling error report
The following text will be sent to our editors: