At this year’s Siggraph event Nvidia’s Jen-Hsun Huang Meets with Wired for an hour-long chat about all things Nvidia, RTX, and AI. Among the topics covered, including the recognition that AI training and inference are huge power demands, was Huang’s claim that more computers will be needed for AI systems in the future — three times as many, to be exact.
Siggraph is an annual conference that usually covers computer graphics and interactivity-related tech (think AR and VR, that sort of thing), but it was only a matter of time before AI became a major topic of discussion. To that end, the Nvidia CEO was interviewed by Wired’s Lauren Goode for an hour-long stream that covered GPUs, RTX, and ray tracing, but mostly AI.
If you’ve been following Nvidia’s push to make generative AI everywhere, there’s nothing in this discussion that will interest you. However, at one pointHuang mentioned that the world of AI is now moving away from its pioneering phase and towards another one, which the Nvidia CEO called the “enterprise wave.”
Then comes the “physical wave,” which Huang said is “really, really extraordinary.” He explained the claim by saying that three computers would be needed: one to create the AI, another to simulate and refine the AI, and finally a third to run the AI itself.
“It’s the three-computer problem. You know, the three-body problem, and it’s so incredibly complicated, and we created three computers to do it.”
Jen-Hsun is of course referring to Nvidia’s entire range of hardware and software packages, from its DGX H100 Servers to create artificial intelligence, Jetsen Embedded Computers to simulate artificial intelligence and then workstations and servers using Omniverse and RTX GPUs to support artificial intelligence.
Siggraph is one of my favorite tech events, and I’ve been watching the presentations and reading the research papers presented at the conference for years. I have to say it’s a bit of a shame that a rather garish advert for Nvidia’s AI sales systems took center stage this year —Huang’s Conversation with Meta’s Mark Zuckerberg was another example of promoting AI without any substance — and while there will still be a lot of discussion about computer graphics, and AI will naturally be a part of it, Huang didn’t say anything that made me think, “Wow, this is going to be so cool!”
Are we really going to? need three computers? PC gamers certainly don’t, and neither do most businesses. Even those who really want to integrate AI into their core operations may balk at the potential costs and complexities of using and paying for three tiers of Nvidia products.
Nvidia is clearly 100% focused on AI now. The days of it being just a graphics/gaming company are long gone, even though that was a core part of the business, as it transformed into a data company. That’s not to say that PC gamers won’t benefit from Nvidia’s technological advances in AI, as things like RTX and DLSS were arguably the substantial leaps forward in the rendering world.
And I certainly wouldn’t expect the CEO of the world’s most successful artificial intelligence company NO utilize it for every business opportunity we can, but I think we could all utilize a break from the relentless push for AI to take over every aspect of our computerized lives—it gets a little tiring after a while.
