Nvidia CEO Jensen Huang tipped his hat to OpenAI on Sunday.
The chipmaking mogul took to X to congratulate OpenAI’s workforce on the discharge of the corporate’s newest and most powerful model, Astra, and made the notable assertion that “AGI has arrived,” a nebulous term that usually describes AI fashions that match or surpass human intelligence.
He additionally identified that Astra was educated on Nvidia’s chips.
“From ChatGPT to o1 to Astra in 4 years,” Huang stated in a publish on X. “AGI has arrived. Congratulations @OpenAI workforce.”
OpenAI unveiled Astra on Thursday. The ChatGPT maker known as it the world’s “most clever and aligned mannequin,” and stated it’s able to performing “essentially the most demanding skilled work with unmatched velocity, accuracy, and judgment.”
Greg Brockman, the corporate’s president, stated on a name with reporters on Thursday: “Welcome to the AGI period.”
AGI stands for artificial general intelligence, a fuzzy idea that nonetheless has grow to be a core objective for main frontier AI labs.
OpenAI defines it as “extremely autonomous techniques that outperform people at most economically beneficial work.” The corporate pitched Astra as a significant analysis breakthrough and a shift in what’s attainable to delegate to AI. It is rolling out to clients this week.
On the Thursday name, Brockman stated that sooner or later, he thinks individuals will look again and suppose AGI was created “about this time, and I feel it is likely to be about this mannequin.”
“For me personally, I do suppose we’re there,” he stated.
Not everybody agrees.
Gary Marcus, a outstanding AI researcher and critic, stated Huang had jumped the gun.
“Huang gave no proof and no definitions, which feels to me like an effort at a takeover of a scientific query by company fiat,” Marcus wrote on his Substack Sunday night. “Declaring victory with out a definition merely muddies the waters.”
Marcus hooked up his personal 10-point definition of AGI, noting that Astra solely meets one or two of these benchmarks.
“By typical definitions, Astra nonetheless falls brief,” he stated.
Altman, for his half, advised the “Sources” podcast in a latest look that AGI is, at finest, “a really poorly outlined time period. I used to be going to say it is like an irrelevant advertising time period.”
Frontier AI corporations like OpenAI, Meta, Anthropic, and Google have grown increasingly reliant on superior chips manufactured by Nvidia. In a funding announcement in March, OpenAI known as Nvidia “the inspiration of our infrastructure.”
“Our coaching fleet and nearly all of our inference stack proceed to run on Nvidia GPUs,” the corporate stated.
That demand has translated into monumental gross sales for Nvidia. The corporate reported $96.2 billion in quarterly income in August, greater than double the quantity reported in the identical interval a yr earlier. Its information middle enterprise, which includes its AI chips, generated $89 billion in income.
And in his X publish on Sunday, Huang prompt there’s a lot extra computing energy on the way in which.
“400K GPUs coming on-line subsequent,” he stated, referring to graphics processing models, the superior chips used to coach and run AI fashions.
