During Nvidia’s earnings call on Wednesday, chief executive Jensen Huang stated that the company had “achieved AGI”, referring to one of the technology industry’s most sought-after objectives. Moments later, he described the milestone itself as “senseless”, underscoring the lack of agreement over what artificial general intelligence actually means or how anyone would recognise it.
Asked about OpenAI’s pursuit of the goal, Huang said that where Nvidia is concerned, “for many tasks, we could say that we’ve already achieved AGI.” He offered no precise definition or benchmark, adding that such milestones are “kind of senseless at this point.” He pointed to AI moving beyond responding to simple prompts and towards autonomous agents capable of learning new skills and improving themselves “recursively.”
What matters, according to Huang, is that AI is “doing productive and useful work” and “generating profitable tokens”, with greater compute producing more tokens and, in turn, more profit. “This is the exact phase where we’re at. Which is the reason why everybody’s leaning in,” he said.
Not the First Time Huang Has Made the Claim
This is not the first occasion on which Huang has suggested that AGI has been reached. In March, during an appearance on the Lex Fridman podcast, he stated plainly, “I think we’ve achieved AGI,” without specifying what he meant by the term. Fridman proposed his own definition: an AI system able to “essentially do your job”, such as starting, growing and running a successful technology company worth more than £790 million. Rolling back his earlier remark, Huang said that “the odds of 100,000 of those agents building Nvidia is zero percent.”
Over the years, various technology leaders have taken advantage of the term’s vagueness, producing a wide range of definitions and benchmarks. All revolve around the same loosely defined concept: AI capable of matching or surpassing human intelligence across a broad range of domains, even though “intelligence” itself lacks a universally agreed definition.
OpenAI and Rivals Offer Competing Definitions
The definition adopted by OpenAI, a company founded expressly to build AGI, leaves considerable room for interpretation. Its charter describes AGI as “highly autonomous systems that outperform humans at most economically valuable work.” Chief executive Sam Altman has acknowledged that this is hardly a measurable standard, conceding last year that AGI is “not a super useful term.”
Adding further complexity is a separate, financially driven definition worked out between OpenAI and Microsoft, reportedly covering systems capable of generating at least £79 billion in profits. Chief research officer Mark Chen has estimated that OpenAI is “80% of the way” to AGI, while Altman has said the company would have something he would call AGI by the end of the year.
Anthropic chief executive Dario Amodei has called AGI “imprecise” and even a “marketing term”, preferring instead to speak of “powerful AI.” Others have reached for their own terminology to describe broadly similar ideas, reflecting the continued absence of any shared threshold for the concept.
Source
Image: theverge.com
