Nvidia executive chairman Jensen Huang stated on Friday that, According to some reports, artificial general intelligence may emerge in as short as five years.
At a Stanford University financial forum, Huang—the head of the world’s leading producer of AI chips used to power systems like OpenAI’s ChatGPT—was answering a question regarding the duration it might require to realize one of Silicon Valley’s long-standing ambitions of building computers with human-like intelligence.
According to Huang, a lot depends on the way the aim is communicated. Artificial general intelligence (AGI) is going to be here soon, according to Huang, if it truly can survive human testing.
In five years, I’m guessing that we’ll perform exceptionally on every single exam if I handed an AI every test you can think of, Huang said, whose company saw its market valuation reach $2 trillion on Friday. “You create a checklist of tests and present it in front of the information science industry,” Huang added.
AI is currently able to pass exams like the law enforcement bar exam, but it still has difficulty with specialty medical exams like gastroenterology. However, Huang claimed it expected to be prepared to clear any one of them in five years.
However, according to Huang, different definitions could put AGI considerably further off since researchers still need to agree on a common understanding of how minds function.
Because engineers need clearly defined goals, it is therefore difficult to succeed as an engineer, explains Huang.
The subject of the number of additional chip factories referred to as “fabs” in the industry required to accommodate the growth of the AI sector was further discussed by Huang. Based on media sources, OpenAI CEO Sam Altman believes a significant number of additional fabs are required.
Although Huang stated that further development will be required, the quantity of chips required is going to be limited because each chip will improve with time.
He continued by saying that More fabrications would be required. But keep in mind that over time, we’re also making significant advancements in AI processing and algorithms. The current level of computing efficiency is different from what drives this level of demand. I will have improved computing a million times in ten years.
