Reuters reported on March 2 that at Stanford University's (USA) economic forum on artificial intelligence, Nvidia CEO Jensen Huang made some striking predictions about the prospect of creating a supercomputer system capable of thinking like a human.

According to Mr. Jensen Huang, AGI systems will pass all the tests in the academic exams of humans in the next 5 years. Currently, AI can complete the bar exam but not meet the requirements when testing for medical expertise. However, Mr. Huang believes that AGI will soon overcome this challenge.

AGI is a type of super-intelligent artificial intelligence that can perform many tasks as well or better than humans. Notably, this technology has the ability to self-improve, creating an infinite feedback loop of information.

However, Nvidia's CEO said scientists are currently unable to agree on a definition of AGI because of disagreements in describing how the human mind works, which could slow down the process of building supercomputers.

As for plans to expand the AI ​​sector, Mr. Jensen Huang assessed that more chip foundries are needed. However, the billionaire emphasized that the quality of chips will improve over time, helping to significantly reduce production.

“We will need more foundries. But remember that future chips will dramatically improve the speed and processing power of AI. I think computing power will improve 1 million times in the next 10 years,” said Mr. Jensen Huang.

Nvidia is now the world’s most valuable chipmaker, with a market capitalization of about $2.06 trillion. Nvidia shares are up 66% by 2024, currently at $822.79.