Nvidia CEO Jensen Huang Calls General Artificial Intelligence Can Be Realized In 5 Years
Jensen Huang, CEO of Nvidia, said that general artificial intelligence could materialize in the next five years. Huang, who leads the world's leading artificial intelligence chipmaker used to create systems like OpenAI's ChatGPT, is responding to questions in the economic forum held at Stanford University about how long it will take to achieve one of the long-standing goals held by Silicon Valley, namely creating computers that can think human-like.
Huang said that the answer really depends on how it is defined. If its definition is the ability to pass human tests, Huang says that generalized artificial intelligence (AGI) will soon materialize.
"If I give AI... every test you can imagine, you make a list of tests and place them in front of the computer science industry, and I think within five years, we will succeed in every test," Huang said on Friday, March 1.
Until now, AI can pass exams such as legal exams, but it is still difficult in special medical exams such as gastroenterology. But Huang said that within five years, AI should also be able to pass all these tests.
But according to another definition, Huang said, AGI may still be far away because scientists still disagree on how the human mind works.
SEE ALSO:
"Therefore, it is difficult to achieve as an engineer, as engineers need a specified goal," Huang said.
Huang also answered the question of how many chip factories are needed to support the expansion of the artificial intelligence industry. Media reports state that OpenAI CEO Sam Altman thinks more factories are needed.
Huang said that more factories would be needed, but each chip would also get better over time, which limits the number of chips needed.
"We will need more factories. However, remember that we also continue to improve (AI) algorithms and processing tremendously from time to time," Huang said. "Unlike the current computational efficiency, and therefore the demand is as much as this. I increased computing a million times over the 10 years."