Nvidia CEO Jensen Huang says the industry will achieve a form of artificial general intelligence (AGI) within five years. He said that by then, AI will be “quite competitive” with humans and the world will see the emergence of “off-the-shelf AI systems” that companies will adopt and adapt to their needs, from chip design to software creation on drug discovery and radiology. But this prediction is controversial as critics say it only serves to sow ambiguity and fuel the AI bubble so that Nvidia can further increase its revenue through sales of AI chips.
The Nvidia CEO appeared at the New York Times’ DealBook conference on Wednesday and answered several questions from journalist Andrew Ross Sorkin about AI and AGI in particular. When asked if he thinks AGI will ever become a reality, Huang answered “yes.” However, he began with the AGI definition and suggested that the emergence of the technology will depend on how the world defines it. “Depending on how you define it, I think the answer is yes,” Huang replied. It should be noted that there is currently no real consensus on the definition of AGI.
The lack of consensus on this issue leads to a whole range of explanations and predictions. Sam Altman, CEO of OpenAI, was quick to say that his AI chatbot ChatGPT could have been considered AGI 10 years ago. He suggested that the chatbot was a victim of the AI effect. Google DeepMind recently decided to classify OpenAI’s ChatGPT and Google’s Bard as new forms of AGI. The company has published a framework that defines five levels of AGI. The Nvidia CEO’s thoughts on AGI come at a time when some business leaders are sounding the alarm about what they personally consider to be AGI.
At the conference, Huang announced his definition of AGI, which is: “Software or a computer capable of passing tests that reflect a basic intelligence that is fairly competitive with that of a normal human.” According to Huang, that stands Implementation of AGI imminent as the industry adopts this definition. “I would say that of course in the next five years there will be AIs that can pass these tests,” he said. But Huang adds that while AI is developing faster than expected, there is no sign yet that it can demonstrate or surpass complex human intelligence.
Huang says: There is no doubt that progress is high. But there are a lot of things we can’t do yet. The multi-level thinking that humans are very good at cannot be mastered by AI. However, some critics considered his view ambiguous. Others say it’s a new marketing pitch from the CEO of one of the best-known companies in the AI space. At the same time, some believe his view is more “moderate,” saying Huang is trying to reconcile reality with a marketing argument that could help him sell more.
I think Jensen Huang knows very well that current AI models are nothing more than advanced machine learning algorithms and have nothing to do with human intelligence. At the same time, he cannot speak out more clearly on this issue because he risks destroying the hopes of many people and shooting himself in the foot. That’s why he says, “We don’t have AGI yet. But it will be here in five years. So buy as many GPUs as you can and do your own thing so you don’t get left behind.” “He will say anything to sell more GPUs and increase his company’s stock price,” said one critic.
In fact, Nvidia is one of the companies benefiting the most from the AI race. Nvidia’s AI chips are so popular that sales tripled in the company’s third fiscal quarter. Nvidia’s business is booming due to increasing demand for its powerful GPUs, which are needed to train large language models (LLMs) and run heavy workloads in industries such as computing, automotive, architecture, electronics, engineering and scientific research. Meta, Microsoft, OpenAI and other tech giants are competing for access to Nvidia GPUs.
Nvidia’s fiscal third-quarter revenue tripled while net profit rose to $9.24 billion from $680 million a year earlier. In the interview, Nvidia’s CEO recalled delivering “the world’s first AI supercomputer OpenAI” after Elon Musk heard Huang speak about the device at a conference. Musk founded OpenAI in 2015 with Sam Altman and other investors before launching the company in 2018 amid management tensions. Elon saw it and said, “I want one.” He told me about OpenAI. “I delivered the world’s first AI supercomputer OpenAI that day,” says Huang.
Regarding the recent chaos at OpenAI, its board structure and the ouster and subsequent reinstatement of CEO Sam Altman, Huang said he hopes things will calm down. I’m glad things are resolved and I hope they will be. It’s a really great team. Huang. It also reminds us of the importance of corporate governance. Nvidia is here 30 years after its founding, we have been through a lot of adversity. If we hadn’t started our business right, who knows what would have happened? , he defined. Huang later declined to evaluate the successes of AI companies.
I won’t judge my friends. “I admit I want to, but I won’t,” Huang said. Additionally, he said that despite its shortcomings, AI is becoming increasingly indispensable in many, if not all, areas. “Software cannot be written without AI, chips cannot be designed without AI, nothing is possible,” he concluded about AI’s potential. Huang also said that the H-100 chips that Nvidia is shipping today were developed using a series of AIs. But without a clear definition, it is difficult to answer the question of when AGI will be a reality.
Finally, it is important to note that many people are extremely critical of what they see as the media hype surrounding AGI. This would mislead consumers. According to technology journalist Dana Blankenhorn, AGI is the lie of the year. He argues that the intense communication surrounding AGI is linked to the political goals of some major players in the AI sector and their attempts to fully control the development of the technology. According to the journalist, the aim of these players is to significantly increase the hurdles for new players to enter the market.
And you ?
What is your opinion on this topic?
What do you think of Nvidia CEO Jensen Huang’s predictions?
What do you think of the definition he gives to artificial general intelligence?
Do you, like him, have the impression that AI has become indispensable in all areas?
What is your forecast for AGI? Is this technology unattainable?
See also
Google DeepMind has just published a list of AGI levels and defined performance levels. It suggests principles such as focusing on capabilities rather than mechanisms
AGI is the lie of the year. “Artificial general intelligence does not exist,” says critic, who claims lies are being told for financial and political reasons
OpenAI CEO Sam Altman says ChatGPT could have been considered artificial general intelligence 10 years ago, suggesting the chatbot is a victim of the AI effect