Jensen Huang: AI Computing Power to Increase by 1 Million Times in the Next 10 Years

  In two public videos, Jensen Huang detailed how to deal with the ups and downs experienced by NVIDIA, organizational structure management, the development of AGI, the potential growth of AI computing power, and future challenges, among other topics.

 Huang candidly stated that NVIDIA's success lies in the fact that for the past 30 years, almost all of NVIDIA's work has revolved around technology and the market, which is also NVIDIA's core: creating technology before creating demand.

 Discussing NVIDIA's flat management, Huang mentioned the importance of a culture that does not believe "information = power," highlighting the significance of trust and empowerment for employees. NVIDIA has only 30,000 employees, making it one of the smallest large companies in the world, but each employee is endowed with significant power. 

As the top "shovel seller" in the era of generative AI, NVIDIA's market value has now surpassed 2 trillion US dollars. Regarding AI trends, Huang stated that AI will be the "defining technology" of this era, and in the future, AI will become a country's "sovereignty," meaning a country must control its own digital intelligence and must be committed to controlling sovereign AI. The latest breakthrough in AI is deep learning; another important breakthrough is Reinforcement Learning from Human Feedback (RLHF) technology, newly invented for language models, and NVIDIA has already found a system-level solution to implement this technology. 

Huang also believes AI technology has narrowed the technological gap among humans, and the future of information processing will fundamentally change with AI, with generative AI evolving from an information "seed," and the future of computing relying heavily on generation rather than retrieval. Moreover, within the next five years, AGI will pass human tests, including bar exams, gastroenterology, and other professional tests, meaning AGI will arrive soon. However, he also acknowledges that from an engineer's perspective, achieving AGI is challenging because engineers need clear goals, and AGI may take some time since there is still disagreement on how human thought operates. 

Regarding the development of AI computing power, Huang believes that in the next ten years, NVIDIA will increase the computing capability of deep learning by 1 million times, thereby enabling AI computing devices to continuously train, infer, learn, apply, and improve, constantly transforming super AI into reality.