TAIPEI (Taiwan News) – Nvidia CEO Jensen Huang (黃仁勳) delivered the commencement speech at the California Institute of Technology (Caltech) on Friday (June 14), encouraging graduates to work through adversity and continually seek new opportunities. 

Huang said Caltech is known for esteemed graduates such as Nobel Prize winner in physics Richard Feynman, Nobel Prize winner in chemistry Linus Pauling, and semiconductor pioneer Carver Mead, who is credited with coining the term “Moore's Law.” Huang encouraged graduates to engage in AI, which he described as one of the fastest-growing fields of modern computing.

Huang added that Nvidia’s two chief scientists are both from Caltech and reminded graduates that the company continues to recruit new engineers. He joked that Nvidia is a great company to work for and is run by a great boss.

Huang also gave life advice to new graduates, encouraging them to savor their time with their parents. He said his children did not move out of the family home until later in life, and upon reflection, he cherished all of the time he had with his children.

Huang said he was honored to deliver a similar commencement speech at National Taiwan University last year, sharing Nvidia’s journey and the lessons learned along the way. He noted the company’s setbacks as well as its successes, admitting that mistakes are necessary if one wants to pioneer new fields of technology.

He said the early days of the modern computing industry were based upon IBM System 360. Huang said basic computer architecture, ideas, and strategy still come from this computing system, even though it was introduced a year after Huang’s birth.

After the IBM System 360, Huang said the single most important computing innovation was the CPU which fueled the exponential growth of the computing industry. Huang said CPUs led to the development of the software industry, which has become a multi-trillion-dollar industry.

Huang believed society is now in an era of accelerated computing, with computing duties offloaded to GPUs working in parallel. Huang says computing has reached a tipping point, ushering in deep learning and other advancements.

Huang said the modern GPU had to be reinvented at every layer, including the system connections and software. Thousands of engineers had to work on related projects for over a decade to enable deep learning, which cost Nvidia billions of dollars in investment.

“In 2016, we announced DGX-I, our AI supercomputer, which I delivered to a startup that no one knew about, OpenAI. In 2022, ten years later, and with a million-fold increase in computing, the company launched ChatGPT which is well known to many now," said Huang.

He said Nvidia has transformed from a graphics company that builds GPUs to an AI company that makes massive data center-sized supercomputers. The computing stack uses GPUs to process large language models trained by supercomputers instead of CPUs operating on instructions.

In the future, Huang believes computers will reason, plan, and execute a mission. He sees the next wave of AI involving robotics, a physical manifestation of AI that will include small robots as well as whole warehouses.

Huang described a series of setbacks the company encountered when working with industry heavyweights such as AMD, Intel, Qualcomm, and others. After each hurdle, the company transformed, and sought new opportunities such as the production of the first robotics computer.

Ten years later, Huang says he cannot be happier about the company's progress. Ironically, the robot computer doesn't even have graphics function which was one of the company's core strengths.

The future is uncertain and may deal bad cards, he said, encouraging young graduates to “swiftly shake it off.”