Andrew Ng's Scientist Story
Andrew Ng, a revered figure in the field of Artificial Intelligence (AI), was born in London and grew up in Hong Kong. As a Chinese-American, he bears a distinctly Chinese name—Ng. He studied at Carnegie Mellon University (CMU) and the Massachusetts Institute of Technology (MIT) in the United States before teaching at Stanford University. He left a significant mark at all three of these top-tier universities.
Andrew Ng's AI Journey
Approximately 10 to 15 years ago, Ng and his friends discovered a way to employ hundreds of engineers to develop AI software, which was then applied to hundreds of millions, even billions, of users. This sounded immensely valuable, but they realized that outside the internet industry, it was challenging to find such large-scale applications for AI systems.
The Three Waves of Artificial Intelligence
First Wave of AI (1950-1960): Symbolic Logic
Aimed to teach computers the "logical thinking" of humans. This wave ultimately failed because humans could not fully articulate their own thought processes.
Second Wave of AI (1980-1990): Expert Systems
Attempted to encode all human knowledge into computers. This wave also failed as humans couldn't solve every problem and translate them into rules.
Third Wave of AI (2010-Present): Machine Learning
Focused on enabling computers to understand what humans see. This wave is still evolving.
The One Algorithm Hypothesis
The "One Algorithm Hypothesis" posits that different regions (modules) of the brain share the same set of programs (algorithms). The functional differentiation is due to variations in input data. This implies that we don't need multiple algorithms for different functions but should seek a universal algorithm capable of handling various tasks.
Impact in the Tech Industry
Between 2011 and 2012, Ng was the founding leader of the Google Brain (Deep Learning) project. From 2014 to 2017, he served as the Chief Scientist at Baidu. Starting April 9, 2024, he became a board member of Amazon.
The Generative AI Revolution
Ng points out that the advent of generative AI has sparked excitement and anticipation. Fifteen years ago, small AI models showed limited improvement even with large amounts of data. Hence, during his time with the Google Brain team, the goal was to build large-scale neural networks using Google's computing power—continually feeding data to improve the models.
Two Key Factors in Unleashing AI Capabilities:
The Commercial Value of Current AI Technologies Compared to Three Years Ago:
Supervised learning is vast—Google invests millions of developers annually, with a value exceeding $100 billion.
Generative AI:
Although currently smaller in scale, it is expected to double in growth over the next three years.
Despite seeming reservations about generative AI, Ng is optimistic. Three years is a short time, and the pace might be even faster—the key is that developers, large companies, and startups can all find exciting opportunities in this field.
Future Prospects
Supervised learning is vast, with Google investing millions of developers annually, creating value exceeding $100 billion. Generative AI, while currently smaller in scale, is expected to double in growth over the next three years. Ng believes that developers, large companies, and startups can find exciting opportunities in this area.
Notable Quote
"Despite receiving much media attention, it has been proven that the top application layer is more successful, generating costs and economic benefits. Only then can the infrastructure and developer tool layers generate revenue and succeed."
Ng founded Landing.ai and, after discussions with numerous CEOs, compiled common problems and recommended strategies for AI transformation in a publicly available document, the "AI Transformation Playbook."
Comments