Fields Medal winner Yau Shing-tung has talked about AI twice recently, and it is interesting to look at his remarks together.

One thing at a time is doing things. On October 11, the "Qiu Shing-tung Mathematics and Artificial Intelligence Laboratory Alliance" led by him was established in Tsinghua, with him as chairman, bringing together more than a thousand scientists. The alliance should independently develop China's large-scale mathematics-specific models, build infrastructure for AI-assisted mathematics research, and then extend to cross-cutting fields such as quantum computing, life sciences, and financial economics.

He described the impact of AI on mathematics as "like a tsunami coming", and his attitude was to take the initiative to surf, don't hide.

Another time it was a reminder. On November 17, he wrote an article in the People's Daily, saying that China is expected to cultivate a group of leading mathematicians locally in the next 5 to 10 years. The article specifically writes a section about AI: its influence cannot be ignored, but it is always a tool. If you completely follow the conclusions of AI, you will lose your ability to ask independent questions and think deeply.

These words come from a person who has just recruited more than a thousand people to build a large mathematical model, and their weight is different. He is not worried that the AI ​​is too weak, but that young people will hand over judgment too quickly.

In the same article, he also opposed rote memorization and repeated questions, and advocated the establishment of a long-term evaluation system.

While working on a large mathematical model, I say not to be led by AI. It seems contradictory, but it actually means the same thing: AI can do calculations and searches for you, but you have to ask questions yourself.