In a recent dialogue on the Sohu.com video platform's "Zhang Chaoyang's Physics Class" scientist event, Sohu.com founder, Chairman, and CEO Zhang Chaoyang had an in-depth discussion with Liu Ruowei, a physics professor at the University of Pennsylvania and director of the Center for Soft Matter and Living Systems.

Regarding AI and scientific research, Zhang Chaoyang shared his perspective: "After AI provides an answer, you must set aside an equal amount of time to think for yourself, otherwise your brain will atrophy." He believes that when facing truly cutting-edge scientific issues with significant controversy, we cannot simply accept AI's conclusions; instead, we must actively compare different viewpoints and maintain critical thinking.

Discussing the relationship between AI and human cognition, Liu Ruowei noted that AI is very good at providing answers, though these answers are not always correct. However, the most important aspect of physics is asking questions, and the newer and more unique the question, the more valuable it often is. She pointed out that there are fundamental differences in the data received, learning methods, and physical foundations between large language models and the human brain. AI will accomplish many tasks differently from humans, but it will not simply replace them.

When asked about how young scholars should choose their research directions, Liu Ruowei offered a practical and interesting method: "Ask yourself which type of 'problem' you can tolerate the most." If you cannot tolerate experimental equipment breaking down, experimental research may not suit you; if you cannot tolerate repeatedly searching for code errors, computational research may not be for you; if you are willing to spend a long time checking a single symbol in a derivation, then theoretical research might be a better fit.