Why one of the most brilliant AI scientists left the US for China (www.theguardian.com)

Song‑Chun Zhu, a globally respected AI researcher who led a prolific UCLA lab and helped lay foundations for modern pattern‑recognition systems, surprised colleagues by returning to China in 2020 to accept senior roles at top Beijing universities and a state‑backed AI institute. He now sits on a high‑level political advisory body and directs the Beijing Institute for General Artificial Intelligence, where his team shapes curricula and policy. His move highlights a broader shift in research talent and funding: China is offering resources and strategic urgency that some scientists find irresistible, while growing U.S. political pressure and scrutiny of China‑born academics have eroded America’s traditional magnetism for global talent. Technically, Zhu is a prominent critic of the prevailing “big data, big compute” neural‑network paradigm that powers LLMs like ChatGPT. He advocates a “small data, big task” approach emphasizing causal reasoning, goal‑directed planning, physical and social intuition, and resourceful problem solving — qualities he argues are necessary for genuine AGI and that current LLMs fundamentally lack. His leadership in Beijing signals a serious, well‑funded alternative research trajectory that could produce different architectures and benchmarks for general intelligence, and intensifies the geopolitical stakes of who sets the future AI agenda.