Anthropic's 'anti-China' stance triggers exit of star AI researcher (www.yahoo.com)

Yao Shunyu, a high-profile Chinese AI researcher who joined Anthropic in October 2024 and contributed to the Claude 3.7 Sonnet model, has left the startup after under a year to join Google DeepMind’s core Gemini team as a senior research scientist. In a public post he cited Anthropic’s recent “anti‑China” rhetoric — including a policy blocking subsidiaries of Chinese companies from its services and labeling China an “adversarial nation” — as a key reason for his departure, alongside an additional confidential factor he did not disclose. Yao’s background (Tsinghua undergrad, Stanford PhD in theoretical and mathematical physics, Berkeley postdoc) underscores the calibre of talent involved. The move is significant because it spotlights how geopolitical posturing and corporate policy can directly influence talent flows in AI, potentially reshaping where expertise concentrates and which groups steer foundational‑model development. Technically, the transfer of someone who worked on Claude 3.7 to DeepMind’s Gemini effort represents direct cross‑pollination of model expertise at the highest level. More broadly, it fuels concerns that hostile rhetoric could drive researchers away from certain firms or geographies, complicating collaboration on shared challenges like AI safety and accelerating competition between U.S. and Chinese AI ecosystems.