A leading figure in artificial intelligence research has made a big leap from Anthropic to Google DeepMind, and the story behind it is raising eyebrows across the tech world.
Yao Shunyu, a prominent Chinese scientist with a background from Stanford and Tsinghua University, left Anthropic after less than a year, highlighting his concerns over its increasingly harsh stance toward China.
He explained in a personal note that he could not ignore his “strong opposition” to Anthropic’s recent decision to bar any Chinese company subsidiaries from using its technology, tagging China as an “adversarial nation” in the process.
For Yao, the climate grew untenable. “Although, to be clear, I believe most of the people at Anthropic disagree with such a [characterisation], I don’t think there is a way for me to stay,” he shared.
His exit comes amid escalating rhetoric by US artificial intelligence companies against Chinese competitors, with Anthropic’s approach echoing Trump-era policy.
At the same time, organizations like OpenAI and Meta have increased scrutiny of Chinese involvement in artificial intelligence and cast a wary eye at rivals based in Beijing and Hangzhou.
Google DeepMind’s Approach and Yao’s Arrival
In sharp contrast, Google DeepMind has signaled a willingness to pursue constructive engagement between the US and China on AI safety, a stance reflected by CEO Demis Hassabis.
Yao now joins DeepMind as a senior research scientist on the key Gemini team, working at the heart of the company’s most ambitious models.
His resume is hard to overlook, ranging from the Yeh Chi Sun prize at Tsinghua to a doctoral degree at Stanford and postdoctoral work at Berkeley, before guiding the release of Anthropic’s Claude 3.7 Sonnet model.
He noted that the world of artificial intelligence advances far faster than academic physics, admitting, “Looking back, I am surprised by how much has happened in the past year.”
Many in the field are watching closely, as MacroPolo’s study previously showed over a third of top artificial intelligence researchers in the US actually hail from China.
Recent departures and high profile recruitments at Meta and Tencent underscore how international the artificial intelligence talent pool remains, despite the growing divide in rhetoric.
A former OpenAI staffer, asking not to be named, confessed that some researchers felt “uneasiness” over the tone taken toward China by leaders at their organizations.
For now, Yao is already embedded in the Gemini team, bringing his expertise and a unique perspective to DeepMind’s ambitious artificial intelligence mission.
