Moonshot AI創辦人楊植麟於2019年在卡內基梅隆大學僅用四年便完成博士學位，儘管蘋果、谷歌、Meta等科技巨頭紛紛向他伸出橄欖枝，斯坦福和MIT也為他敞開博士後大門，他仍選擇回到中國創業。他的導師Russ Salakhutdinov回憶，楊植麟認為如果不嘗試創業，他將終身遺憾。這一當初看似逆勢而行的決定，如今已被證明極具遠見——Moonshot上週發布的Kimi K3大語言模型已躋身與OpenAI和Anthropic旗艦產品同級別的行列，連馬斯克也稱其為「令人印象深刻的成果」。
楊植麟的回國並非受美國簽證政策所迫，而是出於創業信念與中國AI生態系統日益增強的吸引力。他畢業於清華大學，在CMU期間參與了Transformer-XL等奠基性論文的撰寫，這些工作直接影響了當今大語言模型的核心架構。創立Moonshot僅三年，他的三百人團隊便打造出與全球頂尖模型競爭的產品。多位有類似背景的中國創業者表示，在中國創業更為便利，原因包括更廣泛的人脈網絡、更熟悉的文化環境、統一的監管框架，以及更深厚且高效的人才儲備。（關鍵數字：300, three years）
楊植麟的故事折射出更深層的結構性轉變：中國不再僅僅是美國培養人才的淨進口國，已建立起自己的人才管線。斯坦福胡佛研究所追蹤DeepSeek核心模型背後的356名研究人員，發現53.5%從未在海外學習或工作，而有美國經歷者中70%最終回到中國。2024年斯坦福對1,304名在美華裔科學家的調查顯示，73%感到作為學術研究者不安全。美國數十年來的科技主導地位建立在吸引並留住全球最優秀人才的模式之上，而這一模式正面臨考驗——不僅因為美國變得不那麼友好，更因為中國變得更具磁力。
Yang Zhilin, founder of Beijing-based Moonshot AI, completed his PhD at Carnegie Mellon University in just four years and turned down offers from Apple, Google, Meta, Stanford, and MIT to return to China and start his own AI company. His mentor recalled that Yang believed he would regret it for life if he never tried entrepreneurship. That decision has proven prescient: Moonshot's Kimi K3 large language model, released last week, now ranks alongside flagship models from OpenAI and Anthropic, drawing praise from Elon Musk who called it "impressive work."
Yang's return was not driven by US immigration barriers but by entrepreneurial conviction and the growing appeal of China's AI ecosystem. A Tsinghua graduate who co-authored foundational papers like Transformer-XL at CMU, Yang built a 300-person team that produced a globally competitive model within three years of founding Moonshot. Several Chinese founders with similar backgrounds have noted that starting a company in China offers advantages including stronger personal networks, cultural familiarity, a unified regulatory framework, and a deeper talent pool—despite challenges such as constraints on advanced computing power, geopolitical risks, censorship, and a shallower fundraising market compared to the US.
Yang's story reflects a broader structural shift: China has built its own talent pipeline rather than relying solely on US-trained researchers. Stanford's Hoover Institution found that 53.5 percent of 356 researchers behind DeepSeek's core models had never studied or worked abroad, and 70 percent of those with US experience ultimately returned to China. A 2024 Stanford survey of 1,304 US-based Chinese-descent scientists revealed that 73 percent felt unsafe as academic researchers. America's decades-long formula of attracting and retaining the world's brightest minds is being tested—not only because the US has become less welcoming, but because China has become more magnetic. The critical question is how many future talents will make the same choice as Yang, convinced that the future they want to build lies elsewhere.