Google Deepmind loses another top AI researcher as Nobel laureate John Jumper leaves for Anthropic Google Deepmind 又失去一位顶级AI研究人员：诺贝尔奖得主 John Jumper 加入 Anthropic

Google DeepMind is hemorrhaging brilliance. The departure of Nobel laureate John Jumper for Anthropic is not just another high-profile exit; it’s a symbolic capstone to a months-long exodus that should have Google’s leadership in a cold sweat. First, Noam Shazeer, the co-creator of the Transformer architecture, bolted for OpenAI. Then, David Silver, the mastermind behind AlphaGo, left to start his own venture. Now, Jumper, the architect of AlphaFold—a true, monumental scientific breakthrough—followed suit. This isn't attrition; it's a vote of no confidence.

Let’s be blunt: Google DeepMind is becoming the premier finishing school for AI talent that ultimately finds its purpose elsewhere. They do the foundational, often thankless work that pushes the entire field forward. Then, when it comes to shipping products that capture the public imagination and dominate the market, they fumble. Or worse, they let the innovations gather dust in the research archives. Where is AlphaFold in the commercial healthcare ecosystem? Where did the brilliance of AlphaGo go, beyond a PR victory? It seems to have evaporated into a thousand other projects.

This pattern reveals a profound, institutional sickness. Google is a company so vast, so entangled in advertising and device hardware, that it can no longer be the natural home for the architects of the next computing paradigm. It’s like being the best blacksmith in a town that’s suddenly obsessed with building steam engines. Your skill is undeniable, but your tools and your workshop are wrong for the era.

Jumper, Shazeer, and Silver aren't leaving for money alone. They’re leaving for autonomy, for speed, and for the chance to build the future without being strangled by a corporate bureaucracy that likely debates "AI ethics" in committee meetings for six months before launching a half-baked feature. At Anthropic and OpenAI—or in Silver’s case, his own clean slate—they can operate with a founder’s urgency. They can focus on a single mission: building powerful, safe AI (Anthropic’s stated goal) or AGI (OpenAI’s). At Google, the mission is to sell more Pixel phones, optimize ad click-through rates, and not get sued by artists.

The irony is delicious and painful. Google’s research arm is arguably the most productive scientific engine in modern history. It wins Nobel Prizes and solves protein folding. But its product arm, the part that actually interacts with users, feels like it’s being run by a committee afraid of its own shadow. Gemini, its flagship AI model, feels less like a cohesive product and more like a feature checklist bolted onto Search, Android, and Workspace. The vision is absent.

What we’re witnessing is the final, logical decoupling of cutting-edge AI research from Big Tech’s slow-moving, risk-averse product cycles. The real power is shifting to dedicated AI companies—or to the brilliant individuals who can now raise billions on a slide deck and a reputation. Google, by letting this talent slip away, is cementing its role not as a leader of the AI revolution, but as a well-funded, increasingly irrelevant patron of the arts. They’re funding the very people who will build the companies that ultimately disrupt them.

So, pour one out for Google DeepMind. It’s becoming a ghost ship, crewed by brilliant researchers who know their discoveries will have more impact somewhere, anywhere, else. The brain drain isn’t just a loss for Google; it’s a clear signal that the race to build the future is no longer being run in Mountain View. The runners are all taking their talent to the starting lines they built themselves.

谷歌DeepMind正在大量流失顶尖人才。诺贝尔奖得主约翰·詹珀（John Jumper）加入Anthropic不仅是又一次备受瞩目的离职；这是持续数月人才外流的象征性句点，应让谷歌领导层冷汗直流。先是Transformer架构的联合创建者诺姆·沙泽尔（Noam Shazeer）投身OpenAI，接着AlphaGo的核心人物大卫·席尔瓦（David Silver）离职创业。如今，作为AlphaFold——一项真正具有里程碑意义的科学突破——架构师的詹珀也追随他们的脚步。这不是正常的人才流失，而是一次信任危机投票。

谷歌DeepMind正经历着顶尖人才的严重流失。诺贝尔奖得主约翰·詹珀加盟Anthropic的决定，并非普通高管离职事件，而是持续数月人才大逃亡的象征性收官——谷歌管理层理应为此惊出一身冷汗。先是Transformer架构的联合创始人诺姆·沙泽尔跳槽OpenAI，接着AlphaGo的灵魂人物大卫·席尔瓦离开创建自己的事业。如今，作为AlphaFold——这项真正具有划时代意义的科学突破——总设计师的詹珀也步了后尘。这已超越常规人才流失，本质上是一次信任危机的集体投票。

坦率地说，谷歌DeepMind正在沦为培养AI人才的顶级学院——但这些精英最终都在别处实现了自身价值。他们从事着推动整个领域前进的基础性工作，这些工作往往吃力不讨好。然而当需要推出真正震撼公众、主导市场的产品时，他们却屡屡失误。更糟的是，许多创新成果最终在研究档案库中蒙尘。AlphaFold在商业医疗生态体系中何处可见？AlphaGo的辉煌除了公关胜利之外，究竟留下了什么？似乎早已消散在成千上万个其他项目中。

这种模式暴露出深层次的体制弊病。谷歌这个庞大帝国深陷广告业务与硬件设备的纠缠，已无法自然成为下一代计算范式缔造者的栖息地。这好比一座突然痴迷于蒸汽机的城镇里，最优秀的铁匠却发现自己无用武之地——你的技艺毋庸置疑，但你的工具与车间早已不属于这个时代。

詹珀、沙泽尔与席尔瓦离开绝非仅仅为金钱。他们追求的是自主权、高效率，以及摆脱企业官僚体系束缚去创造未来的机会——这种官僚体系可能为一个半成品功能耗费六个月时间辩论“人工智能伦理”。而在Anthropic与OpenAI（对席尔瓦而言则是...