[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-deepmind-veterans-are-leaving-london-en":3,"article-related-deepmind-veterans-are-leaving-london-en":30,"series-industry-96ad3567-ab75-487a-b9ac-656da06056ef":77},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":22,"views":26,"created_at":27,"published_at":28,"topic_cluster_id":29},"96ad3567-ab75-487a-b9ac-656da06056ef","deepmind-veterans-are-leaving-london-en","DeepMind老兵正在离开伦敦","\u003Cp data-speakable=\"summary\">\u003Ca href=\"\u002Fnews\u002Fgoogle-deepmind-turns-science-into-tools-en\">Google DeepMind\u003C\u002Fa> 正从学术型研究所转向产品和模型竞赛，这让一批老研究员开始离开伦敦。\u003C\u002Fp>\u003Cp>DeepMind 曾经靠 \u003Ca href=\"https:\u002F\u002Fdeepmind.google\" target=\"_blank\" rel=\"noopener\">DeepMind\u003C\u002Fa> 的 AlphaGo、AlphaFold 和强化学习研究，建立起一种近乎学术圣殿的气质。现在，随着 \u003Ca href=\"\u002Ftag\u002Fgemini\">Gemini\u003C\u002Fa> 时代的 \u003Ca href=\"https:\u002F\u002Fdeepmind.google\u002Ftechnologies\u002Fgemini\" target=\"_blank\" rel=\"noopener\">Google DeepMind\u003C\u002Fa> 更深地卷入模型竞赛与产品交付，这种气质正在变淡。\u003C\u002Fp>\u003Cp>这篇讨论的核心并不复杂：当一家研究机构开始同时承担模型能力、产品节奏和商业压力时，最先感到不适的，往往是那些把它当作长期研究家园的人。伦敦办公室还在，但它所代表的那种身份认同，已经不太像过去了。\u003C\u002Fp>\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>时间点\u003C\u002Fth>\u003Cth>相关事实\u003C\u002Fth>\u003Cth>意义\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>2016\u003C\u002Ftd>\u003Ctd>AlphaGo 战胜李世石\u003C\u002Ftd>\u003Ctd>DeepMind 的公众声望达到顶点\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>2020\u003C\u002Ftd>\u003Ctd>AlphaFold 2 取得蛋白质结构预测突破\u003C\u002Ftd>\u003Ctd>AI for Science 叙事被推到主流视野\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>2023\u003C\u002Ftd>\u003Ctd>Google 将 Bard 与 DeepMind 体系整合，Gemini 成为核心模型方向\u003C\u002Ftd>\u003Ctd>研究重心开始向产品与平台靠拢\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003Ch2>从研究圣地到产品机器\u003C\u002Fh2>\u003Cp>DeepMind 最吸引人的地方，从来不只是模型本身，而是它曾经允许研究员做长周期、低噪音、强问题导向的工作。那种环境适合探索强化学习、蛋白质折叠、通用智能路线，也适合培养一种“先把问题想清楚，再考虑怎么落地”的研究文化。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782777770669-33e7.png\" alt=\"DeepMind老兵正在离开伦敦\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>但 \u003Ca href=\"\u002Ftag\u002Fgoogle\">Google\u003C\u002Fa> 现在要的东西更多。Gemini 不只是论文里的分数，它还要进入搜索、Workspace、Android、云服务和广告系统。研究团队一旦和这些产品线绑得更紧，评估标准就会从“这个想法是否漂亮”变成“这个功能能否按季度上线”。\u003C\u002Fp>\u003Cp>这就是很多老员工感到陌生的地方。组织目标没有错，错的是目标变了之后，原来的工作方式还在被期待继续成立。\u003C\u002Fp>\u003Cul>\u003Cli>AlphaGo 让 DeepMind 成为 AI 领域最有辨识度的名字之一。\u003C\u002Fli>\u003Cli>AlphaFold 让它从“会下棋”变成“能改变科学研究流程”的机构。\u003C\u002Fli>\u003Cli>Gemini 时代则要求它同时做前沿模型、平台能力和产品集成。\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>为什么老研究员会走\u003C\u002Fh2>\u003Cp>离开的原因通常不会只有一个。对很多 DeepMind 老兵来说，最直接的变化是研究自主权变少了。过去可以围绕一个科学问题持续推进几年，现在更常见的是围绕产品目标、模型发布节奏和内部优先级调整工作方向。\u003C\u002Fp>\u003Cp>另一个变化是组织气质。早期 DeepMind 更像一个由研究兴趣驱动的团队，大家围绕论文、实验和长期目标建立信任。现在它越来越像一个必须和 Google 其他部门协同的核心业务单元，沟通成本更高，决策链条也更长。\u003C\u002Fp>\u003Cblockquote>“The main thing I’m worried about is that we don’t move too fast and break things.” — Demis Hassabis\u003C\u002Fblockquote>\u003Cp>这句话来自 \u003Ca href=\"https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FDemis_Hassabis\" target=\"_blank\" rel=\"noopener\">Demis Hassabis\u003C\u002Fa>，他在不同场合都强调过速度与谨慎之间的平衡。问题在于，当整个行业都在追速度时，组织内部想保持研究耐心，本身就会越来越难。\u003C\u002Fp>\u003Cp>对研究员来说，最难受的不是忙，而是忙得没有研究上的完整闭环。你可能在做一个很强的模型组件，却很少有机会看到它如何从科学问题一路走到产品价值。\u003C\u002Fp>\u003Cul>\u003Cli>更少的长期自由课题，更多的里程碑和交付节点。\u003C\u002Fli>\u003Cli>更强的产品耦合，更多的跨团队协调。\u003C\u002Fli>\u003Cli>更高的外部关注度，更多的绩效和曝光压力。\u003C\u002Fli>\u003Cli>更清晰的商业目标，也意味着更少的“纯研究”空间。\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>Gemini 时代的 DeepMind 到底变了什么\u003C\u002Fh2>\u003Cp>如果说 AlphaGo 和 AlphaFold 时代的 DeepMind 追求的是“证明某件事能被 AI 做到”，那么 Gemini 时代更像是在回答“这件事如何被稳定地放进 Google 的产品体系里”。这两个问题都重要，但它们对应的是两种完全不同的组织逻辑。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782777771894-t93d.png\" alt=\"DeepMind老兵正在离开伦敦\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>前者更像研究机构，后者更像平台型公司内部的核心引擎。前者奖励新想法，后者奖励可复制、可发布、可维护的能力。前者让研究员享受探索，后者让研究员面对现实约束。对于习惯了前一种环境的人，后一种环境很难不让人想离开。\u003C\u002Fp>\u003Cp>这也解释了为什么外界常把 DeepMind 的变化看成“人才流动”，但更准确的说法其实是“文化再定义”。不是每个离开的人都在否定 DeepMind，很多人只是觉得自己熟悉的那家机构已经不在了。\u003C\u002Fp>\u003Cp>从公司战略看，这种变化并不意外。\u003Ca href=\"https:\u002F\u002Fopenai.com\" target=\"_blank\" rel=\"noopener\">OpenAI\u003C\u002Fa>、\u003Ca href=\"https:\u002F\u002Fwww.anthropic.com\" target=\"_blank\" rel=\"noopener\">Anthropic\u003C\u002Fa> 和 Google 都在争夺同一批前沿模型能力、同一类开发者心智，以及同一批企业客户。研究机构如果不把成果尽快接入产品，就会在竞争中失去回报。\u003C\u002Fp>\u003Cp>但从组织管理看，代价也很明确：一个曾经靠理想主义凝聚人的团队，一旦被要求同时承担商业化和模型竞赛，就很容易出现认同感松动。老研究员离开伦敦，不一定是因为伦敦这座城市，而是因为他们不再确定自己还能在这里做多久“真正属于研究的工作”。\u003C\u002Fp>\u003Ch2>这场变化会把 DeepMind 带向哪里\u003C\u002Fh2>\u003Cp>短期内，\u003Ca href=\"\u002Ftag\u002Fgoogle-deepmind\">Google DeepMind\u003C\u002Fa> 还会继续吸引顶级人才，因为它仍然能提供世界级算力、顶级研究问题和极高的行业影响力。对很多年轻研究员来说，这些条件依然足够诱人。\u003C\u002Fp>\u003Cp>但长期看，组织内部会出现更明显的分层：一部分人接受平台化、产品化和更强的协同节奏；另一部分人则会把职业重心转向更独立的研究实验室、初创公司，或者更纯粹的学术环境。这个分化已经在发生，而且很难逆转。\u003C\u002Fp>\u003Cp>如果你把 DeepMind 当成一家普通 AI 公司，你会觉得这只是正常的人才流动。如果你记得它曾经是什么，你就会明白，真正值得关注的不是谁离开了，而是这家公司还愿不愿意给“长期研究”留出足够大的空间。\u003C\u002Fp>\u003Cp>接下来最值得观察的不是某一位明星研究员的去向，而是 Google DeepMind 是否还能同时满足两种互相拉扯的需求：一边继续产出足以改写学界认知的研究，一边把这些能力快速塞进 Google 的产品机器里。只要这两个目标继续并存，离开的人就不会少。\u003C\u002Fp>\u003Cp>对于外部观察者来说，问题也很直接：当一家机构从“做出惊艳结果”转向“持续交付能力”时，它还能保留多少原来的灵魂？\u003C\u002Fp>","Google DeepMind 正从学术型研究所转向产品和模型竞赛，这让一批老研究员开始离开伦敦。","zhuanlan.zhihu.com","https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F2053547511319639558",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782777770669-33e7.png","industry","en","21e55851-9929-4b8c-86a1-e97fe2524a50",[17,18,19,20,21],"Google DeepMind","Gemini","AI research","AlphaFold","人才流动",[23,24,25],"DeepMind 正从研究型机构转向产品和平台型组织。","Gemini 时代提高了交付速度，也压缩了长期研究空间。","老研究员离开，反映的是文化和目标的变化，不只是个人选择。",0,"2026-06-30T00:02:29.06378+00:00","2026-06-30T00:02:29.056+00:00","e63df91b-385f-44c9-b3f6-44a1a0e4b505",{"tags":31,"relatedLang":36,"relatedPosts":40},[32,34],{"name":17,"slug":33},"google-deepmind",{"name":18,"slug":35},"gemini",{"id":15,"slug":37,"title":38,"language":39},"deepmind-veterans-are-leaving-london-zh","DeepMind老兵為何離開倫敦","zh",[41,47,53,59,65,71],{"id":42,"slug":43,"title":44,"cover_image":45,"image_url":45,"created_at":46,"category":13},"81fa50cf-ee8b-4b76-b017-7dfc45a2dea0","bitcoin-price-page-risk-asset-market-signal-en","Bitcoin’s price page proves the market still treats BTC like a risk a…","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782776869895-brr2.png","2026-06-29T23:47:27.031808+00:00",{"id":48,"slug":49,"title":50,"cover_image":51,"image_url":51,"created_at":52,"category":13},"5408aa94-6f8f-4f20-9629-7c5550859f3b","sora-smash-ultimate-final-dlc-pick-balanced-en","Sora in Smash Ultimate is a strong final DLC pick, not a broken one","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782775073444-djk4.png","2026-06-29T23:17:22.741007+00:00",{"id":54,"slug":55,"title":56,"cover_image":57,"image_url":57,"created_at":58,"category":13},"08cd2ab1-2a2c-4ab6-ab51-4b16a0fed4ab","openclaw-135000-star-saas-security-crisis-en","135,000-star OpenClaw hits SaaS security crisis","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782771534012-jg28.png","2026-06-29T22:17:16.610831+00:00",{"id":60,"slug":61,"title":62,"cover_image":63,"image_url":63,"created_at":64,"category":13},"13701fd7-c4c2-4966-a6e7-db3646d99bd7","anthropic-ipo-965b-valuation-sec-filing-en","Anthropic IPO: $965B valuation and SEC filing","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782770565405-h3hj.png","2026-06-29T22:02:19.831993+00:00",{"id":66,"slug":67,"title":68,"cover_image":69,"image_url":69,"created_at":70,"category":13},"9f3418e2-07ff-4903-a189-6fbe97d079da","hp-openai-frontier-partnership-en","HP and OpenAI expand Frontier partnership","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782766963986-pbe2.png","2026-06-29T21:02:22.652434+00:00",{"id":72,"slug":73,"title":74,"cover_image":75,"image_url":75,"created_at":76,"category":13},"ca076802-bd15-44b3-8236-f1bc2ba89463","anthropic-california-public-sector-ai-deal-en","Anthropic’s California deal makes Claude the default public-sector AI","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782757079985-9z3y.png","2026-06-29T18:17:33.203469+00:00",[78,83,88,93,98,103,108,113,118,123],{"id":79,"slug":80,"title":81,"created_at":82},"d35a1bd9-e709-412e-a2df-392df1dc572a","ai-impact-2026-developments-market-en","AI's Impact in 2026: Key Developments and Market Shifts","2026-03-25T16:20:33.205823+00:00",{"id":84,"slug":85,"title":86,"created_at":87},"5ed27921-5fd6-492e-8c59-78393bf37710","trumps-ai-legislative-framework-en","Trump's AI Legislative Framework: What's Inside?","2026-03-25T16:22:20.005325+00:00",{"id":89,"slug":90,"title":91,"created_at":92},"e454a642-f03c-4794-b185-5f651aebbaca","nvidia-gtc-2026-key-highlights-innovations-en","NVIDIA GTC 2026: Key Highlights and Innovations","2026-03-25T16:22:47.882615+00:00",{"id":94,"slug":95,"title":96,"created_at":97},"0ebb5b16-774a-4922-945d-5f2ce1df5a6d","claude-usage-diversifies-learning-curves-en","Claude Usage Diversifies, Learning Curves Emerge","2026-03-25T16:25:50.770376+00:00",{"id":99,"slug":100,"title":101,"created_at":102},"69934e86-2fc5-4280-8223-7b917a48ace8","openclaw-ai-commoditization-concerns-en","OpenClaw's Rise Raises Concerns of AI Model Commoditization","2026-03-25T16:26:30.582047+00:00",{"id":104,"slug":105,"title":106,"created_at":107},"b4b2575b-2ac8-46b2-b90e-ab1d7c060797","google-gemini-ai-rollout-2026-en","Google's Gemini AI Rollout Extended to 2026","2026-03-25T16:28:14.808842+00:00",{"id":109,"slug":110,"title":111,"created_at":112},"6e18bc65-42ae-4ad0-b564-67d7f66b979e","meta-llama4-fabricated-results-scandal-en","Meta's Llama 4 Scandal: Fabricated AI Test Results Unveiled","2026-03-25T16:29:15.482836+00:00",{"id":114,"slug":115,"title":116,"created_at":117},"bf888e9d-08be-4f47-996c-7b24b5ab3500","accenture-mistral-ai-deployment-en","Accenture and Mistral AI Team Up for AI Deployment","2026-03-25T16:31:01.894655+00:00",{"id":119,"slug":120,"title":121,"created_at":122},"5382b536-fad2-49c6-ac85-9eb2bae49f35","mistral-ai-high-stakes-2026-en","Mistral AI: Facing High Stakes in 2026","2026-03-25T16:31:39.941974+00:00",{"id":124,"slug":125,"title":126,"created_at":127},"9da3d2d6-b669-4971-ba1d-17fdb3548ed5","cursors-meteoric-rise-pressures-en","Cursor's Meteoric Rise Faces Industry Pressures","2026-03-25T16:32:21.899217+00:00"]