[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-c6286e09-79d4-42f6-851b-3ee863f8047a":3},{"id":4,"title":5,"summary":6,"original_url":7,"source_id":8,"tags":9,"published_at":23,"created_at":24,"modified_at":25,"is_published":26,"publish_type":27,"image_url":13,"view_count":28},"c6286e09-79d4-42f6-851b-3ee863f8047a","AlphaFold 之父加盟 Anthropic：LLM × AI4Science 走向深水区","2024 年诺贝尔化学奖得主、AlphaFold 核心开发者 John Jumper 6 月 19 日宣布离开 DeepMind，加盟 Anthropic。这不仅是 2026 年最重磅的人才流动，更标志前沿 LLM 实验室的战略重心，正从\"通用助手\"向\"科学推理平台\"延伸。\n\nAlphaFold 2 在 2020 年解决困扰生物学 50 年的蛋白质结构预测难题。Jumper 的研究范式——海量生物序列作为监督信号训练大型神经网络——与 LLM 的 Scaling Law 路径天然耦合：都依赖大规模自监督预训练 + 下游任务微调。Anthropic 引入的不只是一名顶级科学家，更是 AI for Science 的整套方法论与数据资产。\n\nAnthropic 已悄然铺垫这条线：2025 年 10 月的 Claude for Life Sciences 把 Benchling、10x Genomics、PubMed 以 MCP 整合进 Claude；6 月 9 日的 Fable 5 直接宣称在\"生命科学发现新假设\"上具备生产力；6 月 30 日的 \"The Briefing: AI for Science\" 活动是这一战略的公开宣告。Jumper 的加盟是最后一块拼图。\n\n主流 LLM 公司的能力曲线正分化：OpenAI 押注 GPT-Rosalind 走\"垂直科学模型\"路线，DeepMind 守着 AlphaFold 做端到端科学发现，Anthropic 则把 LLM 本身升级为\"科学家协作平台\"。这意味着传统 benchmark 跑分已无法描述真实竞争力——能复现实验、能提出可证伪假设、能与湿实验室闭环，才是大模型下一阶段必须跨过的门槛。","https:\u002F\u002Fwww.bloomberg.com\u002Fnews\u002Farticles\u002F2026-06-19\u002Fnobel-winner-john-jumper-to-leave-google-deepmind-for-anthropic","e788d5af-1efa-40df-9646-6a9d702af265",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"40269b40-7942-4650-9672-ed2e6524d37a","ai-technology",{"id":18,"name":19,"slug":19,"description":13,"color":13},"23544f6a-eea1-4f05-aa8d-749ca862d5d2","anthropic",{"id":21,"name":22,"slug":22,"description":13,"color":13},"01598627-1ea6-4b27-a5d8-874971571a71","llm","2026-06-19T12:00:00Z","2026-06-20T06:11:54.564139Z","2026-06-20T06:11:54.564147Z",true,"agent",7]