凡人小北@frxiaobei · 6月15日57

纳德拉给微软定调，别卷最强模型，去做模型之上的生态。 逻辑是模型早晚商品化，生态锁定才能更持久。 但他说平台让长在上面的价值多于自己攫取的。这是手里没有 SOTA 的人，最舒服的世界观。 要是微软自己有 SOTA，这篇大概率换个写法。 不过抛开动机这个框架是很有用，把闭环建在模型之上，IP 留自己手里，模型随时可换。

译微软CEO纳德拉明确表示，微软不追求最强AI模型，而是聚焦模型之上的生态建设。其逻辑是：模型终将商品化，生态锁定才能更持久。同时他指出，平台让生长其上的价值多于自身攫取的。这一框架被评论为“没有SOTA”的公司最舒服的世界观，但背后动机是将其作为策略：把闭环建在模型之上，IP留自己手里，模型随时可换。

宝玉@dotey · 6月15日62

微软 CEO Satya Nadella 发了一篇长文，提出了一个新概念：Token 资本。 他的核心论点是，AI 时代每家公司都需要同时经营两种资本。一种是传统的人力资本，员工的知识、判断力、关系网络；另一种是 Token 资本，公司自己构建并拥有的 AI 能力。两者不是此消彼长的关系，人的判断力越强，Token 资本增长越快。没有人的方向引导，算力只是在空转。 这个说法听起来抽象，但 Nadella 给出了一个具体的检验标准：你能不能随时换掉底层的通用大模型，而不丢失公司积累的专有经验？如果能，说明你真正拥有自己的 AI 能力；如果不能，说明你只是在租用别人的智能。 他建议企业把工作流、行业知识、决策经验转化成可以持续改进的 AI 系统，建立私有评估体系来衡量模型在实际业务中的表现，而不是只看公开跑分。这个学习飞轮一旦转起来，就像复利，每次改进的工作流都会产生更好的训练信号，进一步加速知识积累。 Nadella 还发出了一个颇有政治意味的警告。他拿全球化做类比：第一轮全球化时期，GDP 数字看着不错，但整个产业被外包掏空了，后果至今还在显现。如果 AI 时代重演这个剧本，少数几个模型吃掉所有行业的知识和价值，"政治经济体系不会容忍这种结局"。 --- 原文翻译 --- 没有生态支撑的前沿技术，注定无法行稳致远 Satya Nadella 最近，我一直在深思：在由人工智能驱动的经济浪潮中，企业的未来究竟在哪里？ 这次变革与以往任何一次平台更迭都截然不同。过去，我们只是用数字化系统来提升人类的工作效率。但这一次，我们破天荒地在人类与数字系统之间建立起了一个真正的认知循环 (cognitive loop)。这绝对是个颠覆认知的概念，因为它彻底改变了我们对企业内部“工作”本质的定义。 当 AI 模型能够源源不断地吸收人类和组织的专业知识，并将其变成大众化的廉价商品（即将原本稀缺的专业技能变成人人唾手可得的通用能力，从而削弱企业的核心壁垒）时，真正的危机出现了。我们面临的关键挑战，不再仅仅是如何使用某个数字化工具或系统，而是企业该如何在这个全新的世界中持续学习、积累知识产权 (IP)、保持独特性并茁壮成长。 每家公司都必须构建两种资本：一种是我们熟知的“人力资本” (human capital)，另一种我称之为“Token 资本” (token capital)。人力资本包含了员工的知识储备、判断力、人脉关系、创造力以及识别事物规律的能力；而 Token 资本则是指企业自身打造并掌控的 AI 实力（在这里，“Token 资本”一词很形象，因为大语言模型 (LLM) 处理信息的基本单位就是 Token）。 必须强调的是，随着 Token 资本的不断壮大，人力资本并不会因此贬值。相反，它会变得比以往任何时候都更加宝贵！我坚信，人类的主观能动性 (human agency) 将是推动 Token 资本增长的核心引擎。人类负责设定宏大的目标，跨领域地将线索串联起来，建立关系网，并洞察出最关键的规律。如果没有人类在前方指引方向，那些强大的计算力不过是在原地打转罢了。 这就意味着，真正的机遇并不在于你去市面上挑选一个“最好”的模型，而在于如何在模型的基础之上，构建一个能让人力资本和 Token 资本产生复利效应 (compound) 的“学习循环” (learning loop)。你可以把某项任务甚至整个岗位都外包出去，但你绝对不能把“学习能力”给外包了。企业未来的核心竞争力，就在于能否在人类与 AI 之间不断积累并放大这种学习能力。 这需要一种全新的架构思路：每家企业都要能够构建出能随着时间推移自我迭代的 AI 智能体系统 (agentic systems)，同时还要牢牢掌控自己的知识产权。一家公司应该能够随时替换掉底层的某个“通才模型” (generalist model)，而不丢失那些已经沉淀在系统里的、像“公司老兵”一样丰富的专业经验。在未来的时代，这将是检验企业是否拥有数据控制权和技术主权的关键“试金石”。 企业需要将自身的工作流、领域知识以及多年积累的判断力，统统转化为每一次使用都能自我进化的 AI 系统。企业应当建立私有评估机制 (private evals)（即企业内部针对自身真实业务场景定制的模型能力测试标准），用来检验模型是否真正在对企业有价值的结果上取得了进步，而不能仅仅依赖外界的公开跑分盲目自嗨！专属的强化学习 (reinforcement learning) 环境，应该让模型通过吸收组织内部真实的业务数据和工作轨迹变得越来越强大。这样的专属知识库，能让企业的组织记忆变得随时可检索，同时也让 token (tokens) 的运转效率大幅提升。 这种循环，将成为企业全新的知识产权。我把它想象成一台不断向上攀登的机器 (hill climbing machine)。而且与大多数资产不同，它具有强大的复利效应。每一个被优化的工作流，都会产生更优质的训练信号，从而加速这家企业独有的隐性知识 (tacit knowledge) 的积累。那些尽早布局构建这种循环的公司，将会获得一道难以复制的护城河，无论未来市面上又出了什么能力炸裂的新模型，都无法轻易撼动其地位。 我们最不愿看到的局面，就是各行各业的所有公司，都在向少数几个贪婪吞噬一切的巨头模型割让价值。如果所有的经济价值都只被少数几个模型垄断，政治经济体制是绝对无法容忍的。社会也绝对不会允许一个让整个产业被彻底掏空的 AI 未来。 回想一下全球化初期发生的事情吧：大规模的业务外包曾让许多工业经济体被彻底掏空。表面上看 GDP 数据依然光鲜亮丽，但大量产业工人流离失所是血淋淋的现实，其带来的严重后果至今仍未消散。我们绝不能让这种悲剧在 AI 时代重演——决不能让少数几个 AI 系统攫取了所有的经济回报，而一整个行业的从业者却只能眼睁睁地看着自己赖以生存的专业知识被无情地廉价化。 在我看来，我们的当务之急不仅是打造前沿模型 (frontier model)，更要构建一个繁荣的“前沿生态系统” (frontier ecosystem)。只有这样，价值才能像活水一样，广泛地流向每一家公司、每一个行业、每一个国家。在这个生态中，每个组织都能拥有属于自己的学习循环，将组织智慧沉淀其中，让人力资本与 Token 资本共同实现滚雪球式的增长。 这也是伴随我职业生涯一路走来的核心理念：真正的平台，能够让在其之上生长出来的价值，远远大于平台自身所截留的价值。在这样的生态里，每家公司都能持续创新，并构建属于自己的真正价值。 当这一切实现时，企业不仅能为自己、也能为周边的整个经济体创造巨大的红利。员工们将会看到自己的专业技能被无限放大，个人的判断力将被融入系统，变得可以复制和规模化应用。而这一切带来的好处，最终将回馈给企业以及他们所在的广泛社区。 这才是企业为自身和宏观经济创造价值的正确方式。这也是我们应当携手共建的、稳定而持久的生态平衡。

译微软CEO Satya Nadella提出“Token资本”概念，认为AI时代每家公司需同时经营人力资本（员工知识、判断力）和自建AI能力（Token资本）。两者互补：人的判断力越强，Token资本增长越快。检验标准：能否随时替换底层通用大模型而不丢失专有经验？若能，则真正拥有AI能力；若不能，则只是租用智能。他建议将工作流、行业知识转化为可迭代AI系统，建立私有评估机制，形成复利式学习飞轮。同时警告：若少数模型垄断行业价值，政治经济体系将无法容忍，类比全球化外包掏空产业的教训。

Rohan Paul@rohanpaul_ai · 6月15日75

Great article by Satya Nadella on organizational economics of AI and "token capital" The real contest is not model quality alone, its the loop around the model: the workflows, feedback, judgments, exceptions, failures, and private tests that teach a system what matters inside a firm. That requires private evals, private reinforcement loops, and queryable institutional memory

译Satya Nadella 关于 AI 组织经济学和“token capital”的好文 真正的竞争不在于模型质量本身，而在于模型周围的循环：那些教会系统什么对企业重要的工作流、反馈、判断、例外、失败和私有测试。 这需要私有评估、私有强化循环和可查询的机构记忆。

Rohan Paul@rohanpaul_ai · 6月15日51

Satya Nadella on the supply side of the physical economics of AI "Tokens per Dollar per Watt" His energy is something here. 🔥 The new equation for the AI age for every Company or Industry or Country. "And that means Infrastructure, Infrastructure and Infrastructure." --- From "Microsoft India" YT channel (link in comment)

译Satya Nadella 在微软印度频道访谈中提出 AI 物理经济学供应侧新公式：“Tokens per Dollar per Watt”，强调每美元每瓦特获得的 token 数是竞争力关键，并呼吁“基础设施、基础设施、基础设施”。在其关于 AI 组织经济学的文章中，Nadella 指出真正的竞争是围绕模型的循环——工作流、反馈、判断、异常、失败及私有测试，这要求企业建立私有评估、私有强化循环和可查询的制度记忆（token capital）。

elvis@omarsar0 · 6月15日35

Highly recommended reading. Don't offload your learning. Don't offload your creative process. "You can offload a task, or even a job, but you can never offload your learning."

译强烈推荐阅读。 不要外包你的学习。不要外包你的创意过程。 “你可以外包一项任务，甚至一份工作，但你绝不能外包你的学习。”

Satya Nadella@satyanadella · 6月14日65

http://x.com/i/article/2065582894790365184 # A frontier without an ecosystem is not stable I’ve been thinking a lot about the future of the firm in an AI-driven economy. This transition is different than any previous platform shift. In the past, we used digital systems to enhance human capital. This is the first time we can create a real cognitive loop between people and digital systems. That is a mind-bender, because it changes how we even conceptualize work inside an enterprise. What is at stake is not some digital tool or system and its use, but how organizations continue to learn, build IP, differentiate, and thrive in a world where AI models can continuously absorb the expertise of humans and organizations and commoditize it. Every company is going to have to build what I think of as human capital and token capital. Human capital comprises the knowledge, judgment, relationships, ingenuity, and pattern recognition of its people, while token capital is the firm’s AI capability it builds and owns. Importantly, human capital does not become less valuable as token capital grows. It only becomes more valuable! I believe human agency will be the driver of token capital growth. Humans will set ambitious goals, connect dots across domains, build relationships, and recognize patterns that matter most. Without human direction, you have compute running in circles. This means the real opportunity is not in picking the best model but instead in building a learning loop on top of models where human capital and token capital compound. You can offload a task, or even a job, but you can never offload your learning. The future of the firm is the ability to compound that learning across people and AI. This requires a new architectural approach where every business is able to build agentic systems that improve over time, while still retaining control over their IP. A company should be able to switch out a “generalist” model without losing the “company veteran” expertise built into their learning system. This is the key “test” of your control and sovereignty in the era ahead. Companies need to turn their workflows, domain knowledge, and accumulated judgment into AI systems that improve with each use. Private evals should capture whether a model is actually improving against outcomes that matter to the business (not just external benchmarks!). Private reinforcement learning environments should let models grow stronger on real traces from inside the organization. Its knowledge base makes institutional memory queryable and use of tokens more efficient. This loop becomes the new IP of the firm. I think of it as a hill climbing machine. And unlike most assets, it compounds. Every improved workflow generates better training signal, which accelerates the accumulation of tacit knowledge unique to the firm. The companies that build this early will have an advantage that is hard to replicate, regardless of any new individual model capability. The last thing any of us want is a world where every company across every sector is ceding value to a few models that eat everything they see. If all the value is accrued by only a few models, the political economy will simply not tolerate it. There is no societal permission for an AI future that hollows out entire industries. Think about what happened in the first phase of globalization where entire industrial economies were hollowed out by outsourcing. The GDP numbers looked fine on the surface, but the displacement was real and the consequences are still being felt. Let us not bring that dynamic into the AI era, with a small number of AI systems capturing all the economic returns, while entire industries find their knowledge commoditized right out from underneath them. In my view, our priority has to be building a frontier ecosystem, not just a frontier model, so value flows broadly across every company, every industry, and every country. One where every organization can own the learning loop that encodes its institutional knowledge, compounding its human and token capital. This is the ethos I’ve grown up with where platforms enable more value on top than is captured inside, and where every company can continuously innovate and build value of its own. When that happens, companies will create value for themselves and for the economy around them. Employees will see their expertise amplified and their judgment become part of systems that make it replicable and scalable and the benefits accrue to the companies and communities around them. That is how companies drive value for themselves and the broader economy. And it is the stable equilibrium we should build together.

译微软CEO Satya Nadella认为，AI驱动的平台转变首次实现人与数字系统间的认知循环。企业需同时构建人力资本（知识、判断、关系）与token资本（自有的AI能力），且人力资本不会贬值，反而随token资本增长而增值。真正的机会在于建立人力资本与token资本复合增长的学习循环——企业应能替换通用模型而不丢失已内化的专家知识，通过私有评估和强化学习让模型从内部真实轨迹中持续提升。他警告，若所有价值被少数模型吞噬，将重演全球化空心化悲剧，呼吁构建前沿生态系统，让每家企业、行业和国家拥有自己的学习循环。

Microsoft Research@MSFTResearch · 6月13日15

Project Ire examined a timely malware sample and determined its intent through reverse engineering—identifying LOTUSLITE characteristics even as most major EDR tools did not detect it. https://msft.it/6011viy4N

译Project Ire 分析了一个及时的恶意软件样本，并通过逆向工程确定其意图——识别出 LOTUSLITE 特征，即使大多数主流 EDR 工具未检测到它。https://msft.it/6011viy4N

Satya Nadella@satyanadella · 6月10日62

Today in @naturemethods, we shared research on how AI can help us better understand cell behavior, offering new insights into why cancer medicines do not work the same for everyone. By learning more about cell state — how individual cancer cells respond to their surroundings — we have the potential to match therapies more precisely to each patient and improve outcomes. https://news.microsoft.com/signal/articles/why-dont-cancer-medicines-work-the-same-for-everyone-ex-vivo/

译今天在《自然方法》上，我们分享了关于AI如何帮助我们更好地理解细胞行为的研究，为癌症药物为何对每个人的效果不同提供了新的见解。 通过学习更多关于细胞状态——单个癌细胞如何响应周围环境——我们有可能更精确地为每位患者匹配疗法并改善结果。https://news.microsoft.com/signal/articles/why-dont-cancer-medicines-work-the-same-for-everyone-ex-vivo/

Microsoft Research@MSFTResearch · 6月10日63

New research in Nature Methods from Project Ex Vivo shows AI models learn more from diverse cell states than from scaled datasets alone, a finding that could reshape how therapies are matched to patients. https://msft.it/6013vgE8l

译在《Nature Methods》上发表的最新研究来自Project Ex Vivo，表明AI模型从多样化的细胞状态中学到的知识，比仅从规模化数据集中学到的更多，这一发现可能重塑疗法与患者的匹配方式。https://msft.it/6013vgE8l

Satya Nadella@satyanadella · 6月7日64

Great to see NHS England scaling Microsoft 365 Copilot to more than 500,000 staff. In early trials, staff saved an average of 43 minutes per day, helping put more time back into what matters most, caring for patients. https://ukstories.microsoft.com/features/nhs-england-accelerates-ai-adoption-with-microsoft-365-copilot-to-improve-service-delivery-reduce-costs-and-create-more-time-for-care/

译很高兴看到 NHS England 将 Microsoft 365 Copilot 推广给超过 50 万名员工。 早期试验中，员工平均每天节省 43 分钟，帮助将更多时间投入到最重要的事情——患者护理上。https://ukstories.microsoft.com/features/nhs-england-accelerates-ai-adoption-with-microsoft-365-copilot-to-improve-service-delivery-reduce-costs-and-create-more-time-for-care/

Microsoft Research@MSFTResearch · 6月6日49

A BTS look a first timer's experience at the MSR lab at Microsoft Build 2026—featuring the demos, the builders, and the conversations that made it worth the trip.

译幕后花絮——一位新手在微软Build 2026的MSR实验室的首次体验，展示了那些让此行值得的演示、构建者与对话。

Microsoft Research@MSFTResearch · 6月6日60

During the Inside Azure Innovations breakout at Build 2026, Microsoft Azure CTO, deputy CISO and technical fellow Mark Russinovich introduced Project Mosaic, an experimental optical interconnect technology from Microsoft Research Cambridge using micro-LEDs for low-power, high-speed data transmission. A live demo led by senior researcher Kaoutar Benyahya displays individual LED modulation forming letters, proving the concept’s real-time responsiveness. Check out Mark and Kaoutar starting @ 38:38: https://msft.it/6015vdhS9

译微软Azure CTO Mark Russinovich在Build 2026上介绍Project Mosaic，这是微软剑桥研究院的实验性光学互连技术，采用micro-LED实现低功耗、高速数据传输。高级研究员Kaoutar Benyahya现场演示单个LED调制形成字母，证明概念具备实时响应能力。

🚨 AI News | TestingCatalog@testingcatalog · 6月6日56

MICROSOFT 🔥: Early look at Microsoft Scout Agent for Microsoft Frontier users. Scout agent is designed for work use cases and was recently revealed at Microsoft Build 2026. Features 👀 > Models from OpenAI and Anthropic, including GPT-5.5 and Claude Opus 4.7 > Multi-step automations with support for browser use. > Co-Create, an open canvas where users can collaborate with AI and export outputs as documents. > Deep integration with Microsoft Teams > Daily Briefing that summarises work context across connected services. > Skills support, document generation, coding capabilities, and all that. > Both macOS and Windows apps are available. It would be quite a decent tool for Windows users! Yet, it is only for Frontier. Looking forward to testing their super app too.

译微软在Microsoft Build 2026上为Frontier用户推出Scout Agent工作用AI智能体。该Agent可调用OpenAI和Anthropic模型（包括GPT-5.5和Claude Opus 4.7），支持多步骤自动化（含浏览器操作）、Co-Create协作画布（可导出文档）、深度集成Teams、每日简报跨服务汇总工作上下文，以及技能、文档生成和编码能力。目前提供macOS和Windows应用，但仅限Frontier用户使用。

swyx@swyx · 6月5日75

chat is he cooked

译Satya Nadella 在 Latent Space 发布最新访谈，链接见原文。原推文仅评论“chat is he cooked”。

meng shao@shao__meng · 6月4日18

Microsoft：我们发了好多新模型、我们 Copilot 更 NB 了，好像我们又又又进入 AI 领域了？

译推文调侃微软发布大量新模型并声称Copilot性能增强，然而给人的感觉像是“又又又进入AI领域”，暗示其在AI竞争中存在感不足。引用推文描述了其他AI公司现状：OpenAI出问题后重置，Anthropic封禁账号，Google发布新模型却无人关注。整体呈现AI巨头间的混乱与关注度差异。

Chubby♨️@kimmonismus · 6月4日60

Microsoft Build. My personal review. For me, this was the first time I had the chance to attend Microsoft Build, at Microsoft's invitation. To be honest, I didn't really know what to expect, but I was especially looking forward to the keynote. And it wasn't just the keynote: I also visited GitHub HQ, saw the event hall, sat in on numerous sessions, and even met Satya Nadella in person. Holy moly. It truly exceeded all my expectations. 2026 is turning out to be a crazy year for me. It started with NVIDIA GTC in San Jose in March, followed shortly after by a trip to China - Guangzhou and Beijing - then Google I/O in California, and now Microsoft Build, also in California. What a wild ride! I met incredible people and had fascinating conversations late into the evening about LLMs, chips, energy, geopolitical challenges, financial markets, and so much more. What impressed me most was the pioneering spirit, the optimistic atmosphere, the enthusiasm for being at the forefront of this tech-revolution. Optimism mixed with passion and a love of building, that's what I take away from all these trips. Microsoft was no exception. I got a behind-the-scenes look, heard exclusive GitHub sessions, experienced a personal demo of the flagship Surface Laptop Ultra, met researchers, and much more. My honest take on Microsoft Build: Microsoft is taking feedback seriously and is trying to set things in motion and drive change on every front. Seven new AI models - clearly not aiming for the absolute top end, but positioned in the mid-range, roughly at Sonnet level, and affordable; a new laptop with a new chip meant to rival the MacBook Pros, which, frankly, at first glance even seems capable of pulling it off; bold experiments like Project Solaris and the agentic handheld (yes, I've read all the Rabbit comparisons :D); a revamped Copilot app; the rollout of agentic features into enterprise editions with a new quantum chip; and plenty more. It certainly wasn't boring. Time will tell what succeeds, but I'd argue Microsoft is on the right track.

译Kim受邀首次参加微软Build，参观GitHub HQ、参与多场会议并见到Satya Nadella，认为远超预期。微软发布7个新AI模型（定位中端、约Sonnet级别、价格亲民），新Surface Laptop Ultra配新芯片对标MacBook Pro，展示Project Solaris和智能体手持设备等实验项目，推出改版Copilot应用，企业版新增智能体功能及新量子芯片。作者认为微软正认真听取反馈，在各个方向推动变革。

MiniMax (official)@MiniMax_AI · 6月4日65

We are part of @nvidia and @Microsoft ’s Local LLM lineup at #GTC Taipei.🔥 The PC is being reinvented around local, agentic, open-weight models MiniMax-M3 is built exactly for this future: Open-weight. 1M context. Strong coding. Native multimodality. Excited for what comes next!

译我们已加入 @nvidia 和 @Microsoft 在 #GTC Taipei 的本地 LLM 阵容。🔥 PC 正围绕本地、智能体、开放权重模型重新定义。 MiniMax-M3 正是为此未来而打造： 开放权重。 1M 上下文。 强编码能力。 原生多模态。 对接下来的一切充满期待！

Satya Nadella@satyanadella · 6月4日40

Thanks for joining us at Build, Jensen! Grateful for the deep partnership with NVIDIA across cloud and edge.

译感谢你加入我们的Build大会，黄仁勋！感谢与NVIDIA在云和边缘领域的深度合作。

Chubby♨️@kimmonismus · 6月4日65

I took a "behind-the-scenes" tour at Microsoft today, where I was able to inspect the Surface Laptop Ultra firsthand and therefore was able to record those clips. The most obvious takeaway: Microsoft is now aiming to enter into direct competition with Apple and challenge the MacBook Pro. Needless to say, I wasn't able to conduct any real-world testing. However, the build quality, thermal management, the display, and- above all- the NVIDIA chip are certainly impressive. Whether it will truly manage to challenge Apple's MacBooks remains to be seen. But one thing is certain: Microsoft means business.

译微软推出全新Surface Laptop Ultra，定位创作者和AI笔记本，搭载NVIDIA新芯片（RTX GPU），最高提供1 petaflop AI算力、128GB统一内存。配备15英寸mini-LED PixelSense Ultra触摸屏（3:2比例，262 PPI，峰值2000尼特HDR亮度），厚度不足18mm。作者在幕后参观中亲手检测，认为做工、散热、显示屏和芯片令人印象深刻，微软明确将目标对准MacBook Pro，意在直接挑战苹果。

Microsoft Research@MSFTResearch · 6月4日62

A three‑month pilot in a Midwestern bottling plant shows what happens when AI moves beyond chat and into decision-making, where constraints shift, stakes are real, and answers must hold. https://msft.it/6015vjYUN

译一份在中西部装瓶厂进行的三个月试点显示，当AI超越聊天进入决策领域时会发生什么——约束条件变化、风险真实、答案必须可靠。 https://msft.it/6015vjYUN

Chubby♨️@kimmonismus · 6月4日57

First hands-on with Microsoft’s new Surface Laptop Ultra. Microsoft is clearly positioning this as a new class of creator and AI laptop, powered by new NVIDIA silicon with an RTX GPU built for local AI, creative workflows, and gaming. A few standout specs: -New NVIDIA chip with RTX GPU -Up to 1 petaflop of AI compute -Up to 128GB unified memory -15-inch mini-LED PixelSense Ultra touchscreen -3:2 aspect ratio -262 PPI -Up to 2,000 nits peak HDR brightness -Less than 18mm thick

译首次上手微软新的 Surface Laptop Ultra。 微软明确将其定位为面向创作者和 AI 的新品类笔记本电脑，由搭载 RTX GPU 的新 NVIDIA 芯片驱动，专为本地 AI、创意工作流和游戏打造。 几个突出规格： - 带 RTX GPU 的新 NVIDIA 芯片 - 最高 1 petaflop AI 算力 - 最高 128GB 统一内存 - 15 英寸 mini-LED PixelSense Ultra 触摸屏 - 3:2 比例 - 262 PPI - 最高 2000 尼特峰值 HDR 亮度 - 厚度不足 18mm

elvis@omarsar0 · 6月4日66

This SkillOpt paper from Microsoft is a must-read! (bookmark it) I was a bit skeptical of the results reported in the paper when I shared it a few days ago. However, I managed to integrate it into my agent orchestrator and ran a few experiments. The results are mindblowing. Essentially, all my agent skills now have a proper testing framework and a way to self-evolve. I have started to improve all my agent skills with this. One exciting result was when I applied it to my paper-figure-extraction skill, which requires an agent to do multimodal analysis. In particular, it improved quality by +20 points (0.73 → 0.93). I went to see the extracted tables and figures, and I was absolutely stunned by how much better my skill got at the task. Self-improving AI is in the early days, but I think this work is a clear example of the current ability of agents to self-improve. In this case, it was skills, but it's not hard to imagine how this scales to optimizing agent patterns, tool use, context engineering efforts, agentic search, workflows, evals, and even the harness itself. I already started with a few of these ideas inspired by SkillOpt. Stay tuned!

译DAIR.AI的Elvis Saravia将微软SkillOpt论文集成到智能体编排器中后，所有智能体技能获得测试框架与自我演化机制。应用于多模态论文图表提取技能时，质量评分从0.73提升至0.93（+20点），提取结果显著改善。Saravia认为这是自我改进AI的早期范例，该思路可扩展至智能体模式优化、工具使用、上下文工程、智能体搜索及工作流评估等环节。他已基于SkillOpt启动多项后续实验。

Chubby♨️@kimmonismus · 6月3日60

Fantastic in depth guide about Microsoft MAI by @eliebakouch tl;dr about the model: Respect where respect is due. -zero synthetic data or distillation from previous models. -1T model with 35B active, trained on 33.5T tokens

译Microsoft MAI 技术报告公开模型细节：1T 总参数，35B 活跃参数，在 33.5T tokens 上训练。最突出的特点是零合成数据、零知识蒸馏，推理、智能体行为、工具使用全部在后训练中从头学习。报告透明度极高，首次在此规模公开各迭代的 MFU 和完整缩放方案，目标成为前沿实验室。

Rohan Paul@rohanpaul_ai · 6月3日58

Satya Nadella: Microsoft’s latest Wisconsin AI data center keeps yearly water consumption no higher than that of 1 local restaurant. "The cooling loop is filled once and the data centre can operate effectively with zero water consumption. Daily water usage across a year is roughly equivalent to what a single restaurant would use" The mechanism is mainly about replacing evaporative cooling with closed-loop direct-to-chip liquid cooling, so water moves like coolant inside a sealed machine rather than being boiled off into the air. Hot GB200-class AI racks produce too much heat for normal air cooling, so cold liquid is pushed through pipes into the servers and across metal cold plates touching the hottest chips. The liquid enters the rack cool, absorbs heat from the chips through cold plates, then exits the rack at a higher temperature and carries that heat through pipes to a huge cooling system outside the compute floor. Microsoft says Fairwater sends that hot water to cooling “fins” beside the datacenter, where 172 20-foot fans blow air across the fins and dump the heat into the outside air. The important detail is that the air cools the water through metal surfaces, so the water does not need to evaporate the way many older datacenters use cooling towers. The cooled liquid then returns to the servers, repeats the loop, and keeps absorbing heat from the chips. In older data centers, heat is often removed partly through cooling towers. Hot water meets moving air, some water evaporates, and that phase change carries heat away. Effective, but it consumes fresh water continuously. But Firwater is a closed loop because the same coolant keeps circulating through sealed pipes: it absorbs heat from the chips, releases that heat through radiator-like fins, then flows back to the chips again. For Wisconsin Fairwater, Microsoft says more than 90% of the facility uses closed-loop liquid cooling, while the remaining portion uses outside air and switches to water only on the hottest days. ---- From "Microsoft" YouTube channel, (link in comment)

译微软CEO萨提亚·纳德拉在Build 2026上介绍了威斯康星州Fairwater AI数据中心。该设施采用闭环直接芯片液体冷却，冷却液一次性注入后可零水耗运行，年日用水量约等于一家餐厅。超过90%设施使用闭环液冷，仅最热天切换部分外部空气冷却。数据中心采用垂直两层架构，三维密集部署GPU，保持低延迟与高带宽网络，集群如同一台巨型AI机器。

Satya Nadella@satyanadella · 6月3日77

Building a frontier intelligence ecosystem together. Highlights from my keynote at Microsoft Build this morning.

译与我们共同构建前沿智能生态系统。 今早我在 Microsoft Build 上的主旨演讲亮点。

Chubby♨️@kimmonismus · 6月3日64

http://x.com/i/article/2061993838718382080 # What a day. OpenAI turns Codex into a work platform, Microsoft ships an entire agent stack. 06/02/26 What a day. OpenAI turns Codex into a work platform, Microsoft ships an entire agent stack. 06/02/26 recap. Lets start with OpenAI, because it's bigger than one number. OpenAI is recasting Codex from a coding tool into a productivity app for everyone. Today they launched six role-specific plugins that make Codex useful without writing a line of code, from data analytics (Snowflake, Databricks, Tableau) to creative production (Figma, Canva, Shutterstock). 62 apps and 110 skills bundled in. Plus Codex Sites: in preview, Codex can now build interactive, hosted websites and apps (dashboards, planners, review workspaces) and share them by link across a workspace. This is the groundwork for merging ChatGPT, Codex, and the Atlas browser into one desktop app. The numbers behind it are exciting. Per an internal all-hands (via The Information): 5 million weekly Codex users, enterprise revenue up 50% week over week, usage growing 5% a day. And GPT-5.6 is already on the horizon. https://x.com/kimmonismus/status/2061961710823686489 On top of that, the milestone: the ChatGPT app crossed 1 billion monthly active users - the fastest app in history to that mark, in three years. Maps, YouTube, and TikTok each needed five to eight. But the main event was Microsoft Build 2026 in San Francisco. Three hours of Nadella, and the message was clear: Microsoft no longer just resells OpenAI, it ships its own. 7 in-house MAI models. Headlined by MAI-Thinking-1, Microsoft's first reasoning model, trained entirely on licensed data with no distillation from GPT. 35B active parameters, 256k context, and in Microsoft's own blind tests it beats Claude Sonnet 4.6 and matches Opus 4.6 on coding. Plus MAI-Code-1-Flash (rolling out to all GitHub Copilot tiers today), image models (already live in PowerPoint), transcription across 43 languages, and a new voice model. Suleyman claims one is 10x more efficient than GPT-5.5. The full enterprise / agent stack — this is where Build really lived: - GitHub Copilot app (preview): a native desktop app bringing agentic workflows out of the IDE, alongside a new GitHub Copilot CLI for the command line. https://www.youtube.com/watch?v=mv6MMQ2j128&source_ve_path=MjM4NTE&embeds_referring_euri=https%3A%2F%2Fgithub.blog%2F - Microsoft IQ (GA): the unified context layer for agents, combining Work IQ (workplace knowledge inside the M365 trust boundary), Fabric IQ (business semantics), Foundry IQ (enterprise knowledge + retrieval), and the new Web IQ (live web grounding that already powers Copilot and ChatGPT). Build once, reuse across GitHub Copilot, Foundry, and Copilot Studio. - Microsoft Foundry as the agent factory: Hosted Agents with sub-100ms sandbox cold starts and zero idle cost, Toolboxes, tracing and evals, an Agent Optimizer, and one-click publishing of any agent straight into Teams and Microsoft 365 Copilot (GA June). Fireworks AI's open models also went GA on Foundry. - Agent 365: the framework-agnostic SDK went GA (free, supports Microsoft Agent Framework, OpenAI Agents SDK, LangChain, Semantic Kernel). Local Agents (preview) can even discover agents like Claude Code and GitHub Copilot CLI on managed endpoints, and Microsoft 365 E7 now bundles Agent 365 with E5, Copilot, and Entra. - Project Rayfin (preview): a managed backend-as-a-service on Fabric, so developers can take agentic apps from prototype to production. - Azure Agent Mesh (announced, GA Q4): a control plane that federates agent execution across machines and geographies. - Project Solara — Microsoft's bet on agent-first hardware. A chip-to-cloud platform built from the ground up for devices that run AI agents instead of apps. It's based on a fork of Android (the Microsoft Device Ecosystem Platform, MDEP) rather than Windows, with enterprise security baked in via Intune, Entra ID, and Windows Hello, plus "just-in-time UI" that reshapes itself to whatever device it's running on. Microsoft showed two reference designs (not products it plans to ship itself) - A desk companion that signs you in with facial recognition, responds to voice, and surfaces your most pressing items from Outlook, Excel, and M365. Plug in a monitor and it becomes a full cloud-hosted Windows machine. - The agent handheld / wearable badge - a reimagined employee ID card. A fingerprint button wakes an agent in one press, a single tap records and transcribes a conversation, and a built-in camera lets the agent act on what you're looking at. Fully mobile with 5G and a touchscreen. - Microsoft Discovery (GA): an agentic platform for scientific research, already used by BHP, GSK, and Syensqo. Plus Frontier Tuning (private preview), which lets agents learn your business inside your compliance boundary -OpenClaw comes to Windows. Peter Steinberger — the "ClawFather" — was actually on stage. His viral open-source agent (one of the most-starred GitHub projects ever, now MIT-licensed under a foundation) now runs natively on Windows through Microsoft's new containment layer. The live demo leaned into the obvious anxiety: someone asked OpenClaw to wipe a messy desktop, and it couldn't, because its container was set to read-only. Microsoft is promising "very granular" control over what files an agent can touch. The theme over all of it: Microsoft is recasting Windows, Azure, GitHub, and M365 as the operating environment for agents — moving developers from writing code to orchestrating systems of agents. Surface RTX Spark Dev Box. A mini workstation on NVIDIA's new RTX Spark superchip: 1 petaflop of AI compute, 128GB unified memory, running 120B-parameter models locally with a 1M-token context. No cloud call. A direct shot at per-token pricing. Mayo Clinic. Microsoft and Mayo are building a frontier model for healthcare. Mayo owns it; long-term it's meant to support clinicians and improve how Copilot answers health questions. Majorana 2. The new quantum chip, with claims of 1,000x higher reliability and a commercial quantum machine by 2029. Caveat: the claims rest on a non-peer-reviewed preprint, and independent physicists are openly skeptical. I wouldn't celebrate this one uncritically. Copilot Super App? Teased, not shown. Nadella said Chat, Cowork, and Code would land in one Copilot app "come summer." The through-line on both sides: nobody's selling models anymore. OpenAI is turning Codex into the operating system of work; Microsoft is turning its whole stack into an agent platform. 2026's race is officially a platform race.

译OpenAI将Codex从编码工具升级为生产力平台，新增六大角色插件并集成62个应用，还推出Codex Sites功能。关键数据显示，Codex周活达500万，企业收入周环比增长50%，GPT-5.6即将发布，ChatGPT月活已突破10亿。微软在Build 2026大会上发布完整智能体栈，推出自研推理模型MAI-Thinking-1（35B参数，256k上下文），其在编码盲测中超越Claude Sonnet 4.6。同时，微软发布了Agent 365、GitHub Copilot桌面应用、Microsoft IQ上下文层，并公布了专为智能体设计的硬件项目Solara。

Satya Nadella@satyanadella · 6月3日82

With the new MAI models and Frontier Tuning capabilities we announced today, we're focused on helping every company move from just consuming a frontier model to fully participating at the frontier.

译凭借我们今天宣布的全新MAI模型和前沿调优能力，我们致力于帮助每家公司从仅仅使用前沿模型，转变为全面参与前沿领域。

Berryxia.AI@berryxia · 6月3日74

老树开新花了，这个老大哥微软今天发布新模型了😄 刷一波存在感哈哈哈，不然都没有人记得了~ Microsoft AI今天直接甩出七个全新MAI模型。 官方说：不是简单迭代，而是从零开始、干净数据血统、零蒸馏训练的一整个家族。 MAI-Thinking-1主推理、MAI-Code-1-Flash主编码、MAI-Image-2.5主图像、MAI-Transcribe-1.5主转录、MAI-Voice-2主语音，还有各自的Flash版本。 最狠的是MAI-Code-1-Flash，直接在SWE-Bench Verified上干到71.6，比Claude Haiku 4.5高5分，Pro榜单高16分，还省60% token，现在已经在Copilot里逐步上线。 MAI-Image-2.5在Arena图像编辑排第二、文本生图排第三，精准保留人脸、logo和细节，已经直接塞进PowerPoint和OneDrive。 MAI-Transcribe-1.5在43种语言上同时拿准度和速度第一，一小时音频15秒搞定。 MAI-Voice-2能控情绪、支持多语言code-switching，长内容说话人身份也稳。 它们不是各自为战，而是设计成一个能无缝协作的家族。Microsoft这次没玩“一个大模型通吃”，而是把每个任务拆开，用干净数据从头训，公开所有技术细节和学习心得。 这其实把行业当前最主流的路径反过来了。 大家都在卷参数规模、卷蒸馏别人家的输出，Microsoft却在说：真正长期有竞争力的，是从零构建、血统干净、任务专精、还能互相配合的模型家族。 实际效果如何，其实还有待大家的测试~~期待看看实际表现！

译微软在Build大会宣布推出七个全新的MAI模型家族。该家族以“干净数据血统”从零开始训练，旨在任务专精并能无缝协作。其中，MAI-Code-1-Flash在SWE-Bench Verified上得分71.6，比Claude Haiku 4.5高出5分，并能节省60% token。MAI-Transcribe-1.5处理一小时音频仅需15秒，在43种语言上实现速度与准度领先。微软此次发布旨在展示其从零构建、专精且能协同工作的模型发展路径。

Berryxia.AI@berryxia · 6月3日64

微软的新模型MAI-Image-2.5 在图像编辑中斩获第二名的位置。 那么可以看出来还是GPT-Image-2 最强，第一！ Google 的Nano Banana 模型都已经被微软的MAI超越了…… Google 老大哥能不能整点新活儿出来啊，Pro会员都要到期了…

译微软发布新模型MAI-Image-2.5，并在Image Edit Arena（单图编辑）评测中取得第二名，得分为1401。根据评测数据，该模型分数比Nano Banana 2、Grok Imagine Image Quality和ChatGPT-Image-Latest-High Fidelity高出10分。尽管取得了进步，但评测显示当前的第一名仍是GPT-Image-2模型。该消息来源于X用户@berryxia。

Microsoft Research@MSFTResearch · 6月3日27

Day 1 of Microsoft Build is a wrap. A lot happened today, and we brought some of our favorite tools to the floor for developers to get hands-on with. Explore it all: https://msft.it/6019vjO9D

译微软Build大会第一天圆满结束。今天发生了许多事，我们带来了一些最受喜爱的工具供开发者亲身体验。探索全部内容：https://msft.it/6019vjO9D

SemiAnalysis@SemiAnalysis_ · 6月3日50

IMPORTANT: it is important to understand that the CoreWeave & Microsoft photos are still Engineering/Quality Samples, and there is still some time before the software stack bring-up finishes & first production tokens are generated. The VR200 & MI455 rack metric to watch out for is time to first at-scale production token TTF-(ASP)-T. You can clearly see in the CW rack photos that none of the scale-out 800G OSFP cages are even populated.

译重要提示：需理解CoreWeave与微软的机架照片仍为工程/质量样品，距离软件栈启动完成并产出首批生产token尚需时日。VR200与MI455机架的关键指标是达到规模化生产token的时间，即TTF-(ASP)-T。从CW机架照片中可清晰看到，所有横向扩展的800G OSFP笼位均未安装模块。

meng shao@shao__meng · 6月3日72

Microsoft Build 一口气发布了 7 个模型！ 微软，最后再信你一次 (1)(1)(1)(1)(1)(1)(1) 😄

译微软Build大会一口气发布了7个模型！ 微软，最后再信你一次 (1)(1)(1)(1)(1)(1)(1) 😄

小互@xiaohu · 6月3日64

收到Mac mini被开发者追捧的吸引 微软发布了一台类似Mac mini的 台式机： Surface RTX Spark Dev Box 它是一个小盒子，放在桌上就行 配置了英伟达最新的 RTX Spark 芯片，128GB 内存，算力达到 1 petaflop（1000 万亿次运算），能在本地跑 1200 亿参数的大模型，不用连云端 GPU。 外观看起来像一个"压扁的 Xbox Series X"，顶部有类似的散热格栅，只是通风孔是方形的而不是圆形的。整个机身是阳极氧化铝 3D 打印的，顶部有 1000 个通风孔。 定位：给开发者在本地跑 AI 模型、Agent 工作流、模型微调用的，不用什么都往云上送，省钱也快 开箱即用：预装了开发者版 Windows 11 Pro，VS Code、GitHub Copilot、WSL、PowerShell 7 都配好了，开机就能写代码 散热：整个铝合金机身就是散热系统，100W 功耗，顶部有 1000 个通风孔，能扛长时间训练任务不降频 价格：官方还没公布，行业分析师估计在 3000 到 3500 美元之间，同类产品 AMD Ryzen AI Halo PC 和 NVIDIA DGX Spark 大约卖 3999 美元 今年晚些时候在美国上市...

译微软推出Surface RTX Spark Dev Box，一款专为本地AI开发的小型台式机。它搭载NVIDIA RTX Spark芯片、128GB内存，算力达1 petaflop，可在本地运行1200亿参数大模型。其阳极氧化铝机身集成了散热系统，功耗100W。设备预装了开发者版Windows 11 Pro及开发工具链，预计售价3000至3500美元，将于今年晚些时候在美国上市。

小互@xiaohu · 6月3日60

微软宣布 将OpenClaw 引入 Microsoft 和 Windows 生态系统 小龙虾现在可以在 Windows 上原生运行，使用了微软新推出的 MXC安全容器技术，node 和 gateway 都在容器内运行。 Windows 还提供了一个配套应用（companion app），可以直接设置和连接 Claws。 同时微软在 Build 2026 上发布了 Microsoft Scout，这是一个基于 OpenClaw 的"始终在线"（always-on）个人 AI Agent 能连接 Teams、Outlook、OneDrive、SharePoint，在后台自动执行协调工作。 微软把这类 Agent 称为"Autopilots"。 微软没有自己另起炉灶做一个封闭的 Agent 框架，而是直接在 OpenClaw 仓库上构建 Scout，并承诺把企业级的策略控制能力贡献回上游开源项目。 之前 OpenClaw 最大的企业落地障碍就是安全，公司不敢让一个开源 Agent 随便访问内部系统。现在微软把 Defender、Entra、Intune 这套企业安全栈全接上了，等于替 OpenClaw 补了最大的短板。

译微软宣布将OpenClaw引入Windows生态，使其可通过MXC安全容器技术原生运行，并提供配套应用进行设置。同时，微软在Build 2026上发布了基于OpenClaw的“始终在线”个人AI智能体Microsoft Scout，可连接Teams、Outlook等应用自动执行任务。微软没有构建封闭框架，而是承诺将企业级策略控制能力贡献回OpenClaw开源项目，并通过接入Defender、Entra等安全栈，解决了其在企业落地的安全障碍。

Satya Nadella@satyanadella · 6月3日74

With Project Solara, we are building a new platform purpose-built for agent-first devices. Excited to work with @cristianoamon and @Qualcomm on this!

译通过Project Solara，我们正在构建一个专为智能体优先设备打造的新平台。 很高兴能与@cristianoamon和@Qualcomm合作！

Microsoft Research@MSFTResearch · 6月3日72

Weather forecasts thousands of times faster than traditional supercomputers. Hear from Kenji Takeda on Aurora at the Microsoft Research Lab at #MSBuild. Learn more: https://msft.it/6018vjGUA

译天气预报速度比传统超级计算机快数千倍。听听Kenji Takeda在#MSBuild微软研究实验室关于Aurora的分享。了解更多：https://msft.it/6018vjGUA

Rohan Paul@rohanpaul_ai · 6月3日81

Microsoft unveiled MAI-Thinking-1. So Microsoft now has a full in-house pipeline for building stronger reasoning models again and again. Microsoft calls this system a “hill-climbing machine,” meaning it keeps improving the data, training setup, rewards, safety tests, and evaluations as one connected process. Strong for its size, including 97.0% on AIME 2025, 87.7% on LiveCodeBench v6, and 52.8% on SWE-Bench Pro. MAI-Thinking-1 is the first model from that process, using 35B active parameters inside a 1T total parameter mixture-of-experts model, where only part of the model runs for each token. The base model was trained from scratch on 30T mostly human-generated tokens, with Microsoft saying it avoided third-party model distillation during pre-training. After that, the team used reinforcement learning, which means the model practiced tasks and improved from feedback, to teach math reasoning, coding, tool use, helpfulness, and safety.

译微软发布了 MAI-Thinking-1，这是一款采用 MoE 架构的模型，拥有 35B 活跃参数和 1T 总参数。该模型从零开始在 30T tokens 上完成预训练，且未使用第三方模型蒸馏。微软称其迭代优化流程为“爬山机器”。在基准测试中，该模型于 AIME 2025 获得 97.0%，在 LiveCodeBench v6 获得 87.7%，在 SWE-Bench Pro 获得 52.8% 的成绩。

Microsoft Research@MSFTResearch · 6月3日54

Agentic experiences powered by small models that fit on your own device. Hear from Maya Murad on MagenticLite at the Microsoft Research Lab at #MSBuild.

译由可在您自己设备上运行的小型模型驱动的智能体体验。请听 Maya Murad 在 #MSBuild 微软研究院实验室介绍 MagenticLite。

Chubby♨️@kimmonismus · 6月3日50

Just figured out that „Mai“-1 thinking stands for: Microsoft AI-thinking. 🤯

译刚刚发现“Mai”-1 thinking 代表： 微软 AI 思考。 🤯

Rohan Paul@rohanpaul_ai · 6月3日63

Satya Nadella on Microsoft’s Fairwater data center, an AI superfactory. at today's Microsoft Build 2026 keynote. its vertically designed, two-story AI data center architecture. Instead of spreading compute only across a flat floor, Microsoft can place racks in three dimensions, packing far more GPUs densely while preserving fast network access. This helps the cluster behave more like one massive AI machine, with low latency and high bandwidth between GPUs. The other major point is its cooling efficiency: its cooling loop is filled once and can operate with effectively zero ongoing water consumption, using roughly the annual daily-water equivalent of a single restaurant. ---- From "Microsoft" YouTube channel, (link in comment)

译在微软 Build 2026 主题演讲中，Satya Nadella 介绍了 Fairwater 数据中心，这是一个为 AI 设计的“超级工厂”。其核心是垂直设计的双层 AI 数据中心架构，允许在三维空间内密集部署机架，在保持 GPU 间低延迟、高带宽网络连接的前提下，实现更高的计算密度，使整个集群更像一台大型 AI 机器。另一大亮点是其极高的冷却效率：冷却系统只需填充一次，实际运行中水耗几乎为零，其年度总用水量约等于一家餐厅的日用水量。这是微软构建“前沿智能生态系统”硬件基础的一部分。