Dr. Feifei Li, CTO of Alibaba Cloud & Tommy Eastman, Head of Strategy, Nous Research As we orchestrate intelligence at scale, reshaping knowledge work, giving agents autonomy requires reproducible actions—the core secret behind Hermes agent's success.

译阿里云CTO李飞飞博士与Nous Research战略主管Tommy Eastman：当我们大规模编排智能、重塑知识工作时，赋予智能体自主性需要可复现的行动——这是Hermes agent成功背后的核心秘诀。

Alibaba Cloud@alibaba_cloud · 6月3日51

Ship the Next with Dr. Feifei Li, CTO of Alibaba Cloud & President of International Business We are upgrading to be Agent Native—moving past Cloud & AI Native to a full-stack system from Qwen models to silicon, delivering the scale and economy for agent reasoning and action.

译与阿里云 CTO 兼国际业务总裁李飞飞博士一同“Ship the Next”。我们正在升级为 Agent Native——超越云原生和 AI 原生，构建从 Qwen 模型到硅片的全栈系统，为智能体的推理和行动提供规模与经济效益。

X.PIN@thexpin · 6月3日40

As AI strains power grids, China is moving data centers into space! Industrial collabs are rapidly forming across major cities, while Beijing aims to launch its first experimental satellite by 2028 to build a space-ground computing network. This gives China a crucial head start.

译随着 AI 对电网造成压力，中国正将数据中心送入太空！各大城市快速形成产业合作，北京计划在 2028 年前发射首颗试验卫星，以构建天地计算网络。这使中国占据了关键先机。

X.PIN@thexpin · 6月3日55

Alibaba, China's e-commerce and tech giant, has added Group CTO Wu Zeming to its Partnership Committee. His promotion from managing key AI initiatives signals the company's strong focus on AI. In fact, Alibaba expects its AI ARR to reach $4.2 billion by year-end.

译中国电商科技巨头阿里巴巴已将集团CTO吴泽明加入其合伙人委员会。他从管理关键AI项目晋升，标志着公司对AI的高度重视。事实上，阿里巴巴预计其AI年度经常性收入到年底将达到42亿美元。

Alibaba Cloud@alibaba_cloud · 6月3日65

The first-ever Qwen Cloud Global AI Hackathon is LIVE! Join 5 advanced tracks to push AI agent limits on a global stage. $70,000+ total prize pool ($10K for track winners). Sign up on Devpost now: https://click.qwencloud.com/m/20000000281/

译首届Qwen Cloud全球AI黑客马拉松现已启动！参与5大高级赛道，在全球舞台上挑战AI智能体极限。总奖金池超过70,000美元（赛道冠军1万美元）。立即在Devpost报名：https://click.qwencloud.com/m/20000000281/

X.PIN@thexpin · 6月3日66

DeepSeek is launching a massive initial funding round! And the most surprising figure doesn't come from Tencent. According to Reuters, the company aims to raise ~$7.4B, bringing its post-money valuation to between $52B to $59B. This would mark the largest AI funding round in China to date. Among the investors: 🔹 Founder Liang Wenfeng contributes ~$3B personally 🔹 Tencent invests ~$1.5B 🔹 Battery manyfacturer CATL invests ~$0.7B as it expands into supplying power for AI data centers 🔹 NetEase, http://JD.com, and China's national AI fund are in talks to join 🔹 Hong Kong's IDG Capital and Cornerstone Capital are also among the intended investors The deal is expected to close in about two weeks. After operating as a self-funded research lab for years, DeepSeek is finally accepting outside capital, though Liang remains the largest single investor.

译DeepSeek启动首轮大规模融资，目标募资约74亿美元，投后估值520亿至590亿美元，创中国AI行业融资纪录。创始人梁文锋个人出资约30亿美元，腾讯投资约15亿美元，宁德时代投资约7亿美元（同时布局AI数据中心供电）。网易、京东、中国国家人工智能基金等也在洽谈中。交易预计两周内完成。长期自筹资金的DeepSeek首次接受外部资本，梁文锋仍是最大单一股东。（来源：Reuters报道）

Berryxia.AI@berryxia · 6月3日40

字节的Coze 可能是第一个和线下实体联动搞AI的产品吧！刚刚接电话，扭头看到在唐宁书店居然和Coze 做了联动宣传的活动。不知道实体店多少人知道Coze的产品可以做什么用？🤔

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。

Alibaba Cloud@alibaba_cloud · 6月3日28

Join the Qwen & @ModelScope2022 communities in Singapore on June 10! An evening for AI developers covering agent monetization, Qwen updates, the Global AI Hackathon, and sharing sessions from fellow builders. Apply: https://luma.com/4x2srooq #Qwen #ModelScope

译加入Qwen与@ModelScope2022社区，6月10日在新加坡！一场AI开发者晚间活动，涵盖智能体变现、Qwen更新、全球AI黑客马拉松以及来自同行构建者的分享环节。申请：https://luma.com/4x2srooq #Qwen #ModelScope

Alibaba Cloud@alibaba_cloud · 6月3日30

Wan 2.7 is now available on @AskVenice! Your credits go further on every generation.

译Wan 2.7现已在@AskVenice上线！您的积分在每次生成中能发挥更大价值。

Alibaba Cloud@alibaba_cloud · 6月3日46

The AI conversation is evolving from technology breakthroughs to business outcomes. At our recent Alibaba Cloud x TiDB AI Innovation Night, supported by Lumen Technologies, we explored how enterprises can build the foundations for the AI era, covering topics such as intelligent agents, AI-ready data infrastructure, and strategies for achieving measurable AI ROI. Thank you to our customers, partners, and speakers for an evening of insightful discussions and meaningful exchanges. #AlibabaCloudSG #AIInnovation #EnterpriseAI #AgenticAI #Qwen #CloudInfrastructure

译阿里云在新加坡联合 TiDB 举办了 AI 创新之夜活动，探讨企业如何构建 AI 时代的基础设施。讨论核心已从单纯的技术突破转向衡量可实现的 AI 投资回报率（ROI）。活动议题涵盖了智能体、AI-ready 的数据基础设施等方向。推文未提及具体的模型版本号、参数或性能分数等量化信息。

歸藏(guizang.ai)@op7418 · 6月3日49

看来快手确实要分拆可灵单独上市了了。其实早就该搞，太晚了现在。

MiniMax (official)@MiniMax_AI · 6月3日29

🏆 Build with Minimax-M3 → win a Token Plan Plus Our official Discord Showcase Event is live! 🎉 Create a case or project with Minimax-M3, post it in our showcase forum, and you could be 1 of 10 winners. How to enter: 1. Join the Discord 👉 https://discord.gg/mmx 2. Build something with Minimax-M3 3. Post it in the showcase forum + tag Minimax-M3 4. Add a short note on how you used it 🗓 June 2–8, 2026 🎁 10 winners, announced June 9 in our official Discord #MinimaxM3 #AI

译🏆 使用 Minimax-M3 构建 → 赢取 Token Plan Plus 我们的官方 Discord 展示活动现已开始！🎉 使用 Minimax-M3 创建案例或项目，发布到我们的展示论坛，你就有机会成为 10 位获奖者之一。参与方式： 1. 加入 Discord 👉 https://discord.gg/mmx 2. 使用 Minimax-M3 构建内容 3. 发布到展示论坛 + 标记 Minimax-M3 4. 添加简短说明，介绍你如何使用它 🗓 2026年6月2日至8日 🎁 10位获奖者将于6月9日在官方 Discord 公布 #MinimaxM3 #AI

Berryxia.AI@berryxia · 6月3日29

Codex 刚刚遇到多次 exceed retry limit 429的错误，好像挺多人遇到，这又是闹哪出？

Alibaba Cloud@alibaba_cloud · 6月3日47

Narek Hayrapetyan, Head of Video Product, Picsart Launching HappyHorse and Wan models led to a 72% surge in video generations, with users creating over 1M monthly assets and using 6M+ Picsart credits.

译Narek Hayrapetyan，Picsart视频产品负责人推出HappyHorse和Wan模型后，视频生成量激增72%，用户每月创建超过100万个资产，并使用超过600万Picsart积分。

jason@jxnlco · 6月3日39

We’re aware of another codex issue with too many requests.

译我们已知悉另一个 Codex 问题，即请求过多。

Alibaba Cloud@alibaba_cloud · 6月3日36

Desmond Tan, Senior Minister of State, Prime Minister's Office, Deputy Secretary-General, National Trades Union Congress (NTUC) Alibaba Cloud's Qwen, Wan, and Qoder demonstrate how AI applies to real-world work, from development to content creation.

译Desmond Tan，新加坡总理公署高级政务部长、全国职工总会副秘书长阿里云的Qwen、Wan和Qoder展示了AI如何应用于实际工作，从开发到内容创作。

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。

Alibaba Cloud@alibaba_cloud · 6月3日62

Manulife Hong Kong announced a strategic partnership with Alibaba Cloud to establish a collaboration framework focused on advancing responsible AI innovation and accelerating the deployment of AI technologies across the business.

译宏利香港宣布与阿里云建立战略合作伙伴关系，旨在构建一个专注于推进负责任的AI创新，并加速AI技术在业务中部署的合作框架。

Berryxia.AI@berryxia · 6月3日12

笑死，这也算是应援上了。 AI圈可不兴这个啊~~😄 今晚就正式比赛，到时候给大家汇报同步我的Agent的亏损情况😄 感谢朋友们的支持啊~

译AI Agent 不应只消耗 Token，更要尝试赚钱。用户 Berryxia 的 AI Agent “Berry” 参加了 Bloome 官方举办的 Agent 模拟炒股大赛，奖品达数千美金。Berryxia 称其 Agent 已训练得“听书读写，样样精通”，并在比赛中寻求观众投票支持，承诺若获奖将在 X 上抽奖回馈支持者。

Huawei Cloud@HuaweiCloud1 · 6月3日38

Ready for the next wave of AI? Don't miss out Huawei Cloud INSPIRE 2026. Discover new products and innovations shaping the future of AI. See you there! Learn More: https://tinyurl.com/dcyx8bw7 #INSPIRE2026 #HuaweiCloud

译准备好迎接下一波 AI 浪潮了吗？不要错过华为云 INSPIRE 2026。探索塑造 AI 未来的新产品与创新。届时见！了解更多：https://tinyurl.com/dcyx8bw7 #INSPIRE2026 #HuaweiCloud

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。

Rohan Paul@rohanpaul_ai · 6月3日50

Sam Altman reveals that OpenAI’s top “token leader” uses 100B tokens every month, and still falls short of the world’s highest user. Source: axios --- axios. com/2026/06/02/altman-openai-top-token-user

译Sam Altman透露，OpenAI的顶级“token用户”每月使用1000亿token，但仍不及全球最高用户。

Rohan Paul@rohanpaul_ai · 6月3日52

Uber is limiting AI coding tool spending to $1,500 per employee each month. Uber’s CEO said last month that AI agents now submit and build roughly 10% of its code, while legal and marketing teams are also warming up fast to generative-AI tools. --- finance. biggo. com/news/M36WiZ4BLfE1EzqPygfr

译Uber将每位员工每月的AI编程工具支出限制在1500美元。 Uber的首席执行官上个月表示，AI智能体现在提交并构建了大约10%的代码，而法律和营销团队也正在快速适应生成式AI工具。

ginobefun@hongming731 · 6月3日49

#BestBlogs 早报 06-03 BestBlogs 今日早报推荐阅读： Anthropic 博客详解 Claude Code 动态工作流，Claude 能为每个任务即时生成专属编排脚本，告别「智能体懒惰」和「目标漂移」； GitHub 在 Build 同步亮相 Copilot 桌面应用，每个 Agent 独占 worktree、提交量已破 14 亿/月。腾讯云工程师则从控制论视角点出：大模型是史上首个「认知引擎」，工程师的核心职责正在从「写代码」升级为「设计能自我纠偏的 AI 系统」。

译Anthropic 详解 Claude Code 的动态工作流，其能为每个任务即时生成专属编排脚本，旨在解决智能体懒惰与目标漂移问题。GitHub 发布 Copilot 桌面应用，为每个智能体提供独立的 worktree，其月代码提交量已突破 14 亿 tokens。此外，有观点指出大模型是史上首个“认知引擎”，工程师角色正从编写代码升级为设计能自我纠偏的 AI 系统。

DogeDesigner@cb_doge · 6月3日49

Elon Musk just got featured on Forbes’ inaugural Iconoclast 50 2026 list Forbes highlights Elon for his impact across electric vehicles at Tesla, spaceflight at SpaceX, global connectivity through Starlink, and frontier AI at xAI.

译马斯克入选福布斯首届“Iconoclast 50”2026年榜单。福布斯指出，马斯克的影响力跨越多个前沿领域，包括特斯拉的电动汽车、SpaceX的太空探索、Starlink的全球卫星互联网以及xAI的人工智能。引用内容进一步强调，鲜有高管能像他一样在如此多的行业中留下足迹，其公司持续塑造着关于技术未来的讨论。

MiniMax (official)@MiniMax_AI · 6月3日44

MiniMax M3 now on GMI Cloud @gmi_cloud

译MiniMax M3 现已登陆 GMI Cloud @gmi_cloud

Anthropic@AnthropicAI · 6月3日69

This Executive Order is an important step in strengthening America’s leadership in AI. We look forward to collaborating with the White House to support its implementation. https://www.whitehouse.gov/presidential-actions/2026/06/promoting-advanced-artificial-intelligence-innovation-and-security/

译这项行政令是加强美国AI领导地位的重要一步。我们期待与白宫合作，支持其实施。 https://www.whitehouse.gov/presidential-actions/2026/06/promoting-advanced-artificial-intelligence-innovation-and-security/

MiniMax (official)@MiniMax_AI · 6月3日57

Amazing deep dive from the @togethercompute team on serving MiniMax M3 in production. M3 with its 1M context, native multimodality and MiniMax Sparse Attention requires real work across paged decode, index scoring, and multimodal preprocessing to get it efficient. This is what a partnership at the frontier looks like🤝.

译@togethercompute 团队对 MiniMax M3 生产部署的精彩深度解析。 M3 凭借其 1M 上下文、原生多模态和 MiniMax Sparse Attention，需要在分页解码、索引评分和多模态预处理方面进行大量工作才能实现高效运行。这就是前沿合作的样子🤝。

Chubby♨️@kimmonismus · 6月3日36

Ok what? Same Training FLOPs as Gemini 3.1 pro?

译什么？训练FLOPs和Gemini 3.1 Pro一样？

Artificial Analysis@ArtificialAnlys · 6月3日49

We’re hosting a Coding Agent Benchmarks event on Thursday, June 11 in San Francisco with lightning talks and a panel discussion with leading AI researchers, builders, and engineers. If you're building coding agents, LLM tooling, or AI infrastructure, we’d love to see you there! Request to join 👇 https://luma.com/i5zotp6c

译我们将于6月11日星期四在旧金山举办一场编程智能体基准测试活动，包含闪电演讲以及与顶尖AI研究人员、开发者和工程师的小组讨论。如果你正在开发编程智能体、LLM工具或AI基础设施，我们很期待你的到来！申请加入 👇 https://luma.com/i5zotp6c

Satya Nadella@satyanadella · 6月3日31

Great to be back at Microsoft Build today. For us, it is not about any one piece of technology or even the platform. It is about how we can build a frontier intelligence ecosystem together. Sharing some of our big announcements today ...

译很高兴今天回到微软 Build 大会。对我们来说，这不仅仅是关于某一项技术，甚至不仅仅是平台。而是关于我们如何共同构建一个前沿智能生态系统。分享一些我们今天的重要公告……

Microsoft Research@MSFTResearch · 6月3日44

Microsoft Research is at BUILD 2026 this week, giving developers a hands-on look at some of the many AI-based models and tools they can use to accelerate innovation, enhance their capabilities, and quickly transform ideas into prototypes. https://msft.it/6010vjBUe

译微软研究院本周参加BUILD 2026，让开发者亲身体验众多基于AI的模型和工具，以加速创新、增强能力，并快速将想法转化为原型。https://msft.it/6010vjBUe

Epoch AI@EpochAIResearch · 6月3日46

We’re running a short survey to ensure we’re producing the most useful work on AI's trajectory. If you haven’t yet, we'd love your input.

译我们正在进行一项简短调查，以确保我们能产出关于 AI 发展轨迹最有价值的工作。如果您尚未参与，我们很乐意听取您的意见。（您可以在问卷末尾注册，加入我们的有偿用户研究小组。）

swyx@swyx · 6月3日38

uhhh did Mustafa just leak the Mythos FLOP count?? was this public knowledge before, even if its an estimate i dont get what you gain out of this

译Mustafa是不是刚泄露了Mythos的FLOP数据？？这之前是公开信息吗，就算是个估计值我也不明白这有什么好处

Microsoft Research@MSFTResearch · 6月3日50

Welcome to Microsoft Build 2026. Step inside the Microsoft Research Lab, where developers are getting hands-on with the latest from MSR. Aurora’s forecasting models, Trellis 3D generation, Magentic UI, Mindful Memory, and more.

译欢迎来到 Microsoft Build 2026。走进微软研究院实验室，开发者们正在亲身体验来自MSR的最新成果。Aurora的预测模型、Trellis 3D生成、Magentic UI、Mindful Memory等。

Rohan Paul@rohanpaul_ai · 6月3日36

Boston Dynamics’ Spot is patrolling World Cup venues in Dallas, using 360° cameras, thermal imaging, and chemical sensors to detect suspicious packages, scan surroundings, and support security teams live. No facial recognition capability.

译波士顿动力的Spot机器人正在达拉斯的世界杯场馆巡逻，使用360度摄像头、热成像和化学传感器来检测可疑包裹、扫描周围环境，并实时支持安保团队。不具备面部识别能力。

jason@jxnlco · 6月3日43

course guy in search of education lead https://openai.com/careers/developer-education-lead-san-francisco/

译课程专家寻找教育负责人 https://openai.com/careers/developer-education-lead-san-francisco/

SemiAnalysis@SemiAnalysis_ · 6月3日52

Major hyperscalers introduce a variety of GPU, XPU, and CPU chips, leading to more diverse server rack and board designs. To meet different customer needs, ODMs need larger design teams, along with more assembly and testing capacity. ODMs now have a good opportunity to reach new customers and expand their client base. As AI server designs become more customized, early engagement and customer trust will be key factors in securing long-term business. AI server assembly requires heavy upfront investment in capacity and equipment. Vendors with stable customers are better positioned to support expansion and capture long-term growth.

译超大规模云服务商引入多样化的GPU、XPU和CPU芯片，导致AI服务器机架和板卡设计更加多元化。为满足不同客户需求，ODM厂商需要扩大设计团队规模，并增加组装与测试产能。这为ODM提供了接触新客户、拓展客户群的良机。随着AI服务器设计日趋定制化，早期介入与客户信任将成为赢得长期业务的关键。