Scroller Briefings -Lenovo’s “Xiang Bang Bang” Launches Free OpenClaw Deployment Service, Covering More Than 2,000 Stores Nationwide 🚀
编辑
目录
单词学习：![[2026031549657982.mp3]]
新闻播报：![[2026031508987625.mp3]]
联想“想帮帮”上线 OpenClaw 免费部署服务，涵盖全国 2000 余家门店
简报：
- 联想 AI 服务“想帮帮”宣布从 3 月 16 日起推出 OpenClaw 免费部署服务，支持线上一键部署及线下指定门店预约安装。
- 线下部署服务覆盖全国 2000 多家门店，自 3 月 15 日 21:00 起开放预约，每日限量提供 1000 个免费名额，由工程师协助完成配置。
- 部署后的 OpenClaw 内置 51 个 Skills 技能包，提供基础聊天、编程辅助、图片生成、语音转文字等 AI 常用功能。
- 此次活动正值国内“养虾”（OpenClaw 部署）热潮，百度、腾讯、阿里巴巴等科技巨头近期均已推出各自的 OpenClaw 简化部署工具或线下支持服务。
- 除免费渠道外，联想百应还与京东、美团等平台合作，为用户提供 OpenClaw 的远程付费部署与技术维护支持。
相关链接：
Lenovo’s “Xiang Bang Bang” Launches Free OpenClaw Deployment Service, Covering More Than 2,000 Stores Nationwide 🚀
Brief:
- Lenovo’s AI service “Xiang Bang Bang” announced that, starting March 16, it will launch a free OpenClaw deployment service, supporting one-click online deployment as well as appointment-based installation at designated offline stores.
- The offline deployment service will cover more than 2,000 stores across China. Reservations will open at 21:00 on March 15, with 1,000 free slots available per day on a limited basis, and engineers will assist users in completing the setup.
- After deployment, OpenClaw comes with 51 built-in Skills packages, offering commonly used AI features such as basic chat, programming assistance, image generation, and speech-to-text.
- The campaign comes amid a domestic “shrimp farming” (OpenClaw deployment) boom 🦐, as major tech companies including Baidu, Tencent, and Alibaba have recently launched their own simplified OpenClaw deployment tools or offline support services.
- In addition to the free options, Lenovo Baiying has also partnered with platforms such as JD.com and Meituan to provide users with paid remote OpenClaw deployment and technical maintenance support 🤖.
Related link:
性能测试逊于竞品，Meta 新一代大语言模型 Avocado 推迟至 5 月发布
简报：
- Meta 已将其代号为“Avocado”的新一代大语言模型（即 Llama 4）发布时间从原定的 3 月至少推迟至 5 月。
- 内部测试显示，该模型在逻辑推理、编程和写作性能方面虽优于前代及谷歌 Gemini 2.5，但仍逊于 OpenAI、Anthropic 以及谷歌最新的 Gemini 3.0。
- 由于研发进度受挫，Meta 高层已讨论考虑临时授权使用谷歌的 Gemini 模型来为其 AI 产品提供动力，但尚未达成最终决定。
- 为支撑庞大的 AI 算力需求，Meta 宣布了四款自研芯片（MTIA 系列）的路线图，并预计 2026 年资本支出将达到 150 亿至 350 亿美元。
- 有消息称 Meta 正考虑裁减 20% 或更多的员工，以抵消巨大的 AI 基础设施投入带来的资金压力。
相关链接：
Outperformed by Rivals in Performance Tests, Meta’s Next-Generation Large Language Model Avocado Delayed to May 🚧🤖
Brief:
- Meta has postponed the release of its next-generation large language model, codenamed “Avocado” (i.e., Llama 4), from the originally planned March launch to at least May.
- Internal testing shows that while the model outperforms its predecessor and Google’s Gemini 2.5 in logical reasoning, coding, and writing, it still lags behind OpenAI, Anthropic, and Google’s latest Gemini 3.0.
- Due to setbacks in development progress, Meta executives have discussed the possibility of temporarily licensing Google’s Gemini models to power its AI products, though no final decision has been made.
- To support massive AI computing demands, Meta has announced a roadmap for four in-house chips (the MTIA series) and expects capital expenditures in 2026 to reach between dJyH6kDkfC5RlM16.answerText5 billion and $35 billion.
- Reports indicate that Meta is considering cutting 20% or more of its workforce to offset the financial pressure caused by heavy AI infrastructure investment 💸.
Related links:
- Meta Delays Release of New AI Model Over Performance Concerns
- Meta Unveils Plans for Four In-House AI Chips to Support Infrastructure Expansion
- Meta Considers 20% Job Cuts Amid Huge AI Spending and Project Setbacks
宝马德国莱比锡工厂正式引入AEON人形机器人参与汽车制造
简报：
- 继在美国斯帕坦堡工厂成功试点Figure 02机器人后，宝马在德国莱比锡的iFACTORY工厂正式引入由Hexagon Robotics生产的AEON人形机器人。
- AEON机器人身高1.65米，重60公斤，采用轮式腿部设计，主要负责电动汽车“打螺丝”、高压电池组装和零部件生产等高重复性及高精度任务。
- 该机器人搭载基于AI的运动控制系统与22个环境感知传感器，具备360度实时空间感知能力，能够自主评估周围环境、做出决策并避开路障。
- 此次部署是宝马“物理AI”战略的重要一步，机器人此前已通过实验室评估阶段，并计划于2026年夏季在工厂完成全面的试点整合。
相关链接：
BMW Officially Introduces AEON Humanoid Robots into Vehicle Manufacturing at Its Leipzig Plant in Germany 🤖🚗
Brief:
- Following the successful pilot of the Figure 02 robot at its Spartanburg plant in the United States, BMW has officially introduced the AEON humanoid robot, produced by Hexagon Robotics, at its iFACTORY plant in Leipzig, Germany.
- The AEON robot stands 1.65 meters tall and weighs 60 kilograms. Featuring a wheeled-leg design, it is primarily responsible for highly repetitive and high-precision tasks such as screw fastening for electric vehicles, high-voltage battery assembly, and component production.
- The robot is equipped with an AI-based motion control system and 22 environmental perception sensors, giving it 360-degree real-time spatial awareness. It can independently assess its surroundings, make decisions, and avoid obstacles.
- This deployment marks an important step in BMW’s “Physical AI” strategy. The robot has already passed the laboratory evaluation phase and is scheduled to complete full pilot integration at the plant by the summer of 2026. ⚙️
Related link:
谷歌上线Gemini“任务自动化”功能，手机可跨应用自主执行订餐打车
简报：
- 谷歌与三星联合宣布基于 Gemini 的“任务自动化”（屏幕自动化）功能进入 Beta 测试阶段，首批适配三星 Galaxy S26 系列及后续 Pixel 10 系列设备。
- 该功能通过在虚拟窗口中模拟真实用户操作（而非单纯 API 对接），可自主执行跨应用任务，如在 Uber 中填入目的地叫车，或在星巴克菜单中滚动寻找并添加特定单品。
- 自动化流程设有严密的“人工复核”机制：用户可实时监控 AI 的每一步操作并随时接管，且在最终支付或提交订单前，系统会强制要求用户手动点击确认。
- 目前已优先适配 Uber、Lyft、DoorDash、Starbucks 等主流外卖与网约车应用，未来计划引入 Instacart 等更多第三方 App。
- 针对不同订阅等级设置了每日请求限额：免费用户每天 5 次，AI Pro 用户 20 次，AI Ultra 用户最高可达 120 次。
- 此外，Gemini 深度集成至 Google Workspace，支持在 Docs、Sheets 和 Slides 中通过简单指令调取 Gmail 或 Drive 信息，自动生成完整的初稿或数据表格。
相关链接：
- Gemini’s task automation is here and it’s wild - The Verge
- Gemini can now order your lunch as Android app control rolls out on Galaxy S26 - 9to5Google
- Google just gave Gemini the power to control apps on the Galaxy S26 - Android Central
Google Launches Gemini “Task Automation” Feature, Allowing Phones to Order Food and Hail Rides Across Apps 🤖📱
Brief:
- Google and Samsung have jointly announced that the Gemini-powered “Task Automation” (screen automation) feature has entered the Beta testing phase, with initial support for the Samsung Galaxy S26 series and later the Pixel 10 series.
- Instead of relying solely on API integrations, the feature simulates real user actions in a virtual window, enabling it to carry out cross-app tasks autonomously, such as entering a destination in Uber to book a ride or scrolling through the Starbucks menu to find and add a specific item.
- The automation workflow includes a strict “human review” mechanism: users can monitor every step the AI takes in real time and intervene at any moment, while the system will require a manual user confirmation before any final payment or order submission.
- It currently prioritizes support for major food delivery and ride-hailing apps such as Uber, Lyft, DoorDash, and Starbucks, with plans to add more third-party apps like Instacart in the future.
- Daily request limits vary by subscription tier: free users get 5 requests per day, AI Pro users get 20, and AI Ultra users can receive up to 120.
- In addition, Gemini is now deeply integrated into Google Workspace, allowing users to pull information from Gmail or Drive with simple prompts inside Docs, Sheets, and Slides, and automatically generate full drafts or data tables ✨.
Related links:
- Gemini’s task automation is here and it’s wild - The Verge
- Gemini can now order your lunch as Android app control rolls out on Galaxy S26 - 9to5Google
- Google just gave Gemini the power to control apps on the Galaxy S26 - Android Central
Claude 插件重大升级：实现 Excel 与 PowerPoint 跨应用数据无缝协同
简报：
- Anthropic 宣布对其 Claude 的 Microsoft Excel 和 PowerPoint 加载项（Add-ins）进行重大升级，引入“共享上下文”与“可复用技能”两大核心功能。
- 此次更新打通了跨应用工作流，使 Claude 拥有统一“记忆”，能直接引用 Excel 中的财务分析逻辑在 PowerPoint 中自动生成对应的图表和总结。
- 新增的“Skills（技能）”功能允许用户将复杂、高频的工作流（如审计财务模型或生成清洗报表）保存为快捷方式，实现一键调用与后台自动化运行。
- 该功能已面向 Mac 和 Windows 的 Claude 付费用户开放，企业可通过 Amazon Bedrock、Google Cloud Vertex AI 或 Microsoft Foundry 等云基础设施部署，确保数据符合合规标准。
- 出于隐私安全考虑，所有输入输出数据将在 30 天内从后端自动删除，但目前插件仅能操作已手动打开的文件，无法自主创建、打开或关闭本地文件。
相关链接：
Major Upgrade to Claude Plug-ins: Seamless Cross-App Data Collaboration Between Excel and PowerPoint 🚀📊
Briefing:
- Anthropic has announced a major upgrade to its Microsoft Excel and PowerPoint add-ins for Claude, introducing two core features: “Shared Context” and “Reusable Skills.”
- This update connects workflows across applications, giving Claude a unified “memory” that allows it to directly apply financial analysis logic from Excel to automatically generate corresponding charts and summaries in PowerPoint.
- The new “Skills” feature allows users to save complex, high-frequency workflows—such as auditing financial models or generating cleaned reports—as shortcuts for one-click use and automated background execution.
- The feature is now available to paid Claude users on both Mac and Windows. Enterprises can deploy it through cloud infrastructure such as Amazon Bedrock, Google Cloud Vertex AI, or Microsoft Foundry, helping ensure data compliance standards are met.
- For privacy and security reasons, all input and output data will be automatically deleted from the backend within 30 days. However, the plug-ins can currently only operate on files that have been manually opened, and cannot independently create, open, or close local files. 🔐
Related link:
中国批准全球首款商业化侵入式脑机接口装置，助力瘫痪患者康复
简报：
- 国家药监局（NMPA）于2026年3月13日正式批准博睿康医疗科技（上海）有限公司研发的“植入式脑机接口手部运动功能代偿系统”创新产品注册申请。
- 该系统实现了侵入式脑机接口（BCI）医疗器械的全球首发上市，标志着该技术正式从实验室研究迈向临床商业化应用阶段。
- 产品主要针对因颈段脊髓损伤（C2-C6级）导致四肢瘫痪、且病情稳定至少6个月的18至60岁成年患者。
- 技术核心采用硬脑膜外微创植入与无线供能通信，通过采集大脑运动意图并由解码软件驱动气动手套，辅助患者实现手部抓握功能。
- 临床试验数据显示，受试者在植入该系统后，手部抓握能力及日常生活自理能力均有显著提升，填补了国际临床空白。
相关链接：
- China Approves First Brain Implant for Commercial Use
- China approves world’s first brain implant for commercialisation
- 全球首个!中国侵入式脑机接口医疗器械获批上市
China Approves the World’s First Commercialized Invasive Brain-Computer Interface Device, Supporting Rehabilitation for Paralysis Patients 🧠🤖
Briefing:
- On March 13, 2026, China’s National Medical Products Administration (NMPA) officially approved the innovative product registration application for the “Implantable Brain-Computer Interface Hand Motor Function Compensation System,” developed by Borui Kang Medical Technology (Shanghai) Co., Ltd.
- The system marks the world’s first commercial launch of an invasive brain-computer interface (BCI) medical device, signaling that the technology has officially moved from laboratory research into the stage of clinical commercialization.
- The product is primarily intended for adult patients aged 18 to 60 who suffer from quadriplegia caused by cervical spinal cord injuries (C2-C6 level) and whose condition has remained stable for at least six months.
- The core technology uses minimally invasive epidural implantation and wireless power supply and communication. By capturing motor intentions from the brain and using decoding software to drive a pneumatic glove, it helps patients regain hand grasping function.
- Clinical trial data show that after implantation of the system, subjects experienced significant improvements in both hand grasping ability and activities of daily living, filling an international clinical gap. 🙌
Related Links:
- China Approves First Brain Implant for Commercial Use
- China approves world’s first brain implant for commercialisation
- World’s First! China Approves Market Launch of Invasive Brain-Computer Interface Medical Device
绿联 NAS 原生接入 MiniMax 大模型，一键开启私有云 AI 助手服务
简报：
- 绿联科技宣布与人工智能巨头 MiniMax 达成深度战略合作，成为 NAS 行业首家原生内嵌 MiniMax 大模型的品牌，首发支持绿联 DXP 系列及即将上市的 iDX 系列设备。
- 用户可通过绿联 UGOS Pro 应用中心一键安装“OpenClaw 龙虾应用”，无需复杂配置即可调用大模型实现文档总结、创意写作和智能问答等功能，推动 NAS 从存储工具向“家庭智脑”进化。
- 配合此次上线，MiniMax 为绿联用户提供限时优惠，即日起至 2026 年 4 月 12 日，内置大模型可享受 30 天全量免费使用权限。
相关链接：
UGREEN NAS Natively Integrates the MiniMax Large Model, Enabling a Private Cloud AI Assistant Service with One Click 🤖☁️
Briefing:
- UGREEN Technology has announced a deep strategic partnership with AI leader MiniMax, becoming the first NAS brand in the industry to natively embed the MiniMax large model. Initial support will be available for UGREEN’s DXP series, as well as the upcoming iDX series devices.
- Through the UGREEN UGOS Pro App Center, users can install the “OpenClaw Lobster App” with one click. Without complicated setup, they can leverage the large model for document summarization, creative writing, intelligent Q&A, and more, accelerating the evolution of NAS from a storage tool into a “home intelligence hub” 🧠.
- To mark the launch, MiniMax is offering a limited-time promotion for UGREEN users. From now until April 12, 2026, the built-in large model will come with 30 days of full-featured free access.
Related Links:
微软发布 Copilot Health：深度整合 5 万家医疗机构数据，打造个人 AI 健康空间
简报：
- 微软正式推出 Copilot Health 平台，作为 Copilot 内的独立安全空间，旨在通过 AI 技术将用户的个人医疗记录与穿戴设备数据转化为“连贯的健康故事”。
- 该平台已实现大规模数据互通，通过 HealthEx 可调取美国超过 50,000 家医院的病历、处方及化验结果，并同步 Apple Watch、Fitbit 等 50 多种主流穿戴设备的健康指标。
- 核心功能包括智能解读体检报告、分析睡眠及运动趋势、根据保险计划和地理位置筛选匹配医生，并协助用户准备诊前咨询问题。
- 为确保信息权威性，AI 回答优先引用 50 个国家权威机构及哈佛健康的专家资料，且健康数据与普通 Copilot 严格隔离，承诺不用于 AI 模型训练。
- 微软强调 Copilot Health 定位于辅助理解数据的“健康管家”，不会取代专业医生提供诊断或治疗建议，目前已面向首批用户开放候补名单申请。
相关链接：
- 你的私人 AI 医生来了：微软发布 Copilot Health，深度打通医疗大数据
- Microsoft's Copilot Health can use AI to turn your fitness data and medical records 'into a coherent story'
- Microsoft launches Copilot Health
Microsoft Launches Copilot Health: Deeply Integrating Data from 50,000 Healthcare Institutions to Build a Personal AI Health Space 🏥🤖
Brief:
- Microsoft has officially launched the Copilot Health platform, a standalone secure space within Copilot designed to use AI to transform users’ personal medical records and wearable device data into a “coherent health story.”
- The platform has already achieved large-scale data interoperability. Through HealthEx, it can access medical records, prescriptions, and lab results from more than 50,000 hospitals across the United States, while also syncing health metrics from over 50 mainstream wearable devices, including Apple Watch and Fitbit.
- Core features include intelligently interpreting medical examination reports, analyzing sleep and exercise trends, filtering suitable doctors based on insurance plans and geographic location, and helping users prepare questions for pre-appointment consultations.
- To ensure information authority and reliability, AI responses prioritize references from expert sources such as 50 national health authorities and Harvard Health. Health data is also kept strictly separate from general Copilot data, with Microsoft promising it will not be used to train AI models 🔒.
- Microsoft emphasized that Copilot Health is positioned as a “health concierge” to assist users in understanding their data, rather than replacing professional doctors in providing diagnoses or treatment recommendations. The waitlist is now open to an initial group of users.
Related Links:
- Your personal AI doctor is here: Microsoft launches Copilot Health, deeply connecting healthcare big data
- Microsoft's Copilot Health can use AI to turn your fitness data and medical records 'into a coherent story'
- Microsoft launches Copilot Health
如果对你有用的话，可以打赏哦
打赏
本文作者:topwind
本文链接:
版权声明:本博客所有文章除特别声明外，均采用 BY-NC-SA 许可协议。转载请注明出处！
目录