单词学习：![[5da3a1f7-a632-4f16-90be-afec23025fa8.mp3]]

新闻播报：![[65b97a3a-b00b-4cb8-a5ec-85f8d8a11964.mp3]]

微软Copilot全面融入Windows 11，新增语音唤醒与本地自动化

简报：作为其AI战略深化的一部分，微软正将Copilot深度集成至Windows 11核心，新增了“嘿，Copilot”语音唤醒、可分析屏幕内容的Copilot Vision以及用于处理本地简单任务的Copilot Actions等一系列功能，旨在让Copilot成为用户与PC交互的主要方式。

相关链接：

• 微软深化AI战略
  核心集成 Copilot，支持语音控制、屏幕分析和本地自动化

Microsoft Copilot Fully Integrated into Windows 11, Adds Voice Activation & Local Automation

Brief: As part of its deepening AI strategy 🚀, Microsoft is deeply integrating Copilot into the core of Windows 11. This includes new features such as "Hey, Copilot" voice activation 🗣️, Copilot Vision (which can analyze screen content), and Copilot Actions for handling simple local tasks. The goal is to make Copilot the primary way users interact with their PCs 🤖.

Related Links:

• 微软深化AI战略
  核心集成 Copilot，支持语音控制、屏幕分析和本地自动化

李飞飞团队推出RTFM模型，单卡H100即可实时生成持久3D世界

简报：

• 由李飞飞团队发布的RTFM（Read The Field Model）模型，可在单个NVIDIA H100 GPU上以交互式帧率实现实时3D世界生成。
• 该模型生成的3D世界具备持久性与3D一致性，在交互过程中能保持几何结构、物体位置与外观的稳定。
• RTFM引入了“持久性记忆机制”，支持用户在生成的虚拟空间中进行无限时长的探索，实现了“可持续存在”的AI世界。
• 模型渲染效果逼真，支持反射、阴影、镜面高光等复杂视觉效果，被认为是AI世界模型迈向高保真实时渲染的关键一步。

相关链接：

• 李飞飞团队发布 RTFM:单卡 H100实现实时3D 世界生成

Fei-Fei Li's Team Unveils RTFM Model: Persistent 3D Worlds Generated in Real-Time on a Single H100 GPU

Briefing: 👇

• The RTFM (Read The Field Model) developed by Fei-Fei Li's team enables real-time 3D world generation at interactive frame rates on a single NVIDIA H100 GPU.
• The 3D worlds generated by the model boast persistence and 3D consistency, maintaining stable geometric structures, object positions, and appearances during interaction.
• RTFM introduces a "persistent memory mechanism" that allows users to explore the generated virtual spaces for an unlimited duration, realizing an "ever-present" AI world. 🌍
• The model produces highly realistic rendering effects, supporting complex visual features such as reflections, shadows, and specular highlights, marking a crucial step for AI world models towards high-fidelity real-time rendering. ✨

Related Link:

• Fei-Fei Li's Team Releases RTFM: Real-Time 3D World Generation on a Single H100

Sora 2上线Azure开启商业化，同步更新功能并处理内容争议

简报：

• OpenAI的视频生成模型Sora 2已在微软Azure AI Foundry国际版上线并进入公共预览阶段，定价为每秒0.1美元，中国区用户暂时无法使用。
• 新功能已上线，所有用户可生成最长15秒的视频，Pro用户在网页端可生成长达25秒的视频并使用新增的故事板功能。
• 因用户在Sora社交平台上生成了关于马丁·路德·金等历史人物的不当视频，OpenAI已应其遗产管理处请求，暂停了对金博士肖像的生成。

相关链接：

• OpenAI视频生成模型Sora 2上线微软Azure平台：定价每秒0.1美元，进入公共预览阶段
• OpenAI 暂停Sora生成马丁・路德・金视频，保护历史人物形象
• ​OpenAI Sora 2全新功能上线，Pro用户生成视频长达 25 秒

Sora 2 Launches on Azure, Kicking Off Commercialization, Simultaneously Updating Features and Addressing Content Controversies 🚀

Brief:

• OpenAI's video generation model, Sora 2, has launched on Microsoft Azure AI Foundry International and entered public preview, priced at $0.1 per second. Users in mainland China are temporarily unable to access it.
• New features are now live: all users can generate videos up to 15 seconds long, while Pro users on the web can create videos up to 25 seconds and utilize the newly added storyboard function. 💡
• Due to users generating inappropriate videos of historical figures like Martin Luther King Jr. on the Sora social platform, OpenAI has paused the generation of Dr. King's likeness at the request of his estate. 🛑

Related Links:

• OpenAI video generation model Sora 2 launched on Microsoft Azure platform: priced at $0.1 per second, entering public preview
• OpenAI pauses Sora generation of Martin Luther King Jr. videos to protect historical figures' images
• OpenAI Sora 2 new features launched, Pro users can generate videos up to 25 seconds long

GPT-5解决数学难题？乌龙事件引发业界批评

简报：

• OpenAI管理层成员凯文·韦尔（Kevin Weil）及其他研究员在社交平台X上发文，声称GPT-5解决了10个“此前未解的”埃尔德什（Erdős）数学问题。
• 数学家托马斯·布鲁姆（Thomas Bloom）澄清此为“严重误解”，指出GPT-5只是检索到了他本人未曾注意到的已有研究成果，并非解决了学术界真正的未解难题。
• 该事件引发DeepMind CEO德米斯·哈萨比斯、Meta AI负责人杨立昆等业界人士批评，后者讽刺OpenAI“被自己的炒作所反噬”。
• 涉事OpenAI研究人员随后删除了相关推文并承认错误。

相关链接：

• OpenAI 研究人员夸大 GPT-5 数学能力后删帖，遭杨立昆等业界人士批评

GPT-5 Solves Unsolved Math Problems? Faux Pas Triggers Industry Outcry

Brief:

• OpenAI management member Kevin Weil and other researchers posted on social platform X, claiming GPT-5 had solved 10 "previously unsolved" Erdős mathematical problems. 🤯
• Mathematician Thomas Bloom clarified this was a "serious misunderstanding," explaining that GPT-5 had merely retrieved existing research findings he hadn't noticed, rather than solving genuine unsolved academic problems. 🤥
• The incident drew criticism from industry figures including DeepMind CEO Demis Hassabis and Meta AI head Yann LeCun, who satirized OpenAI for "being consumed by its own hype." 📉
• The OpenAI researchers involved subsequently deleted their posts and admitted their error.

Related Link:

• OpenAI researchers delete posts after exaggerating GPT-5's math capabilities, draw criticism from Yann LeCun and others

苹果发布M5芯片，多款新品性能与价格齐升

简报：

• 苹果公司发布M5系列芯片，并推出搭载该芯片的新款iPad Pro、MacBook Pro及Vision Pro。
• 新款iPad Pro性能据称比M1版快5.6倍，起售价约新台币32,000元。
• 14吋MacBook Pro续航力长达24小时，起售价约新台币51,000元。
• 升级版Vision Pro同样采用M5芯片，售价约新台币119,900元起。
• 苹果宣称M5是自M1芯片问世以来最大幅度的性能跳跃，新品于10月22日起全球开卖。

相关链接：

• AI效能炸裂！蘋果M5系列性能翻倍、價格更驚人

Apple Launches M5 Chip, New Products See Performance and Price Hikes

Briefing:

• Apple Inc. has launched its M5 series chips, introducing new iPad Pro, MacBook Pro, and Vision Pro models powered by them. 🚀
• The new iPad Pro reportedly boasts 5.6 times faster performance than its M1 predecessor, with a starting price of approximately NT$32,000.
• The 14-inch MacBook Pro offers an impressive 24-hour battery life and starts at around NT$51,000. 🔋
• The upgraded Vision Pro also features the M5 chip, with prices starting from approximately NTdJyH6kDkfC5RlM16.answerText19,900.
• Apple claims the M5 represents the biggest performance leap since the M1 chip's debut. The new products will be available globally starting October 22. ✨

Related Link:

• AI Performance Explodes! Apple's M5 Series Doubles Performance, Prices Even More Shocking

爱诗科技获1亿元B+轮融资，ARR突破4000万美元

简报：

• AI视频生成公司爱诗科技（PixVerse）于10月17日宣布完成1亿元人民币B+轮融资，投资方包括复星锐正、同创伟业和顺禧基金。
• 公司披露其年度经常性收入（ARR）已突破4000万美元，注册用户数超过1亿，月活跃用户超过1600万。
• 自2024年11月启动商业化以来，公司收入在不到一年内实现了十倍增长。
• 其自研的视频生成大模型已迭代至PixVerse V5版本，并引入了Agent创作助手功能。
• 公司采用双品牌策略，运营面向全球市场的PixVerse和主要服务国内用户的拍我AI。

相关链接：

• AI视频公司爱诗科技完成1亿元B+轮融资

PixVerse Secures RMB 100 Million Series B+ Funding, ARR Surpasses $40 Million 🚀

Brief:

• AI video generation company PixVerse announced on October 17th the completion of its RMB 100 million Series B+ funding round, with investors including Fosun RZ Capital, Tongchuang Weiye, and Shunxi Capital.
• The company disclosed that its Annual Recurring Revenue (ARR) has exceeded $40 million, with registered users surpassing 100 million and monthly active users exceeding 16 million. 📈
• Since launching commercialization in November 2023, the company's revenue has achieved a tenfold increase in less than a year.
• Its self-developed video generation large model has iterated to PixVerse V5, introducing an Agent creative assistant feature. ✨
• The company employs a dual-brand strategy, operating PixVerse for the global market and Paiwo AI (拍我AI) primarily serving domestic users.

Related Link:

• AI Video Company PixVerse Completes RMB 100 Million Series B+ Funding

清华长庚与北电数智联手，共研国内首个药学大模型

简报：

• 北京清华长庚医院与北京电子数智科技有限责任公司于10月16日宣布达成战略合作，将联合研发国内首个药学领域的专用大模型。
• 该项目旨在通过AI技术，聚焦并提升对老年人、儿童和孕产妇等特殊人群用药安全评估的效率与精准度。
• 模型将基于北电数智的“星火·医疗底座”平台，并结合清华长庚医院拥有的高质量临床实践数据进行训练。
• 合作初期将在清华长庚医院落地应用，建立技术与临床的反馈闭环，并计划共同参与行业标准制定。

相关链接：

• 清华长庚医院与北电数智合作研发国内首个药学大模型：聚焦特殊人群用药安全评估

Beijing Tsinghua Changgeng Hospital and Beidian Shuzhi Join Forces to Develop China's First Pharmaceutical Large Model 🤝

Brief:

• Beijing Tsinghua Changgeng Hospital and Beijing Dian Shuzhi Technology Co., Ltd. announced on October 16th the formation of a strategic partnership to jointly develop China's first specialized large model in the pharmaceutical field.
• This project aims to leverage AI technology to focus on and enhance the efficiency and accuracy of medication safety assessments for special populations such as the elderly, children, and pregnant women. 💊
• The model will be trained based on Beidian Shuzhi's "Spark · Medical Foundation" platform, combined with high-quality clinical practice data from Beijing Tsinghua Changgeng Hospital.
• Initially, the collaboration will be implemented at Beijing Tsinghua Changgeng Hospital, establishing a feedback loop between technology and clinical practice, and planning to jointly participate in the formulation of industry standards. ✨

Related Link:

• Beijing Tsinghua Changgeng Hospital and Beidian Shuzhi Collaborate to Develop China's First Pharmaceutical Large Model: Focusing on Medication Safety Assessment for Special Populations