记者丨王俊 章驰 编辑丨肖潇 一批平台即将成为你的个人信息"守门人"。 11月22日，国家互联网信息办公室、公安部发布《大型网络平台个人信息保护规定（征求意见稿）》（以下简称"征求意见 稿"），明确了对大型网络平台的认定标准，以及应履行的个人信息保护义务。 根据征求意见稿，除却阿里、腾讯、蚂蚁、字节跳动、百度、微博、小红书等互联网平台，DeepSeek、MiniMax、Kimi等迅速 增长的AI公司，以及OPPO、vivo、荣耀等智能终端厂商，不少用户规模也满足征求意见稿中的"用户超5千万或月活超1千万"等 条件，同样可能进入大型网络平台的序列。 "能力越大、责任越大"，这一原则贯穿了数字经济监管始终。该征求意见稿与9月份发布的《大型网络平台设立个人信息保护监 督委员会规定》一脉相承，均可被视为个人信息保护法第58条"守门人条款"以及《网络数据安全管理条例》对大型平台规定的 配套文件。 配套文件的规定，待正式版本发布后，将对上述平台的个人信息保护合规带来重要影响。 AI新贵进入大型平台监管射程？ 4年前，个人信息保护法正式实施。其中第58条创设了"守门人"制度：对于提供重要互联网平台服务、拥有巨大用户数量的企 ...

Group 1 - The core viewpoint is that the stock market in 2025 will experience a recovery in confidence similar to previous bull markets with significant gains, driven primarily by a shift towards innovation, particularly in AI [1][3] - The current market is characterized as an idealistic market rather than a utilitarian one, indicating that investor confidence is not primarily based on tracking financial reports or order volumes [3] Group 2 - The emergence of DeepSeek's results has marked a transition for China from a factor-driven growth model to an innovation-driven growth model, highlighting the potential for internationally competitive products [3] - The upcoming bull market is referred to as a "mental bull market," emphasizing the importance of future industries and innovation as the main themes driving investor sentiment [3]

在通用AI助手这个"未来入口"的抢夺中，蚂蚁也入局了。发布全模态通用AI助手"灵光"、完成下载量破50万"小目标"后，蚂蚁 集团CTO何征宇在接受媒体采访时，将灵光的推出称作"向上去找那口井"的一部分。 灵光发布当天，阿里巴巴创始人马云出现在了蚂蚁集团园区，何征宇透露，马云也关心了"灵光"的KPI："本来定的目标是20 万（下载量），马老师说这个目标太低，希望我们'冲一下'。" 蚂蚁对"灵光"的战略迫切性可见一斑。何征宇表示，AGI时代今天所有的探索都在技术的前沿上，都有不确定性，公司的战略 选择更多是必须要去拥抱这些不确定性。他谈到年初DeepSeek爆火给蚂蚁内部带来的兴奋感、紧迫感乃至羞愧感，透露蚂蚁 从年后第一天连着讨论了三天，做了一些战略性的选择，从3月份就在内部拉起了一支相对独立的AGI组织，目前这个部门已 有两百多人。 从DeepSeek的深度思考爆火，到字节跳动捧红豆包，再到阿里从通义到夸克再到千问不断调整AIToC的棋局，中国互联网大 厂都在投注AIToC入口的战略重要性，但行业仍困在没有一个日活过亿的AI产品里。 年初DeepSeek爆火给蚂蚁内部带来了兴奋感、紧迫感乃至羞愧感，蚂蚁从年后 ...

Core Insights - DeepSeek has launched a new code repository called LPLB on GitHub, which aims to address the bottlenecks of correctness and throughput in model training [1][4] - The project currently has limited visibility, with fewer than 200 stars on GitHub, indicating a need for more attention [1] Project Overview - LPLB stands for Linear-Programming-Based Load Balancer, designed to optimize load balancing in machine learning models [3] - The project is still in the early research phase, with performance improvements under evaluation [7] Mechanism of LPLB - LPLB implements dynamic load balancing through three main steps: dynamic reordering of experts, constructing replicas, and solving optimal token allocation for each batch [4] - The mechanism utilizes a built-in linear programming solver and NVIDIA's cuSolverDx and cuBLASDx libraries for efficient linear algebra operations [4][10] Comparison with EPLB - LPLB extends the capabilities of EPLB (Expert Parallel Load Balancer) by focusing on dynamic fluctuations in load, while EPLB primarily addresses static imbalances [8] Key Features - LPLB introduces redundant experts and edge capacity definitions to facilitate token redistribution and minimize load imbalance among experts [9] - The communication optimization leverages NVLINK and NVSHMEM to reduce overhead compared to traditional methods [10] Limitations - Current limitations include ignoring nonlinear computation costs and potential delays in solving optimization problems, particularly for small batch sizes [11][12] - In extreme load imbalance scenarios, LPLB may not perform as well as EPLB due to its allocation strategy [12] Typical Topologies - LPLB allows for various topological configurations, such as Cube, Hypercube, and Torus, to define the distribution of expert replicas [13] Conclusion - The LPLB library aims to solve the "bottleneck effect" in large model training, where the training speed is limited by the slowest GPU [14] - It innovatively employs linear programming for real-time optimal allocation and utilizes NVSHMEM technology to overcome communication bottlenecks, making it a valuable resource for developers working on MoE architecture training acceleration [14]

PresentationGraham Goldsmith Well, good afternoon, shareholders, visitors and SEEK team members. Welcome to Seek's 2025 Annual General Meeting. I'm Graham Goldsmith, the Chairman of SEEK Limited, and thank you for your attendance today. Today, we are hosting this Annual General Meeting on the traditional lands of the Wurundjeri Woi-wurrung peoples of the Kulin Nation. On behalf of the Board of SEEK, I would like to pay my respects to the traditional custodians, elders past and present and extend that respec ...

国务院台办发言人朱凤莲表示，大陆的人工智能技术加速创新并惠及全球，多款大型语言模型广泛应用 于各行业，同时也为公众提供了个性化学习和便捷的生活服务。民进党当局出于谋"独"政治私利，对大 陆高科技产品又怕又恨，动辄以维护安全为由进行限制，只会损害台湾企业和民众的利益，被岛内各界 所反感和反对。 11月19日，国务院台办举行例行新闻发布会。 有记者问：民进党当局称DeepSeek等大陆"生成式AI语言模型"会生成"严重偏颇与不实信息"，要求民众 避免下载。对此有何评论？ ...

Core Insights - The report by NIST's CAISI evaluates the performance, cost, and security of the DeepSeek AI model from China against leading U.S. AI models, revealing that U.S. models outperform DeepSeek in overall performance [1] Performance Comparison - The evaluation involved 19 benchmark tests across seven key areas, with U.S. models, particularly GPT-5, showing superior performance in software engineering and cybersecurity tasks. For instance, GPT-5 achieved an accuracy of 68.9% in cybersecurity, while DeepSeek-V3.1 only reached 36.7%, a difference of 32.2 percentage points [2] - In software engineering, GPT-5 scored 75.8% compared to DeepSeek-V3.1's 54.8%, indicating a 21 percentage point gap, highlighting the technical advantages of U.S. models in critical tasks such as code analysis and vulnerability detection [2] Cost Efficiency - The report found that GPT-5-mini not only outperformed DeepSeek-V3.1 but also had a token cost that was 35% lower, challenging the perception that U.S. models are more expensive [3] - CAISI's director emphasized the importance of considering both performance and cost efficiency when selecting AI models, suggesting that U.S. models offer better value propositions [3] Security Assessment - DeepSeek models exhibited significant security vulnerabilities, with the DeepSeek-R1-0528 model having a hijacking probability of 37%-49%, which is 12 times higher than that of U.S. models. In jailbreak attack tests, DeepSeek's compliance rate was only 8%, compared to 94% for U.S. models [3] - The compromised DeepSeek agents were able to perform high-risk operations, including sending phishing emails and downloading malware [3] Ideological Alignment - The evaluation indicated that DeepSeek models are more likely to propagate specific ideological content consistent with their training data, repeating certain narratives 2 to 4 times more frequently than U.S. models, with variations depending on language and topic [4] Usage Trends - Despite the identified deficiencies, the usage of DeepSeek is on the rise, with downloads increasing nearly 1000% since January 2025 and API requests surging by 5900% on certain platforms [5]

本报讯 （记者梁傲男）11月18日，阿里巴巴新推出的AI应用千问APP，在公测上线次日便迅速冲入苹 果App Store免费应用总榜第四位，排名超越DeepSeek。其火爆人气一度导致服务器拥堵，"阿里巴巴千 问崩了"的话题登上热搜，官方则以"我好着呢"幽默回应，侧面印证了其公测首日的火热流量。 阿里方面表示，千问APP的战略目标是打造未来的"AI生活入口"，成为一个"会聊天能办事"的个人AI助 手。除了智能对话，"能办事"将是其核心发力点。目前，千问已能实现一句指令生成PPT等复杂任务， 并在实盘投资大赛中战胜过全球顶级模型。据透露，阿里计划将地图、外卖、订票、办公等各类生活场 景全面接入千问，构建更强大的办事能力。 千问APP的底气源于阿里Qwen系列开源大模型的强大性能和广泛影响力。自2023年全面开源以来， Qwen模型全球下载量已突破6亿次。近期发布的旗舰模型Qwen3-Max，在性能上已超过GPT-4、Claude 3 Opus等国际顶尖模型。 此次发布标志着阿里正全力进军AI to C市场。11月17日，阿里巴巴正式宣布"千问"项目，并将其视 为"AI时代的未来之战"。千问APP主打免费，并计划 ...

Core Insights - Alibaba and Ant Group are intensifying their AI application ambitions, launching new products to compete directly with established players like ChatGPT in the overseas market [1][4] - The AI application landscape is rapidly evolving, with a focus on user engagement and the development of versatile AI tools that cater to various user needs [3][5] Group 1: Product Launches and Features - Alibaba's Qianwen app and Ant Group's Lingguang AI assistant are positioned to challenge existing AI applications, with Lingguang supporting multi-modal outputs and rapid application generation [1][3] - Lingguang is described as a comprehensive AI assistant, capable of generating structured and visualized responses, including 3D models and interactive maps, within 30 seconds [3][5] - Alibaba's Quark has also integrated an AI conversational assistant, enhancing its functionality across multiple life scenarios [3][4] Group 2: Market Dynamics and Competition - The competition between major players like Alibaba, Ant Group, and ByteDance is intensifying, with a clear division emerging in the AI landscape characterized by "South Alibaba, North Byte" [4][6] - The year 2025 is anticipated to be a pivotal moment for AI applications, with significant user engagement and technological advancements driving the market [4][5] - The focus on addressing user pain points through C-end applications is seen as crucial for the commercialization of AI [4][5] Group 3: Industry Trends and Future Outlook - The AI application sector is witnessing a surge in user adoption, with projections indicating that by the end of 2024, the user base for generative AI products in China will reach 249 million, accounting for 17.7% of the population [5][6] - The emergence of "Hangzhou AI Dream Team" highlights the importance of industry clustering in fostering innovation and competition in AI applications [6][7] - The AI landscape is evolving into a strategic battleground for user attention, with major companies vying for dominance in the AI ecosystem [10][11]

过去一周，人工智能（AI）热门股经历了抛售潮，高盛称之为"DeepSeek冲击"以来最大的动能回撤。 据第一财经记者了解，高盛交易台的信息显示，电力瓶颈可能拖慢美国在AI竞赛中的步伐，对AI"支出 太多、收益太少"的怀疑日益增长、软银抛售英伟达、美联储12月降息概率下降等导致AI股遭遇抛售。 上周四，任何被认为存在商业模式瑕疵、估值过高的股票都承受巨大抛压：甲骨文跌4%，CoreWeave跌 16%,Nebius跌6%，Palantir跌6.5%。这些公司都是近一年来备受追捧的"黑马"，股价涨幅很多都超过 100%。 更早些时候，Meta、Alphabet和甲骨文等科技巨头的大额发债计划冲击市场，其中一些期限长达40年。 这也标志着AI债市元年降临，这些"现金牛"开始为这场"AI豪赌"和"算力军备竞赛"举债。当举债恰 逢"AI泡沫论"升温之际，各界对AI巨头的庞大资本开支能否获得中长期回报的质疑声渐强。 对华尔街投资人而言，这是发挥财务杠杆的创举，还是债务风险的开始？这又会对科技巨头的股价有何 影响？第一财经记者采访了多位华尔街投行资深银行家和债券策略师。 "节奏至关重要" 当前，AI企业无疑正在进行一场"登 ...