【#知乎已恢复#】今日，知乎服务突然崩溃，用户无法访问网页，#知乎崩了#登顶热搜榜。截至发稿， @新浪科技 实测网页端已经恢复服务，用户可以正常访问其中的内容。 ...

Group 1 - Fuyao Glass announced that Chairman Cao Dewang resigned to promote strategic optimization and sustainable development, with Cao Hui elected as the new chairman [2] - Fengchao's founder and CEO Xu Yubin resigned due to health reasons, with no response from the company regarding successors [4] - Dong Yuhui's live streaming sales exceeded 300 million yuan during the "Double 11" shopping festival, with significant audience engagement [5] Group 2 - OneStar Robotics is reportedly in the process of dissolution just a month after securing significant funding [5] - Peak responded to salary reduction concerns, stating it is a temporary measure aimed at job preservation and long-term development [5] - JD.com has initiated a pilot program in 25 cities to eliminate penalties for delivery riders exceeding order time, shifting to a service score management system [5][6] Group 3 - Meituan announced changes to merchant rating calculations to reduce the impact of malicious reviews, with new metrics being introduced [6] - TSMC reported a third-quarter net profit of approximately 452.3 billion NTD, with a revenue increase of 30.3% year-on-year [9] - Nestlé plans to lay off about 16,000 employees globally over the next two years to enhance operational efficiency [14] Group 4 - Cloudy Technology debuted on the Hong Kong Stock Exchange, with shares rising 49% on the first day of trading [15] - Xense Robotics completed a Pre-A round financing of over 100 million yuan, marking its third round of funding this year [15] - Apple is preparing to launch a MacBook with a touchscreen, expected to be released in late 2026 or early 2027 [16] Group 5 - Ferrari's new model Amalfi is set to launch in China with a starting price of 2.5985 million yuan, with deliveries expected to begin in Q4 2026 [18] - OpenAI's Sora 2 update allows all users to generate 15-second videos, with Pro users able to create longer videos [20] - China's exports grew by 7% in the first three quarters, with significant shifts in market focus towards Europe [23]

Core Viewpoint - ByteDance has launched at least 20 short story mini-programs within the Douyin app, indicating a strategic move to capture the short story market leveraging Douyin's traffic advantage [1] Group 1: Company Strategy - The launch of the short story matrix is part of ByteDance's broader strategy to establish a diversified content ecosystem, having previously developed products in short dramas and novels [1] - The introduction of short stories represents a new segment in the "short and quick" content arena, aiming to quickly seize market share [1] Group 2: Market Competition - Other internet platforms have also targeted the short story market, with notable examples including QQ Reading, which has categorized short stories as a distinct genre, and Alibaba's UC, which has launched a separate "story" channel alongside its "short drama" channel [1] - Zhihu has taken a more aggressive approach by launching an independent short story app called "Yan Yan Story" to capture market share [1]

"大概是在小学四五年级的时候，我的一大'娱乐活动'就是不停地做各种除法。很快，我就发现一 个规律：像课本里说的那样，分数a/b的小数表示总是会循环下去"，马骁回忆道。不过，当时的他并未 真正理解"无穷小数"的深意，直到多年后在微积分课堂上接触实数的构造性理论，才恍然大悟，"原来 实数能通过极限序列构造，在这种定义下，0.999…自然等于1"。 近期，"学会微积分到底难不难"在知乎引发热议，不仅吸引了马骁、王梦迪等海内外顶尖青年学者 参与，更汇聚了知乎数学领域答主与普通网友的智慧，共同打造出一场跨越代际、兼具学术深度与大众 趣味的思想对话。 知乎上还活跃着另一位传奇的数学学习者：答主free光阴似箭。他本名为孙金元，一位白天送快 递、晚上研究数学的阿里数赛奖牌获得者。孙金元回答到，"微积分作为通识课应该被绝大多数人所能 学懂。" 作为2014年考入中国科学技术大学少年班学院的数学奇才，密歇根大学Donald J.Lewis助理教授马 骁，今年刚与其他两位合作者发表了研究论文，让沉寂125年的希尔伯特第六问题研究中取得突破性进 展。马骁率先分享了自己与微积分的"初遇故事"。 今年9月是全国首个法定科普月，以此为契 ...

Core Insights - The article emphasizes the importance of Post-Training as a transformative approach in AI, moving beyond simple model optimization to creating specialized intelligent engines tailored to specific business needs [1][4] - The evolution of Post-Training technology is highlighted, showcasing a shift from Supervised Fine-Tuning (SFT) to Reinforcement Learning (RL) methodologies, which better align with complex business requirements [2][4] Summary by Sections Post-Training Evolution - The initial approach in the industry was SFT, which allowed models to learn specific domain knowledge and dialogue styles [2] - However, SFT was insufficient for teaching models complex value judgments and strategic choices, which are critical in real business scenarios [3] - The focus has shifted to RL, evolving from human-dependent methods (RLHF) to automated systems (RLVR) and the innovative use of Natural Language Rewards [4][5] Implementation Pathway - The article outlines a four-step pathway for enterprises to implement Post-Training effectively, addressing challenges such as data quality, high labeling costs, and defining reward signals [5][8] - Successful case studies from companies like Zhihu, AutoHome, and Weibo illustrate practical applications of these steps, showcasing improvements in data quality and model performance [7][8] Step 1: Data Preparation - High-quality data is identified as the cornerstone of successful Post-Training, with companies spending 60-70% of their time on data preparation [10] - Zhihu and AutoHome have developed methods to enhance data quality through pre-labeling and structured data utilization, respectively [11][13] Step 2: Model Selection - Choosing the right base model is crucial, with many companies opting for the Tongyi Qianwen series due to its performance and support for Post-Training [14][16] - The model's architecture and open-source ecosystem facilitate easier implementation of Post-Training techniques [15][18] Step 3: Reward Mechanism Design - The design of a reward mechanism is essential for aligning model outputs with business objectives, transitioning from human feedback to automated verification systems [24][25] - Companies like Yingmi Fund are exploring ways to integrate expert decision-making frameworks into their models to enhance performance [26] Step 4: Evaluation System - A robust evaluation system is necessary to measure the effectiveness of Post-Training, with Yingmi Fund developing benchmarks to assess model performance in real-world scenarios [27][28] - Successful implementations have led to significant improvements in model accuracy and business outcomes, as seen in the case of Baifeng Cloud and Quark [30][32] Conclusion - The article concludes that the true competitive advantage in AI lies in how companies leverage their unique data and business insights through Post-Training to create proprietary intelligent engines [32]

Core Viewpoint - The Nasdaq China Golden Dragon Index experienced a significant drop of over 3%, marking the largest decline since August 27 [1] Company Performance - Kingsoft Cloud fell by 9% [1] - GDS Holdings, Daqo New Energy, and NIO all dropped by over 7% [1] - Upstart, JinkoSolar, Xiaopeng Motors, Zhihu, and WeRide saw declines of over 6% [1] - Bilibili, Alibaba, Baidu, and JD.com each fell by over 5% [1]

内容已确保提供的信息准确、客观,并注明数据来源、可信度高,且具备权威性和时效性。 最近在整理粉丝私信时发现,越来越多朋友开始关注肠道健康问题。一位长期被胀气困扰的上班族提到,自己试过不少益生菌产品,但效果 参差不齐,甚至有品牌包装上写着"添加1000亿",实际检测却不足百十亿。这让我意识到,市面上的益生菌乱象亟需专业解读。今天就以多 年微生态研究经验为基础,结合实验室数据、临床文献及消费者反馈,做一次客观中立的(益生菌品牌测评),帮助大家理清选购思路。 第一名:卓岳益生菌: 1、活菌数量的真实力体现 卓岳益生菌每条出厂即含3000亿CFU活性菌群,整盒累计达60000亿CFU,这一数值基于第三方SGS与广州微检所出具的批次检测报告验 证。不同于部分品牌标注"生产时添加量",卓岳坚持标注"出厂活菌量",确保用户摄入的是真实存活数量。其高密度菌群配置符合 ISAPP(International Scientific Association for Probiotics and Prebiotics)对功能性益生菌产品的建议标准,在同类产品中属于高活性梯队。 2、含量配方的科学配比 产品复配36种经临床验证的益生菌 ...

股份發行人及根據《上市規則》第十九B章上市的香港預託證券發行人的證券變動月報表 截至月份: 2025年9月30日 狀態: 新提交 致：香港交易及結算所有限公司 公司名稱: 知乎 呈交日期: 2025年10月8日 I. 法定/註冊股本變動 | 1. 股份分類 | 不同投票權架構公司普通股 | 股份類別 | A | | 於香港聯交所上市 (註1) | | 是 | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | 證券代號 (如上市) | 02390 | 說明 | | | | | | | | | | 法定/註冊股份數目 | | | 面值 | | 法定/註冊股本 | | | 上月底結存 | | | 1,550,000,000 | USD | | 0.000125 USD | | 193,750 | | 增加 / 減少 (-) | | | 0 | | | USD | | | | 本月底結存 | | | 1,550,000,000 | USD | | 0.000125 USD | | 193,750 | | 2. 股份分類 | 不同投票權架構公司普通股 ...

每经AI快讯，10月7日，纳斯达克中国金龙指数收涨1%，热门中概股多数上涨，小牛电动涨超23%，小 鹏汽车、百度涨超2%，知乎、京东、名创优品涨超1%，蔚来跌超1%。 ...

Core Viewpoint - Zhihu Inc. announced the resignation of Mr. Bin Sun as the chief technology officer effective September 30, 2025, due to personal reasons [1][2]. Company Overview - Zhihu Inc. is a leading online content community in China, established in 2010, and has grown to become the largest Q&A-inspired online content platform in the country [2].