Group 1 - The founder of Xibei, Jia Guolong, has been very active recently, frequently trending on social media [1] - Jia announced he would respond to "Luo Yonghao's major slander against Xibei" at 10 PM on January 16 [1] - Luo expressed that he would try to hold back but Jia encouraged him not to, indicating a tense atmosphere [4] Group 2 - Both Jia and Luo's Weibo accounts were silenced around 9:50 PM, surprising the public who expected a significant revelation [4] - Despite the silencing, Jia was not satisfied and posted a response through the account @XibeiRenXinSheng, which was also handled shortly after [5] - The initial response had garnered 12,000 comments before being taken down, with many media outlets and netizens having saved screenshots [6] Group 3 - The reason for the silencing was clarified by Sina Weibo's CEO, Wang Gaofei, who stated that engaging in online disputes is not advisable and that the authorities would intervene [7] - A "negative behavior list for online celebrity accounts" was officially released by the Cyberspace Administration on December 26, 2025 [8] - Although both Jia and Luo were silenced, discussions about them continued on platforms like Weibo [8]

Core Insights - The article discusses how a father, KiddFlash42, utilized AI to create an educational game with his 5-year-old son, transforming the learning process into a fun experience [1][4][17] Group 1: Project Initiation - The project began as a solution to the father's concern about his son's reluctance to engage in learning activities like reading and math [4] - After discovering that AI could write code, the father asked AI if it could create video games, leading to the generation of a game structure and example code [5][6] Group 2: Development Process - The development process was characterized by minimal documentation, relying heavily on the child's imagination for game features [8] - The father primarily handled copying, pasting, and troubleshooting with AI, while the son acted as the "product manager," directing the game's design [7][10] Group 3: AI Interaction - Initially, the AI (Gemini) performed well, providing a friendly interaction that made the project feel like a game rather than work [9] - As the codebase grew to around 1200 lines, the AI began to struggle, leading to frequent errors and crashes, prompting a switch to GitHub Copilot for improved productivity [10][13] Group 4: AI Model Performance - The father ranked various AI models based on their performance, with Claude Opus 4.5 being the most effective, followed by Gemini and GPT-5.1 [14][15][16] - Claude Opus 4.5 was noted for its ability to understand the child's creative ideas without destabilizing the project, while Gemini was friendly but technically unstable [15][16] Group 5: Final Outcome - The educational game successfully integrated reading, spelling, and math into engaging gameplay, reversing the father's role from enforcer to facilitator in his son's learning [17][18]

Core Insights - The article highlights the significant financial success of Huanfang Quantitative, which is projected to earn 5 billion RMB in 2025, allowing for the training of 2,380 DeepSeek R1 models [1] - Huanfang Quantitative, led by Liang Wenfeng, ranks second among large quantitative funds in China with an average return rate of 56.6% and manages over 70 billion RMB [1] - The revenue generated from Huanfang's management fees and performance fees has provided DeepSeek with substantial funding for its AI research, enabling it to operate independently without external financing [2] Financial Performance - Huanfang Quantitative's earnings of approximately 5 billion RMB last year surpassed the pre-IPO fundraising of AI unicorn MiniMax [1] - The average management fee of 1% and performance fee of 20% contributed significantly to Huanfang's revenue [1] AI Development - DeepSeek's training costs are relatively low, with the R1 model costing only 294,000 USD and the V3 model costing 5.576 million USD, allowing for extensive model training with the funds available [2] - The financial model creates a symbiotic relationship where profits from quantitative trading support AI research, while AI technology enhances quantitative strategies [2]

但怕你又收不到 这个笑话小程程以前发过，估计还不止 1 次。但是，我最近首次看到有网友把它用在面试中。 好了，下次大家遇到 UDP 面试题时，要不要就用这个笑话开场呗。拿到 offer 后，记得回来留个言咯，立帖等 你们 ...

新智元报道 【导读】 一个大模型持续写代码，能写多久？一小时？一天？还是像大部分AI编程工具那样，完成一个任务就结束对话？Cursor的CEO MichaelTr uell决定搞一次极限压力测试！ Michael Truell 让Cursor中的GPT-5.2连续运行了 整整一周 。 不是一小时，不是一天，而是不眠不休，昼夜不停，168小时持续写代码。 结果？ 300万行代码。数千个文件。 AI 完全从零构建出一个全新浏览器。 而且，还是Chrome那种浏览器。 HTML解析、CSS布局、文本渲染、还有一个自研的JavaScript虚拟机——全是AI自己写的。 Michael Truell轻描淡写地发了条推文：它基本能跑！简单的网页能快速且正确地渲染出来。 一个模型究竟能跑多久 传统的AI编程工具，比如Github Copilot和早期的其他IDE，都是一问一答模式。 对话长度有限，上下文有限，任务复杂度有限。 后来出现了所谓的Agentic编程——Claude Code、Cursor Agent、Windsurf等工具让AI可以自主执行多步任务，读取文件、运行命令、修复错误。 这已经是很大的进步，但大多数情 ...

Core Insights - Claude Cowork is a general-purpose intelligent agent designed for work scenarios, built on Anthropic's advanced self-developed model [2] - The development of Claude Cowork took approximately 10 days, with Claude Code writing all the code, although human intervention was still necessary for planning and design [3][5] - The tool is aimed at non-technical users, allowing them to leverage AI capabilities without programming knowledge [8] Development Process - The initial version of Claude Code was in internal testing by the end of 2024, originally named Claude CLI, and was not fully mature in programming capabilities [16][17] - The unexpected usage of Claude Code by data scientists and other professionals led to its evolution beyond just a coding tool [20][22] - The development team realized the need to make the intelligent agent more accessible to non-programmers, resulting in the creation of Claude Cowork [23] Team Dynamics - The development team operated under a tight deadline, forming a small group to release an early version of Claude Cowork [25] - Developers managed multiple instances of Claude to implement features and fix bugs, indicating a collaborative approach to AI management [25][29] - The team prioritized obtaining user feedback early in the process to refine the product [30] Comparison with Competitors - Users have noted that while Manus is suitable for more complex workflows, Claude Cowork is still in its early stages and may be seen as a less sophisticated alternative [31][32] - Caution is advised regarding the trust placed in AI for coding and file operations, emphasizing the need for human review [33][34] Safety Measures - The Claude team has implemented measures to ensure safe operations, particularly when granting file system permissions [36]

以下文章来源于算法爱好者 ，作者小蒜 算法爱好者 . 算法是程序员的内功！「算法爱好者」专注分享算法相关文章、工具资源和算法题，帮程序员修炼内 功。 1 月 15 日，网上流传"西贝初步规划关闭 102 家门店"的消息。消息称西贝于 1 月 11 号在呼和浩特基地开 会，初步规划关闭全国 102 家门店，覆盖超 30 个城市。 | 切步沟通闭店102家门店，清单如下 序列 所在博士 所在城市 印度 | | | | | --- | --- | --- | --- | | | 原创业分部 | 内容 | | | b HOR 9 | | | 保障用明生建议 | 15 日下午， 贾国龙 在朋友圈发文回应称，"（网上）截图信息基本是准确的，遭到了铺天盖地的污蔑 125 天"。 贾国龙强调西贝门店从不用预制菜，称是网红恶意混淆了 "中央厨房备菜" 和 "料理包预制菜" 的概念 ，还恶 意诋毁西贝的急冻保鲜有机西兰花以及进价高的草原羊等食材。他直言自己不懂公关，西贝能连续 11 年蝉联 中式正餐营收第一，靠的是多年来的服务品质，而非公关手段。 "嗯嗯，西贝不是预制菜，继续作，大大方方承认不好吗，我之前也经常吃，后来觉得价格高 ...

Core Viewpoint - The "Did It Die?" app has been taken down, with the developer stating it was "requested to be removed" but did not disclose specific reasons for the action [1]. Group 1 - Speculation suggests that the app may have voluntarily taken down to avoid penalties due to incomplete compliance filing processes [2]. - The app has experienced a brief but notable lifecycle, having been launched in June 2025, renamed, and subsequently taken down within a short span, leading to humorous commentary on its rapid rise and fall [2]. - According to Qimai data, the "Did It Die?" app has faced one previous takedown and three instances of being removed from rankings since its launch, occurring on June 28, September 20, December 12, and December 15, 2025 [2][3]. Group 2 - The core mechanism for app ranking removal on the Apple App Store involves a combination of algorithmic risk control and manual review, triggered by factors such as malicious traffic manipulation, review violations, data misjudgment, and compliance or quality issues [2].

以下文章来源于伯乐在线 ，作者伯小乐 伯乐在线 . 携程净利率是 32%，排第七位。其他互联网大厂没进前十，腾讯拼夕夕网易分别是 29.4%、28.5%、28.2% 手握流量王牌，携程对商家的掌控堪称"生杀予夺"。云南民宿老板吐槽，基础佣金 12%-15%，叠加推广费后 实际成本超 30%，还被强制"二选一"，不独家就遭流量屏蔽。郑州、贵州监管部门早就约谈过它的强制调价问 题，但始终没拦住其垄断套路。 2025 年 11 月，云南民宿协会干脆公开征集 OTA 平台不正当竞争的证据，直指携程等个别平台，为后续反垄 断调查埋下伏笔。 截至 15 日早晨，微博科技热搜 Top50 话题中 12 个和携程相关。 "天下苦携程久矣！" 这句话频繁出现在网易跟帖和公号文章标题中。还有网友点评说" 在携程帝国前面，其他 互联网大厂都是弟弟 "。这话可不是随口吹的， 携程的"底气"来自 碾压级的市场垄断 ，它是实打实的行业霸 主 。 2024 年它的酒旅 GMV 市占率高达 56%，加上控股的同程旅行，"携程系"直接霸占近 70% 的市场。美团和飞 猪分别是 13% 和 8%，这种一家独大的格局，在互联网垂直赛道里少见。 更 ...

Core Viewpoint - A developer named Armas has created a collection of five UEFI-based mini-games that transform the computer booting process into a challenge, requiring users to win the games to successfully start their systems [1][2][7]. Group 1: UEFI Games Overview - The UEFI games are designed to run during the UEFI boot phase, utilizing its capabilities to create an interactive experience before the operating system loads [4][5][6]. - Armas spent approximately 10 months developing these games, which test various skills such as intelligence, reaction time, memory, and cultural knowledge [7]. Group 2: Game Descriptions - The first game, "User Evaluation for Ineptness," presents a math problem that users must solve correctly to avoid being shut down [9][10]. - The second game, "Insult Sword Fighting," requires players to select the correct comeback to a taunt from the system to proceed [13][16][18]. - "Fall To Boot" is a fast-paced game where players navigate a falling cursor through a tunnel, with failure resulting in an immediate shutdown [20][22][23]. - The fourth game, "Age Verification," tests players on 1980s pop culture trivia, with incorrect answers leading to a shutdown [26][29]. - The fifth game, "UEFI Says," challenges users' memory skills, where they must replicate a sequence of colors to boot the system [30][31]. Group 3: Technical Aspects - UEFI serves as the first program that runs after powering on the computer, acting as a bridge between hardware and the operating system, allowing these games to operate with higher privileges than typical applications [34][35]. - The games can directly shut down the computer due to UEFI's power management capabilities, but they require disabling "Secure Boot" to run, which poses security risks [36][37]. Group 4: Community Reactions - The project has sparked discussions on platforms like GitHub and Reddit, with opinions divided between those who appreciate the creativity and those who criticize it as unnecessary [38]. - Supporters argue that it adds a layer of security and fun, while detractors express concerns about practicality and potential risks associated with disabling security features [39][41].