Core Viewpoint - The sudden emergence of a new model named "MODEL1" from DeepSeek has generated significant excitement in the domestic large model sector, indicating potential advancements in AI technology and competition in the market [1][3]. Group 1: MODEL1 Features - MODEL1 has been revealed to include advanced technologies such as optimized KV cache layout and support for FP8 sparse decoding, which are expected to greatly enhance inference efficiency and reduce memory usage [2]. - The model integrates a long-context optimization mechanism, addressing the common issue of large models struggling to retain long texts [2]. Group 2: Speculations and Expectations - There is speculation among the community that MODEL1 could either be the long-awaited R2 model, which has faced delays due to chip shortages, or the upcoming V4 model, following the naming convention after V3.2 [3]. - DeepSeek has not officially responded to these speculations, but there are indications that the new model may be released around the Chinese New Year [3].

转自：新智元 【导读】 昨天，Claude刚刚被曝要有永久记忆，今天就被开发者抢先一步。一个叫Smart Forki ng的扩展，让大模型首次拥有「长期记忆」，无需重头解释。开发者圈沸腾了：难以置信，它真 的能跑！ 昨天，一篇Claude要获得永久记忆的爆料，震惊整个AI圈。 这种「知识库」的全新记忆方式，可以让Claude在自己的永存大脑中，自动记住一切。 巧的是，就在今天，就有开发者抢先「截胡」。他在现有工具上实现了一个扩展能力——Smart Forking detection，让大模型第一次拥有了「可继承的长期记忆」！ 这个所谓的Smart Forking，通过给Claude Code会话嵌入向量数据库，让它能从成百上千次的历 史对话中，找到最相关的前文。 因此，你不必再重新解释，Claude可以直接「接上文」搞开发了。 这个功能一出，开发者社区直接炸锅了。大家纷纷表示：「不敢相信，它居然真的跑起来了！」「这 一刻，我是真的被震撼到了。」 强烈推荐每一个Claude Code用户，都把这个思路直接加入自己的工作流！ 英雄所见略同，但路径不同。当官方还在设计永久记忆的形态时，开发者已经用Smart For ...

Core Viewpoint - Anthropic is set to revolutionize AI with the introduction of "permanent memory" in Claude Cowork, transforming it from a simple chatbot into a powerful AI collaborator capable of long-term memory and task execution [1][3][44]. Group 1: Permanent Memory and Knowledge Base - Claude Cowork will feature a new "knowledge base" that allows it to retain information over time, categorizing data instead of relying on a chaotic general memory [12][14][19]. - This knowledge base will enable Claude to actively retrieve relevant background information when responding to queries, enhancing its contextual understanding [16][20]. - Users will have the ability to manage multiple distinct knowledge bases, allowing for tailored interactions based on specific tasks [19][20]. Group 2: Integration of Cowork and Chat Modes - The Cowork mode will become the primary interface for Claude, integrating chat functionalities while emphasizing workflow and productivity [21][28]. - Traditional chat features will still be available but will be folded into the Cowork mode, which will serve as the default workspace for users [22][23]. Group 3: User Interface and Automation Enhancements - The user interface will undergo significant changes, including a dedicated Artefacts sidebar for managing and reusing outputs, moving away from transient interactions [29][30]. - Enhanced automation capabilities will be introduced through the MCP Registry, allowing Claude to dynamically manage remote connectors and improve task execution [33][39]. Group 4: Experience Layer Upgrades - New features such as a web voice mode and upgraded Pixelate functionality will enhance user experience, indicating a shift towards multi-modal and frequent usage [40][41][44]. - These updates suggest that Claude is evolving into a more interactive and capable AI partner, rather than just a question-and-answer tool [44]. Group 5: Industry Context and Future Implications - The introduction of permanent memory aligns with a broader industry trend towards continuous learning in AI, with expectations that 2026 will be a pivotal year for advancements in this area [47][51][65]. - The competition among AI assistants may enter a new phase as companies like Anthropic push the boundaries of what AI can achieve with persistent memory and task execution capabilities [45][56].

Core Viewpoint - The article emphasizes that a healthy public opinion environment is crucial for economic development, and emotional online disputes should not undermine businesses [2] Group 1: Incident Overview - The controversy began with a social media post by Luo Yonghao, which highlighted issues at Xibei, a restaurant chain with over 6 billion in annual revenue, leading to scrutiny over food quality and practices [1] - Xibei's CEO, Jia Guolong, responded strongly, claiming that Xibei does not use pre-prepared dishes, which intensified the conflict and resulted in the announcement of the closure of 102 stores in Q1 2026, with a year-on-year business decline of 50% [1] Group 2: Public Response and Corporate Communication - The first commentary from People's Daily pointed out that the core issue is not about the quality of pre-prepared dishes but the need for companies, influencers, and platforms to reflect on the chaotic public discourse [2] - Jia Guolong acknowledged the commentary, expressing a commitment to improve and serve customers better, indicating a shift towards a more proactive approach in crisis management [2] Group 3: Crisis Management and Transparency - The second commentary stressed that the real issue lies in Xibei's lack of sincerity, suggesting that honesty is the best public relations strategy during a crisis [3] - It was noted that pre-prepared dishes align with industry trends, and consumers are not opposed to them if transparency is maintained [3] - Jia Guolong accepted the commentary's viewpoint, indicating a willingness to adapt and improve in response to the evolving internet landscape [3] Group 4: Role of Influencers and Media - The third commentary focused on the responsibilities of influencers like Luo Yonghao, stating that their public criticisms should be based on facts and reason, rather than personal emotions [5] - It highlighted the need for a responsible approach from influential figures in the media to avoid damaging businesses and to contribute positively to the public discourse [5] - The commentary also referenced legal provisions to protect private enterprises from malicious online attacks, calling for a collaborative effort to maintain a healthy online environment [5]

Group 1 - The core issue is that Pinduoduo's customer service chat function experienced a sudden outage, displaying a message that the "chat function has been disabled" [1] - Most users across the platform were unable to access the customer service feature, with only a few reporting normal functionality [3] - The outage is speculated to be due to regional server fluctuations or a phased restoration of services, with no official response from Pinduoduo at the time of reporting [6]

Core Viewpoint - The company has announced the open-sourcing of its recommendation algorithm for X, aiming to enhance transparency and invite community collaboration in its optimization efforts [1]. Group 1: Algorithm Overview - The recommendation algorithm consists of components such as Home Mixer, Thunder, and Phoenix, focusing on sorting information from "followed accounts" and "content mined from the web" using a Grok-based Transformer model [3]. - A notable feature of the algorithm is the elimination of all manually designed features and heuristic rules, relying solely on the model to learn correlations from user interaction history [5]. Group 2: Algorithm Functionality - The algorithm's workflow involves capturing user behavior data, sourcing candidate content from two channels, and applying data augmentation and multi-round filtering to predict user engagement probabilities (likes, shares, etc.), ultimately calculating a final score for content presentation [5]. - The implementation of X's recommendation algorithm is done using Rust and Python, and it is licensed under the Apache-2.0 license. As of now, it has received 1.1k stars and 195 forks on GitHub [5].

2019 年 3 月，996.ICU 开源项目突然在 GitHub 爆火，从程序员群体出圈成为全民性话题。这事也是当年互 联网标志性事件之一。 那时候，小程程只在关注互联网，看到不少大佬都发表过关于 996 的观点，比如： "996 是修来的福报" 等。 直到西贝预制菜风波后，我才偶然刷到，原来西贝创始人贾国龙也发表过对 996 的看法。 2020 年 9 月初，他公开宣称， "996 算个啥，我们是「715、白加黑、夜总会」，我们就是这么拼，经常是每 周工作 7 天、每天 15 小时，白天加晚上连轴转，夜里还总开会。" 西贝贾国龙 V to 45分钟前 来自 iPhone 11 Pro "715、白加黑、夜总会"在嘴边念叨很多年了，认真一写，无意当中，引起争论。 好多人问我怎么看： "不符合劳动法的事儿，不干！违背员工意愿的事儿，不干！自愿奋斗获得高回 报,支持!" 当然了，在被吐槽和批评后，贾国龙找补说"不符合劳动法的事，不会干"， 西贝官方解释说"715 不是强制性 的"。 当年他这个暴论一出，立刻引发了很多吐槽，还被诸多媒体批评了。 ...

Core Viewpoint - The article discusses Games Workshop's (GW) strong stance against the use of AI in content creation and design processes, contrasting it with the industry's trend of embracing AI as a core business element [1][4]. Group 1: Company Policy on AI - GW has implemented a strict internal policy prohibiting employees from using AI to generate any official content or participate in design processes, including concept setting, illustrations, writing, and sculpture [5]. - Only a few executives are allowed to research AI out of curiosity, and none of them are particularly excited about it [6]. Group 2: Investment in Human Creators - Contrary to industry trends of layoffs, GW is increasing its investment in human creators by hiring more concept designers, illustrators, writers, and sculptors, emphasizing the importance of human creativity and intellectual property protection [6][7]. - The CEO stated that passionate and talented individuals are essential for creating the rich and engaging IP that is Warhammer [7]. Group 3: Community Sentiment and Brand Integrity - GW's zero-tolerance policy towards AI reflects a keen understanding of the community's sentiments, as even minor AI-like elements can trigger trust issues among fans [9][10]. - The company is aware that any association of its artwork with AI could lead to significant backlash, especially given the high value of its products like the Warhammer 40K Codex, which contains premium illustrations and lore [11][12]. Group 4: Legal and Ethical Considerations - GW's cautious approach is rooted in concerns over the copyright and data source issues associated with generative AI, as the current legal framework does not guarantee the transparency of training data or the non-infringement of output content [13]. - The company prioritizes maintaining the integrity of its IP over adopting potentially faster but riskier AI solutions, indicating a preference for a slower, more deliberate approach to innovation [14][15].

Core Viewpoint - The traditional IDEs may become obsolete by 2026, as engineers who do not adapt to new AI development methods will fall behind in productivity and competitiveness [1][5]. Group 1: Future of Development Tools - The transition from traditional IDEs to new AI-driven development tools is likened to moving from handheld tools to CNC machines, emphasizing the need for precision and automation in coding [2][11]. - Current AI models are inefficient, akin to sending a single diver into a vast codebase, suggesting a need for multiple specialized roles to enhance productivity [3][19]. - The future of software development will involve task decomposition, gradual refinement, and collaboration among multiple intelligent agents rather than relying on a single AI model [22][24]. Group 2: Productivity Disparities - There is a significant productivity gap among engineers, with those using AI tools like Codex potentially being ten times more productive than their peers who do not [5][27]. - Management is concerned about the implications of this productivity divide, fearing that they may need to reduce their workforce by up to 50% if engineers do not adapt [5][27]. - The situation mirrors the decline of the Swiss watch industry, where traditional craftsmanship was rapidly overtaken by quartz technology [8]. Group 3: New Development Paradigms - The concept of "Vibe coding" is introduced, where coding is driven by dialogue and AI-generated outputs, moving away from traditional manual coding practices [29][35]. - The economic implications of Vibe coding are significant, allowing for more modular and experimental approaches to development, thus increasing the value of options available to teams [38]. - Examples from companies like Travelopia and Capital One illustrate how AI-driven development can drastically reduce the time and resources needed for project completion [41][49]. Group 4: Organizational Transformation - The integration of AI tools is reshaping technical organizations, enabling non-developers to contribute to coding and deployment processes, thus democratizing software development [44][47]. - Trust in AI increases with usage, suggesting that engineers need to practice and engage with these tools to fully leverage their capabilities [42]. - The rapid pace of code submission and the ability to resolve legacy system issues with AI-generated solutions indicate a transformative shift in how teams operate [49].

很多朋友也都跑来问，如果真是这样，他们能提前做哪些准备？ 有一位在大厂做程序员的朋友老张，之前联系过一段时间，然后现在他做的这条业务线已经开始不赚钱，说他们 团队被集体裁员了 。 从去年开始，身边这种情况越来越多，互联网上也是呜呼哀哉一片，大批人哭诉降薪、失业、裁员，生活质量难以保障。 真的，不知道哪一天，失业就是一夜间的事。 可能你不相信，其实更多的，是你不愿意相信。2008年的时候，也没人会相信诺基亚会消失。 悲剧在有些人身上重复上演，就是因为他们不愿意相信。 所以劝大家，这个时候一定要给自己提前谋出路，多一手准备。 之所以跑来问，是因为他们的这个情况，几年前走过一遭，而且当时的情况比现在还要难，毕竟现在还有时间。 好在后来找到了解法。所以这几年，其实不论外面怎么变，虽说不是稳如泰山，但起码每次都是有对策的。 有对策，心里就不慌。不慌，事情就会慢慢有起色。 这是那一遭之后，我最大的体会。 那一遭，真的，痛得我这辈子也忘不了，不过也成了我这辈子最大的财富。 你说人生好玩吧。 01 早些年啊，我在阿里，每天工作十几个小时，有时候每周 6 天，各种难题到了我这儿，我都能想办法解决。 玩得好的几个同事都说我"活好 ...