编辑｜冬梅 北京时间昨日深夜，OpenAI 正式发布了新一代智能体编码模型 GPT-5.2 Codex，并在官网同步发布 技术博客，对该模型的定位、能力改进及部署方式进行了说明。 OpenAI 重磅发布 GPT-5.2 Codex 据介绍，GPT-5.2 Codex 基于通用模型 GPT-5.2 构建，并针对"智能体编码"（Agentic Coding）场景 进行了专门优化，主要面向复杂的软件工程任务。相较此前版本，新模型在长程任务执行、大规模代 码变更、Windows 原生环境支持以及网络安全相关能力等方面进行了系统性改进。 在工程能力层面，OpenAI 表示，GPT-5.2 Codex 通过引入原生的上下文压缩（compaction）机制， 提高了对超长上下文的理解与利用效率，使模型在跨文件、跨模块的长期编码任务中具备更稳定的表 现。同时，该模型在代码重构、迁移等涉及大规模改动的场景下，整体可靠性和一致性有所提升。 安全能力也是此次更新的重点方向之一。OpenAI 在博客中提到，随着模型在推理与工具调用能力上 的增强，其在网络安全领域的适用性也随之提升。 官方披露，就在上周，一名安全研究人员使用 GPT- ...

Core Insights - Snowflake's BUILD event represents a significant platform for discussing cloud architecture, large-scale parallel computing, and data processing in the Data + AI field [3][4] - The introduction of BUILD in China signifies the integration of international technology with the local developer ecosystem, highlighting the importance of Chinese developers in global innovation [5] Group 1: Event Overview - The BUILD event is a tribute to the core developer activity of building, evolving into a leading platform for cloud and data discussions [3] - BUILD is not just a conference name; it symbolizes extreme performance and limitless scalability in the Data + AI sector [4] Group 2: Developer Empowerment - The event showcases practical cases from top global experts, focusing on topics like Agentic AI and multi-modal data processing, providing developers with insights from concept to prototype [7] - BUILD aims to accelerate the transition from development to production for Chinese developers, offering tailored solutions for various application scenarios [7] Group 3: Future Engagement - The upcoming BUILD 2025 event will further engage the Chinese technical community, providing insights into how Snowflake empowers business growth in the AI era [9] - Snowflake's AI data cloud is already utilized by 766 of the Fortune Global 2000 companies, indicating its significant role in helping businesses innovate and derive value from data [10]

Core Insights - The article discusses the launch of Doubao Model 1.8 by Huoshan Engine, which is optimized for multi-modal agent scenarios, featuring a context window of 256k and various token limits for input and output [2][3]. Model Performance - Doubao 1.8 achieves a processing speed of 5000k tokens per minute (TPM) and 30k requests per minute (RPM), leading to significant improvements in various benchmarks, surpassing competitors like Gemini 3 [3][4]. - In specific benchmarks, Doubao 1.8 scored 94.6 in AIME-25 for mathematics and 85.7 in GPQA-Diamond for reasoning, indicating its strong performance across multiple tasks [4]. Multi-modal Capabilities - The model has enhanced multi-modal understanding, excelling in visual judgment, spatial understanding, document parsing, and video motion recognition, positioning it among the global leaders in these areas [3][7]. - Doubao 1.8 can efficiently process long videos, quickly identifying critical moments, which has applications in various sectors such as online education and safety inspections [5][7]. Business Applications - The model's capabilities allow for complex agent construction, which can create significant value across various industries, with a reported daily token usage exceeding 50 trillion, marking a 417-fold increase since its launch [6][16]. - Huoshan Engine introduced the "Doubao Assistant API," enabling businesses to utilize core agent capabilities easily, with plans to expand functionalities [16][17]. Cost Efficiency Initiatives - The "AI Savings Plan" offers unified pricing for enterprises using large models, allowing for cost savings of up to 47% based on usage [17]. - The "Inference Outsourcing" service allows businesses to upload encrypted model parameters without managing GPU infrastructure, potentially halving hardware and operational costs [18][19]. Creative Tools - The article highlights advancements in Doubao's image and video generation capabilities, including the new Seedream and Seedance models, which enhance creative processes in various applications [8][9]. - Seedance 1.5 Pro introduces features like synchronized audio-visual output and multi-language support, significantly improving content creation efficiency [9][13].

Core Insights - The article emphasizes that the most valuable opportunity in the industry currently lies in AI Coding, with the no-code tool "秒哒" (MiaoDa) generating significant value and user engagement globally [2] - The tool has been widely adopted, serving over 10 million users and creating more than 5 billion yuan in value across various application scenarios [2] User Demographics and Needs - 81% of MiaoDa's users are non-programmers, primarily from the workforce and academic sectors, highlighting the challenge of articulating complex application needs [3] - The team has designed a unique approach to address this issue, differentiating MiaoDa from similar products by enhancing the "demand communication phase" [4] Demand Communication and User Experience - MiaoDa employs a "Product Manager AI" to facilitate deeper conversations with users, transforming vague requests into structured product documentation [4][6] - This design significantly lowers the barrier for expression and reduces the risk of rework due to unclear requirements [6] Strategy and Development Focus - The current strategy prioritizes building a robust foundational capability to ensure a smooth user experience, with plans to develop vertical versions tailored to specific business needs in the future [7] - MiaoDa also offers a deep research mode for complex business requirements, integrating with mainstream AI platforms for enhanced functionality [7] Backend Challenges and Solutions - The article discusses the challenges in backend capabilities, leading to the emergence of the BaaS (Backend as a Service) concept, with MiaoDa being recognized as a leading player in this space [9] - Key challenges include ensuring stable and efficient cloud database operations, integrating AI with databases, and managing underlying resources effectively [9][10][11] Product Experience and User Accessibility - MiaoDa aims to simplify complex database capabilities into an easy-to-use product experience, achieving a significant advantage by allowing users to complete integrations in one conversation without external configurations [12] - The article highlights the importance of maintaining code quality and preventing "code bloat" through careful review and structured development processes [15] Competitive Landscape and Future Outlook - The article concludes that while China has started its Vibe Coding development 1-2 years later than abroad, the competitive gap is narrowing, with Chinese products expected to catch up rapidly [17] - The core competitive advantages in Vibe Coding products are identified as strong code generation capabilities and superior user experience [17]

Core Viewpoint - Google has launched its latest AI model, Gemini 3 Flash, which emphasizes speed and efficiency, aiming to enhance learning, building, and planning capabilities for users globally [3][4][6]. Group 1: Model Performance and Features - Gemini 3 Flash is touted as Google's fastest AI model, showcasing advanced reasoning and knowledge capabilities, achieving a score of 90.4% in GPQA Diamond and 33.7% in Humanity's Last Exam without using tools [8][12]. - The model significantly outperforms its predecessor, Gemini 2.5 Pro, with a score of 37.5% in Humanity's Last Exam compared to 11% for Gemini 2.5 Flash and 34.5% for GPT-5.2 [8][13]. - Gemini 3 Flash is designed to operate efficiently, achieving a 3x speed increase over Gemini 2.5 Pro while maintaining lower costs, with input pricing at $0.50 per million tokens and output pricing at $3.00 per million tokens [15][16]. Group 2: Target Applications and Use Cases - The model is suitable for complex applications such as video analysis, data extraction, and visual question answering, making it ideal for developers seeking intelligent applications [22]. - Gemini 3 Flash can perform near real-time multi-modal reasoning, enabling interactive experiences from static images and facilitating rapid A/B testing of designs [25][27]. Group 3: Market Position and Competitive Landscape - Google positions Gemini 3 Flash as a primary model for large-scale tasks, aiming to lower the barrier for enterprise adoption and overall costs [28]. - The launch of Gemini 3 Flash comes amid increasing competition with OpenAI, as Google aims to capture a larger share of the consumer market, which has reportedly affected ChatGPT's traffic [30][31].

前几天，今年软考考试出成绩了，我报的「系统架构设计师」成绩还不错~ 话说回来，现在报名软考的程序员越来越多了，为此政策都发生了重大调整：程序员爱报考的 「系统分析师」和「系统架构设计师」由一年考一次 改成一年考两次；系分跟系规都发生了重大 改版 。 | 考试日期 | 级别 资格名称 考务要求 | | --- | --- | | | 系统分析师 | | | 系统架构设计师 高级 | | | 系分辨和时要构 E 31 | | | 在春亚洲球队或在两次 | | 考试日期 | 级别 | 资格名称 | 考务要求 | | --- | --- | --- | --- | | | | 系统分析师 | | | | | 系统架构设计师 | | | | 高级 | | | | | | 网络规划设计师 | | 软考全称计算机技术与软件专业技术资格（水平）考试，是计算机领域唯一一个国家级证书，私 企、企事业单位都非常认可，招聘时是大大的加分项。 并且，证书不限制报名门槛，一经考取，永久有效，全国范围内通用。 然而，想考取软考证书，可没有那么容易。官方教材上千页，怎么看？教材已改版，哪里搞最新 重难点？真题年年有，答题有没有套路…… 正好 ...

Core Insights - The article highlights the evolution of AI applications from showcasing capabilities to becoming integral roles in daily life and work, emphasizing the shift towards practical utility and reliability [4][10]. Event Announcements - An important AI ecosystem event titled "AI Shining in China" will take place on December 28, 2025, in Xiamen, where the annual "AI Application Ecosystem Research" results will be revealed [1]. - The first AI hardware exhibition of the year will be held from January 8 to 11, 2026, in Shenzhen, featuring over 76 categories, 200 exhibitors, and 1000 cutting-edge products [3]. AI Application Trends - The latest AI application rankings reveal three main trends: the specialization of productivity tools, the industrial leap in design tools, and the rise of AI in everyday life and education [4][10]. - AI applications are increasingly focusing on delivering specific outcomes rather than emphasizing technical specifications, with tools becoming more integrated into existing workflows [10]. Featured Applications - Notable applications include Youdao Translation AI Writing, which integrates translation and writing processes, and a gamified homework assistant for children that transforms assignments into engaging activities [6][7][8]. - Other applications mentioned are SocialEcho for managing multiple overseas social media accounts, and tools like Chuangkit and Gaoding AI that focus on design and content production [7]. Application Ranking Mechanism - The ranking of AI applications is based on community feedback, with core metrics including comment counts, collections, and recommendations from registered users [10][11].

Core Viewpoint - Xiaomi has introduced its new large model MiMo-V2-Flash, which emphasizes efficiency and practical deployment over sheer size, marking a significant step in its AI exploration journey [4][9]. Group 1: Model Overview - MiMo-V2-Flash features a total parameter scale of 309 billion, with only about 15 billion parameters activated during inference, utilizing a MoE (Mixture of Experts) architecture [8]. - The model incorporates Multi-Token Prediction (MTP) technology, designed for high-speed inference and agent workflows, aiming for efficiency rather than just increasing parameter size [8][21]. - Xiaomi's approach to MiMo-V2-Flash is driven by the need for models that are not only intelligent but also practical and deployable in real-world scenarios [21][22]. Group 2: Performance Metrics - During inference, the model achieves a throughput of 5000 to 15000 tokens per second in a single-machine environment, with a single request output speed of 150 tokens per second, representing a speed increase of approximately 2-3 times compared to models without MTP [24][47]. - MiMo-V2-Flash has entered the first tier in seven mainstream evaluations, particularly excelling in the SWE-Bench test with a 71.7% accuracy rate [27][28]. Group 3: Future Directions - The next generation of intelligent agents must be capable of continuous interaction with the real environment, moving beyond mere language processing to a unified, dynamic world model [30][32]. - Xiaomi emphasizes that true intelligence arises from interaction rather than just textual understanding, indicating a shift towards models that can engage with the physical world [52][53].

Core Viewpoint - The focus of enterprises is shifting from "whether to use AI" to "how to effectively implement AI" and "how to continuously create value within organizations" as AI moves from concept to practice [1] Group 1: Event Overview - The Global AI Development and Application Conference and Shijingshan Digital Transformation Summit (AICon Beijing Station) will be held on December 19-20, 2025, in Beijing [1] - The conference aims to promote the integration of AI technology with enterprise digital transformation, featuring a project launch meeting, a national AI digital forum, and an AI industry summit [1][2] Group 2: Project Launch Meeting - The project launch meeting will focus on the practical challenges faced by enterprises during their digital transformation and will officially kick off the regional AI service implementation [2] - The meeting emphasizes a "project-based, service-oriented, and executable" approach, aiming to clarify transformation paths for enterprises [2] Group 3: National AI Digital Forum - The national AI digital forum will take place on the afternoon of December 19, focusing on "Elevating Enterprise Strategy and Industrial Innovation Waves in the AI Era" [3][9] - It targets C-level decision-makers and aims to help enterprises navigate uncertainties by discussing AI technology evolution, industry trends, and organizational changes [3][10] Group 4: AI Industry Summit - The AI industry summit on December 20 will focus on practical applications of AI in business scenarios, emphasizing "making good use of AI" and "executing AI effectively" [4][13] - The summit will feature case studies from frontline enterprises, helping managers understand how to integrate AI into organizational operations for efficiency and capability enhancement [4][13] Group 5: Agenda Highlights - The agenda includes keynote speeches and discussions on various topics, such as AI's impact on industry, enterprise AI practices, and the evolution of enterprise operating systems [7][8][11][12] - Specific sessions will cover themes like policy support for digital development, AI-driven industry trends, and practical AI applications in business processes [11][13]

Core Viewpoint - The article discusses the recent IPO of MuXi Integrated Circuit (Shanghai) Co., Ltd., marking a significant event in the domestic GPU industry, following the successful listing of Moore Threads. This indicates a shift from the technology validation phase to a market and capital verification phase for domestic GPU companies [2][23]. Company Overview - MuXi officially listed on the Sci-Tech Innovation Board with a stock price of 679 CNY per share, achieving a market valuation of over 270 billion CNY [3][5]. - The company was founded in 2020 and has established R&D centers in multiple cities across China, with a core team averaging nearly 20 years of experience in high-performance GPU development [10][18]. Financial Performance - MuXi's revenue grew from 42.64 thousand CNY in 2022 to 743 million CNY in 2024, reflecting a compound annual growth rate (CAGR) exceeding 4000% [15]. - Despite significant revenue growth, the company reported a cumulative net loss of over 3 billion CNY over three years, attributed to high R&D investments and talent incentives [15]. Fundraising and Investment - The company raised approximately 4.197 billion CNY through its IPO, with funds allocated for the development of high-performance GPUs and AI inference GPUs [8][16]. - Notable investors include Sequoia China and other prominent venture capital firms, indicating strong market interest and support for MuXi's growth [16]. Product Development and Strategy - MuXi aims to provide a full-stack GPU chip and system solution, focusing on various applications including AI training, cloud computing, and data centers [10]. - The company has outlined a clear product evolution strategy, transitioning from validating GPU architecture to deploying GPUs capable of handling AI training and high-performance computing tasks [12][13]. Market Position and Competitive Landscape - The entry of MuXi and Moore Threads into the public market is seen as a pivotal moment for the domestic GPU industry, reshaping investor perceptions and expectations [21][24]. - The competition is shifting from merely producing GPUs to ensuring sustainable sales and market presence, highlighting the industry's long investment cycles and high failure rates [25].