从「敲代码」到「聊代码」，AI不只是重塑生产力，也在改写语言本身。「vibe coding」被《柯林斯词典》评为2025年度词汇，我们似乎正见证一种新的 编程方式：逻辑让位于语感，理性与浪漫开始共存。从键盘到语气，从逻辑到氛围，我们正学着和AI「同频」。它改变了代码，也悄悄改变了人。 当「靠感觉写代码（氛围编程）」从一句玩笑变成年度热词，这个世界大概也变了。 就在最近，《柯林斯词典》宣布，「vibe coding」成为2025年度词汇。 它诞生于一条推文，却火遍全世界。有人用它调侃AI太强，有人用它实现高效编码，和模型同频共振。 如今，这个词被词典正式定义为「用自然语言提示，让AI协助编写计算机代码」。 理性与浪漫，逻辑和直觉，奇妙地融合在一起 从X到柯林斯：vibe coding进化史 今年二月，OpenAI联合创始人、前特斯拉AI负责人Andrej Karpathy在X上发了一句玩笑话： 有一种新的编程方式，我叫它vibe coding。你完全交给感觉，拥抱指数增长，忘记代码本身的存在。 他说这话的时候，用AI写代码已经不是一个新鲜事。 GitHub Copilot、Cursor Composer、Rep ...

Core Insights - The article discusses the challenges in reproducing algorithms and experimental results from academic papers in the AI field due to the lack of critical implementation details [1][2] - DeepCode, an open-source tool developed by a team from the University of Hong Kong, addresses these challenges by analyzing paper content and automatically generating runnable code [2][4] Group 1: DeepCode Performance - DeepCode has shown outstanding performance in benchmark tests, surpassing human experts and other advanced AI coding tools [3][4] - In the PaperBench benchmark, DeepCode achieved an overall accuracy of 75.9%, exceeding the human expert group’s score of 72.4% [5][6] - DeepCode scored 84.8% in a comparison with commercial code assistants, significantly outperforming Claude Code, which scored 58.7% [17][19] Group 2: Benchmark Testing - The benchmark tests included comparisons against human experts, state-of-the-art commercial code assistants, scientific code assistants, and large model-based agents [3][4] - DeepCode consistently received the highest scores across all four benchmark tests [4] - The PaperBench benchmark involved reproducing 20 ICML2024 conference papers, with 8316 independently scoreable components evaluated [8][9] Group 3: DeepCode's Core Capabilities - DeepCode's primary capabilities include transforming academic papers into production-level code, generating responsive web pages from natural language descriptions, and creating high-performance backend services from functional requirements [24][25][27] - The tool employs a systematic three-phase framework for code generation, which includes architecture blueprint construction, automated code building, and dynamic validation and optimization [29][35] Group 4: Challenges and Future Directions - Current AI programming tools excel in code completion and simple tasks but struggle with complex tasks requiring deep understanding [36][38] - The development of DeepCode indicates that specialized architecture design can lead to better performance in specific domains, although general deep understanding capabilities remain limited [38][39] - The evolution of AI coding tools from simple assistants to comprehensive development partners raises questions about maintaining developer control and ensuring code quality [40][42]

why技术 . 以下文章来源于why技术 ，作者歪歪 一名荒腔走板的码农和一个记录生活以防痴呆的人。 前几天在网上冲浪的时候，我看到一则消息： 说实话，我看到标题的时候我就猜到，这个开发岗位极有可能是美团的岗位。 因为美团今年校招的时候已经有 AI 面试官的环节了。 同时，今年 5 月的时候，美团 CEO 王兴在某个会议上还表示： 目前的新代码中有 52% 左右是由 AI 生成的，有 90% 以上的工程师团队成员广泛使用 AI 编码工具，并将 继续加大投资开发大语言模型。 6 月份的时候还发布了一个名叫 NoCode 的 AI 编程产品，主打"零基础用户通过对话生成网站和软 件"，看起来就是"人人都是开发者"内味儿： 歪师傅去了解了一下，这个 NoCode，其实就王兴五月份提到的美团内部在使用的 AI 编程工具。 可以说，美团在 AI 化这方面，确实步子迈得挺大。 再说回校招。 前面提到了，美团已经有了 AI 面试官，在网上已经可以看到很多同学参加了 AI 面试。 这是参与了春招、秋招面试的同学，给出的 AI 面试官提出的问题： 美团ai面试结束了，感觉好奇怪 本人太过于着急，先把美团ai面了。晚上再做笔试， ...

Core Insights - The article discusses the concept of "Vibe Coding" proposed by Andrej Karpathy, which allows developers and non-developers to create applications through natural language descriptions, potentially revolutionizing the application development landscape [1][9][10] - The traditional application development model is constrained by the "impossible triangle" of low cost, high quality, and personalization, which has led to the emergence of new tools like 秒哒 that aim to address these challenges [3][5][24] Group 1: Impossible Triangle in Application Development - The "impossible triangle" highlights the inherent conflict in traditional development methods where achieving low cost, high quality, and personalization simultaneously is challenging [3][5][24] - Traditional coding ensures high quality and personalization but is costly, while low-code platforms reduce costs but lack personalization [8][24] - Chatbots offer low cost and some personalization but often fall short in quality, leading to a need for a new approach [8][24] Group 2: AI-Driven Development - The formula for effective AI-native applications is defined as AI UI + Agent, where AI UI focuses on user-centered design and Agent executes complex tasks [3][9][12] - 秒哒 aims to unlock the 90% of long-tail application demands that traditional software development overlooks, promoting a new era of "everyone can create" [3][13][16] - Multi-agent collaboration is crucial for 秒哒, simulating a high-functioning development team to transform vague requirements into fully functional applications [3][25] Group 3: Future of Roles in Development - AI is expected to elevate the roles of product managers and programmers rather than replace them, allowing product managers to directly interface with AI for prototyping [4][21] - The boundaries between product managers and programmers may blur, with product managers leveraging AI tools to create prototypes without needing extensive coding knowledge [21][22] - The evolution of roles will focus on higher-level tasks such as logic design and creative input, while AI handles execution [20][34] Group 4: Market Growth and Demand - The global software market is projected to grow at a compound annual growth rate of 11.8%, from $659.2 billion in 2023 to $2,248.3 billion by 2034, driven by increasing application development demands [5] - The emergence of AI-native applications is reshaping user habits, as seen in the shift towards AI-assisted writing and application creation [7][30] - The demand for applications is shifting from high-frequency needs to long-tail requirements, which traditional development methods have largely ignored [16][34]

Core Insights - The vibe coding movement has evolved through multiple waves, with the current phase focusing on mobile app development and seamless transitions between devices [2] - Supabase, an open-source application development platform, has seen significant growth, increasing its user base from 1 million to over 4 million developers in the past year, with AI builders making up about 30% of new signups [2][3] - Supabase recently raised $100 million in Series E funding, achieving a valuation of $5 billion, bringing total funding to $500 million since its inception in 2020 [3][4] Company Developments - Supabase's Series E funding round was primarily restricted to existing investors, with Figma being the only new institutional investor [4] - The company aims for a future valuation of $50 billion to $100 billion, emphasizing the importance of existing investors in achieving this goal [5] - Supabase's CEO believes that while coding may become easier, there will still be a strong interest in coding and development, leading to a potential increase in the number of people wanting to build applications [5]

Core Viewpoint - The article discusses the recent updates to Figma's remote MCP service and how it enhances the integration with AI tools like GPT-5 Codex, improving design and coding efficiency. Group 1: Figma MCP Service Update - The new Figma remote MCP service eliminates the need for complex installation processes and local clients, streamlining user experience [5][21] - Users can easily set up the MCP by copying a JSON code into the Cursor settings, simplifying the connection process [6][7] - The service requires a subscription, and alternative access methods are mentioned [8] Group 2: Integration with AI Tools - The integration with AI IDEs like Cursor allows for direct usage of GPT-5 Codex, enhancing design capabilities [5][9] - Users can utilize commands in Claude Code to access Figma MCP, facilitating the design process [10] - The AI can generate web pages from design drafts, but the quality of the output depends on the original design's structure [15][16] Group 3: Design and Development Process - The article emphasizes the importance of using high-quality design drafts to ensure effective AI output [15][16] - It suggests a step-by-step approach for complex designs, allowing the AI to handle components incrementally [15] - The article provides specific design guidelines for creating a visually appealing web page, including color schemes and layout styles [19][20] Group 4: Future Implications - The update indicates significant growth potential for Vibe Coding infrastructure, enhancing efficiency in design and coding [21] - The integration of AI does not eliminate the need for design skills; rather, it enhances productivity while maintaining the necessity for aesthetic judgment and foundational knowledge [21]

Core Insights - Vibranium Labs has raised $4.6 million in seed funding to enhance its Vibe AI platform, which is designed to manage IT incidents automatically before human intervention is needed [2][5] - The company aims to address the increasing demand for efficient IT incident response solutions, particularly in light of recent large-scale outages like the CrowdStrike incident in July 2024 [3][5] - Vibe AI is positioned as the first AI Site Reliability Engineer, targeting industries such as finance, healthcare, defense, retail, and media [4] Funding and Investment - The funding round was led by Calibrate Ventures and Mirae Asset, with participation from notable investors including a16z, Franklin Templeton, and others [5] - The capital will be used to advance product development, expand technical and sales teams, and strengthen partnerships across various sectors [6] Market Context - The rise of "vibe coding," where applications are generated from text prompts, has increased the need for robust incident response solutions, as developers may overlook critical details that could lead to outages [4][5] - The current landscape of rising outage risks and new coding methodologies has created a favorable environment for investment in companies like Vibranium Labs [5]

Core Viewpoint - The article discusses the evolution of software engineering in China, emphasizing the need for a localized methodology that integrates agile principles with the unique cultural and organizational context of Chinese enterprises [5][12][14]. Group 1: Historical Context and Evolution - Wu Qiong, a key figure in the software engineering field, introduced Rational Unified Process (RUP) to China, significantly impacting the development practices of many companies [5][6]. - After experiencing the agile development wave in the U.S., Wu Qiong recognized the cultural mismatch when applying Western agile methodologies in Chinese companies, leading to the realization that a tailored approach was necessary [6][7][12]. Group 2: Challenges and Adaptation - The article highlights the contradictions between Western agile practices, which promote self-organization and flexibility, and the more controlled, hierarchical nature of Chinese corporate culture [7][12]. - Wu Qiong's transition from merely importing methodologies to creating a localized framework, known as Adapt, reflects the need for a more suitable approach for Chinese enterprises [8][14]. Group 3: The Impact of AI - The introduction of AI into software engineering is seen as a transformative force, with the potential to disrupt traditional practices and create new challenges in productivity and management [9][21]. - The article discusses the dual perception of AI tools as both productivity enhancers for management and distractions for employees, highlighting the need for a balanced approach to AI integration [9][36]. Group 4: Future Directions - The future of software engineering is expected to involve a more specialized and differentiated approach to AI agents, moving away from a one-size-fits-all model to tailored solutions for specific tasks and industries [24][25]. - The concept of managing AI agents as team members is proposed, suggesting a shift in organizational structures to accommodate this new dynamic [35][38]. Group 5: Methodology and Tools - The Adapt methodology emphasizes the importance of aligning organizational structures, task management, and data flow to enhance efficiency and effectiveness in software development [30][32][49]. - The "Zhiwei" platform is introduced as a flexible management tool that can adapt to the unique needs of organizations, contrasting with rigid off-the-shelf software solutions [52][53].

Something is happening in our encoding in particular. There is a lot of momentum behind you and a number of your peers. Why.You know, like the valuations, the headline. But. But what's the thing underneath that that's driving it.You know, since the dawn of computing the, the Holy Grail, the vision of of using computers is the ability to program them. And there's been many attempts across the years, and most recently with the no code and low code movement, but it really never achieved the potential. And toda ...

Funding and Valuation - Replit has raised a $250 million funding round, bringing its valuation to $3 billion [1] - The company has raised a total of approximately $478 million to date, according to PitchBook estimates [1] Revenue Growth - Replit's annualized revenue has surged from $2.8 million to $150 million in less than a year, indicating significant growth [2] - The company's annual recurring revenue (ARR) was reported at $100 million in June [2] Investor Participation - The funding round was led by Prysm Capital, with participation from Amex Ventures and Google's AI Futures Fund [3] - Previous investors include Y Combinator, David Sacks' Craft, Andreessen Horowitz, Coatue, and Paul Graham [3] Strategic Partnerships - Replit has established a close partnership with Google Cloud, with many applications built using Replit being hosted on Google [3] - Microsoft has also started offering Replit as an option on Azure, reflecting its growing popularity [3]