作者 | Alain 译者 | 平川 策划 | Tina 本文最初发布于 Alain 的个人博客。 我不经常发表博文。当我这么做时，是因为我觉得还没大有人把我注意到的事情说出来。 我一直在从头开始构建一个产品。不是那种"我启动了一个 Next.js 模板"的从头开始。我的意思是 从网络配置到产品设计再到定价决策。真正的端到端。我一直在使用比较前沿的模型和编码代 理，每天花费数小时，无论是在这个项目上还是在我的全职工作中。我一直在努力远离混乱和炒 作，努力筛选出真正有价值的东西。 自 2025 年 12 月以来，事情戏剧性地变好了。许多人都注意到了，但很少有人得出过正确的结 论。 自动化编程 Antirez 喜欢称之为"自动化编程"，我真的很喜欢这种表述。与"氛围编码"这个肤浅、几乎带有轻 蔑意味的标签相比，它更能抓住本质。在人类历史上，自动化一直是大多数工作和文化革命的核 心所在。印刷机、织布机、装配线，这一次并没有太大的不同。 我的大部分工作还是那样。对于我想要构建的东西，我仍然需要深入地思考每一个重要的方面 ——架构、权衡、产品决策、凌晨 3 点反复思考的边缘情况。消失不见的是那些需要逐行敲击代 码的令人 ...

Core Insights - Vercel has appointed Mitchell Hashimoto, co-founder of HashiCorp and creator of Terraform, to its board of directors, bringing significant expertise in developer communities and infrastructure [1][2][3] Company Developments - Hashimoto's appointment comes at a pivotal time for Vercel, following a $300 million Series F funding round that valued the company at $9.3 billion, alongside a GAAP revenue run-rate of $340 million and an 84% year-over-year growth [3][4] - Hashimoto has a strong history with Vercel's products, having been a customer and user of Next.js since its inception, which aligns with his commitment to the company's mission [3][4] Industry Context - Hashimoto is recognized as a leading figure in the open source community, having developed widely adopted tools like Terraform, which is now the standard for infrastructure-as-code [2][4] - His experience includes leading HashiCorp to a $6.4 billion acquisition by IBM in 2024, showcasing his capability in scaling technology companies [2][3] Strategic Vision - Hashimoto expressed enthusiasm for Vercel's focus on building a powerful developer presence and industry-defining products, emphasizing the importance of trust and quality in software development [4][5] - Vercel aims to enable rapid, secure, and scalable development for AI applications, positioning itself as a key player in the evolving tech landscape [5][6]

Core Insights - The article discusses the development of vinext, an AI-driven framework that serves as a plug-and-play alternative to Next.js, built on Vite, which significantly improves build speed and reduces client bundle size [2][4][8]. Group 1: vinext Overview - vinext was created by an engineer using AI to reconstruct popular front-end frameworks, achieving a build speed increase of up to 4 times and a client bundle size reduction of up to 57% [2][12]. - The project cost approximately $1,100 in Token fees [3]. - vinext allows for easy deployment to Cloudflare Workers with a single command, handling all aspects of application building and deployment [17]. Group 2: Next.js Challenges - Next.js, while popular, faces deployment challenges in serverless environments due to its highly customized toolchain, requiring modifications for compatibility with platforms like Cloudflare and AWS Lambda [4][5]. - OpenNext, another project aimed at addressing similar issues, has limitations and requires reverse engineering of Next.js outputs, leading to unpredictable changes [5][6]. Group 3: Performance Benchmarking - Initial benchmarks show vinext outperforming Next.js in build times, with vinext achieving a production build time of 1.67 seconds compared to Next.js's 7.38 seconds, marking a 4.4x speed improvement [14]. - The client bundle size for vinext is significantly smaller, with gzip sizes of 72.9 KB compared to Next.js's 168.9 KB, representing a 57% reduction [14]. Group 4: Features and Capabilities - vinext supports incremental static regeneration (ISR) and is designed to work seamlessly with Cloudflare's KV caching processor, allowing for flexible caching strategies [18][26]. - The framework is built to be compatible with various hosting services, not limited to Cloudflare, and aims to foster collaboration with other platforms [21]. Group 5: AI Integration in Development - The development of vinext was expedited by AI, which played a crucial role in coding and testing, achieving a high quality of code with extensive testing coverage [40][41]. - The project utilized AI to automate code reviews and testing processes, enhancing efficiency and maintaining quality standards [40][41]. Group 6: Future Directions - vinext is still in the experimental phase, with ongoing testing and user feedback indicating positive results, including successful deployment in production environments [23][24]. - The framework aims to evolve by incorporating traffic-aware pre-rendering, optimizing build times by focusing on high-traffic pages [30][31].

Core Insights - Vercel has appointed Susan St. Ledger, former President of Worldwide Field Operations at HashiCorp, to its board of directors, bringing over two decades of enterprise go-to-market leadership experience [1][3] Company Overview - Vercel recently completed a $600 million Series F funding round, achieving a valuation of $9.3 billion, as it scales its AI Cloud platform [3] - The company provides AI-native infrastructure for both emerging AI startups and global enterprises, powering popular frameworks like Next.js and AI SDK [3][5] - Vercel has doubled its user base over the past year and achieved 82% year-over-year top-line growth [3] Leadership and Expertise - Susan St. Ledger has a proven track record in scaling go-to-market organizations, having previously helped Okta and Splunk triple their revenues, with Splunk's revenue growing from approximately $700 million to nearly $2.5 billion [2] - She has served on the board of directors at Klaviyo and was involved with HashiCorp from pre-IPO through its acquisition by IBM in 2025 [2][3] Strategic Vision - St. Ledger emphasizes the evolution of cloud computing and the role of AI in automating manual tasks, allowing developers to focus on innovation [4] - Vercel aims to enable customers to build with the speed, security, and scalability necessary to succeed in the AI landscape [4]

Network News - A critical vulnerability in React Server Components, tracked as CVE-2025-55182 and nicknamed React2Shell, is actively exploited by multiple threat groups, putting thousands of websites, including crypto platforms, at risk of having users' assets drained [1] - The flaw allows attackers to execute code remotely on affected servers without authentication, with widespread exploitation observed shortly after its disclosure [1] - The bug affects React versions 19.0 through 19.2.0, including packages used by popular frameworks such as Next.js, and merely having the vulnerable packages installed can allow exploitation [1] Ripple Developments - Ripple is expanding its U.S. dollar-backed stablecoin, RLUSD, to Ethereum layer-2 (L2) blockchains, including Optimism, Coinbase's Base, Kraken's Ink, and Uniswap's Unichain, aiming to deepen its integration into the multichain ecosystem [2] - The company is starting with a test phase ahead of a wider rollout expected next year, pending regulatory approval from the New York Department of Financial Services (NYDFS) [2] - The pilot integrates Wormhole's Native Token Transfers (NTT) standard, allowing RLUSD to move natively across chains without wrapping or synthetic assets, which helps maintain liquidity and regulatory control [2]

Core Viewpoint - The article introduces Next AI Draw.io, an AI-powered tool that automates the process of creating and modifying diagrams in draw.io, significantly enhancing efficiency and user experience [2][7]. Group 1: Product Overview - Next AI Draw.io allows users to generate diagrams by simply describing what they need, such as "draw a Transformer architecture diagram with animated connectors" [7]. - The tool can also reconstruct existing diagrams by uploading an image and requesting modifications, such as changing components or adding new elements [9]. - It features a history tracking system that allows users to revert to previous versions of their diagrams, providing a safety net for users [10]. Group 2: Key Features - The tool utilizes large language models (LLMs) to directly generate draw.io XML, enabling users to focus on verbal instructions while the AI handles the drawing [10]. - Users can upload images to automatically recreate editable diagrams, ensuring that lines and layouts are neat and organized [10]. - An interactive chat interface allows for real-time updates and modifications to diagrams, such as adding nodes or changing database types [10]. Group 3: Technical Details - Next AI Draw.io supports various LLMs, including AWS Bedrock, OpenAI, and Google AI, which can be configured through a local environment file [17]. - The application is built using Next.js for the frontend and integrates with Vercel AI SDK for streaming AI responses [19]. - Installation options include a one-click Docker setup or a manual installation process, providing flexibility for users [24][26].

Core Concept - Introduces the concept of "human-in-the-loop" middleware for Langchain agents, allowing human review and intervention in agent workflows [5][18] - Explains the agent's reasoning loop: reason, act, observe, and how human intervention fits into this loop [3][5] - Highlights the three decision types for human reviewers: approve, edit, and reject, and how these decisions guide the agent's subsequent actions [7] Technical Implementation - Demonstrates the integration of a Langchain agent with human-in-the-loop middleware in a Nextjs application for sending emails [2][17] - Emphasizes the importance of a checkpointer (using Redis database) to store the agent's state and enable resuming the workflow after human intervention [13][14] - Describes how the middleware intercepts tool calls (e g, sending emails) and pauses the agent's execution, awaiting human input [5][6] Benefits and Use Cases - Positions human-in-the-loop as a way to combine agent autonomy with human oversight, especially for actions with risk or requiring judgment [18][19] - Suggests use cases such as sending emails, updating records, or writing to external systems, where human review is valuable [19] - Underscores the flexibility of the middleware, allowing customization of interruption logic based on tool name, arguments, or runtime context [19][20] Practical Example - Provides a practical example of using the middleware to allow a human to revise an email drafted by the agent before it is sent [2][16] - Showcases how to reject a proposed action and provide feedback to the agent, influencing its subsequent behavior [16] - Mentions a publicly available repository (github com/christian broman/lunghat) for users to experiment with the human-in-the-loop concept [20]

Core Insights - Vercel's V0 has evolved from a simple "AI webpage builder" to a comprehensive Agent capable of automating planning, research, construction, and debugging across front-end, back-end, and content creation [1][2] - The company emphasizes that every organization will undergo an AI transformation, which is fundamentally about large-scale production of intelligence [1][4] - V0 serves as a "zero patient" to validate the value of AI Cloud infrastructure [1] V0 Growth and User Engagement - V0 can generate 7 new applications per second and has produced over 100 million applications to date, surpassing Vercel's total user base from the past decade in less than a year [2][3] - The concept of "Vibe coding" allows users without engineering backgrounds to create applications easily, reflecting a shift towards democratizing software development [2][3] Product Development and Team Structure - Initially developed by a small team of four, V0's rapid launch was facilitated by a dedicated group focused on creating a "zero patient" application to test the AI Cloud infrastructure [3][4] - Vercel has implemented a "GM model" for V0, where a general manager acts as the CEO of V0 while also being a user of the Vercel platform [4][5] Market Position and Competitive Landscape - Vercel aims to expand its user base from 1.4 million active developers to potentially billions, as the need for full-stack development becomes accessible to smaller teams [4][5] - The company positions itself as a foundational platform for the future of AI-driven applications, similar to the models of Microsoft and Amazon AWS [5][6] Future Directions and AI Integration - Vercel plans to enhance V0's capabilities by integrating with third-party platforms like Supabase and Salesforce Commerce Cloud, enabling users to build full-stack applications [11][12] - The company envisions a future where AI Cloud will provide interfaces for Agents, allowing for seamless interaction between AI systems and traditional applications [19][20] User Experience and Community Engagement - V0 is designed to be user-friendly, allowing anyone with a computer to start building applications without the need for extensive technical knowledge [8][9] - The company emphasizes the importance of community feedback and examples to help users effectively utilize V0 [12][13] Financial Performance and Growth Metrics - Vercel reported significant revenue growth, reaching approximately $180 million, representing an 80% increase year-over-year, driven largely by the success of V0 [33][34] - The company is exploring monetization strategies for individual users, similar to an App Store model, to create a sustainable creator economy [33] AI Cloud Vision and Infrastructure - Vercel's AI Cloud aims to automate traditional cloud computing challenges, providing solutions directly through AI Agents rather than merely reporting issues [25][26] - The company is developing a self-healing system that can automatically detect and resolve infrastructure issues, enhancing operational efficiency [26][27] Conclusion - Vercel is at the forefront of transforming how applications are built and deployed, leveraging AI to create a more accessible and efficient development environment for users across various backgrounds [1][2][4]

Core Insights - OpenAI claims that GPT-5 excels in front-end coding, outperforming its predecessor in 70% of internal tests [2] - Mixed reviews from developers indicate that the initial excitement around GPT-5 may be overstated, with some users reporting a decline in performance [3][4] - A poll conducted by AI engineer Shawn Wang revealed that over 40% of respondents rated GPT-5 as "average" or "poor" [4] Developer Experiences - Influential developer Theo Browne initially praised GPT-5 but later expressed disappointment, stating that its performance had worsened over time [3] - A GitHub Copilot user criticized GPT-5 for its weak summarization and explanation capabilities, comparing it unfavorably to Claude Sonnet 4 [3] - Developers are exploring the potential of GPT-5 to create applications without traditional frameworks like React, suggesting a shift in front-end development practices [7][8] Performance Comparisons - The ability of GPT-5 to create websites without frameworks has impressed some developers, raising questions about the necessity of tools like React [8] - Differences in performance between various versions of GPT-5 have been noted, with some users experiencing less impressive results with non-premium versions [10] - A study by Sonar highlighted the varying coding styles and effectiveness of different AI models, indicating that GPT-5's coding personality is still being evaluated [11]

Core Insights - OpenAI's GPT-5 shows impressive performance in front-end web development, outperforming its predecessor in 70% of internal tests [5][6] - User experiences with GPT-5 are mixed, with some developers expressing disappointment compared to earlier expectations [6][7] - A significant portion of users rated GPT-5 as average or poor in a poll, indicating that OpenAI's promotional claims may be overly optimistic [7][8] Group 1: Performance and Reception - GPT-5 is supported by Vercel, which claims it to be the best front-end AI model [6] - Influential developers have had varying opinions, with some initially praising GPT-5 but later expressing dissatisfaction with its performance [6][7] - A GitHub Copilot user reported that GPT-5's summarization and explanation capabilities were lacking, favoring competitors like Claude Sonnet 4 [6] Group 2: Development Capabilities - Developers are exploring the potential of GPT-5 to create applications without relying on frameworks like React, using only HTML, CSS, and JavaScript [13] - GPT-5's ability to generate complete technical stacks and working prototypes has been highlighted by users [11][13] - The emergence of AI tools like GPT-5 raises questions about the necessity of traditional frameworks in front-end development [13] Group 3: User Experience and Variability - User experiences with GPT-5 vary significantly, with some using less powerful versions leading to disappointing results [14][15] - Different models of GPT-5 exhibit distinct coding styles, which may affect user satisfaction and performance [15][16] - The ongoing evaluation of GPT-5's coding personality is crucial for developers to understand its capabilities and limitations [17]