转自：InfoQ 一位谷歌资深工程师透露，Anthropic 的 Claude Code 用一小时跑出了她团队打磨了一年的系统。 "我不是在开玩笑，这件事一点也不好笑。" 她补充说，这个元旦假期里自己终于第一次有时间拿一些"玩具项目"随手试一试。由于无法分享任何 专有信息，提示里也不可能包含真实的内部细节，所以她只是基于一些已有思路搭了一个"玩具版"来 评估 Claude Code；整个问题描述也就三段文字。 "它还不完美，我也还在不断迭代和打磨，但现实就是这样。"Dogan 写道，"如果你对编码 Agent 持 怀疑态度，就把它拿到你已经非常熟悉的领域里试一试：从零开始做一个足够复杂的东西，由你亲自 来判断它产出的这些'成果物'到底靠不靠谱。" 她同时说明，在谷歌内部，Claude Code 目前只允许用于开源项目，不得用于公司内部代码。 补充一点背景是，Dogan 说自己做编程语言相关工作以来，几乎没见过开发者社区对同一件事出现 如此两极化的反应。她在 1 月 3 日晚发文感叹：围绕 coding agents 的讨论里充斥着大量"炒 作"和"空话"，这些噪音常常把真正扎实的工作淹没掉；她强调，事情本不必 ...

Coding agents are becoming one of the most active areas in applied AI, yet many teams keep rebuilding fragile infrastructure every time models or providers change. We believe there is a better way. By anchoring on a stable abstraction layer like Codex, we can stop worrying about harness rewrites and focus on the parts of the stack that create lasting value. We treat models as interchangeable sub-agents, plug into shared primitives, and let upstream improvements flow through without breaking products. This l ...

RT eric zakariasson (@ericzakariasson)turns out, senior engineers accept more agent output than juniors. this is because:- they write higher-signal prompts with tighter spec and minimal ambiguity- they decompose work into agent-compatible units- they have stronger priors for correctness, making review faster and more accurate- juniors generate plenty but lack the verification heuristics to confidently greenlight outputshows that coding agents amplify existing engineering skill, not replace it ...

Hi there, this is Christian from Lchain. If you build with coding agents like cursor, you probably recognize this. The first few turns with the agents are great.But then as you keep continuing talking to the agent in the same thread, the quality slides, the decision get more fuzzy and the overall code quality drops and then cursor drops this system line context summarized. That's the moment you know you've crossed the context boundary line. So why is summarization such a big deal for context engineering.Eve ...

AI Coding Evolution - 软件开发正经历从终端到图形界面，再到AI辅助的快速演变 [1][2][3][4] - Cursor 旨在通过AI 自动化编码流程，重点在于模型和人机交互 [46] - Cursor 的目标是让工程师更专注于解决难题、设计系统和创造有价值的产品 [47][49] Context Engineering & Coding Agents - Context Engineering 关注于为模型提供高质量和有针对性的上下文信息，而非仅仅依赖 Prompt 技巧 [16][17] - Semantic Search 通过自动索引代码库并创建嵌入，提升代码搜索的准确性和效率 [19][20] - Semantic Search 将计算密集型任务转移到离线索引阶段，从而在运行时获得更快、更经济的响应 [22] - Cursor 发现用户更倾向于使用 GP 和 Semantic Search 相结合的方式，以获得最佳效果 [22] Cursor's Products & Features - Tab 功能每天处理超过 4 亿次请求，通过在线强化学习优化代码建议 [7] - Cursor 正在探索多种 Coding Agents 的管理界面，包括并行运行和模型竞争 [38][39][42][43] - Cursor 正在探索为 Agent 提供计算机使用权限，以便运行代码、测试并验证其正确性 [44] - Cursor 允许用户通过自定义命令和规则，共享 Prompt 和上下文信息，实现团队协作 [32][33]

Are coding agents really just creating a bigger problem for production. I think the problem in my opinion with VIP coding is that it isn't going far enough, right. Um, and yeah, you know, you've gotten to this place where you can build code quickly.Um, and then you deploy it, something breaks, but now you don't have that deep understanding of like what you actually built, right. And somebody else has to go and fix that and then that learning doesn't sort of transfer back through, right. Um but maybe one way ...

Core Insights - The article discusses the transformative impact of AI on software development, emphasizing that AI is not replacing developers but rather reshaping the coding process and the role of developers [1][2][4] Group 1: Evolution of Software Development - The introduction of AI tools like GitHub Copilot has changed the starting point, process, and purpose of coding, moving from traditional coding practices to a more collaborative and creative approach [1][2] - AI is enabling a shift from "vibe coding" to "agentic DevOps," where developers act as orchestrators rather than mere code writers [1][2][4] - The initial skepticism about AI's ability to generate code has been replaced by recognition of its effectiveness, with early data showing that Copilot wrote approximately 25% of the code in enabled files [5][6] Group 2: User Experience and Interaction - The integration of features like Tab completion has significantly lowered the learning curve for developers, making coding more accessible [7][8] - Developers have adapted to using AI tools by leveraging existing coding habits and learning behaviors, such as modifying code snippets from various sources [9][10] - The user feedback for Copilot has been overwhelmingly positive, with a net promoter score of around 72, indicating high satisfaction among users [6] Group 3: The Role of Developers - The role of developers is evolving to include validating the outputs generated by AI agents, ensuring that the code meets business objectives and maintains security standards [13][14] - Learning programming is still essential, but understanding how to effectively use AI tools is becoming equally important in the software development landscape [11][12] - Developers must continuously adapt their skills to incorporate AI and new models into their workflows, as the landscape of software development is rapidly changing [15][16] Group 4: Open Source and Collaboration - GitHub's decision to open-source Copilot reflects a commitment to the developer ecosystem and aims to foster innovation and collaboration within the community [17][18] - The open-source nature of Copilot allows developers to learn from the code and potentially create competing products or integrate similar functionalities into their own tools [19][20] - The integration of multiple models and tools is expected to drive further innovation in software development, allowing for more tailored solutions [22][23] Group 5: Future of Software Development - The boundaries between deterministic and non-deterministic code are becoming blurred, with future software engineering requiring the ability to navigate both realms [24][25] - There is potential for a future where software systems are generated in real-time, with AI agents assisting in various tasks, leading to a more seamless user experience [26][27] - The concept of interconnected agents that can manage both personal and work-related tasks is emerging, suggesting a future where AI plays a central role in daily life [40][41]

AI Development & Application - The industry is currently using outdated software development techniques for AI features, hindering the full potential of AI, which should enable users to program software using natural language [1][18] - AI application development is often approached by trying to fit AI into existing applications, rather than redesigning tools from the ground up to automate repetitive tasks [18][62] - The industry needs to move beyond the chatbot paradigm and focus on AI's capability to automate work and accomplish tasks on behalf of users [58][60] - A key element is providing users with "tools" that agents can use to accomplish tasks, such as labeling emails, archiving them, or writing drafts [53][54] System Prompts & User Control - Current AI applications often hide the "system prompt" (instructions given to the AI) from the user, limiting customization and personalization [1][11] - The industry should allow users to view and edit system prompts, enabling them to tailor the AI's behavior to their specific needs and preferences [8][10] - Allowing users to control system prompts shifts the responsibility for the AI's output from the developer to the user [35] - While not everyone may want to write system prompts from scratch, the option should be available, and AI could assist in generating and customizing prompts based on user history and feedback [41][42][48] Future of AI Development - The industry needs to develop better tooling and UI conventions for interacting with and teaching AI, potentially including AI-assisted system prompt writers [45][46] - AI models are good at processing instructions and turning them into text output, making them particularly effective for coding agents [31][32] - Founders should rethink existing tools from the ground up with AI, focusing on offloading repetitive work from users [61][62]