Core Viewpoint - Linus Torvalds emphasizes that AI is just another tool in software development, similar to compilers, which enhances productivity without eliminating the need for programmers [1][24]. Group 1: Role and Contributions - Over the past two decades, Torvalds has transitioned from being a programmer to a technical leader and maintainer, primarily observing the progress of Linux and Git rather than actively coding [5][7]. - The core work of long-standing projects like Linux is maintenance and continuous support, with ongoing modifications to the kernel for better organization and stability [9][10]. Group 2: AI and Software Development - AI's role in the Linux kernel is still experimental, with ongoing efforts to utilize AI for patch management and code review, but it has also caused significant disruptions [21][20]. - Despite the hype around AI, Torvalds believes that it will not replace programmers but rather enhance their efficiency and open new development opportunities [24][25]. Group 3: Hardware and Industry Trends - The rise of companies like Nvidia and AMD has shifted focus from traditional CPUs to accelerated processing units (APUs), although Linux remains integral to system maintenance and operation [17][18]. - The involvement of Nvidia in the Linux kernel has increased due to the growing importance of AI in cloud services, marking a positive shift in collaboration [19]. Group 4: Personal Insights and Management Style - Torvalds admits to spending most of his time reviewing emails rather than coding, often not responding unless there are issues, which can lead to a perception of being unapproachable [30][31]. - He encourages finding hobbies outside of work to manage stress, highlighting the importance of engaging in activities where failure is acceptable and can be enjoyable [26][27].

Core Insights - Linus Torvalds emphasizes that he has transitioned from being a programmer to a system maintainer, focusing on overseeing the development of Linux rather than directly coding [4][6][7] - The introduction of AI in software development is viewed as a tool that enhances productivity without replacing programmers, similar to how compilers transformed programming practices [21][25] - The rise of Nvidia and AMD has shifted the hardware focus away from traditional CPUs to accelerated processing units, yet Torvalds believes that general-purpose CPUs remain crucial for Linux [17][18][19] Group 1: Role Transition and Development - Torvalds states that for nearly 20 years, he has not been a programmer but rather a technical leader and maintainer of Linux [4][6] - He notes that the core work of long-standing projects like Linux is maintenance and continuous support, rather than reaching a final completion point [8][9] - The development model of the Linux kernel has remained stable over the past 15 years, although Torvalds has shifted from primarily saying "no" to sometimes saying "yes" to new ideas [10][11] Group 2: AI and Software Development - Torvalds has not personally used AI to assist in coding but acknowledges that others are exploring its application in kernel development [21][23] - He highlights that AI's impact on the Linux community has been largely experimental, with some disruptions caused by AI crawlers affecting kernel.org [20][21] - The potential for AI to enhance productivity is recognized, but it is believed that the need for skilled programmers will persist as new development areas emerge [25][26] Group 3: Hardware Evolution - The conversation notes a significant shift in hardware focus from CPUs to GPUs and APUs, driven by companies like Nvidia and AMD [17][18] - Torvalds argues that while AI and accelerated processors are gaining attention, the role of Linux in managing systems and user interfaces remains vital [18][19] - The involvement of Nvidia in the Linux kernel development is seen as a positive outcome of the AI boom, indicating a growing interest in Linux from hardware manufacturers [19][20]

Core Viewpoint - Linus Torvalds emphasizes that AI is just another tool in software development, similar to compilers, which enhances productivity without eliminating the need for programmers [2][27]. Group 1: Role and Contributions - Linus Torvalds states that he has transitioned from being a programmer to a system maintainer and observer, highlighting that the real contributions come from others in the community [6][9]. - Over the past 35 years, the core work of maintaining and supporting the Linux kernel has become increasingly important, with ongoing modifications to improve code cleanliness and stability [10][11]. Group 2: AI and Software Development - Torvalds has not personally used AI to assist in coding and believes that while AI can enhance productivity, it will not replace programmers [27][25]. - The Linux community is exploring AI's potential in code review and maintenance, but most applications remain experimental and face challenges, such as AI interference with infrastructure [23][22]. Group 3: Hardware and Industry Changes - The rise of Nvidia and AMD has shifted focus from traditional CPUs to accelerated processing units (APUs), although Torvalds maintains that general-purpose CPUs remain crucial for Linux [19][20]. - AI's growth has led to Nvidia's increased involvement in the Linux kernel, which is seen as a positive development for the open-source community [21]. Group 4: Personal Insights and Hobbies - Torvalds shares his experience of building guitar pedals as a hobby to relieve stress, emphasizing the importance of having interests outside of work [29][30]. - He acknowledges that while he spends much of his time reviewing emails, he rarely responds, which may lead to misconceptions about his demeanor [32][34].

"过去将近 20 年里，我其实已经不是程序员了。" "至于我发明的 Git，我现在也只是旁观者的角色。" "我以前常说我的工作就是拒绝（提案），但如今反而要在一些长期维护者的反对声中，对新东西说'同意或者是'。" "Vibe Coding 让人做到了以前做不到的事情，但从维护者视角来看，要维护它生成的代码'可能糟糕透顶了'。" 这些话并非是玩笑话，也非自嘲，而是 Linux 之父、Git 的发明者 Linus Torvalds 在面对技术浪潮时的清醒自白。 本月早些时候，Linus Torvalds 与 Verizon 开源负责人 Dirk Hohndel 在韩国首尔举办的 Linux Foundation 开源峰会上进行了一场对谈。他谈到了自己角色的 转变、谈到了 AI 如何重塑软件开发，以及在越来越多的硬件更多依赖 Nvidia 的专有 GPU 和 CUDA 而不是开源 Linux 时的想法，也谈到了 Rust 在内核团 队引发的冲突，聊到了 kernel.org 被各种 AI 爬虫工具严重干扰的现实困境，还有自己日常面对的压力以及缓解方式。 在 AI 热潮几乎重写开发者命运的当下，Torvalds 坦言 ...

Core Insights - Rust has become a controversial programming language, with government agencies in the U.S. advocating for its adoption over C/C++ due to its memory safety features, while some developers express skepticism about its complexity and perceived overhype [1][2]. Group 1: Rust's Impact on Android Security - Memory safety vulnerabilities in Android have dropped below 20% for the first time, according to Google's 2025 data [2]. - Rust has reduced the density of memory safety vulnerabilities by 1000 times compared to existing C/C++ code in Android [4]. - The introduction of Rust has not only improved security but also enhanced software delivery efficiency, with rollback rates decreasing by 4 times and code review times reduced by 25% [4][15]. Group 2: Adoption and Trends - Since 2021, Google has been integrating Rust into the Android system as a safer alternative to C/C++ [5]. - The usage of Rust is rapidly increasing, while new C++ code is declining [6]. - Rust's new code volume is now comparable to that of C++, indicating similar development efficiency [9]. Group 3: Performance Metrics - Google utilized the DORA framework to assess performance, focusing on throughput and stability [10]. - Rust code requires approximately 20% fewer modifications than C++ code of similar scale [11]. - Rust's rollback rate is about one-fourth that of C++ in medium to large changes, indicating higher stability [18]. Group 4: Broader Applications of Rust - Google is expanding Rust's use in various areas, including system services, libraries, and applications, due to its safety and productivity advantages [22]. - Specific implementations include Nearby Presence for Bluetooth device discovery, RCS security messaging, and various parsers in Chromium [23]. Group 5: Addressing Concerns and Future Outlook - Google acknowledges that while Rust does not guarantee zero vulnerabilities, it significantly reduces vulnerability density, estimating 0.2 vulnerabilities per million lines of Rust code compared to 1000 per million lines of C/C++ [32][33]. - The company believes that Rust allows for a balance between speed and safety, potentially restoring performance and productivity previously sacrificed for security [37][38].

Core Insights - The article emphasizes that technology selection, particularly programming languages, often masks deeper issues related to personal identity and emotional attachment rather than purely technical considerations [1][4][18] - It highlights the importance of recognizing the hidden conversations that influence decision-making processes in technology choices, which can lead to significant financial implications for companies [17][19] Group 1: Case Studies - The first case involves a company, Takkle, where a new CTO's decision to switch from PHP to Perl resulted in a 9-month delay in product launch and a doubling of monthly burn rate from $200,000 to $500,000, ultimately leading to financial distress [5][6] - The second case at Google illustrates a similar pattern, where a vice president's push for Rust over Go was based on emotional and identity-driven reasoning rather than a thorough analysis of technical merits [7][8][11] Group 2: Decision-Making Dynamics - The article distinguishes between visible conversations focused on technical attributes and invisible conversations centered on personal identity and professional aspirations [9][10][18] - It argues that decisions driven by identity can lead to substantial costs, as technology stack choices account for 40% to 60% of total development costs over a product's lifecycle [17][19] Group 3: Recommendations for Improvement - Companies are encouraged to shift the focus of technology discussions from "which language is best" to "what are the costs associated with this language," encompassing all dimensions that affect survival and growth [19][20] - A framework is suggested to make hidden costs visible, allowing for more rational and economically driven decision-making in technology selection [19][20]

出于「安全」考虑，越来越多的组织加入了 Rust 语言的阵营。 近日，Debian 社区开发者 Julian Andres Klode 发布了一则声明，宣布从 2026 年 5 月开始， Debian 的 APT 软件包管理工具将强制要求使用 Rust 工具链。 他强调，"如果你维护的端口没有可用的 Rust 工具链，请在接下来的 6 个月内确保配备，否则就淘汰该端口。" 这一消息让不少 Debian 用户和开发者感到担忧。因为这意味着 Debian 的所有架构都必须支持 Rust，而目前一些尚未具备 Rust 支持的端口，要么需要尽 快开发支持，要么可能淘汰。 Debian 强制使用 Rust 的原因 Julian Andres Klode 在公告中写道："在 APT 中引入 Rust 硬依赖和 Rust 代码。最初涉及 Rust 编译器和标准库，以及 Sequoia 生态系统。特别是我们解析 .deb、.ar、.tar 文件的代码，以及 HTTP 签名验证相关代码，将大大受益于内存安全语言和更完善的单元测试方法。" 他还表示："对整个项目来说，能够向前发展并依赖现代工具和技术非常重要，而不是试图在老旧设备 ...

Core Viewpoint - The choice of programming languages is becoming increasingly important in the AI era, as it directly impacts the quality of code generated by AI agents [19][28]. Group 1: Programming Language Comparison - Go is favored in AI scenarios due to its thin abstraction layer and structured nature, making it easier for models to understand and rewrite code. In tests, Go outperformed Python and Rust in generating code for similar small programs [2][27]. - Python remains essential for any company, especially for tasks involving machine learning or data processing, even if it is not used for core services [12][16]. - JavaScript and TypeScript are also unavoidable in the current landscape, with TypeScript often accompanying JavaScript [12][17]. Group 2: Language Evolution and Future Trends - The industry is witnessing a trend towards creating "next-generation languages" designed for human-agent collaboration, as existing languages may not be optimal for this new paradigm [3][29]. - The migration from Python 2 to 3 serves as a cautionary tale for future language transitions, highlighting the complexities involved in such changes [4][6][7]. - Rust has learned from Python's migration challenges by implementing an "edition system" that allows for incremental feature adoption without breaking compatibility with older versions [7]. Group 3: Practical Considerations in Language Choice - The choice of programming language should be pragmatic, focusing on the product being built rather than the code itself. Early-stage companies should limit their technology stack to three or four languages [11][18]. - The emergence of AI tools has shifted the focus from the necessity of a unified codebase to maintaining clear boundaries between systems, enhancing development efficiency [18][20]. Group 4: AI's Impact on Software Development - AI tools are significantly changing the software development landscape, allowing for more efficient coding and problem-solving. A substantial portion of code (over 80%) in some companies is now generated by AI [21][24]. - The role of human developers is shifting towards creative and thoughtful tasks, while AI handles more routine coding responsibilities [21][24]. - The democratization of programming is occurring as AI lowers the entry barrier, enabling more individuals to engage in coding without extensive prior knowledge [25]. Group 5: Error Handling and Language Design - Different programming languages exhibit varying error handling characteristics, which can significantly impact system reliability and user experience [34][35]. - The design of programming languages often involves trade-offs between performance and error handling capabilities, which can affect the overall robustness of applications [40][42].

Core Viewpoint - A good "taste" in technology is crucial for developing AI products, and it is distinct from technical ability. Cultivating a good technical taste can lead to results that exceed one's technical capabilities [2][5]. Group 1: Importance of Engineering Taste - Engineering taste is defined as the ability to choose appropriate engineering values for current projects, as most decisions in software engineering involve trade-offs between different goals [6][11]. - The essence of technical taste lies in understanding that every decision in software engineering is a trade-off, and recognizing the balance between conflicting engineering values is a hallmark of maturity in the field [11][15]. Group 2: Characteristics of Good and Bad Taste - Good taste is difficult to identify compared to technical ability, as it involves selecting suitable engineering values for specific technical problems. Success in projects can indicate good taste [16][17]. - Bad taste often stems from rigidity, where engineers advocate for methods that worked in past projects without considering their suitability for current projects [13][15]. Group 3: Engineering Values - Key engineering values that define technical taste include resiliency, speed, readability, correctness, and flexibility. Each engineer prioritizes these values differently based on the project requirements [11][12]. - Other important values include portability, scalability, and development speed, which can influence preferences for programming languages and architectural decisions [14]. Group 4: Developing Good Taste - To cultivate good taste, it is recommended to try different types of work and observe which aspects of projects are easy or challenging. Flexibility in thinking about software development is also essential [17][18].

Core Insights - The GOSIM HANGZHOU 2025 conference highlighted the integration of open-source and AI technologies, showcasing their potential across various industries and emphasizing the importance of community collaboration in driving innovation [1][3][4]. Group 1: Conference Overview - The conference attracted over 200 global leaders in open-source and AI, along with more than 1500 developers, featuring keynote speeches, high-end forums, and specialized discussions on AI models and infrastructure [1][3]. - Keynote speakers included influential figures from organizations like the United Nations and major tech companies, discussing the significance of open-source in AI development and global collaboration [3][6][7]. Group 2: Community and Collaboration - The event emphasized community engagement, with forums dedicated to the Rust programming language and hands-on workshops that fostered interaction among developers [4][5][15]. - The conference featured a strong focus on practical applications, including hackathons that encouraged developers to create innovative solutions in real-time [22][24]. Group 3: AI and Open Source Integration - Discussions on the future of AI highlighted the need for high-quality training data and the challenges of integrating AI into real-world applications, stressing the role of open collaboration in overcoming these hurdles [8][12]. - The conference explored various AI themes, including embodied intelligence, intelligent agents, and the next generation of AI technologies, showcasing advancements and potential applications [10][12][14]. Group 4: Workshops and Practical Engagement - A total of 14 workshops were organized, allowing developers to engage in hands-on learning and collaboration on cutting-edge technologies [17][20]. - The workshops covered a range of topics, from AI inference to cross-platform development, providing participants with practical skills and insights [18][20]. Group 5: Future Directions and Closing Remarks - The conference concluded with a call for continued collaboration in the open-source AI community, setting the stage for future events and innovations [33][34]. - GOSIM HANGZHOU 2025 served as a platform for fostering connections between academia and industry, promoting ongoing dialogue and exploration in the tech community [29][31].