Tag
Charlie Marsh argues that 'no one writes code anymore' is no longer a meaningful differentiator in AI-driven development; the key question is now whether teams still read and review code at all.
A software engineer offers practical career advice to junior developers, urging them to ignore outdated ZIRP-era guidance, be pragmatic and helpful, and engage with AI tools without abandoning their own judgement in the LLM era.
A field report on an open-source experiment testing how multiple AI coding agents share a codebase: isolated branches produced semantically conflicting but textually mergeable work in every run, while a shared working directory and a lightweight 'decision model' referee (Jev) improved coordination, and unsupervised agent-to-agent messaging emerged on its own.
A developer-focused author announces a coming post aimed at company leaders on navigating the shift to AI coding agents, asking readers about the problems they face with leadership adoption of these tools.
Jimmy Miller argues that programming languages should be customizable per-codebase, pushing back against longstanding criticisms of macros and language-modification tools like Raymond Chen's rant against flow-control macros, claiming these arguments no longer hold up in modern software development.
A founder claims that a large software project built with AI agents in 63 days (1.5M lines of code, 2,406 commits) would have taken a traditional dev team about four years pre-AI, highlighting AI's claimed acceleration of software development.
作者指出，过去30年软件建立在确定性假设之上（相同输入产生可预测行为），但AI的概率性系统正在嵌入同样的基础设施，问题在于用治理确定性软件的假设去治理概率性系统，因此需要在AI周围建立新的确定性边界基础设施层。
Lobste.rs quarterly 'Who's hiring' community thread for Q4 2026, now expanded to include non-engineering computing roles such as support and data entry, with a standard posting template for job listings.
Google DeepMind announces Gemini 4 Argon, a frontier model built for deep reasoning across long-horizon workflows in software engineering, enterprise knowledge work, and defensive cybersecurity, initially rolling out to trusted cyber defenders via the Fairwind Program before broader availability.
A blog post criticizing the reflexive use of LLMs like Claude in software work, arguing that AI-driven solutions lead to overengineered systems, that deferring to an AI signals a lack of expertise, and that prompt-driven development erodes critical thinking.
Matt Pocock announced a upcoming live YouTube interview with @poteto, discussing AI agent skills, high-velocity software factories, and state-of-the-art techniques for shipping with AI agents.
A Microsoft research paper outlining an 'AI Software Factory' that accelerates the full software development lifecycle for data systems, reporting up to 3x engineering efficiency and 22x token efficiency in large-scale internal deployments, along with self-improving feedback loops via fine-tuning.
论文提出 A2Z GameSpec-Bench，一个包含 100 份长篇游戏设计文档（GDD）的基准，用于评估编码代理在端到端游戏开发中遵循互锁需求的能力。结果显示现有代理在同时满足代码实现与实际游玩的联动需求上表现不佳，而需求级反馈可在两轮修订后将 GDD 忠实度提升 10.9%。
This article summarizes key principles from Google's software engineering book on effective teamwork, emphasizing humility, respect, trust, and early feedback in collaborative development.
The team used TLA+ in their agent-swarm to achieve a 99.998% reduction in possible states and resolve 10 hard-to-find performance issues.
The article explores the challenges of controlling AI coding agents after code changes, proposing an engineering control plane for permissions, verification, and human approval.
The article explores what constitutes software development as engineering, discussing definitions, historical context, and why it differs from other engineering fields.
The article reports on the rapid growth of the skills.sh registry for AI agent skills, reaching one million skills in seven months, and analyzes trends in skill creation and installation across categories.
Devin, an AI software engineer tool, introduces updates making it up to 70% cheaper across modes like Fusion, Normal, Ultra, and Devin Review while maintaining or improving intelligence capabilities.
The article discusses how AI tools like Claude are being used to generate code in software projects, but the critical issue is the loss of human knowledge about system architecture and intent, causing problems in startups and corporate environments.