Tag
Sonar CEO Tariq Shakir’s talk points out that AI agents have caused code output to surge (+2,450% in lines of code, +240% in commits), yet releases have grown only 30% — the “scaling from production to delivery” problem remains unsolved. He proposes a “code confidence formula” built by multiplying intent, guardrails, verification, and evidence, helping software factories deliver confidence rather than blind trust.
gohawk 是一款面向 Go 语言的工业级静态分析工具，专注于资源管理和并发问题检测，已用于发现并修复 Docker、Kubernetes 和 Caddy 中的真实 bug。
Part 2 of a blog series arguing that Modern C++ is a mindset rather than a new standard: writing correct, fast, safe code by default via algorithms like std::equal and C++20 ranges instead of classic loop-based, error-prone constructs, and warning that AI tools are scaling up the generation of unsafe classic C++ code.
Qodo launched a Codex plugin that injects organizational codebase context and independent AI code review into developers' coding sessions, helping catch cross-repository P0 bugs before pull requests are opened.
The open-source tool ai-design-check breaks down the "AI look" into 70+ countable check rules (spacing grid, type scale, gradient buttons, emoji icons, 'Welcome back' greetings, etc.). It's zero-dependency, requires no AI or API key, runs on React/Vue/Svelte/HTML/SwiftUI/Jetpack Compose, and can be hooked into Claude Code so the agent automatically fixes design issues itself.
A team enabled CodeRabbit's slop detection to label low-effort AI-generated PRs, sparking a debate over whether publicly marking work as 'slop' helps quality or shames contributors. They ultimately required human review before the label appears.
noalloc 是一个用 AI 编写的 Go 工具链，可通过 `//noalloc:root` 等指令检查函数调用树中的堆分配，并提供固定容量的零分配数据结构（含 JSON 编解码器）以帮助编写无堆分配代码。
Uncle Bob (Robert C. Martin), a senior software engineer with 50 years of experience, has released the ub-mode Skill for coding AI agents, transforming the engineering principles of Clean Code into agent constraint rules, and establishing quantitative gates for cyclomatic complexity, code duplication, and mutation testing before delivery through three tools: crapper, dryer, and mutator. This article fully breaks down the rule system and tests the toolchain's performance on a real open-source project.
一篇关于 AI 编码代理（如 Codex）承担大部分开发工作后，工程师仍需保留何种技能的思考文章，作者以 SonarQube 报错引发的排查经历为例，提出「知道何时出错并能深入排查」比亲手写代码更重要。
The author explains the common mistakes when iterating backwards in C and presents the clean idiom `for(size_t i = count; i--; )`, which avoids signed-index and wrap-around pitfalls, especially as LLM-generated code repeats the buggy versions.
A developer reflects on how AI evolved from a 'rubber duck' debugging companion into a capable coding assistant, weighing its benefits against risks like code entropy, skill atrophy for juniors, and unhelpful assumptions.
作者分享了在一周内用 AI 辅助但不盲信 AI 的方法，安全删除约 19000 行编译器死代码的经验：先用静态分析界定范围，再让 LLM 标记 RED/BLUE 函数，并用独立的调用图工具交叉验证标记完整性，最终实现零回滚、零修复提交的清理。
A Hacker News discussion sparked by a Meta engineer's LinkedIn post questioning the growing practice of shipping AI-written code without fully understanding it, asking how developers cope with this new standard.
A practical guide on how to trust AI coding agents more, emphasizing an 'inverse pyramid' of effort: codebase improvements, static analysis, prose guidance, then human review. It links to Cursor's official plugins repo for installing agent-friendly tooling and workflows.
开发者并未被 AI 取代，但 AI 取代了「理解自己代码」的环节：代码由模型生成后无人真正读过，调试退化为把报错喂回模型。文章引用 Veracode 对 100+ 模型的测试（45% 生成代码存在已知安全问题），质疑这种「用理解换速度」的权衡是否会反噬整个行业。
一篇探讨 AI 时代代码“churn”（频繁重写）观念转变的随笔：当编写代码的成本大幅下降，可抛弃的高质量“一年生”代码与稳定的“多年生”核心架构应被明确区分，churn 本身不是问题，问题在于承重部位的无计划变动。
FrontierCode 1.1 is now available on Artificial Analysis; built with open-source maintainers, it is the first benchmark measuring end-to-end code quality and mergeability — whether a maintainer would actually merge the PR.
开发者讨论在 AI 代理规模化工作流中如何验证代理自称的“已完成且测试通过”，提出需要外部独立的验证步骤与可留证的调用记录，而非信任代理的自测结果。
Kerno 1.0 正式发布，定位为 AI 运行时后端 QA 工程师，可映射代码库、自动测试全后端（含用户流程、安全检查、后台任务、队列、MCP 服务器和 AWS 无服务器），并在 PR 进入 CI 前捕获并复检 Bug。
Matt Pocock argues that when AI coding agents write bad tests, the fault lies with the codebase structure that permits such tests, not the agent itself.