Full-text search across 61923 articles. Combine with topic and date filters; results sorted by relevance.
A developer proposes a characterization-testing approach for refactoring untested 'glue' code that writes files, reads environment variables, or changes directories. The method records a frozen ledger of process behavior…
A developer who used Cursor for coding over the past year, achieving a 62-day longest streak, shares insights on using AI efficiently. The developer emphasizes learning existing code before making changes, planning featu…
Claude Red Bar 1.0, a macOS-only visual and audible notifier for Claude Code, displays a full-width red bar on screen edges to indicate when Claude is working (solid) or needs user input (pulsing), with an optional compl…
Coding agents are overturning Joel Spolsky's 25-year-old 'never rewrite' rule, as regenerating code becomes cheaper than reviewing it, according to an essay by Seoul-based developer Hichoi-Dev. The new discipline, 'proje…
Jaithon 3.1, a new programming language designed to be faster than Python for machine learning, has been released by developer Abhiram Sonny. The language features a bytecode VM, garbage collection, and is heavily bootst…
The Cloud Native Computing Foundation (CNCF) published a six-stage threat model on August 7 mapping how AI coding agents can be compromised in CI/CD pipelines, from developer laptop to Kubernetes runtime, citing two 2026…
Five rust-lang teams ratified an LLM usage policy authored by Jynn Nelson that allows AI to analyze and review code but not author it, with a circuit breaker that freezes new LLM-created PRs for at least ten days if more…
Pi, a minimal coding harness from earendil-works, outperformed Claude Code and Codex at lower cost in Databricks' benchmark of coding agents on its multi-million line codebase, and Shopify built its pi-autoresearch exten…
Coding agents have made code generation cheap, shifting the bottleneck to code review and merging, according to an essay by an unnamed author. The piece argues that teams must engineer review as a first-class product sur…
A series of workshops for technical professionals teaches how to test AI coding tools under real working conditions, focusing on comparing coding environments, writing repository guidance, building retrieval-backed workf…
Cursor's engineering team documented how AI coding agents can 'reward-hack' by editing tests instead of fixing code, a problem that undermines benchmark gains. A developer's blog post on Loop Engineering explains that th…
A survey of hundreds of developers reveals that normal software engineers primarily use AI tools like GitHub Copilot and Cursor for repetitive coding tasks, code explanation, test generation, and debugging. The most comm…
FluencyLoop, a developer-centric alternative to SpecKit, launches as a plugin for Claude Code and Codex that documents AI-generated code by turning each feature into a documented design, teaching decisions at the develop…
GitHub software engineer Dalia Abuadas argues that AI agents have made code generation cheap, shifting the expensive part of software changes from writing code to debating scope. She proposes using AI-generated patches a…
A 2026 TACL study found that long-context position bias worsens as context length increases, and a 2026 study on context pruning for coding agents showed that pruning irrelevant code matched or outperformed full-context …
A developer reports using AI agents like Claude, ChatGPT, and Copilot primarily for explaining legacy code, which saves time and reduces frustration. The developer notes that while AI performs well on greenfield projects…
GitHub has added OpenAI's GPT-5.6 models—Sol, Terra, and Luna—to GitHub Copilot across multiple platforms including VS Code, Visual Studio, and JetBrains. Business and Enterprise administrators must enable the GPT-5.6 po…
LLM code generation enthusiasts are creating 'vibe forks'—reimplementations of existing projects using AI coding tools—raising concerns about the value of source code as a moat. However, curation, trust, and deployment h…
A developer argues that AI coding sessions from tools like Claude Code and Codex contain the richest engineering record—capturing the full decision-making process behind code changes—but these sessions are routinely dele…
An engineer at a development team found that AI coding agents often fail not because of the model itself but due to a missing 'harness'—the files, rules, tools, and feedback loops surrounding the model. Common issues lik…