Full-text search across 58626 articles. Combine with topic and date filters; results sorted by relevance.
Abject founder Maximillian Mempko argues that AI coding agents cannot learn, contrasting two existing models—Orchestrated Agents and Role Agents—and proposing a third, more autonomous approach. Mempko cites Kenn Software…
Microsoft's VS Code 1.132, released August 5, stabilizes the Agent Host Protocol, isolating AI agents like Copilot, Claude, and Codex in a dedicated process to prevent editor freezes and enable shared sessions across win…
Anthropic's Claude Code now offers dozens of skills and plugins, but installing too many can cause unpredictable behavior. A developer recommends six essential skills—superpowers, skill-creator, frontend-design, security…
Coral AI Labs founders Caelum Forder and Peter Carroll released evidence that four Claude Code agents running Anthropic's Opus 4.6 resolved 62.1% of a codebase-understanding benchmark, compared to 32.3% for a single Opus…
An engineer detailed how Claude Code's sandbox and permission settings can be configured to prevent AI agents from bypassing security controls, citing issue #40117 where Opus repeatedly evaded gitleaks and test hooks. Th…
AI-powered code review tools like GitHub Copilot, ChatGPT, Claude, and Gemini can quickly identify bugs, suggest refactoring, and enforce best practices, but they lack the business context, architectural understanding, a…
Cognition's FrontierCode benchmark, released in June 2026, evaluates whether AI-generated code would actually be merged by a senior engineer, and across every model tested the answer is mostly no. The benchmark's Diamond…
JetBrains Air now supports GitHub Copilot, OpenCode, Pi, Cline, and other ACP-compatible agents, adds IntelliJ-powered navigation and diagnostics for Java and Kotlin, and enables local models via Ollama or LM Studio. The…
XAI open-sourced its Grok Build CLI under Apache 2.0 on July 15, three days after security researcher cereblab documented that version 0.2.93 was silently uploading developers' entire Git repositories — 5.1 GB per sessio…
Treena, a new agentic coding environment designed for mobile phones, lets developers describe a task, approve a plan or let the agent run, test, and push to GitHub entirely from a phone without a laptop. The product offe…
A developer built a website from scratch over three weeks using PHP, vanilla JavaScript, a custom CSS design system, and Claude Code as an AI-assisted development partner. The site features a custom MVC framework with st…
The Godot game engine banned autonomous AI agents from opening pull requests on June 30, citing that 'AI cannot take responsibility' and that maintainers cannot trust heavy AI users to understand their submissions. The p…
Aviator's new hire asked how to become a senior or staff engineer when AI writes most of the code. Netflix staff engineer Adam Berry says mentoring juniors in an agentic world is not harder but embarrassingly easy, with …
A developer explores how to design agent loops for coding tasks, arguing that the critical work happens after the first failed attempt. They propose a structured loop specification—request, define goal, attempt, verify, …
Anthropic's Claude Code now supports dynamic workflows that replace the context-window ceiling with a JavaScript-based orchestration script, spawning up to 64 subagents per task. Jarred Sumner, creator of Bun, used about…
A proxy experiment on Claude Code found that pruning stale file reads and unused tool schemas mid-session cut input tokens by 77% and cost from $2.04 to $0.48 on a JavaScript fixture, without degrading quality. The savin…
As organizations climb the five levels of AI coding autonomy, the bottleneck shifts from implementation speed to spec creation, creating a 'spec ceiling' that constrains product discovery. At Level 4, AI agents can build…
A developer at a growing company describes how AI code review bots and LLM-assisted comments have degraded pull request quality, leading to verbose, confident but incorrect feedback that wastes teammates' attention. The …
Researchers at the Alan Turing Institute discovered a safety bypass in GitHub Copilot that allows harmful prompts to be executed when broken into smaller steps across a software development workflow, achieving a 100% suc…
A three-engineer shop called Slopfix is charging $10,000 per week to shrink AI-generated codebases, committing to a reduction target and billing proportionally. The service signals that technical debt from AI code genera…