Full-text search across 61857 articles. Combine with topic and date filters; results sorted by relevance.
Cost.dev (YC W21) launched a new CLI designed for coding agents that reduces Claude's output token usage by up to 79% and API costs by up to 67% compared to a bare-Claude baseline. The tool runs locally to provide determ…
Tamarillo's analysis of approximately 400,000 public GitHub repositories reveals that five coding harnesses—Claude Code, Codex, Copilot, Cursor, and Hermes—dominate roughly 80% of repositories with harness configurations…
A developer has built shadcncraft, a design-to-code workflow that connects Figma directly to production React code using AI. The system uses four components—a Figma kit, a React registry, and two MCP servers—to let AI to…
The article introduces the concept of "AI slop debt," which refers to the accumulation of functional but soulless code generated rapidly by AI tools like Claude Code, outpacing human ability to review or maintain it. Unl…
DltHub's commercial AI harness and Snowflake's Cortex Code (CoCo) coding agent operate on opposite sides of a data pipeline, with dltHub's harness handling ingestion into a lakehouse and CoCo taking over once data lands …
AWS announced AWS Continuum, a service that discovers, prioritizes, validates, and remediates code vulnerabilities including agent-generated and agent-touched code, at AWS Summit NYC 2026. As of September 2026, penetrati…
Archify, an agent skill for Claude Code, Cursor, Codex, and OpenCode, generates architecture diagrams from code repositories via a coding agent that creates a typed JSON description, which Archify validates and renders a…
Bun 1.4, released last week, includes a rewrite from Zig to Rust with over 1 million new lines of Rust code, driven by one developer, Jarred Sumner, using a pre-release version of Fable 5 and an unlimited token budget, p…
Open source libraries are evolving into reference implementations for coding agents, according to an analysis by a software engineer. The shift is driven by the rise of coding agents that can generate code from scratch, …
Large language models (LLMs) are increasingly able to flag poor algorithm choices in existing code, a step beyond their current role as coding assistants, according to Derek M. Jones, author of the Shape of Code blog. Jo…
Claude Code, Cursor, and Windsurf are competing AI coding tools with distinct strengths, according to a developer who tested them on a 15,000-line repository. Cursor Pro costs $20/month with fast local RAG indexing, Wind…
A developer explains how SNS filter policies can silently drop messages when publishers omit required message attributes, causing undetected failures in event-driven AWS architectures. The post describes how this issue e…
A developer spent three days using AI to build an authentication system, only to realize the user flow was fundamentally wrong because no one had defined the requirements beforehand. Industry observers warn that AI accel…
A developer used Claude to generate a landing page in a minute and deployed it without reading the code. Three weeks later, when asked to change a button color, they had no mental model of the code, realizing AI-generate…
Meituan released LongCat-2.5-Preview, a preview multimodal model with image understanding for cross-modal question answering, content summarization and complex visual reasoning, plus coding performance across code genera…
Developer Ziad Shalaby has released Code Merge, an open-source VS Code extension that merges selected files, folders, or code selections into a single Markdown preview with a project tree and per-file headers. The TypeSc…
A senior developer reported that using Claude Code reduced task completion time from a week to two days, but the AI-generated code caused production failures twice due to subtle defects that were hard to detect. The deve…
A Google Cloud employee ran a controlled experiment with 310 prompts to test whether shaping search results could steer AI coding agents toward Google Cloud products, after an Armature study of 16,893 coding-agent sessio…
A developer demonstrated how tool calling works in practice, showing that a foundation model does not execute code itself but instead emits a structured request that application code reads and runs. Using Amazon Bedrock'…
An engineer at Cyclopt argues that AI-generated code does not eliminate the need for system design, warning that no-code AI tools can produce prototypes that look finished but harbor architectural flaws. The post highlig…