Articles for people who
ship the stack
Rewrites on React, Node.js, TypeScript and AI. A translated body is shown when it exists. If it does not, you see the English text and a note that says so.
Tagged: software-engineering
Why Most Developers Still Get Little From Coding Agents
What JetBrains survey data and 400,000 Claude Code sessions say about delegation, expert versus novice outcomes, and a seven-stage loop for agent work.
8 min readAgentic Engineering in Five Layers: From Specs and Parallel Agents to Rules and Oversight
Karpathy named the shift past vibe coding. See the five layers of agentic engineering: spec, orchestration, codified rules, human review and observability, plus a week plan.
19 min readTwelve Patterns for Writing Idiomatic Go With Claude Code
Twelve Claude Code patterns for idiomatic Go: interface-first prompts, test-driven loops, CLAUDE.md rules, error wrapping, concurrency and Plan Mode limits.
7 min readWhen AI Writes the Code: Engineering Skills That Gain Value
Coding agents now draft features end to end. What to study, how to review generated code, and why judgment, systems knowledge and verification matter more than typing.
5 min readHow AI Coding Agents Are Reshaping the Developer's Job in 2026
AI coding agents moved from autocomplete to whole tasks. Adoption data, context engineering, typed code, verification and MCP, and what changes for junior and senior devs.
13 min readA RAG Project Is Not One Giant Python File: A Folder Layout That Scales
A RAG project shouldn't be one giant Python file. A folder layout for ingestion, chunking, embeddings, vector store, retrieval, prompts, API and tests.
9 min readFour AI Agent Roles for Engineers: Coordination, Creativity, Clarity, Coaching
Four AI agent roles engineers can hand real work to: coordination, creativity, clarity and coaching, from a morning briefing to mock interview practice.
6 min readHarness, Loop and Graph: Three Nested Layers of a Coding-Agent System
A failure symptom points to the layer to fix. How harness, loop and graph nest, where multi-agent helps, what adversarial checks can't promise, and a checklist.
15 min readWhy Coding Agents Still Struggle With Private Company Code
Real-SWE tests eight coding agents on licensed private codebases. The best resolves 38.8 percent, and the five failure classes show where agents go wrong.
9 min read
About these articles
Request a 24h estimate
Need the same stack in a production operator layer? Send the brief — estimate within 24 hours.