Blog
Blog
Practical notes on AI Spec-Driven Development, AI engineering, and how teams put AI to work.
Who Decides What, and How Does AI Execute? — Reading the Division of Labor from Anthropic's Claude Code Analysis
Anthropic's Claude Code usage analysis reveals a division of labor: humans set the goal, constraints, and definition of done; AI handles exploration, editing, and execution. From official data on ~400,000 sessions, here's what engineers should hold onto in the AI era.
- Claude Code
- Anthropic
- AI Agents
- AI Spec-Driven Development
Claude Code in Large Codebases: Best Practices and Where to Begin
Anthropic's guide to running Claude Code in large codebases is a practical starting point for organizations adopting AI coding at scale. We break down Agentic Search, the harness around the model (CLAUDE.md, Hooks, Skills, Plugins, MCP, LSP, Subagents), three design patterns, and where to begin.
- Claude Code
- Anthropic
- AI Agents
- AI Coding
Why AI Spec-Driven Development Was Born — Its Originator on the Origin and the Road to Proof
I used to think 'letting AI build software is impossible.' Feel Flow CTO Futoshi Okazaki looks back on why he changed his mind — the origin of AI Spec-Driven Development and the numbers that proved it out on our own SaaS.
- AI Spec-Driven Development
- AI Development
- AI Coding
- AI Agents
The Development Revolution of the AI Era: Why Documentation Is Better When There's Less of It
In 2025, software development has hit a major turning point. With AI tools generating most of the code, why does the old, human-centric documentation system quietly sabotage efficiency — and what does an AI-optimized structure look like?
- AI Development
- Documentation
- AI Spec-Driven Development
- Context Engineering
Follow blog articles
Subscribe to the English blog feed for practical engineering and AI articles from Feel Flow.