Articles from Dominik Fretz.
From spicy autocomplete to dark factories. The 5 levels of agentic software development explain why most teams are stuck — and what Level 4 and 5 actually look like in production.
Vibe coding ships demos. Agentic coding ships products. Here's why the distinction matters for your career as a software engineer, and how to be on the right side of it.
Agentic software engineering is the discipline of building production software with AI agents. Not vibe coding. Not autocomplete. Here's the methodology that actually works.
A step-by-step guide to building production-grade AI agents using Claude Code. Subagents, MCP servers, worktrees, and real-world deployment patterns from an Anthropic Claude Ambassador.
An honest comparison of Claude Code, Cursor, and OpenAI Codex from someone who uses all three daily. Which is best for what? A practitioner's take.
A practitioner's guide to OpenClaw (formerly ClawdBot/MoltBot) for Australian businesses and developers. Setup, security, use cases, and real-world deployment from an Anthropic Claude Ambassador.
OpenClaw has real security gaps. Here's what you need to know before deploying it in your business, from an AI agent expert and Anthropic Claude Ambassador.
Most AI coding fails because there's no spec. Here's the PRD-driven development framework that turns AI tools into engineering leverage instead of technical debt generators.
Your meetings aren't the problem. Your team size is. AI didn't change the brain's coordination limits — it changed the cost of getting the number wrong.
Stripe is merging 1,300 pull requests a week from autonomous agents. Here's what the architecture decisions behind Minions actually mean for how we build agentic systems.