My Exact AI Coding Workflow, Step by StepThe practical version of the system: how I go from a rough feature idea to reviewed, tested, production-ready code with AI agents in the loop.Apr 10, 2026 · 10 min read
7 Principles for Getting AI to Write Production CodeThe model matters less than people think. These seven workflow principles make AI output precise, testable, and safe to ship.Apr 7, 2026 · 5 min read
Power User Mode: Running 3 Agents in ParallelThe real power move is not one brilliant prompt. It is running multiple AI agents in parallel without losing review quality or architectural control.Apr 3, 2026 · 8 min read
What Is Pretext and Why It MattersPretext turns text layout into something application code can know about ahead of DOM measurement, which makes text-heavy interfaces faster to build and much easier to control.Mar 31, 2026 · 9 min read
The Intent Layer: How Prompts, Specs, and Validation Make Engineering DeclarativeReact moved frontend work above manual DOM mutation. Prompts, specs, and validation loops now do the same for AI-assisted engineering.Mar 30, 2026 · 7 min read
Complete the Loop: Why Your AI Writes Bad Code (and How to Fix It)Most bad AI code is not a model problem. It is a feedback-loop problem: if an agent cannot see, test, and correct its work, it codes blind.Mar 27, 2026 · 9 min read
Stop Typing, Start Talking: Why Voice-First Changed EverythingVoice-first prompting is not about being able to type faster. It is about making rich context cheap enough to provide, which improves AI coding output.Mar 22, 2026 · 11 min read
Code Is No Longer Written for HumansMaintainability in code still matters, but AI shifts the real source of truth upward: from hand-written code to specs, scenarios, docs, and tests.Mar 18, 2026 · 11 min read