A firsthand record of configuring a site using Astro, Starlight, and Vercel alongside AI. Covers the role of each tool and a breakdown of which parts of the setup were delegated to AI versus confirmed personally.
The criteria I use to decide whether to work with Claude Code or Codex in the same project, with a specific example of handing off mid-task.
Three axes for deciding how much to delegate to AI — reversibility, scope of impact, and ease of verification — explained with concrete scenarios.
A step-by-step record of building a website from scratch in collaboration with AI. What I delegated to AI and what I decided myself, explained through the actual stages of the project.
A record of five specific patterns where AI-generated Mermaid diagrams failed to render correctly — including special characters, arrow direction, label length, Japanese text, and nested structures — along with how each was resolved.
A record of four consecutive Git conflicts on a feature branch, the attempt to resolve them with AI assistance, the decision to abandon the branch, and the lessons I took from it.
A record of setting up automatic deployment on Vercel with AI assistance, starting from manual publishing after every code change. Covers the configuration steps, errors encountered, and how they were resolved.
A record of the new problems that emerged once the Vibe Coding build phase ended and maintenance began, and the harness, tests, and validation I introduced to address them.
Is the time spent building CLAUDE.md and rule files really necessary? This article explains concretely what changed before and after setting up a harness, based on my own experience.
A record of three specific problems encountered when starting Vibe Coding — vague instructions, unintended design output, and lost context — with explanations of why each occurs and how to address it.
What should humans do when AI handles implementation? A concrete breakdown of tasks that can be delegated to AI versus tasks that humans must verify themselves.
AI gives different answers to the same question, and it does not remember previous sessions. This article explains Harness Engineering, a design approach developed to address these challenges when using AI in ongoing projects.
An explanation of the Vibe Coding concept. What does it mean to build sites and features by giving instructions to AI rather than writing code yourself, what can it do, and what are its limits?
A look back at three months of building and running AI Learning Playground with Vibe Coding, organized around insights from the technical, design, and operational phases.
Reviewing every line of AI-generated code is not practical at scale, but accepting it entirely without review is uncomfortable. This article presents a framework for deciding what to verify based on code type.
A step-by-step guide to setting up Google Analytics in a Vercel environment, covering measurement IDs, environment variables, and an Astro component.
Announcing the launch of AI Learning Playground — an engineering fundamentals site built for non-engineers: AI developers, PMs, marketers, and designers navigating the age of AI.