Post

Over the Christmas holidays, I built three apps to replace my Google Sheets workflows: - 𝗙𝗶𝗻𝘁𝗿𝗮𝗰𝗸: A budget tracker that imports bank transactions and allows us to categorize them - 𝗣𝘂𝗺𝗽𝗲𝗹: A gym tracker - 𝗛𝗮𝗯𝗶𝘁𝘀: A daily habit tracker

All three started as spreadsheets I'd been using for months or years. Now they have proper UIs and additional features. Why the upgrade? AI made it possible 🥰 But here's what I learned:

I approached each project with different levels of "vibe coding." For Fintrack, I wrote most code myself or carefully reviewed every line. I wanted to understand exactly where data is stored and how it flows through the system (for obvious reasons 😅).

For Pumpel, I ran a pure conversation experiment: How far could I get without even looking at the code? And yes, it works. But I don't like it. Worse, adding new features through prompting gets progressively harder.

Components bloat, there's no clear structure for sharing logic or components, and the codebase becomes increasingly fragile.

Fintrack, on the other hand, is a joy to work with. It has a robust foundation and architecture that makes it easy for both me AND the AI agent to know where features belong and how to reuse existing logic.

The lesson: AI coding tools work best when you establish clear structure upfront. The time you invest in architecture pays dividends in maintainability and makes AI assistance more effective, not less.