Vibe-coding lulled me into a trap.
Starting new projects felt effortless. Just prompt, generate, ship. But each project ended up with a different architecture, different conventions, different patterns — because I was letting the model make those decisions fresh every time. The maintenance debt stacked up fast, and the inconsistency across projects was making my life harder, not easier.
So I took a weekend and built my own boilerplate.
The Problem
When you start every project from scratch with AI, you get whatever patterns the model feels like generating that day. One project uses repository patterns, another uses service classes, a third has everything in controllers. They all work in isolation. None of them work the same way, which means context-switching between client projects costs more cognitive overhead than it should.
I lost myself in different paradigms, inconsistently applied. The robot was happy to build whatever I asked for, but it couldn't maintain a consistent vision across projects because that's not its job. That's mine.
The Solution
Over a weekend, I built a Laravel and Vue boilerplate that codifies my architectural preferences. Not a framework — a starting point. The decisions I make on every project, pre-made so I don't have to remake them.
It's not a new concept. Developers have been building starter kits forever. But the scope of what you can build in a weekend with AI is genuinely mind-boggling. The boilerplate includes a full CMS, a cinematic home page template, social proof components, queued jobs for async processing, Redis cache configuration, SEO metadata structure, and more.
All of that in a weekend. A year ago, that's a month of work.
What It Changed
Now when I spin up a client project, I'm not deciding on architecture — I'm executing on architecture I've already thought through. The AI helps me build faster within a consistent structure instead of generating a new structure every time.
The irony isn't lost on me: I used AI to build the thing that constrains how I use AI. But that's the engineering instinct — take the thing that's causing entropy and systematize it.