When I returned to my computer five minutes after giving Gemini a lengthy prompt, I had two things: a functional app in a preview window, and a message about a bug. My yard was dying, so I made an app for that—but the real challenge lay ahead.

The project began humbly enough as a yard work to-do list. But Gemini's first attempt at vibe-coding produced a dark mode nightmare with purple and brick red. After a few iterations, the app gained a more appealing interface and a cheeky greeting: “Welcome Back, Gardeneer!”

The basic structure stuck, but I had to refine it myself. Integration issues and missing features added a new layer of complexity, turning my unruly yard into an unruly app. Yet, the AI plant doctor was a standout feature, quickly diagnosing plant problems with ease.

Years ago, our landscaper’s solution for weed control was simple: cover everything in fabric and rocks. Now, after years of neglect, something was clearly off. Enter Gemini—the digital helper that turns gardening dilemmas into tech challenges.

The lesson? Sometimes the biggest work is not with plants but with code—and the physical world can sometimes be a bit more complicated than theoretical algorithms suggest.