Text prompting
Text prompting is now a legitimate gateway to complex, functional software. The question is whether you understand enough to do something with what comes out.
My latest project started as a simple Phonebook App and evolved into a Personal Project Planner, with forms, modal popups, authentication, search, profile photo uploads, and a project stats dashboard.
I built the foundation the traditional way, incorporating authentication, search, and CRUD operations, and integrating TypeScript as I went. I wanted to understand what I was building, not just have it built. That took a few weeks.
Once the foundation was solid, I used AI-assisted prompting to add the rest. What would typically take days was done in 1 day through iterative prompting, with Tailwind CSS for frontend styling.
But here’s what that process actually looks like in practice:
My role shifted from writing code to orchestrating it, reviewing what’s generated, creating files, wiring components, and installing dependencies. I wasn’t writing every line, but I still needed to understand the recipe.
Could I have prompted the entire app from the start? Yes.
But building feature-by-feature first gave me a much deeper grasp of how to use TypeScript with React, not just what to build. That made the prompting sharper and the integration faster.
With this practical foundation, the next time I build a TypeScript/React App, I’ll likely start by building using AI-assisted prompting rather than the traditional approach.
🔗 Live app: https://fullstackopen-f0ek.onrender.com/
💻 GitHub: https://github.com/osalvatierra/fullstackopen/tree/main/personal-project-planner
Note
The app runs on a free Render server, so it takes a few seconds to spin up after you’ve registered for an account. Once it loads, try creating an account, adding contacts, projects, and a profile. Feedback is welcome.