Building Apps with Lovable AI

Interesting new project tested - using Lovable AI to create applications.

Quick background - I'm not a coder. I have a basic understanding of how code works, structure and components, inputs and outputs, design and flow, but if you ask me to create a something, a simple html page is about as far as I can get. Maybe I can even host it, with some effort and trial & error.

So when I decided to test out Lovable, I was... fairly skeptical. I knew I could probably get a decent 1st page up with some structured content, but anything beyond that... I was expecting a great deal of broken code, misunderstood instructions, and the entire project careering off the rails as a few initial bugs trigger more breakdowns as I try to fix them, until the whole thing is an unrecoverable catastrophe.

I'd also had a not-optimal recent experience with Tasker - a few years ago I could configure it to run a few basic automations, trigger actions via NFC cards and such, my latest attempt to pull in website data using APIs, run content cleanup + calculations and render as a mobile widget ended in a mess, with me and multiple AI agents struggling to make sense of outdated source data, OS restrictions, app updates, and many more, finally ending up with a product that worked only in the 1st half of the day. (Don't ask!)

Anyway, back to Lovable. I casually threw out the idea to it, fully expecting a detailed, step-by-step struggle for each component, content piece, UI element and user action to be a bitter struggle over the next few day. It produced a completely working prototype in... under a minute.

I literally had a WTF moment.

So I threw out more complex asks. UI settings by feel. AI integrations. Database creation. Exports. Registration.

It. was. flawless.

The only place where it started getting a little wonky, was where I realized I had used similar page titles and element names (in the interest of continuity) and it was having difficulty in telling them apart - but that's more a me-problem, than an it-problem.

(Maybe soon I'll share the finished product here - still working on a few tweaks, and I want to check the AI credits usage implications before I expose it to all.)

In fact, the only thing holding me back, was the 5 credits a day limit in the free tier. I'd whip out 5-6 instructions each evening to refine the product, and within 15 minutes (mostly in testing) it would be ready. I had a fully functional working app in under a few hours total.

And I've worked with web dev agencies. I know that what I asked for, if I were to break it down into a traditional Business Requirement Document, RFP, design+dev man-hours approach, with a system architect, a UI + design guy, front-end and back-end devs, and an infra support person, would easily be anywhere in the range of a few weeks.

I got it in hours. And I'm not even a coder.

It was... thrilling, a wow moment, profoundly empowering, but also a disquieting one. I was seeing years of study and experience requirements, evaporate.

Tomorrow, everyone's a developer... yet nobody is.