My journey with vibe coding began because I had a challenge.

I needed a lead magnet for our website — something more than a simple download or form. I wanted it to have a little bit of “magic” for the user, providing immediate value for them while quietly delivering insights to me about their level of AI awareness and adoption.

In my head, I saw it already: a short, engaging survey with multiple-choice questions. Something lightweight and not intimidating. Friendly. The user would get a score from 1 to 10, along with words of encouragement and specific tips based on their answers. Sounds simple enough.

There was just one small catch…I’m not a developer. I don’t have — or pretend to have — those chops.

But I am vibe-code adjacent. A few friends have built useful tools this way and were pleased with the results. Not a candidate for the web coding hall of fame (is that a thing?) but good enough. And this felt like a good moment to learn by doing.

Vibe coding is an approach to building software using AI — where you describe what you want in plain language, and a large language model (LLM) tuned for code generation helps you build it. The term was coined by Andrej Karpathy, one of the original cofounders of OpenAI (the folks behind ChatGPT). It’s a blend of prompting and prototyping — equal parts imagination and iteration, kind of like an old-school hackathon for one.

And that turned out to be exactly the right mindset for what came next.

I didn’t need to get it perfect — I needed to get it started. I leaned on a team member’s earlier version of the form, which was built using a tool called Lovable using my original Google Form structure. (The code version of the March of Progress.) That gave me something to react to, edit, test, and eventually evolve. In other words, I gave myself permission to play and had a tool to start with.

Traditionally I would’ve outlined all of this for a developer — UI, logic flow, budget, etc. Then I’d need to find someone who had both an understanding of what I wanted and the bandwidth to be available. But this time, I became the one pecking at my keyboard, tweaking logic, testing the UX and learning applicable skills for future projects. And I didn’t need to align and manage anyone else. Instead, I could operate within my own curiosity without any information being lost in translation.

There’s that phrase about best laid plans… At one point, I got stuck trying to push form responses into a Google Sheet, just like I had with my original version. Lovable and Google weren’t playing nice. Bewilderment and frustration crept in. Head scratching. Chin rubbing.

But then I remembered the actual goal: getting the data, not building a spreadsheet. So I changed my approach and set the form up to allow the visitor to send themselves their results — and email me each response directly. That annoying friction led to a better outcome. Instead of hunting for data, it now came to me and provided benefits for both parties. That was unexpected — and even more magical.

So what did I learn from all this?

When you start with “what am I building and why?” you uncover the right paths and priorities. You stop obsessing over process and start focusing on outcomes and value. Enjoy the journey, sure, but remember the destination.

Knowing I didn’t need to wait on a developer* to validate my idea was freeing. Yes, I invested time. But the return was greater confidence, sharper instincts, and a clearer understanding of what’s possible next time.

And if something doesn’t work? That’s the universe providing feedback. Look for the workaround. Often, that detour is where the real insight lives and taking the scenic route to get there is fine.

This experience rewired how I think about prototyping and has empowered me to think bigger. I’m already looking for my next opportunity. What does yours look like?

*For the record, I still love and appreciate great developers. You are true sorcerers.