A problem at work turned into my first product in 7 days

I’m not a developer. I shipped my first product in a week.

It started with a problem at work.

I needed a simple way to create polished screenshots and device mockups. I went looking for something that did exactly what I wanted, but everything I found was either too complicated, too limited, or just didn’t fit the workflow I had in mind.

So I decided to see if I could build it myself.

There was one obvious problem: I don’t really know how to code.

What I do know is products. I use way too many productivity and design tools, and I spend a lot of time thinking about why some software feels great to use and some doesn’t.

A few years ago, that probably wouldn’t have been enough to ship software.

Now it is.

I basically token-maxed my way through building it.

I used Claude until I hit limits. Then ChatGPT. Then Antigravity. For smaller tasks that didn’t need an expensive model or much context, I used free models through OpenRouter.

When one tool became the bottleneck, I moved to another.

It was less like traditional programming and more like managing a tiny, slightly chaotic engineering team where every engineer happened to be an LLM.

The interesting part was that I could stay focused on the thing I actually cared about: the product.

What should this button do? Does this interaction make sense? Is this feature necessary? Why does this screen feel wrong? What would make this faster?

I started building it at work to solve my own problem, then kept going after work and during some time off.

About a week later, I had something I could actually ship.

That became Product Shot.

https://getproductshot.com

It turns screenshots into polished device mockups and product visuals directly in the browser.

It’s now live and this is the first product I’ve shipped.

The biggest thing I learned from the process is that the definition of who gets to build software is changing very quickly.

You can be bad at writing code and still be good at understanding a problem, recognizing what makes a useful product, and knowing what you want to build.

The implementation used to be the wall between those things and actually shipping something.

That wall is getting much smaller.

I’m curious how many other people here are building this way: not really learning to code first, but learning to build by shipping alongside a rotating stack of LLMs.