When building is quicker than searching

Something slightly odd has started happening when I need a piece of software at home.

My first thought used to be to search for something that already existed. I’d find a few options, work out which one did most of what I wanted, sign up and then generally adapt what I was trying to do around the limitations of the product.

Increasingly my first thought is: I could probably just build this.

Obviously I’m talking about the small applications I build as a hobby, not enterprise software. But even so, that feels like quite a significant change.

There must be a point where searching for a product, evaluating it, configuring it and learning how it works takes longer than creating something specifically for the problem you have.

AI-assisted development seems to be moving that point surprisingly quickly.

The enterprise version is much more complicated. Building the code is only part of building a service; security, accessibility, integration, support, maintenance and ownership don’t disappear just because producing software gets easier.

But it does make me wonder what happens to the traditional build-versus-buy decision if the economics of one side of it change significantly.

Perhaps software becoming easier to build doesn’t just change how we develop software. It changes what we decide is worth building in the first place.

Something odd has changed in how I use software. If I need a small tool, increasingly I don’t search for one. I think: I could probably just build this. That’s hobby coding, not enterprise IT, but what happens to build vs buy if AI keeps making the “build” part easier?

POST, but as a short discussion post rather than an article. Keep the personal hobby-coding observation and the enterprise qualification.

This entry remains a draft. Nothing here claims publication.