Building Your Own Toolbox

Software is built to scale. That's how it works. A tool designed for a wide audience has to cover a wide range of needs, which means it carries weight: features you'll never open, menus you'll never click, workflows shaped around assumptions that don't quite match yours. None of that is a flaw. It's just the natural result of building something that needs to work for many different kinds of use.

AI coding tools have changed what's feasible to build for an audience of one. Not to zero effort, but enough that if you understand your own workflow and can describe what you need, building it yourself becomes a reasonable option. Before this, even developers would rarely do it. The time it took to build, maintain, and keep a personal tool updated was hard to justify when you were the only one using it. That's changed.

I've been building my own tools for a while. A finance tracker that works exactly how I think about my money. A personal hub that ties different parts of my life into one place. A spam filter that started as something I needed for my own sites before it turned into a product. None of these existed in the shape I wanted. Every app I tried either did too much or not quite enough, and I was always adjusting how I worked to fit the tool instead of the other way around. So I started building them myself, one at a time, each one shaped around how I actually work.

That's the part that changes when you build your own tools. They fit. Not because they're more sophisticated, but because you defined what they needed to do. You're not paying for features you'll never touch. You're not adapting to someone else's idea of how things should be organized.

There's a quieter benefit too. When you own the tool, you own the data. You feel more comfortable putting more into it, trusting it with the details that matter. And when your tools are yours, they can talk to each other. They become a small ecosystem instead of scattered logins across services that don't know about each other.

This isn't about subscriptions being bad. Software that solves problems creatively, earns trust, and prices fairly isn't going anywhere. The bar is just higher now. Generic tools that charge monthly for a long list of features, where you only use a fraction of them, are harder to justify when building the specific thing you need becomes an option.

It's not just about what you build, either. It's also about how you use it. When AI is part of the tool itself, you spend less time filling in fields and more time reviewing what the system suggests. You describe what you want, check the output, adjust. I explore this more in AI as a Thinking Layer on the RARO blog, but the idea is simple. Software that can reason about your input before storing it works differently than software that just records what you type.

The relationship between people and their software is quietly changing. The tools you use and the tools you build are starting to overlap. And the ones you build tend to fit better.