Skip to main content
camdenwebster

The Hidden Cost of “One More Project”

• 2 min • 353 words

Starting projects is easy. Not starting them in the first place is a little more difficult. How do you know when to start something new or just improve what you already have? When AI lets you do or create anything, how do we re-define “enough”? This is a question I ask myself almost daily in recent years.

Go deep

One perspective I try to keep top-of-mind is this: now that everyone has access to amazing AI tools, having a wide breadth of projects is more a liability than a differentiator.

If I can vibe-code a new productivity app and have it on the App Store in a few hours, anyone else can do the same. Does the world really need your take on a CRM? Instead, I try to go deep on what I’m already doing. Where do I already have an advantage that anyone with an AI tool can’t easily surpass? Where’s my domain knowledge strongest? For lack of a better term, where’s my moat?

I’m grateful that MovingBox has so much opportunity to go deep. Tweaking workflows to make advanced features intuitive, Apple integrations like Shortcuts and Widgets and a million other things currently sitting in my backlog. I don’t start many new projects. Choosing what to build next is the hard part, and as a solo developer, I don’t always have an outside perspective to help me zoom out.

Improve what you already have

Many online are saying that distribution is the new “moat”. I’d argue that customer retention, satisfaction, and making something that’s actually useful long-term is equally as important. Some basic analytics, addressing feature requests and actually talking to customers can go a long way. Where can you provide value to your customers, or to potential new customers?

Starting something new is a tempting distraction. But it’s usually just that—a distraction from what really needs to be done.

AI enables us to do anything, but not everything. Even if you’ve built a lights-out software factory to build all your projects for you, your attention is still a limited resource. How do you decide where to deploy that resource?