Software can be finished. Not abandoned, not dead, just done. Strangely that’s seen as controversial today. Software – a digital expression of an idea, a tool has no physical bounds or qualities. It can, and often does expand exponentially. There is always another version. Another feature. Another market. Another reason for the product to become slightly bigger than it was yesterday.
Operating systems change. Bugs appear. Security issues need fixing. People's expectations grow.
For most of its existence, software was expensive to make. Teams were expensive and time consuming to assemble. Distribution was difficult. Building something narrow, with focus often meant the economics didn’t work.
If you had a successful piece of software there was a pretty strong incentive to keep it going. More customers, more features, more use cases, more money.
Turn a tool into a product.
Turn a product into a suite.
Turn a suite into a platform.
But something has changed. It’s getting much cheaper to build software. Does this mean we can build more software? Without doubt – Github is on pace for a 14x increase in commits this year compared to 2025.
Maybe it also means we can afford to build less.
A tool might solve a small problem for a small number of people. And that might be enough.
It won’t need to find adjacent markets, enterprise tiers, collaborative features, or AI assistants. It just needs to be damn good at the thing it does.
There’s a confidence required in deciding something is done. Adding features are easy to justify. Stopping requires you to have an opinion on what the thing actually is.
Knowing what not to add is as important as knowing what to build. Software can still be cared for, can still improve. But, it doesn’t need to continually expand its reason for existing.
It can be finished.
Not abandoned.
Not dead.
Just done.