The most-used apps change little from year to year. Meanwhile the interesting software is being built by very small teams, often for narrow audiences, and often with AI as a feature rather than a premise.
Three patterns worth watching
- Local-first: data lives on your device and syncs as an option, not a requirement.
- Single purpose: one task, done well, no dashboard.
- Model as plumbing: AI inside the workflow rather than the interface.
Why the edges move faster
A small team can ship a change in a week. A platform with a billion users cannot, because every change is a negotiation. That asymmetry means capability tends to appear at the edges first and arrive at the centre years later.
How to find them
Watch what professionals in a narrow field actually use, not what app-store charts show. The best tools are usually recommended by one person to another rather than advertised.
The risk
Small tools disappear. Prefer apps that export your data in a usable format, and treat any cloud-only small app as a temporary arrangement.
Bottom line: innovation is not on the home screen anymore. It is in the utility you install for one job and keep for years.
