Somewhere along the way, a large part of the software industry started building mainly for itself. Tools for developers, platforms for platform teams, dashboards to monitor the dashboards. It is a comfortable loop, and it is not where most of the real problems are.
The self-referential trap
It is easy to build software for software people. They are reachable, they are online, they buy things with a company card, and they will tell you exactly what they want. The trouble is that this audience is a rounding error next to the rest of the economy.
The startups we find most interesting have escaped that loop. They are pointed at logistics, healthcare, construction, agriculture, local government — the parts of the world that run on clipboards and phone calls and institutional memory.
Harder problems, better moats
Real-world problems are harder, and that is the point. You cannot solve them with a clever front end. You have to understand a domain, earn trust, and survive contact with regulation, unions, weather, and people who have done the job for thirty years.
That difficulty is also the moat. A polished SaaS clone can appear in a weekend. A relationship with a hospital system cannot.
Signs a startup is actually solving something
We look for a few tells that a company is solving a real problem rather than manufacturing one:
- The customer is not a software company. The value has to travel outside the industry.
- The founders can describe the old way in painful detail. They know the clipboard because they held it.
- Success looks physical. Something ships sooner, someone waits less, a cost that was real goes down.
Why it matters
None of this is a moral argument. It is a coverage argument. The story of software eating the world is only interesting where software meets the world — and that is almost never in another developer tool. It is in the parts of daily work that rarely make the tech press at all. That is the beat we care about.