Software is becoming a much weaker moat. For years, owning great software was one of the hardest advantages to copy. AI coding tools are changing that faster than most companies realise.
Build vs. Buy Used to Be an Easy Call
For a long time, the build-vs-buy decision leaned heavily towards buying. Building serious software meant large teams, specialised knowledge, long coordination cycles and a lot of management overhead. Even when you knew exactly what you wanted, getting it implemented could take months.
AI coding tools have changed that equation dramatically.
Near-Unlimited Building Capacity
Companies now have access to something close to unlimited software-building capacity and technical know-how. You can point Claude at a research paper, a new product or a tool you admire, ask it to extract the useful techniques, features and ideas, and work out how to bring them into your own products.
Work that used to take days of analysis, planning and coordination now happens in minutes.
Differentiation Gets Cheaper to Copy
This also changes what software differentiation is worth. When a competitor ships an interesting feature, understanding how it works and building something similar has become far easier. The barrier to reverse-engineering software ideas is collapsing.
So where does the moat move?
The Moat Moves to Operating Software
When the cost of writing code approaches zero, companies will produce much more of it: more services, more automations, more integrations, more internal tools, more dependencies and more paths through the system.
Someone still has to run all of that reliably. And this is not just an infrastructure problem. It takes:
- Engineering discipline and good processes
- Solid security practices
- Observability, so you know what is actually happening
- Careful change management
- Above all, a deep understanding of how systems fail
The expensive part of software is shifting from producing code to taking responsibility for what that code does in production.
You Cannot Prompt Your Way to Operational Discipline
That kind of discipline is hard to generate with a prompt. It comes from strong technical cultures that have spent years learning how to build, deploy, operate and recover real systems.
Which creates an interesting inversion: if your company was already good at building and operating custom software before AI coding tools arrived, you may now be in an unusually strong position.
What This Means at Wolkk.
This is exactly where we focus. Our engineers use AI to build faster, but our real value is in what comes after the first release: running, monitoring, securing and evolving software so it keeps working long after the demo.
The ability to build has become abundant. The ability to operate what you build has not.