Ideas
Sustainable competitive advantage, when AI writes the code
As AI writes more and more of the code, companies need to rethink where their sustainable competitive advantage will come from
AI writes a growing share of the code, and that share will keep growing. For many companies this is good news in the short term: features arrive faster and cost less. It is also a medium-term question they rarely ask: if anyone can produce software at this pace, what still constitutes a sustainable competitive advantage?
For twenty years, the ability to build software was an advantage in itself. A good team, a good stack, a good delivery rhythm set a company apart from its competitors. That advantage is eroding, because it is becoming available to everyone, at the same price.
What was scarce is turning into a commodity, and an advantage built on a commodity does not last
What cannot be copied in a prompt, on the other hand, is gaining value. Domain knowledge: what the business knows about its customers, its constraints and its edge cases, and that nobody has written down anywhere. Data, when it is clean, connected and understood. Judgement: knowing which problem is worth solving, and recognising when a plausible solution is wrong. The ability to verify what is produced, because producing no longer costs anything but trust still does. And the speed at which the organisation learns from what it ships.
None of this is new. What changes is that until now these were the invisible part of the advantage: they expressed themselves through the code.
Now that code can be generated, they are the advantage itself
So the question is not “how do we produce more code with AI?”, since everyone will manage that. It is “what do we know, decide or verify that our competitors cannot generate?”. That is where to invest, and it is a subject I intend to keep digging into here.