Given how agentic programming platforms can now spin up surprisingly complex applications in minutes instead of months, it's tempting to say that the value of code is headed toward zero. But the situation is more complicated.
As we scale from small tools to sprawling systems, AI begins to struggle. To be fair, so do humans, but we still seem to have better heuristics for handling large-scale complexity and maintaining context. AI still needs guidance when navigating massive, tangled codebases. In fact, the need for human involvement increases as the codebase grows, unless the complexity is well-managed through architecture. So, in a messy 100k-line monolith with a dozen dependencies, a simple change might require ten times the human oversight compared to a 10k-line app.
This also means that small, clean applications are much easier to clone with AI. If a company’s core value lies in a basic app, that value might genuinely be eroding. Brand equity and first-mover advantage still matter, of course. But the software component of a simple mobile app is probably worth close to nothing these days.
What about larger, more established platforms? Are they safe? Not really. The only significant moat is the cost and time needed to replicate them. However, AI is rapidly shrinking that advantage. Competitors can now catch up with far less money and a much smaller team. In fact, starting from scratch might even be a benefit. New challengers can build with AI-native architecture and move more quickly once they reach parity.
So no, the value of code isn’t zero. At least not yet. But thanks to AI, it is clearly heading in that direction.
