One of the hardest things for a programmer to do is to NOT build a system. But sometimes (everytime?) the avoidance of a system (if possible) is the best option.
An example: Our company just implemented a huge ERP system from a well known vendor. As expected the problems we were trying to solve (resource management, etc) became secondary to keeping the system up and running. The system itself becomes the problem.
Which implicates this "law" is also relative, making a more useful description:
If you make something which seems complex to you, it will not work. Begin with something that is simple to you, and than evolve it in to something complex.
This is obviously closely related to Richard Gabriel's "Worse is Better", and the general philosophy of starting with a simple design, and iterating under the scrutiny of users.
The inverse proposition would be that if you take a simple system that works and evolve it into a complex system, the complex system will work.