It probably should not be surprising that this new mechanism is as circuitous as a Rube Goldberg device. After all, the mechanism just happened to happen, and so is not exactly elegant. In fact, it consists of a rather unlikely series of steps, as follows:
1. Gene are sometimes duplicated, for reasons we don't fully understand (somehow evolution did that even though there was no reason, but it worked really well in the end).
2. Duplicate genes lead to excessive quantities of the protein. (bad)
3. Too many copies of the protein leads to dosage imbalance. (bad)
4. Small population size means inefficient selection. (usually bad, but good in this case)
5. Inefficient selection means the duplicate genes are not deleted quickly. (bad, but later good)
6. The duplicate genes become mutated. (good)
7. Some of these mutations affect expression levels via microRNA interactions, alleviating dosage imbalance. (good)
8. Some other mutations affect the protein structure, causing less compact, and less stable proteins. (bad, but later good)
9. These proteins are fortunately stabilized by binding with other proteins. (good)
10. These protein-protein interactions cause higher complexity. (good)
As you can see the process got off to a bad start. It did not look promising, but evolution has a way of finding a way to produce, one way or another. That's how evolution works--it creates complexity (see Step 10).
In this case, the complexity of the process almost matches the complexity of the designs it created. And interestingly, it dispenses with the outdated idea of natural selection driving the design. As one evolutionist explained:
the origins of some key aspects of the evolution of complexity may have their origins in completely nonadaptive processes.
Fortunately, evolutionists are rapidly determining how everything came about.