Pretty simple. This outgrowth relies on simplification, fragmentation, and a psychological trap.

To work with the real world, we use forgiveness and analogies that give a partial or slightly distorted understanding of things.

When we look at something, we look at the general model, and when we look in more detail, it turns out that there are other mechanisms behind the result, the results of which you predicted empirically, and partially incorrectly.

Trying to explain something to ourselves, we actually strive for a certain result, the key to which is the analysis of something. Often, a simplified or incorrect model is enough to achieve your goal, and it gives results that seem to work. And you are satisfied.