How do you explain how it knows the output of the first Python expression in this article, if not by understanding it?

Let alone the fact that’s it’s running in a simulated virtual machine, and the other examples presented in this article, which add layers and layers of complexity on top of that.