If we realize that a neural network can learn to recognize objects in less than one hundred lines of codes and also that after training it is characterized by millions of weights that contain its knowledge, we can see that it is easier to understand the initial rules than the final result. This idea is useful in strategy. For example, in Go Game we use tactics like ladder, snapback and net which are easy to understand but at the same time when they are used in life and death problems like tsumego, their complexity can be higher. The same idea can be applied at the level of strategy to understand that some mental schemata are enough to capture the essence of a situation which is rather complex in geopolitics. This is clear in the case of an invasion. That’s why the propaganda is so poor when it tries to convince the others with a fake narrative. In this sense the neural network is a powerful tool in strategy, because it’s robust. It’s like thickness which is better used from a distance. Because it is able to support other actions but it is also important to use light groups which are hard to attack. We can see it in the Ukrainian forces.