If electrons are also networks then they to have network functionality. Then I wonder if there exist something like trained atomic objects like electrons or ions.
So the idea is to use electrons full efficiency capacity. Electrons subjected to obstacles should then have prior obstacle training. Electrons that need to ?cooperate should then have this kind of training.
So maybe the electric car could have groups of different electrons working in areas of their previous training. So you classify the path into groups. From inside the parts of battery to the parts of the motor. Where different trained electron abilities are needed.
Then I got some inspiration from Blender. An open source 3D rendering program. I speculated before that noise could be a factor in the performance of machine learning. So in blender I saw that they used noise reduction for every function. Here I wonder if the same could be done with the electric car battery system.
What I speculate is. If you noise reduce the output functions of a physical object like the ion or electron. This might show defined filtered solutions resembling differential equation solutions but perhaps more complex. Like biological evolution could have been some basic filter of random.