Set Of Imaginary Problems – Could A model() learn relative motion? Then Could Light Be Smart As Water As It Hits Every Kind Of Surface?

Suggestion. Think of old thought experiments i terms of what the network could learn. By supervised learning for instance.