why include physics parameters and classes for objects but not include physics formulas

Why do the prime model and the physics pixel include explicit physics parameters of objects and pixels and classes for objects but do not include explicit physics formulas?

Human gets the understanding of real world not only by seeing but also by feel (like pressure, temperature, pain of harm, etc), taste (yes, many objects are eatable), smell and sound, and also in history human accumulated the knowledge about this real world through countless observations and experiments, and for both aspects AI model cannot learn the physics nature of this real world like human does through only visual flow and text without proper labeling. The physics nature includes not only many physics parameters but also classification of objects which is truth of this world found by human, so they are included in the prime model and the physics pixel.

Physics formulas are basically in nature reprenting correlations in between different physics parameters, and approxiamting correlation especially complicated correlations of many different physics parameters in high dimensinal scene is exactly what neural network stack can do much more accurately and faster than formula based calculation especially in real time or complicated applications. So the prime model and the physics pixel make the ai model to learn the correlations in between different physics parameters of objects and physics pixels to approximate physics nature of real world without calcualtion of explicit physics formulas. Although the prime model and the physics pixel do not include explicit physics formulas originally, I guess there may be some methods to let them work with explicit physics formulas together in future.