Most of the intuitions we build in 2D and 3D break in higher dimensions, a core problem for most machine learning problems. So where do they break?