Satisfying Occam's Razor is not a necessity, it's more a rule of thumb. Based on Occam's Razor alone, modern DNN based models are complete garbage because of t

Satisfying Occam's Razor is not a necessity, it's more a rule of thumb. Based on Occam's Razor alone, modern DNN based models are complete garbage because of the amount of complexity involved in their models, yet they still hold useful predictive power. Sometimes we have to throw out these general rules of thumbs and test the waters outside when we hit walls we can't seem to get around. Interest and pursuit in the philosophy of science is something too largely brushed aside that, IMHO, needs an adrenaline shot these days. Unfortunately, few want to pursue that work because it doesn't pay the bills.