A similiar idea related to learning is how we build knowledge, how we actually learn and what are those steps. Consider learning as a procedure to build a tree

A similiar idea related to learning is how we build knowledge, how we actually learn and what are those steps. Consider learning as a procedure to build a tree of ideas and relationships between them that you want to use. We can take a bottom-up approach where we start by learning the thoery and then you apply it to build a product, to physics, to engineering... The top-down instead focuses on obtaining a result first, you learn tools and frameworks to solve a problem and you fill the holes you need as you encounter them. Permanent is theory, foundamentals. Expirating are tools, libraries and frameworks. Learning statistics + probability vs learning keras or tensorflow i.e.