> I think this is correct, if you consider college level linear algebra and an intuition for applying it to novel problems to be non-trivial knowledge
Yes, in the context of "a self-taught researcher", I think I intuitively meant anything that precisely requires a degree, typical academic knowledge. E.g. you can become a great business person who won't feel hindered by lack of academic knowledge, you definitely can't do that as a surgeon or lawyer.
I guess I was wondering where math fit in this picture for AI research. (which I should explicitely relate to "#2" in user ineedasername's post, i.e. "AI Research as examining the theoretical frameworks & approaches to ML/DL in a way that may itself lead to shifts in the understanding of ML/DL as a whole and/or develop fundamentally new tools for the purpose of #1 [AI Engineering Research]. What might be termed "basic" or "pure" research.")