5) You are asking a general question and get a "not possible, see paper X" from pretty much every expert you ask. You look up the paper or book and see that it'

5) You are asking a general question and get a "not possible, see paper X" from pretty much every expert you ask. You look up the paper or book and see that it's giving a specific result, NOT a general answer at all. For example, in the late 1970's I asked "what about neural nets for AI" and got the uniform answer "No possible, as Minsky and Papert have proven that mathematically. Forget about it." When I looked into that... the proof wasn't general at all, it only applied to one quite specific method of training neural nets. So I set about programming a function to generate perfect tic-tac-toe players that were neural nets, successfully, using another method (but couldn't get any traction even so, Minsky had made the topic far too toxic.)