I agree with you in that there exist such a class of problems, but Occam's razor apply in far more circumstances than it does not. When confronted with novel information under pressure of quick decisions, you have to take the most likely correct path.
If the presenter can't divide a problem until it reaches a series of independently simple concepts, then there's usually something fishy going on.