The point of the article is that there is a 4th category that is even worse, things that you know, but are wrong or have been made irrelevant or obsolete.
Think of the experiment with the 5 monkeys, the problem is we as experts in a field are afflicted with scores of preconditionings.
I thought it was more about the things you know and focus on, which are right in their narrow domain, but which prevent you from looking at the wider problem and therefore prevent you from perceiving, possibly better, alternatives.
I know I've spent days perfecting a certain programming paradigm, be it function composition or object orientated design, only to realise that I've lost sight of the overall problem domain and therefore how best to tackle it.
