I think the real reason monads have a reputation for being difficult to grasp is that they are a big change for someone who is firmly entrenched in imperative /

I think the real reason monads have a reputation for being difficult to grasp is that they are a big change for someone who is firmly entrenched in imperative / object orientated programming. That doesn't mean that understanding a particular monad, like the IO monad, is harder than learning imperative / object orientated programming ab initio. Of course, being too different from the status quo can make a suboptimal solution stick around - the QWERTY vs Dvorak situation is a good example (although as someone who tried and failed to go from QWERTY to Dvorak, and also learned Haskell from a background of OO and imperative programming, I can say the former is much harder than the latter).