For me I think it’s been mostly OTP/concurrency, but that’s because I’ve never truly grasped these concepts from a theoretical standpoint as well as in other languages. Although Elixir has been a pleasant surprise to learn about this subject.

My biggest problem is the OOP-Thinking in my head. I know that Elixir instead is processing a stream of data with transformation functions, instead of modeling and modifying objects.

Also i was in lack of patterns to use like i know from OOP. Even when these will not be the same, there must be a “common knowledge” useable for Elixir.

The most confusing thing for me is the syntax. I find it mostly natural until I want to do something in particular, then I am constantly stumbling about where to put parens or commas or colons, and the compiler error messages are too vague. I think I need to read that Metaprogramming in Elixir book to really know what’s going on, since most of the language constructs are really macros, e.g. “if”. Many of the peculiar usages are not covered in the “Getting Started” guide. For example, did you know you can |> pipe into a case construct, or pattern match its output to a var? This article on with was particularly helpful to show these syntactic peculiarities. Maybe I will get a better hang of it while going through Programming Elixir 1.2.