[zozbot234]

I wouldn't say that Haskell is "trying very hard" to enable efficient compilation. It might be trying hard given what the existing language semantics looks like, but we still don't have e.g. an elegant model that can equally account for both strict and lazy code (even though there's quite a bit of research in this area, based on fairly natural considerations).

Reasoning about what code patterns actually require the use of general GC (because they involve creating references that might outlive their pre-defined scope, as with general closures) and what patterns might dispense with this is another source of potential efficiency improvements.