Just something we haven’t tackled properly yet! Some challenge/language combinations do have “editor’s choice” code examples along with screencasts but most don’t.
We’re experimenting with ways to solve this. With 2000+ stage-language combinations we can’t really rely on humans for this, has to be an automated ranking system. More details on the common complaints with code examples that we’re trying to address: Feedback on Code Examples?