I would argue that you need to be able to "handle" both types of complexity, but that you should avoid "producing" incidental (I've always seen it called accidental) complexity. The reason I say this is you need to be able to evaluate documentation and determine what parts of the stuff they're telling you is needed and what's just going to add complexity to your code to no benefit (I have written loads about the amount of chaff recent versions of React bring to the table). And also if you're a fairly senior developer, you're likely going to be tasked with taking someone's overly-complex solution and understanding what it essentially does and refactoring/rewriting it to the simpler form or coaching a junior developer, who is going to create complexity just through lack of understanding of how to _make_ things simpler, to pare that down to something more elegant.