[creer]

I'd add "learning deliberately" in the category of "necessary". It's not a matter of hacking, secrets or shortcuts but of ramping up faster than by merely passively doing. If a new situation or tool "shows up", think for yourself first, then actually look up a more solid answer. If you are going to use a tool more than 3 times, actually learn to use it. For everyone's favorite example, if you are thrown some Perl, learn enough perl that you won't be the person whining about line noise. Not "everything" about perl, just enough for the task at hand (then more for the next one). In all these cases, (1) it's really not much to learn each time, (2) it will speed you up for the rest of your career.

This does not mean "working on a book for 6 months, doing its projects". You don't need to put in that much time to progress. And when working on a book, skim. Don't spend 6 months on that book. Learning on the job, right in front of your project, can be plenty. If you want to do stuff "on idle time", doing exploratory search engine searches can be a pretty good use of time.

I'd also add mentors. It's work to cultivate a relationship with a mentor. But there is so much that they know and see, that you can't see yet. It would be a shame to ignore that option.