lazily learning
It depends. There's much out there---too much for any one to learn it all---so it's important to prioritize what to learn (whether to learn it), and to know whether, at a particular moment, it's better to learn, or to implement what you have learned.
At first, it's almost always to learn---unless your learning style is more ``hands-on'', wherefore it's better to (guided by a teacher, mentor, vel cetera, or merely a good course (be it static or dynamic, old-style media (text, video, audio,,) or hypermedia,,), or whatever; even yourself, if your intuition about it is sufficiently good) make (but for the sake of learning). (My favourite initial learning style is imagining, in full detail, making.) and:
What to learn? There are some things which are so useful, or so often used, so powerful or promising, or just so beautiful, that one tends to consider them necessary for any such specialist. Often, one hears ``a real'' or ``a true''---whatever professional designation---would/should/oughtta// know. Sometimes, this sentiment is misguided; too-generally applied. (Often, it would be true, were it about a subspecialization (of which the speaker, probably, is) of the specialization about which he is, apparently, speaking. The proper interpretation, in my experience, of such statements, tends to be, having attempted to understand to which specialization (of those that the speaker may be (reasonably considered to be) a part) that statement could truthfully apply, that that statement is about that subspecialization, which, probably, the speaker has confused with that surspecialization.) i.e. A basic ``common'' (among these specialists, that is) knowledge/skill set. But sometimes, there isn't, or it's not well known what would constitute, a basic ``common'' knowledge/skill set; wherefore one would merely start wherever, within such a sea, and progress through as might seem best. (Or, he might be wise to choose a different specialization.)
When it's time to make something, one must muster up all he has learned that may be relevant to the task at hand, to apply all applicable knowledge and skills, to make that which is to be made. Sometimes, his extant knowledge is not quite enough, or could be improved slightly; thus there is a complement to his knowledge, his skills,, which he would do well to learn prior to the actual making. Such complementary, for this learning, I suspect, is where such ``lazily learning''---learning them just before needing them---applies. (It's important to be able to know which things you'll need to learn, so that you'll be better able ``lazily learn'' that complement; this, naturally, would be part of that ``common'' knowledge/skills set.)