The "lead or be led" trope is apt, and certainly became the paradigm of my doctoral years. I might, however, amend this slightly. There's another idea of going rogue too soon or too late (https://matt.might.net/articles/ways-to-fail-a-phd/).

Students, new employees, and other inexperienced folks need to be led initially, and then rapidly, transition into a self-directed paradigm. Success emerges if and only if the advisor and student recognize the need for this transition at a similar moment. The alternative is either the student who runs down rabbit holes repeatedly despite being guided elsewhere (those students tend to at least get SOMETHING done and while they take forever to graduate, do find some interesting results along the way) or the student who after a couple years is still just reading papers and waiting to be told what to do (these students often fail outright as advisors get fed up with the hand-holding).