I strongly like the idea of pioneers as compared to "settlers" and "town planners" (see Simon Wardley's writing). But the idea that pioneers have and/or benefit from no process is BS.

The best pioneer is structured and has process. A pioneer's job is to efficiently conduct tests to derisk some space. This involves the creation of lots of artifacts which do not, usually and probably for the best, include running, trustworthy foundational code.

But that doesn't mean that work wasn't full of process. It's just often a little harder to scale and comes intuitively to some individuals. The process includes things like (a) rapidly reintegrating learning, (b) listing out assumptions and attacking them with an invalidation-minded perspective, (c) brainstorming novel ways to uncover new information cheaply, and (d) rapidly building new mental models and sharing them.

That's process, too, but not process that creates a functioning machine. Just one that proves such a machine could exist.

It's critical to not run the wrong process at the wrong time, so leaving when the pioneering is over can be a good move, but so might finding a new place for pioneering on the edges of the "settling" that's begun to move in.