It's not a Go thing. I've inherited a number of large linear functions of the author's favored style in several C-syntax family languages, they all become increasingly incomprehensible as they grow longer and older. For any advocate of that style, the only way to maintain them (and retain their supposed clarity) is to extract functions and then re-inline those functions after a comprehensive refactoring. Otherwise, they accrue so much cruft over time that their legibility is completely lost.
Your only other option is to freeze them and never make changes, that doesn't happen much in real-world code (though it probably should).
Literally extracting the functions and then re-inlining them makes no sense. Having that as a sort of mental model while you're working on the code does make sense.
It’s to enable refactoring when it grows large. Most effective way I have found for 1k SLOC or larger functions. I usually don’t re-inline because the result after refactoring is almost always clearer.
Trying to in-place refactor those things is an exercise in frustration. That’s part of why they grow so large, from observing their proponents in action. They don’t actually know what the functions do, only where to add a new path and repeat themselves.
Your only other option is to freeze them and never make changes, that doesn't happen much in real-world code (though it probably should).