I gave a fuller answer elsewhere in the thread (https://news.ycombinator.com/item?id=28669204), but it's highly reliant on the specifics of your idea/industry/bizmodel.
B2C and '10x improvement' style ideas tend to need to lead with a prototype (although you ideally first do enough discovery conversations to understand what they're already doing and why). The early product iteration is what Uber got right and Segway got wrong, and it's why the best videogame studios (like Blizzard 10+ years ago) always begin with prototypes of isolated core game mechanics, iterating those ruthlessly until they feel "crunchy" (the game industry term for immediately rewarding interactions).
Whereas ideas that are solving a well-defined, unsolved problem (which is way more common in niche B2B and the sorts of ideas that you'd bootstrap toward) can be pretty fully validated with ONLY conversation, which is obviously extremely quick and quite advantageous.
Apart from those two big categories, there are a million edge cases where the whole suite of approaches you mention can feel like a (somewhat situational) superpower.