I don't see how you can "know your market" without some sort of idea on who your user is, what their behaviors are, and have some sort of empathy with them.
Regarding pivoting and "throwing out" the work... each discipline of a team would do that. That's your UX designers job, just like a programmer would have to throw away code or a graphic artist would throw away design collateral for the new direction.
The "realize nobody cares" is exactly what good UX personas prevent. I guess if a startup doesn't know how to incorporate or conduct good UX you could run into those problems.
I guess I just don't agree with your premise. You're basically describing what happens when something goes wrong with your UX and personas - so of course your argument makes sense since you're highlighting a worst case scenario. But you're neglecting to acknowledge that every day, many startups effectively use and get tremendous value from personas and UX, and you can't do real UX without personas.
Sometimes, you are entering an established market in which the UX is the problem, and in that case, focusing on that right out of the gate makes sense.
Maybe we're talking about different types of personas. The question I'd ask from the very beginning is "who are we building this for and what problem is it going to solve?" And adding a little definition to the "who" and what problem they have is the first step in building anything, and making sure you don't waste time.
In the end, from day 1, from the first utterance of what your startup is about, you're building something for a human. Without empathy, you can't build the right product. And you can't empathize without personas. It's not even really a question of user experience, or user interface, it's just about nailing down the human problem.