No product is truly tested until it faces the marketplace. Until then, even the best ideas and plans may be wrong. So why invest deeply in things that might not work? It'd be better to just know.
It bothers me seeing people giving up too easily. Pisses me off, even. That's what struck a chord for me in Mark's post. It doesn't compute for me when I see people get really excited about an idea, have the first iteration not stick and then just stop. And that seems to be really common. I think that's the rub with Mark too.
No idea where the notion "fail fast" = easily give up comes from.