Look at some of the customer development stuff that's come up recently, though; I've seen companies with very very little actual code/product (in some cases, none!) iterate extremely quickly through wizard-of-oz smokescreens and the like.
I agree that fast execution and the ability to turn around and change what you've done (some hackers I know are very change-resistant, note) is important - that's why my last startup didn't do as well as it could. But a lot of that could have been saved with more customer-development methods as well as faster, better code.
My point is that you are right to say that building a business is the hardest part, but you are wrong to suggest that "building the business" is something separate from "building the product", or that success in the former does not crucially depend on the latter.