I agree that I should be looking for people who produce quality work, but I also think there's some flexibility to the 'interchangeable pieces' part of the argument.

Here's the thing: Building anything ends up being a very iterative process. From my personal experience, by the time one moves from the prototype/beta phase upwards of 90% of the original code/design base is thrown away, redone, etc. Since what I'm really looking for here is the first-cut that allows for the validation of the idea and the on-boarding of the first few (hundred-)thousand users, I think there could be some advantages to a little shorter-term thinking; if-and-only-if that means I'm able to deliver that first cut to market sooner than if I were to focus on the stability of the team upfront.

Don't get me wrong, I'm still wavering on my opinion of this (hence this entire thread). But I am really enjoying / appreciating the various methodologies.