If you know the market is dying for your product, take your time and build it right the first time. Maybe someone else will get to Techcrunch faster (boohoo), but when they're doing their 3rd rewrite, you'll be sailing towards millions in golden doubloons on your beautifully crafted ship, laughing maniacally all the way.
You may know that the market wants your product, but the exact feature set and usability requirements are probably still a mystery, and will almost certainly need to be refined anyway after launch no matter how long you preemptively perfect it.
In the rare instance you describe, I would still look to launch quickly, just maybe not recklessly quick.
http://gigaom.com/2012/12/03/marc-andreessen-not-every-start...