Releasing early means getting more exposure and possibly force you to realize your a writing something no-one earlier. What if though you are turning potential customer's away who are unimpressed with your first couple releases? Are those just considered write-offs and the cost of bootstrapping?
As a business person it's fiscal suicide to release a product for which there is no market. Most people see more benefit from a mediocre mechanic than a world-class buggy whip manufacture.
If the people who will see your product on the first iteration are the majority of your potential customers then chances are your business is a failure anyway.
How many of us think less of PG because HN uses tables?
"How many of us think less of PG because HN uses tables?"
I understand your point, but I think that is more of a implementation of features, rather than which features make it in for first release.
Its a valid point that if the first few users are the majority then you are in trouble. I would hope that the first few users/adopters however would become my biggest cheerleaders thus contributing to a snowball effect of new users.