That's true, but that's exactly why this strategy can be so detrimental to the long-term health of a project. The core product has to be just barely good enough

That's true, but that's exactly why this strategy can be so detrimental to the long-term health of a project. The core product has to be just barely good enough to solve a problem for enough people; once it gets popular, it can coast on that momentum for a while, even as it ossifies. I can think of a number of projects in the "big data" space that seem to have fallen into this trap.