Agreed, although the issue is that it's damn difficult to find something real that is at the intersection of: (1) pleasing users, (2) making them pay, (3) actually being useful, (4) being possible to get off the ground without a multi-year investment in time or money, and (5) remaining profitable or even revenue-generating for at least say three years before competition or evolution gets to it. It's a lot easier to hack something nice when you don't have to sell it.