Chris Boesch covered an interesting startup idea at the python user group meeting tonight.
Basically, save yourself time and money by writing tests for the code you don't want to write by shipping a zipfile of the project off to an Amazon AWS instance that someone someplace writes to your test spec.
So, that's a misleading pitch because it ignores the real problem: you want to time-bound and price-bound the development. That relies on 1) you having a market-realistic price and schedule, and 2) the market having depth and breadth. So, it seems that only experienced developers could use this as it stands; guesstimating the time involved for the "average" programmer times the average rate on the site. Incenting quick delivery seems quite important as well.
In the worst case, this can help programmers understand their managers, and we get more tests written. ;)