I think Cloud Code is a good example of why this is not the case. The more flexibility and customization they allow in the platform, the fewer long tail features they need to implement because people will be able to do it themselves.
I imagine the argument is "once you add Cloud Code and people start using it to solve problems, you have pretty much re-built Google App Engine"; it means the only thing that separates Parse/StackMob/CloudMine (which now all have this feature) is that they come with a simple default app which provides an HTTP API to the data store that is used by an even simpler client library.
Right, because they provide the two things I stated: a default app on the other side that provides fairly liberal access to your schema, and a library that can be used on the client that provides a fairly reasonable API (maybe even CoreData integration; I know some of their competitors do this) to access that API.
Neither of these, however, are that complex to do, and the value of both of these drops dramatically once you start writing server-side logic in Java.
