There's no weekend project that's going to be "fully scalable". Scaling isn't just grabbing a NoSQL backend and declaring victory. A lot of scaling is figuring out how you're going to use your data and creating data structures that allow for fast, efficient, and scalable access. Many NoSQL data stores force one to think about how you're going to access the data more than SQL databases since the NoSQL stores usually don't support things like joins. One still has to do a lot of work to design to implement it right.
Doesn't NoSQL require the same, only it shifts from being the job of a db engineer/sysadmin to that of the developers?
The problem is that the people who actually handle this correctly charge so much that I've never even seen their products in live use. Meanwhile free SQL databases are still really bad at it. But choosing an API that deliberately restricts you to only the worst possible query plans is not going to be part of the solution.
Until someone builds a cloud hosting platform with RDBMS that can do the necessary magic bits, that is.
[1] Yeah, yeah ... MySQL and their commercial licensing BS. It's been awhile but if memory serves and things haven't changed, you can't host a commercial web site backed my MySQL without A) releasing all your source or B) buying a license. IANAL.