So we're bootstrapping, also in Boston, and also use GAE. We've been able to manage to stay afloat by keeping day jobs. One of us works full-time and the other

So we're bootstrapping, also in Boston, and also use GAE. We've been able to manage to stay afloat by keeping day jobs. One of us works full-time and the other part-time, while working on the startup all nights and all weekends. Obviously, it takes a toll on you, and isn't very sustainable in the long term. But we've also learned to delegate: we hired two excellent interns, and we pay them in munchies and free lunches. $30-$40 in CPU costs seems very high. I'm not sure if you have a technical background, but there might be room for optimization, especially in datastore entity structures and queries. Just the other day, we refactored the way we were storing a certain entity to be able to compute a list of key_names and then fetch them from memcache or the datastore in parallel rather than seriallly. That resulted in a 16x performance boost (measured by response time and CPU cycles). There are some other tricks, like "relational" index entities to speed up datastore lookups so that you're not doing too much processing in Python code. I would be happy to chat more about our experience with GAE. mahmoud [at] thesponty