While I agree that reinventing the wheel every time you start a new project is not an efficient method of development, I've grown to become weary of volatile APIs. A service which has only been operational for a short time and is under constant development means that my implementation may no longer be supported 6 months down the road if the service iterates significantly or worse yet, the company goes belly up. I understand this isn't going to happen to the likes of heroku overnight but unless there is an obvious, well supported api to use for a particular functionality, one should still take the time to consider the advantages and disadvantages of doing so.

Good point. I think it's wise to due your diligence on the APIs you're considering. Do they have a good track record of availability? Supporting backwards compatibility when they release features? Are they secure? Is it well documented?

Here at Stormpath, these items are a higher priority than any new feature.

I see someone has received millions in funding. Paying for APIs that nickel and dime you will add up eventually.

In this case it is a lot of services to nickel/dime you. And your user account list is often the most valuable part of you applications data. I don't want that in the hands of another company, or subject to DNS outages.