Well, the point is that the benefits don't outweigh the drawbacks because of the vendor lock-in. Whenever you hit a limitation there is NOTHING you can do. No duct tape! Also on long term the vendor roadmap may not align with your business(i.e. pricing, features etc). Do you have a migration roadmap? Of course you don't... you are already too busy dealing with the BaaS restrictions/APIs. Definitely BaaS have their place along with PaaS, DBaaS and IaaS but if you are locked-in to a specific vendor your options become quite limited very quickly. Maybe when something like Docker for BaaS appears I would reconsider this.
The BaaS is not really more than a framework with various constrains so there is nothing stopping you to implement your own BaaS on top of an existing PaaS(to avoid managing servers). I'm actually doing just this and it works well but I can see how Lambda or any other vendor offer would not work due the amount of customisation required.
I can't speak for the collective. As far as I'm concerned the current offers of BaaS are dead on arrival. Time will tell. I could see some value in the iPhone but I don't see much value in the current BaaS offers. Not enough to switch from the current alternatives. They are not transformative like the iPhone was. To complete your analogy we had smart phones before just like we have various BaaS services now.
You should understand a platforms general limitations before committing to it for any given piece of functionality. If you can't use it for something, don't. No one claims you can't mix and match. Use something like AWS Lambda only where it makes sense. Deploy your own hosts for more complex use cases.
