I really hope Evernote's take-away from this is that they need to scale back development on all their auxiliary stuff - hello, food, whatever, as well as all but the most critical feature requests, and focus as much as possible on making the core experience bulletproof. I would _hate_ to have to give up Evernote, but like others here, am extremely apprehensive about the possibility of losing data.

One stop-gap they might be able to implement quickly would be a scale-up of their version control. They could throw money (storage space and bandwidth) at the problem, increasing the number and frequency of revisions stored. Certainly not as good as preventing loss in the first place, but reliable versioning would help minimize catastrophic loss in the meantime, and would still continue to be valuable once things are more stable.