OpenStreetMap is great, but it's a bit like Linux among OSes - one source of data, but many implementations and not a single one to be dominant enough to be a real threat to google, hence no street view and the search is lousy just to name a few significant drawbacks.
That's because OpenStreetMap is just that -- a source of data. It doesn't aim to offer the kind of utility you expect. Rather, it can serve as a data source for developers that might want to take on that task. The ODBL license they use has some unfortunate terms and the committee seem more interested in license-wanking than progress which is why I think there haven't been a lot of major players creating rich and powerful applications and services with it. For instance, mixing your own proprietary data with OpenStreetMap data isn't allowed. This hasn't been a complete show stopper as some great companies like MapBox have popped up, but I do think its had a chilling effect in general.
Another analogy might be ARM vs. Intel? Lots of small players don't have to go toe-to-toe with a giant in order to be effective competition. I agree though that it needs to keep up momentum to be competitive, a lot is still missing.