It will be interesting to see, Apple has done this many times before and often turned out right, changing the game despite the howling. I remember the howling when OSX came out and replaced 9, and Apple moved to Intel. Hell I remember when someone told me Apples Macintosh GUI would never replace the Windows command line structure. The iPad was going to be an underpowered failure said many. Everyone howled when Jobs refused to support Flash - well it looks like Adobe have thrown in the towel there as well. There have been very few Apple products in recent years that within an iteration or two, were not hugely successful. But admittedly, they have also failed.Having read Jobs' biography one can really understand the mindset; he did not believe in market research, rather anticipating what the public did not know they wanted yet. So time will tell if Apple knew what the pros wanted before they did.
Flash failed because it's processor intensive and requires dual processors to run "properly" AND the iApple has killed on the apps. Playbook tanked and Android wasn't enough to support the money, especially since Apple refused to allow anything Flash on their mobile products. So yes, Apple choice to dump flash was coincidentally correct. The market and Apple playing hardball killed Flash. And if HTML5 becomes a security mess, we can all blame Apple! 
Now with NLEs, the story may be different. Apple jumped into the pro NLE market because Adobe wouldn't make a Premier Pro back in the 90s. They sold a great package, along with getting the right professionals people to support it. Which was great because they threw in some great packages from Shake to Color. But the pro market doesn't make much money compared to the bigger prosumer market, thus the swing towards fewer key features and more a different approach is a good market switch, just not one the professionals who put money into their businesses with Apple products will like, albeit at a discount 
Cinema Tools, Color and Shake are still left in 32bit UB, while Resolve, AEFX and PP are running away.
Now FCX 10.0.1 update has some great features that will rewrite the way some things work. Roles in place of tracks is very interesting. But that's the point. They released a half ready program. What did they think was going to happen? Round tripping from Motion to FCX then to Compressor still isn't smooth. And that magnetic timeline clip lining up is still atrociously imprecise.
I'm pretty sure that the FCX metadata approach will trickle down to the other NLEs, because Roles is a nice way to work. But one major feature in a sea of mess that is FCX won't impress too many people. And they could introduce things slowly instead of pulling the rug from under the customer. XSAN projects can still only be used by one user at a time. You can imagine how unhappy clients will be at things like that.
