There's a lot to by critical about Angular... least of which is that it's own dependency system doesn't play well with others. The reliance of jquery is pretty heavy (since most wind up including the full version, and the light version is loaded). There's a lot of extra boilerplate that the new design will tear down.
With node's tooling in current JS development, package management and modularization have become really easy to do. Testing has also gotten easier with this. Yes, browserify and traceur add some overhead to the mix, but it's still lighter than using the mega tooling stuff like jQuery. And don't get me wrong, I like jQuery a lot, and it's a really nice toolkit... but sometimes you only need a screwdriver, and an adjustable wrench for a job, not the whole huge craftsman toolbox.
Another issue is simply that the current angular would be unlikely to have a story around server-side rendering, and I think that's especially important for any next generation tools.
quoted just to remind people that Angular as we know it is older than most other single-page app frameworks - even if it only became popular a year or two ago.
I think it really comes down to more developers actually embracing JavaScript and the tools that the system offers. I think the JS renaissance is upon us, and that's what has people looking to Angular, Flux/React, Meteor and other new frameworks and tooling.
It's so incredibly designed, even without the ES6 optimizations of 0.12.0 (coming soon) you can code today in ES6 (I use 6to5 with JSX transformer harmony just in case).
With React somehow the whole data flow and how components work and even JSX make a lot more sense.
Facebook also has members on TC39 and is quite involved with evolving the language.
The good news is that if your code is clean, a migration shouldn't be too painful. The danger comes when you have unmodular code that sticks everything on $rootScope (for example).
We've been looking at React at work over the last week, and were considering moving some of our angular directives to react. Now there is a larger case for building a complete app in Flux.
So this is the first massive API change in probable well over five years, with lots of technology changes happening around (most importantly ES6).
The problem is of course that this will lead to a big divide between 1.x and 2.x that eventually will undermine the entire project like what happened in python and Volapück.
This makes me think that Angular didn't think through their API for the 1.0 release. I would have thought Angular could have been written to improve the performance where needed without such drastic changes to the framework.