I'll reserve final judgement until I see it. Let's face it: Mearls has a tendency to butcher these sorts of things. Because of that, I am willing to believe the reality is different enough from his description.
In a recent survey, I complained about feats, primarily along the lines of "There are a lot of useless, but character narrative defining feats, lot of melee-type feats, and a few crappy spellcaster feats." I went on to point out that spellcasters can usually pick and choose feats that they can become somewhat competent in melee but melee types can never use feats to become somewhat competent at spellcasting.
So, I think there might be some hidden nuggets in Mearls butchery:
Feats as "prestige class" means spellcaster dips into melee and melee dips into spellcaster can be done without multiclassing. That could be nice if done right (4E multiclass did this feat dip, but because powers were tied to ability scores so tightly, it often was a very poor dip.).
I think Fighter/rogues getting more feats, for example, might address the melee feat bloat. Many of those options are critical for certain build archetypes (2w fighter, grappler, etc). You *could* roll those into class builds (ie, make a grappler fighter build), but why couldn't a rogue, or a cleric have those melee options? Divorcing them from the class and leaving them in feats allows for more customization. But to spend a bunch of feats to achieve your "build" means you can't take all the power-boosting feats your vanilla fighter buddy did. So, I am guessing here, but I think there will be some classification of feats along the lines of "Build" chains, and "Prestige" chains and "Magic" chains. And fighter/rogue will get another chain rather than more unlimited...maybe. I hope.
