Li Shenron
Legend
An improved D&D what though. We have 5-someodd editions of D&D and a dozen variants. Which D&D are we trying to improve?
It's not at all simple to answer

If I look back at the editions I know, there is always a reason that gives me problems:
- I like the old-school style of BECMI, but I hate the character rigidity and the combat mechanics (also for AD&D)
- I like the flexibility of 3ed, but I hate how heavy it is for the DM
- I like how easy 4ed is for the DM, but I hate everything else of it
As you can see, I'm not really concerned about small things (which are always the easiest to HR anyway) but about general and practical matters that can't be solved with a couple of HR but at least require some lengthy work.
In theory, 5e is supposed to work a bit more like a toolbox which makes you feel much safer when you take a module in or out of the game, compared to HRuling large areas of a previous editions.
So if 5e can let me have the style of BECMI + the character flexibility of 3ed + the ease of DM preparation of 4ed, then I can have just the edition I want right now, but the point is that if it can let me also have other combinations of stuff, then I'll also have more solutions for the future, since I am not always wanting to play the same thing!
I am not expecting perfect customizability, and anyway once you start customizing the details, that is where actually house rules become easier. I don't need 5e to tell me that I can make some changes to a class spells/maneuvers list for instance, or that I can change individual spells/feats/etc if I don't like them, because doing so is almost always very easy. But changing something major like spellcasting mechanics, multiclassing or skill mechanics, has larger consequences on the game, and thus would really benefit everyone if 5e supports changing those things, if not with modules then at least through good guidelines.