It is very important when designing a game to figure out how you are going to attract new players. There are many games that do this remarkably badly and the standout example of this is bridge. I am by no means a competitive bridge player and yet if someone who had only twenty hours of bridge playing under their belt tried to play with me and others like me it would be bit of a trainwreck. When I talk about the game I generally suggest that it has a one hundred hour learning period (which won't make the player excellent, just able to play the game) which is probably an exaggeration... but not that much.
Compare chess and quoridor for example. I have watched kids playing chess and they absolutely can't do it. They can't figure out how to checkmate opponents, they constantly capture the king and call me over to inspect the board, and moves like en passant and castling are impenetrable to them. While they can use a chessboard they aren't playing chess. Quoridor on the other hand they understand. They play incredibly badly but the basic goal of getting your piece to the other side is easily grasped and they finish games without anyone breaking the rules. Both games have a very deep and interesting endgame for advanced players but only quoridor can introduce a new player and have them rightly feel like they are playing the game properly.
This problem becomes magnified with games that have a large investment at the outset like tabletop RPGs. To play DnD you want to have dice and several hardcover books and you will be spending many hours getting started in the game. To overcome this issue the new player experience needs to be really good and new players need to feel that there are things about the game they can master with greater experience but that they are playing the same game as the veterans. I tend to reduce this to a simple percentage that expresses how much better an optimized character is than one built strictly along roleplaying lines. That is, how much better is
"I will built a perfect melee killing machine"
"My character will be big and strong and use a sword and shield and hit stuff!"
"My character will be fast and smart and use a big spear and stab people!"
In DnD 3rd edition the percentage was easily 200%. In extreme cases at certain levels it could go as high as 500%. In DnD Next it looks like the difference is more like 50%. Heroes By Trade has a lower number yet, likely more in the 25% range. Note that these numbers only reflect the character sheet, not choices in combat.
I clearly don't want this number to be 0%. There is definitely a fun element to building a character that works well together and trying to make the best choices. However, when the number is too big new people feel overwhelmed and their characters feel hopeless or useless - not ideal for having a good first experience. I think Next is actually a totally reasonable game in this regard.
However, Next has forgotten that sometimes you have GMs that are novices too. Their current theory is to have multitudes of options for the GM to choose from. The game is essentially a set of dozens of optional rules modules. The difficulty here is that this presents a new GM with a huge problem - how do they decide which modules are in and which are out? This is compounded by the fact that in every edition there has been public play that will use a standard ruleset and also standard rules that everybody adopted and ones everybody ignored. (Okay, there was some group out there that used Weapon Type vs. Armour Type modifiers and ignored the Hovering on Death's Door rule... but they were nutters.)
The current system is fine for a veteran GM who wants to tailor a simple campaign for a group of new players. It is also fine for a veteran GM who wants to make a hideously complex campaign for the grognards. What concerns me is its functionality with a completely new group who pick up the rules. Mearls does talk about how a campaign could use the early levels to figure out which rules and modules they are going to use but I can't see how that works. Just randomly bolting on feats, backgrounds, or other such core elements at an arbitrary time seems questionable, and same goes for ripping them out.
What the game needs and what I am attempting to create with HBT is a game that has a great baseline experience. It won't be all things to all players but that is okay. If they want to houserule their game they can and they don't need the Wizard's or my permission to do that. To bring new players into the hobby what is needed is a complete game that is recognizable right from the outset to the end, from beginner to pro. A mashup of difficult and obscure choices is fine for the veterans but they will do that on their own anyway - what you need is a gold standard for play and a comfortable starting point for the new folks.