Except the 4e DM is NOT supposed to be arbiter of PC powers. That was a class 1 design goal of 4e...
You mean its a system that rewards player ingenuity but doesn't hide its power behind rules bloat that only rules lawyers/min-maxers with all six other sourcebooks get to cherry pick from.
Sure, that's a way of looking at it, but IME there were still people who were a LOT better at doing it than others, even if it was a fairly simple system. Fleet Command was a few pages of rules, yet I had players in the campaign games who ran rampant by exploiting fairly subtle aspects of the rules. This is OK in a wargame, particularly one where the subject matter is essentially fantasy, but it isn't quite as good in an RPG.
I'd also dispute that 4e's power system is THAT exploitable. Now, maybe a 'point system' can be made to be no worse than that, but then how flexible is it?
There are easy ways around that spamming, the easiest one being monster variety. But also spell resistance, spell reflection, variable resistances (such as those of demons), immunities, spell mimicking, spell absorption, interupts, etc.
I think the objection there is that it runs the real danger of being 'puzzle monster of the day', and you have the question of how to make sure that most of the PCs can contribute meaningfully to most of the monsters every time. What we see is that simplicity on the power side starts to turn into complexity on the monster side, and on the encounter DESIGN side (which is where we don't want it).
But another question would be HOW would a given wizard automatically know whats best to use in every given situation and when they find whats best (against a certain monster) why penalize them from using it more than once?
Its OK to say "you're clever, here's your reward", but it can start to go down hill if the reward is too big. It just gets more complicated.
Which gives the DM a chance to both reward player specialization ("Bert's playing a Fire Wizard") but also occasionally throw the players a curveball (Fire Immune monsters vs. Bert).
I'm thinking more like 'area attacks plus enlarged areas plus metamagic to miss your ally' or something like that. I mean, this is all kind of speculation in the sense I can only think about what you can do in stock 4e with a different power system put in place. I think players will generally stick to what they know, and then its just not that big a deal to have a whole design space you don't really use.
The 4E power lists were unmanageable, unintuitive and did nothing for player ingenuity. Personally I just hate rules bloat and I think it really just benefits rules lawyers and gives the DM more headaches.
Meh, I don't know. I ran 4e for more than 10 years and I didn't really have a lot of 'headaches' with powers. I largely left it to the players to fool with those and let me know what they came up with when they unleashed one. I agree that the 4e power list grew organically, that it wasn't entirely consistent, that a lot of powers were kind of crap, or became useless over time, or were really only useful to very niche builds, etc.
Still, if you were to cut back each class to say 100 powers, you can build good lists.
Anyway, this starts to go beyond the realm of this thread, but when I designed my own '4e-like game' I fixed these issues quite simply.
1. The game has 20 levels, that means 10 less levels worth of powers to have to fill, meaning instantly 1/3 less powers!
2. Certain core powers are placed in power source lists. This is tricky, but it is possible to put a pretty good set of powers here. It removes a little diversity, but also makes it easier for players to handle since now their cleric and paladin can use the same basic healing power, for example.
3. Other powers are moved into smaller lists that are attached to 'boons', which in this case work like mini-themes almost. Again, commonality of powers.
4. Just plain getting rid of unneeded stuff and building lists over from scratch.
Honestly, it works pretty well. I think I'm up to about 12 classes now, and most of them actually have maybe 20 powers each. I think I can build a game with something like 600 powers and cover virtually the entire breadth of what 4e has in effect. Mainly, because I've made power sharing/reuse a possibility, now its easy to design for and aim at.
I mean, I agree with you that 4e is flawed in having non-shareable lists of powers attached only (essentially) to classes. I'm just not personally sure that you can go to the opposite extreme and still have something like the same game.
) doesn't have to take that lying down. Now it becomes a 'playing the GM' situation, and it just goes down hill from there really. 
