The problem is you're going to just shift that to other choices. Even if you remove 'feat tax' feats... the min maxer is still going to min-max regardless of the feat choices before him.
They ALWAYS do. So let's say you take the feat choices a minmaxer always takes, and make that your criterion for the 'performance feats' silo. This is going to include 'interesting feats' as well by their nature.
Now you have the exact same problem you had before, just it exists in two silos instead of one.
Yes and no. Divvying it up into two resources will make the discrepancies less severe. Right now, we've got Utility Powers that help in combat and ones with more skill-based effects. I feel like if I went with more flavorful Utility powers, I might be less effective than an optimized character in combat, but not by an unreasonable amount.
But if you asked me to give up, say, Encounter Attack powers in order to take my Utility powers... that would be a much harsher choice. And I feel that is what we have right now with feats.
We've got these really potent combat-centric options, alongside the purely flavor options. And the difference between the two is pretty sizable.
If we put them in seperate categories, I'm sure we'd still end up with optimized vs non-optimized choices. But the difference would not be as extreme.
A system can allow min-maxing to exist while still keeping it somewhat restricted. It comes back to the 'gap' - the possible difference between your optimized character vs your average character.
With the PHB, how much of a difference could you get between an optimized vs non-optimized character? In terms of attack rolls, for example... maybe 2 points from ability score mods, 1 point from proficiency. Another 1 point from Epic Destiny. Potentially 1 point from your class. A few other elements - one or two paragon paths, temporary power or item bonuses, etc - might expand that, but they were much rarer.
Thus, by level 30, your typical optimized character is likely at +5 to hit compared to your non-optimized character. Now, that definitely can be felt. But the characters are in the same field. A DM can toss out an enemy that one of them hits on an 8, the other hits on a 13. Both feel like they can contribute, even if one is more effective at doing so.
But with feats and items these days, the gap has widened. I'm an Essentials character making basic attacks - let's say I'm an archer. I can get +3 to hit from Eagle Eye Goggles. +3 to hit from Expertise. Maybe another +1 or +2 from various feats/items with easy to meet conditions. And another feat lets all my attacks target Reflex instead of AC.
Suddenly I'm looking at being close to +15 to hit compared to a character who hasn't invested in these things. And that's a world of difference. Enemies that he can barely hit are almost automatic hits for me. A DM really can't challenge both of us at the same time via one opponent.
There are ways around this, of course. But it still isn't an ideal situation. It all comes back to the gap. I'm cool with there being a difference between optimized and non-optimized characters. In any system that allows for choice and customization, you can't avoid that.
But I want that gap to remain much more limited. I want the optimizer to be more effective, but not so much so he is fighting in a completely different league.
Less potent options, along with a system that would let characters invest in flavorful choices without having to pass over the mechanically powerful options... that's really what I would prefer.