OK. But when's the last time you had to "rely on items or rare abilities" or "enlist the help of the rest of the party" to deal with a powerful fighter or rogue?
Well, given that equipment and spell scrolls are technically items...
What you're doing here is basically accepting the game's status quo of "mages are the most powerful, the primary challenge in this game is killing mages, deal with it".
Why does it matter how well the player classes are represented among enemies? Honest question. It's one thing to have a concern for how the different classes can interact with enemies and complete the content, but their representation among enemies, more specifically high-difficulty encounters seems a little unimportant.
Hypothetically a Fighter or Thief could be made very dangerous (as the player can easily do), but making them dangerous would really be a pain for a lot of reasons.
Making a Fighter dangerous would involve giving them loads of resistances/immunities via martial-class equipment a la player character Fighters (and be beneficiaries of buffs from any casters on their side). And Thieves would need to make excessive use of Traps or Backstabs (possibly with Invisibility; like the SCS2 mod has no problem doing) as well as lots of sneaky equipment (oh hello Cloak of Non-Detection how are you?).
Basically, this makes both of them huge pains in the ass to integrate into the game. Loading up enemy warriors with equipment equivalent to what a typical Fighter a few hours out of Chateau Irenicus could have would result in a huge monty haul campaign and even if you dodged that by making their items not drop it wouldn't be much fun because you can't use Knowledge: D&D to guess their capabilities as you would a typical monster out of the manual. How are you supposed to know that Lord Dingleberry is wearing Dragon Plate +5 that makes him immune to mind-effects and the elements or that Baron Crustybum the Itchy, through equipment munchkinizing, has stacked his immunities to all damage types into the high 90-percent area besides Acid because *gamey reasons*.
And with Thieves...geez. Most people would flip shit if Thieves did anything close to what something like SCS2 does (actually makes them play well) because their waifus would get chunked by hideous backstab damage. Plus, Traps are cheesy enough when used by the player; imagine enemies being able to randomly mine the battlefield with powerful snares. Imagine how difficult that would be to script.
It's not hard to see why Bioware just went, "Fuck it" as far as making dangerous high-level Fighters and Thieves and just went with casters and monsters as the primary high-level threats. It would have been far too cumbersome. Fighters and Thieves can be powerful in player's hands, but making them powerful in the "hands" of a CPU gets messy fast.
What do you think of the updates you've seen so far?
Eh, ambivalent really.
A lot of the mechanics sound like they could be excellent or terrible very easily depending on how the game-as-a-whole is. Theory-crafting is one thing, but there's no substitute for playtesting as far as understanding and evaluating a system goes. Plus, content is kind of an important thing as well and that's probably my biggest concern. I do
really like the idea of scaling encounters based on difficulty level...though I have no illusions that it will be only applied in some encounters and likely applied systemically once they realize how much of an additional burden it will become (not only content creation, but testing each difficulty).