While Karl Kaster obviously will find the optimal path no matter what, there is a huge difference between recognizing this fact and not freaking out if he notices something really good and throwing up your hands saying 'whelp guess we don't have to care.'
A huge problem with 5e magic is that it didn't take Karl Kaster for things to get out of hand. "Combos" (as far as they could be called combos) as simple as 'Combine sustaining spells with the thing that reduces sustaining penalties and then sustain lots of spells' resulted in very unfun table situations where either the mage just was way too good at everything or mages who didn't sustain and adepts would be cruuuuushed by the attempt to balance magic 'in universe' against sustaining.
In essence, despite the fact that optimizers are going to optimize, it doesn't let the design off the hook when the balance is just bad. The logic fails because you can't apply it in other identical situations and sound sane: "Because Sammy Samurai is going to find the optimal path anyway we can just make things really broken and release a gun that does 20p and automatically hits and costs 200 nuyen, is avail 4R, and is concealment +2." That is a hyperbolic example, but obviously the goal SHOULD be that things are fair.
A huge problem with 'in universe' attempts to 'fix' magic is that magic usage is completely under the control of the PC with no inherent downsides to the PC. Like if an aspect of magic is not a good tool for a specific situation you can just choose not to use that aspect, and in many cases the 'downsides' of magic can be trivially avoided, like drain just via resting (Though the healing changes may fix that). If X buff spells are worth the penalty but X+1 isn't, people will just use X, if you can sustain Y drain safely without really losing anything people will. If combat magic is worse than a gun mages will just buff and use a gun. None of these are INHERENTLY bad things but you want the numbers involved to be good numbers and want to respect the fact that there is a strong emotional attachment to mundanity among many players and the perception that mages get special treatment from the designers has caused... amazingly bitter feelings that have destroyed groups.
Balance isn't really important to keep things balanced, because Karl Kaster is going to find the best thing to do anyway, but you need to make sure that A: Being any role has an upshot that feels like its yours and B: No one feels like the designers are playing favorites or that any one thing has so much upside not picking it actively makes you feel bad.
There are indications the problem might be solved (For example, sustaining penalties don't really work conceptually to balance buffs, buffs are either efficient dicepool wise or they aren't and if they are they REALLY are, so getting rid of them and balancing on a different axis makes total sense) and drain plus the new healing changes might really work to prevent mages from just casting a bajillion small effects before summoning one big spirit and steamrolling everything. But the problem that severe imbalances create really shouldn't be laid at karl Kaster's feet. Minor imbalances get smoothed out by the fact some archetypes have strengths others can't emulate, and while Karl Kaster will still exploit those he probably won't ruin anyone's day with them.
The issue is when the strength of an archtype is so huge it warps the entire game and setting around it, like 5e mages with their summoning mini-gods, once your starting PC can consistently summon a force 12 and survive the fact they can't personally get good soak sorta pales as Spiritzilla solos an indefinite amount of HTR opponents unless they summon their OWN spiritzilla that no one else can touch and suddenly the only people who can really affect the outcome of the conflict are Pokemon Battler mages who aren't even on site and everyone else might as well go home. That isn't Karl's fault, that behavior can be done by Danny McNevercastaspell on his first mage just because "Push drain resistance and casting and summoning to as high as you can go" isn't exactly a nuanced breakage of game mechanics so much as being a thing rookie players often instinctively do because its easy to understand. The 'in universe' cure is inevitably as bad as the disease (See: High Background Counts that can turn off force 12 sustained quickened spells absolutely DEMOLISHING adepts as a playable concept) and it really just is smarter (and also literally the job of a game designer) to ensure THAT doesn't happen.
Ideally (and it seems like this may be the case, though rumors of spirits STILL having hardened armor are spooky) mages should be tossing out big spoopy scary effects that make everyone turn their head, as well as small tricky effects, but should suffer in terms of sustaining big effects for a long time, or personally being the biggest badass in the room, and they shouldn't just be smashing through problems by tactically limiting their expected drain to be 1-3 stun and just laughing off a -6 sustaining penalty because their combined buffs give em +9 to everything. You should have SOME idea what is healthy behavior for an archetype and should make pretty significant effort to ensure that healthy behavior is the natural 'lane' the archetpye falls into, without specifically designing the archetype to be defined by what they can't do (like 5e hackers kinda were).
I think things that would be a red flag for 6e balance for magic as far as we know with public information would be combat buffs that are as selfish and potent as they used to be in 5e, strong tools for supporting buffs that can compete with augs head to head, spirits of even moderate force being unkillable by mundanes, infinite force scaling, ways to nullify drain, and tools that make drain hit softer like centering foci. As the power level for many mundane things has drastically fallen due to how edge is going to work, mages kinda need to feel that drain pinch, need to essentially lose access to augmentations now that the big upshot of augmentations are likely weaker, and definitely shouldn't be able to essentially summon 5e style sams. That seems like what is actually happening, but it really just takes one stinker to bring us back to the status quo where Little Billy's first mage may accidently crack their campaign's spine over their knee.
Also, I am hecka interested in POSITIVE elemental effects on things like healing, just to end on a somewhat positive non-sequitur. Like color me intrigued in the extreme, as weirdo side boons and buffs and debuffs always seemed like a great way to make mages feel magicky while not making their effects just superior to mundane ones.