One big thought I've had that informs my view of the debate (I'm an optimizer, except I'm actually mainly the GM) is the Legolas problem.
The Legolas problem is that in a situation without deliberate optimization, two players sit down to DND and both come up with nornal good faith concepts-- player A wants to play against type and be an orc wizard, while player B decides to be an expy of Legolas.
Player B will end up in a situation where they're playing an elf ranger or fighter, with a bow, and say "oh, Legolas would totally have these Sharpshooter and Elven Accuracy things because duh, he's super cool with his bow, oh and he's really sneaky, so lets do high dexterity"
Player A is in a situation where depending on what book they're using, they have an intelligence penalty and no actual benefits, nor even a bonus to dex to help their clothy AC.
What the Legolas problem illustrates is that differences in power follow thematic lines in RPGs (until a sufficiently high level of optimization) and so you need system mastery to avert optimization to some degree.
A stoner player, stoned during play and character building, yeeting fireballs is incredibley potent, because fireball is overtuned.
I sort of detest that the game makes that whole to do necessary.
Totally Unrelated, but I sure do love Pathfinder 2e ; )
I call this the Druid problem, because of my experiences with 3.5 Druids. I watched as a new player looked over the classes and thought "Druids sound neat!" and she proceeded to make a high Charisma, Disney Princess type character. She fooled around a little with wild shape, but she really liked casting spells, and it annoyed her that she couldn't do that as an animal. Flash forward to level 6, where, without any outside influence, she notices "oh, there's a Druid only feat that lets me cast spells in Wild Shape! I'll take that!".
And to be fair, it took her a bit to really grok what this actually meant, but inevitably, she became a Druidzilla. Not because she was going out of her way to optimize (remember that high score she put in Charisma), but because the system itself was lighting the way down that path. I was only a player in this game, and it wasn't long before the DM came to me and said "I don't understand. Why doesn't anyone think the Druid is OP?". I had to laugh, because unlike myself, I've found most players in the wild don't go on forums and learn what people are saying about the game (infamously, I played a Rogue in Pathfinder 1e, and the GM started grumbling about how powerful I was- when the Rogue had often been labeled one of the weaker classes online for years!).
Magic the Gathering, in their set design, include "signpost cards", usually in the uncommon rarity, that are some sort of payoff for adopting a particular strategy so that, when drafting the set, you can quickly realize "oh, so Red/White in this set wants to make Treasures and sacrifice them for things other than mana!"*. Thus the player starts looking for cards that support that strategy, and lo and behold, they exist in abundance.
*Whether or not this has actually happened isn't relevant to the example.
There's both good and bad here, obviously- on the one hand, it helps newer players create things that will work well in the game. On the other hand, it takes away a little of the discovery and squashes creativity- you end up with a lot of players doing the same things.
This is why, for example, I've heard many a DM over the years gripe about "guides" found online, as if they are some sort of Game Genie designed to break their game, when in reality, they mostly just exist to tell you "these options good, these options not so good" or "you want to do X in the game? This is the best way to do it".
The poster I quoted, for example, used Fireball as an example of something they feel is overtuned for the game. After all, everyone knows the DMG thinks a spell like it should do 6d6 damage instead of 8d6! Busted! But really, if you examine the spell in detail, and compare it to the things you're most likely going to be using it against, it's actually not that great. It's upcast is pretty sad, so it's probably always going to do that same 28-ish damage, save for half. And when you look at the hit points of monsters, you find there are CR 1's that can survive a Fireball, and CR 2's that won't even be brought to half hit points by it. If anything, the spell is probably weaker than it was in 3.x (where a CR 2 Ogre has the hit points of a 5e CR 1), and far weaker than it was in AD&D- it still does it's job, but outside of lucky rolls or damage bonuses, it's actually more like a C+ spell in reality.**
**I know this isn't universally agreed upon, as there was a thread about it on this forum a few months ago. And it's probably campaign-dependent to some degree- if you run old school hordes of weaker monsters and dungeons with smaller rooms that provide a target-rich environment, ie, what the spell was originally made for- it probably performs very well. But if you, like me, are tired of rolling the attacks for 10 enemies every turn just to whiff again and again because +4 to hit is a joke, the spell will have a very different impact.
But once again, despite the actual impact of the spell, it's obvious that Wizards deliberately made it a signpost for those classes that can use it "wow, 8d6 damage! It's so much!" along with internet memes make it an obvious choice. In the game I play in, my party still wants me to cast Fireball in just about every fight, believing in their heart of hearts it will end the fight so much faster than my control spells. And I can see the DM's eyes light up, hoping I'll do just that instead of debuffing his enemies into oblivion (I think he may have cried a little when I cast Synaptic Static for the first time).
And this signposting can go horribly wrong or right, depending on how good or bad the things you're being steered towards are. In the case of the 3.5 Druid, it went horribly right, making an already potentially powerful class that much stronger. But it can backfire as well, like the 3.5 Fighter, who was signposted into choosing from a list of "combat" Feats, many of which are difficult to qualify for and very situational. Some even obsolete themselves- once you have Spring Attack and can move around without provoking, how useful is that Mobility? Or worse, once you've got your Improved (insert combat maneuver here), are you really going to reduce the chances of success by invoking your Combat Expertise?
5e, by narrowing it's range of options, actually makes the impact of signposting worse (or better, depending on your point of view), because there's a lot less paths a character can viably take. It's inevitable that a Cleric is going to find their way towards Spirit Guardians- because the spell isn't just signposted, it's a lush oasis in a parched wasteland! With no other equivalent option in sight (or even one close to it), steering around it is a huge downgrade for your character. So of 99%*** of all Clerics are going to end up using the spell, to the annoyance of DM's everywhere (myself included!) because what else are they going to do? Upcast a healing spell?
***totally made up number, of course. I'd be surprised if I wasn't in the ballpark though.