Thank you for asking - it's always the case you could use more words. On the other hand, you would never post anything if you polish your sentences endlessly, so it's sometimes better to be brief and then answers questions as they come up.
I'm having two "attack vectors" in mind.
First off, any game that obsesses with balance to the degree that it meticulously shuts down any avenue to gain that extra +1 or +2 bonus is a game I consider inherently distrustful of the GM. The contrast is 3E or 5E: games that happily will supply you with a magic item that gives large, almost game-breaking effects. A game like that trusts the GM to have a will of their own to know which bonuses are appropriate for their campaign. Both 4E and PF2 come across as incredibly cautious, defensive and controlling in this regard. Most magic items in these editions are incredibly bland and boring because the devs don't dare give out any really impactful bonuses and effects.
The other one is admittedly not one I've spent a lot of time explaining, at least not in this thread. The way I find PF2 actively hostile to good gamesmastering, defined as saying "yes" or at least "yes, but" instead of "no". Whenever a player asks to do something that his character's abilities does not quite allow, my tendency is to allow it, but perhaps asking for a check, or mandating a drawback for failure (the "but" part).
My problem here is that
by your own definitions 4e trusts the DM more than either 3.0, 3.5, or 5e. And therefore your logic leads to the idea that Pathfinder 2e is making the right choice by not trusting the DM in the slightest.
The game that meticulously shuts down any avenue to gain that extra +1 or +2 bonus is pretty obviously and pretty clearly 5e. Unless you are handing out magic swords, a +1 or +2 bonus is outside the design space of DMing in 5e - it's Advantage or nothing. There isn't even an advantage to be gained from flanking a foe (except by an optional rule where it's full Advantage, thus exhausting the entire GM suggested toolbox). Now
that's controlling.
Meanwhile 3.5 (and 3.0 and Pathfinder) hedge in the bonuses that can be handed out by using things like Touch and Flat Footed AC that are an absolute pain to work out, and it hedges in the DM with a huge collection of rules and modifiers to the point the designers admit that part of the point was controlling the DM because they were excessively cautious at the time of creation (and looked back later to ask "What have we done?"). Whether this is more or less controlling than "Advantage or nothing" is an open question.
4e by contrast has the Combat Advantage status effect that can be handed out like candy in a way Advantage can't. It's only +2 for one thing and for another it's considered a common thing; the rogue (or anyone else) can get it through flanking or through attacking when they are hidden. Or so many other ways. And want to hand out something that isn't the vanilla Combat Advantage? 4e provides DMs with
an entire mark-up language and benchmarks to do this. That's what the powers system is about and what the page 42 of the DMG/improvised rules and damage by level are about.
As for game-breaking items, 4e had almost from the beginning a portable hole that works as in a Bugs Bunny cartoon rather than as an alternative to a bag of holding. Among the list of abilities that is simply handed out is a Monk's optional ability to wire-fu about 40 feet, counting as a flight action once per encounter. At second level. And the DMG came with artifact creation guidance.
Now you probably are (understandably) looking at the endless tedious pages of +1-+6 items that are every bit as much 4e's version of filler as pages of prestige classes almost no one will ever use are 3.5's. The basic fact is that a +2 sword has
never been interesting - and it would have saved a lot of time and boredom to start with inherent bonuses and not have a table under each sword showing six different possible levels so the book could be padded out. This is a place where 4e's presentation was (to put it bluntly) terrible.
However, to my dismay I've found that Paizo has reserved almost the entire space for flexibility for their own use. In other words, I find over and over again that the thing I allow for free turns out to be allowed by a feat down the line (possibly for a totally different character class).
I'm not sure whether your problem here is that they didn't follow the lead of 4e in removing the gates that were present all over the place and getting rid of the "Air breathing mermaid" feats - or whether your problem here is that the players get to control what sort of cool stuff their characters can do.
I find this very unfriendly to gamesmasters, who apparently are expected to load the entirety of the thousands of feats into their brain and account for each and every single of them before making an off the cuff decision.
Welcome to D&D 3.X. It is this that 4e emphatically turned its back on - a 4e DM does not need to know what
any feat does or the rules for
any spell. That's all player side. The DM just needs the MM3 on a business card because PC and NPC rules are different.
5e on the other hand returned closer to the 3.X way of doing things where Spell Like Abilities are a thing even if you don't need the over-proliferation of feats. And fortunately Spell Like Abilities are rare and normally (although not 100% of the time) the SLA is in the monster statblock and it's only to run a spellcaster you need to actively look the rules (in this case the spells) up in the PHB rather than having it right there in front of you. As is so often the case in the metrics you claim to care about 4e is the edition that actively empowers the DM.
It comes across as the game not trusting the GM: every little interaction (including all variants) are locked down by the game, telling the GM "this particular shortcut or flexibility is appropriate for a level 16 Bard" (or whatever) "but certainly not your level 3 Fighter, or the feat would have been made available that much earlier".
Compare that to a game design like 5E where this issue simply doesn't exist.
Yes, of course, there are no level gated class features or spells in 5e. Right.
I know that "this issue simply doesn't exist" is intended as hyperbole - but it is to me absurd hyperbole. In any level based system the character level is a measure of appropriate power. You are literally here criticising Pathfinder 2 for being a level-based game in the same way as literally every edition of D&D has been since 1974.
And on another note magic items and feats do not have to be the same level at all. I can imagine it being a very high level barbarian ability to make their skin as tough as steel - meanwhile the fighter can wear plate armour from level 1 that in practice does the same thing. Technology can make you special without needing magic or ridiculous levels of training.
To me, Paizo needs to explain what the benefit of their design is that warrants such a (for me) huge drawback. Especially since just about the only benefit I can see myself is "it lets us spam character crunch over our dozens of splatbooks" - replacing really different crunch with a deluge of a thousand feats, nearly all of which just rearrange the existing numbers: very easy and safe to design, very cheap to publish, but empty calories compared to a real meal (where new supplements actually add new subsystems to the game).
Tl;dr:
Pathfinder 2 is a game that asks you to care for Talismans.
Pathfinder 2 is a game that asks you to care about literally hundreds (and soon thousands) of little niggly feats.
Not sure I can explain it more clearly than that.
And your opening post might be a bit of a clue. You say that Paizo should have looked at the failure of 4e. Then you give as a critique a list of ways that Pathfinder 2 is not just closer to 5e than 4e but is the far side of 5e from 4e.
Pathfinder 1 was a game that asked you to care about literally hundreds (almost 1500 not counting third party material) of little niggly feats. And it was that because 3.5 was a game that asked you to care about literally hundreds (just over 1500 in all) of little niggly feats. Your critique here is literally "Pathfinder 2 is following in the footsteps of Pathfinder 1".
4e meanwhile was a game which
had hundreds of feats (it ended up with just over 1500 in all as well), some of which were little and niggly (and no one ever took because of it) but the only feats it ever asked you to care about were those that helped you build your character the way you wanted; if it wasn't on your character sheet it wasn't relevant. There was no list of "feat prerequisites" to get into prestige classes or paragon paths and there was no common list of feat chains like Dodge/Mobility/Spring Attack/Expertise/Whirlwind Attack required to let you get at the good stuff. And above all the DM wasn't required to use feats to build monsters let alone know what those feats did because the statblock wouldn't tell you.
And this is why you are getting such strong pushback when you decide to tubthump against 4e. Everything you claim you want was something 4e did - and in most cases it was something 4e did
better than 5e. Meanwhile from your description the problem with the design of Pathfinder II is that it appears to treat the design (rather than the presentation and the way the game was severely undercooked at launch) of 4e as toxic.
