I mean it comes down to personal preference, obviously, but let's imagine two systems.
One system is called "heroic effort". It's described as digging deep into your reserves of strength, a surge of adrenaline, the occasional ability for someone to perform feats that far exceed their normal capacity- the hysterical strength that allows a mother to lift a car off her child.
Another is called "call upon faith". It's described as letting of one's ego and calling upon the supernatural for succor in a time of need- the power inherent to faith, or actually calling upon a divine (or infernal) entity.
Now, one could say that there's no real difference here- both are abilities a character could employ, which are controlled by the player- the player decides that this moment is when their character enjoys a burst of power or gives in and prays to an uncaring deity.
However, one has both a roleplaying element and a described action that needs be taken- uttering a pray or oath, falling to one's knees, steepling one's hands, whichever.
Some people will view the second ability more favorably in that it's the result of something occurring in-universe, despite the fact that both are abilities invoked by the player.
If a game lets you spend an "action point" to do a thing, we can certainly make up a fiction as to what's going on, but it's still a resource that most people can't call upon at will, yet the player can decide the when and where.
And that the player can arbitrarily decide the when and where to invoke these meta-bennies is the crux of the debate.
Further, there's three types of meta-bennies:
1. those that obviate a die not yet rolled by forcing a success
2. those that add a bonus or benefit to a die not already rolled
3. those that do something to change or negate the results of a die that has already been rolled, or force a re-roll.
Of those, 2 is by far the least objectionable. 1 is fairly uncommon but bad in that it can trash an otherwise exciting will-it-or-won't-it moment.
And 3, IMO, simply should not exist. Once a die is rolled, it's rolled, and you shouldn't get to see the results before deciding whether or not to make changes to that roll. It's baked-in player-side fudging with lipstick on.
What gets these long running debates going is the fact that our individual preferences get in the way- some would much rather a game that closely hews with reality and aren't concerned with balanced mechanics- or any mechanics that get in the way of such things!
I think it was Yogi Berra who once said regarding baseball something like "You're gonna win 1/3 of the games no matter what you do. You're gonna lose 1/3 of the games no matter what you do. It's what you do with the other 1/3 that makes you a good or bad team."
The same holds true in D&D design. A certain amount of rules are going to be dissociated/unrealistic/etc. no matter what you do. A certain amount are going to be realistic no matter what you do. It's what you do with those rules where you have a choice that makes the end-result difference....
Some options are better than others in reality, so it should be for the game. If a fire spell doesn't say it ignites objects, too bad, if you're in a bakery, that flour is going up!
...and here's a good example of where there is a choice. A fire spell could be ruled to potentially ignite flammable objects in its AoE (which is realistic) or not (which is unrealistic). For my part, in these cases I say go with the realistic option whenever you can.
Others think the game rules should be functional and balanced, even if that means "realism" has to take a back seat from time to time- D&D, as a game, is full of these, long since enshrined. While some of us accept that, others only tolerate it- sure, the nature of initiative and turn order is an abstraction, but if it results in bizarre or illogical results, they might feel that these should be ignored in the favor of the narrative, even if everything is working as intended.
...
Now, back to healing power creep! Whether increased healing is good or bad is based on a few factors, including, but not limited to:
1- is having more "uptime", where characters are able to make actions in the game more important than "downtime", where characters are unable to act. Characters at positive hit points usually can act, characters who are dying usually cannot. Most conditions in the game impose penalties on actions or prevent them outright. Some feel they should be used sparingly, so that the amount of time players are unable to actually play the game is minimal, while others feel that them's the breaks.
I think the rules have focused far too much on balance (which is a fool's errand at the micro level and fairly easy to achieve at the macro level) and character uptime, and in so doing have done something of a disservice to actual roleplay and immersion.
2- if players really want to cast healing spells before players go down. Some players like the support role and would rather have more efficient healing spells. Others would rather do anything at all but heal their fallen allies. A Light or Storm Cleric might be happiest blasting enemies and will grumble at having to cast an emergency healing word- they're not going to really care much whether said spell heals 1d4 or 2d4, unless the difference is enough to ensure they won't be forced to heal you again next turn.
3- how often and common short rests are. While the healing resources given to characters in 5e are significant, if you can't stop to rest when you need to, you have to rely on magical healing more. If you can't access healing surges, they basically don't exist, and it's not insane to think that something needs to take up the slack. On the other hand, if characters are always resting, then they really don't need more healing outside of combat.
4- the prevalence of healing spells in the game. Somewhat related to point 2, but it's important to note that groups vary. A group with one Cleric, two Clerics, or no Clerics (or Bards, Druids, etc. etc.) will play very differently. The amount of potential resources that can be used to heal cannot be understated here! A group with no magical healing doesn't care about more efficient healing spells- unless I missed hearing it, it's not like healing potions heal more hit points in 2024- the only change is that they can be used as a bonus action, which is a potential buff to action economy, but I can tell you having used this rule for much of this year, only to a point, since WotC absolutely adores giving some characters other uses for that action!
A group with two Clerics and a Druid are less likely to need this buff than the party with one. I don't think WotC expects the "default" adventuring party (if there is such a thing) to have more than one character with healing spells, and that no doubt plays a large role in why healing spells are buffed.
5- encounter difficulty. Some enemies hit like a truck. The CR 1 Brown Bear can deal almost 30 hit points in a turn. This is, of course, limited by their chance to hit, but as enemies grow stronger, so too does their accuracy and damage output (generally speaking). If a campaign consists of a few "marquee fights" with strong enemies or many smaller combats per session, will affect how much "burst" healing you need at the moment.
I argued in the past that if a character is taking 30 damage in a round, then a max-level healing spell has to exceed that damage to be worth that action. I know a lot of people disagree with that notion, but the simple truth is, healing is not helping you win an encounter as efficiently as damaging or disabling enemies. As an MtG player might say, healing is just "losing slower".
D&D has this interesting scenario where damage spells are not terribly efficient at dispatching foes- CR 1 foes who might survive an 8d6 fireball exist, for example. But somehow, healing spells don't even keep pace with damaging spells, let alone with incoming damage!
Now some might prefer the offense beats defense paradigm, as it leads towards faster, deadlier combats. And you know what, if that works for you, great! But the idea that a character can select one of several strategies, but some are just generally better than others bothers me- I don't like it. There should never be a "best choice", only the best choice for the moment.
Now before anyone accuses me of ignoring "heal from zero"- I assure you, I'm not. But if my Fighter took 30 damage and fell down, a healing spell that does less than 30 us not going to prevent him from falling down again- some other defensive option needs to be employed, which now has the party using up two actions to deal with one Attack action. If the party outnumbers their foes, that's not too bad, but if they don't...
The other issue underlying all this is one of pacing. 4e-5e design wants the characters to be going like little energizer bunnies from one long rest to the next, with (in 5e) a few shorts in there to prop up some classes. This paradigm is why healing has been dialled to eleven, and is also utterly unrealistic.
Realism, however, suggests the characters (if at all wise) are going to want to be considerably more cautious and long-rest whenever they can; a more realistic game design would assume and embrace the idea of the 5-minute workday in the knowledge that, given the choice, that's what the characters would try to do.
Realism also suggests that a character who is close to death (i.e. making death saves) shouldn't be up-'n'-at-'em after just one cure and right as rain the next morning. In other words, realism calls for a much slower pace where getting hurt is a bad thing and where curing hit points before someone is down (or just not taking the damage in the first place!) is far preferble to waiting until after someone is down.
And yes, natural recovery in 1e as written is unrealistic the other way: it's too slow.