I was thinking about how it's a huge pain in the ass to generate leaders for brigand camps, because they might reasonably be expected to have ability scores, class and level, proficiencies (feats), and magic items. And that's just for the fighters.
And I was thinking, what if we halved the work, and either picked standardized proficiencies for all 9th level brigands and then rolled the magic items, or picked standard magic items and then rolled the proficiencies?
For a second there they seemed sort of equivalent, with the major difference being that you can steal magic items (or pry them from cold dead hands).
And that brought to mind something that had confused me about OSE: the Basic magic item generation rules, which generate an abundance of high-powered magic items like crystal balls, but from a very limited list. That didn't make a whole lot of sense to me, giving these powerful items to low-level characters. But if there are no proficiencies, I could sort of see it. If each PC ends up with one significant magic item by 3rd level or so, then that's sort of like the class proficiency they'd be getting around 4th level. And then the restrictions on which classes can use which items also start to look a bit like class proficiency lists - wizard-only item, wizard-only proficiency, tomato, tomahto.
The balance between proficiencies/feats and itemization is interesting to consider. In ACKS, magic items dominate proficiencies in terms of numerical bonuses and eventually availability. A sword +3 is like... six proficiencies? Sort of, since you only get the bonus with it and not other weapons? Arguably ACKS' proficiencies provide smaller but broader bonuses, like +1 to hit with a variety of weapons.
In 3e, magic item bonuses were also mostly bigger numerically than feats', but I'm not reaching any good conclusions on breadth (a +2 sword is a bigger but narrower bonus than Weapon Focus+Weapon Specialization, a circlet of intellect +2 is a comparable but broader bonus than Spell Focus, Cloak of Protection +1 is basically Luck of Heroes).
I could definitely see shifting the balance between these two back further towards magic items (though 5e may have, for all I know - making feats optional might have that effect if they're really seldom used). Under the conditions that the magic item market is very small and you seldom get them during character generation, a magic item is a reward for something that happened in play. It's like the "backstory is what happened to you before 6th level" line - instead of writing something on your sheet during levelup, go out in the world and kill some brigands and take their stuff. Want a horn of blasting? Go talk to a sage and find out about a local hero who had one, and went to a particular dungeon and never returned.
(It's a bit funny to contrast this with my position from 2011 on getting rid of "standard" magic items and replacing them with inherent bonuses. I suppose I would rather be inconsistent with my past self than stuck and not evolving. There is a common thread here though - any standard component of a "build" formulated and achieved off-table, be it feat or magic item, really is boring)
Magic items are also sort of heritable proficiencies, if your next character was your next of kin. So it's a form of progress which can be preserved across death, much like henchmen. Having preservable progress lessens that sting.
Looking at magic items like proficiencies might also help explain sentient swords, with their suites of weird powers. They're a way to give a whole package of special abilities to people other than clerics and MUs (most importantly, fighters and thieves). It looks sort of like having a higher rate of proficiency acquisition for sword-using classes if you tilt your head just right, squint real hard, and have two drinks (above and beyond the much higher availability of magic swords than other magic weapons, of course). It's also interesting that in both ACKS and OSE 70% of sentient swords are Lawful - they're clearly meant more for PC hands than the hands of their opponents. It's funny, because the prevailing culture and expectation around sentient swords tends to be in the Stormbringer vein, where they're usually evil. I think I gave my players a Lawful sentient sword in the Bjornaborg campaign, but as a player all of the sentient weapons I've found were evil (and I'm 2 for 2 as "Designated Bearer of the Evil Sword" in our parties, which always ends well). It's also not a 3.x-ism, by rules as written; checking the 3.5 SRD, it looks like 50% of sentient weapons were good, 15% evil, and 35% neutral on the good-evil axis (20% chaotic, 25% lawful, 55% neutral on law-chaos).
This whole thing reminded me of a post from long ago where I noted that henchmen in OSR games sort of fill the role of multi-classing in newer games - they slow down your progression in your main class, but get you a broader range of abilities. It's sort of funny how that also moved from "a thing that you have to find in the world" to "a thing that you choose during character generation and level-up", in parallel to the shift from magic items to feats.
I can't escape the impression that it has been players, aided and abetted by publishers, pushing to sacrifice No Homework in exchange for control over their customization.