I'll try to be concise for once in response here. I will go into more depth if wanted.
1. Agreed that certain systems may need to break first (e.g. instanced theme park and/or universally mandatory trinities) in order to open up numerous more opportunities.
2. I'd like to know what you mean by "substantial difficulty levels of content". Looking at my time guild-raiding with my main and pug-raiding with my alts up to Mists, I can't honestly say that XIV feels that much harder content-wise than many a WoW fight, for instance. In many cases I preferred the aesthetics and mechanics of those WoW fights, finding them more unique, but that comes down to personal preference. Learning times between the two often felt similar when considering that XIV only releases a wing's worth of bosses at a time where WoW was releasing a raid, or four wings worth, if at a somewhat lesser frequency. What strikes me most is small differences in the community. I came from one of the smaller WoW servers, but it still felt much more populous and able to quickly throw together a PF raid, even a couple 25-mans simultaneously, than my current medium sized one, and just have fun—though usually progressing faster than more serious / nervous newer guild parties. Somehow, I haven't felt that much from XIV. I'm sure we could blame DF for some of this on the server-by-server basis, or even housing for the reduced value of shout chat, but to me a lot of that seems to come from the nature of the fights themselves, and how they're flowed into.
But if you mean the sheer number of different levels of difficulty, I'm in agreement. I feel like we could get a lot out of things in the range of challenge modes, achievements, or mythic dungeons, or even just a Diadem-like experience without the obvious oversights. Aquapolis and the Deep Dungeon, I hear, at least, are / should be a step in that direction. I have yet to try Aquapolis, but I can't say I'm sold on the idea going by Youtube footage, and look forward to the other's release.
3. There will always be certain constraints. I just hope that they stop seen as the tertiary elements such as combo systems, cross-class skills, and/or an arsenal of mitigation cooldowns, and bring it down to the simple question of "is this composition worthwhile?" where jobs within the same role, range, or line-up may differ somewhat individually, but always have a place in some similar, even if therefore adjusted, party. It can favor a trinity-based setup for a majority of content, but at least gives other options for part of it. I'd love to be able to skillfully no-tank or no-heal a heroic dungeon. But that doesn't mean I'm okay with a certain job, especially with untranslatable gear, being left on the sidelines for cutting-edge content, either. I want to start bringing out the differences, but, depending on the context, those can come out more flatly or numerically than one would think. Designing with that in mind requires both intentional and generally-constrained development for class viability in the "hard" or "fixed" content, while pushing the availablity of options both intentionally (Can a 4-DPS party, working perfectly, clear this? How do different compositions compare?) and generally (allowing CCs on most mobs or giving enough of a delay in certain attacks to make certain mobs kite-able, etc., maybe even eventual changes to enmity systems, who knows?) for lighter content as well.
4. While I do think this game is basically a theme park MMO, with scarcely any sandbox elements at all, and I feel that it often under-challenges and under-instructs the player in terms of combat classes, while allowing little synergy or innovation between DoL and DoH classes and the combat side of the game. Progression in general feels weirdly dislocated precisely because of how little different DoW/DoM classes intersect outside of certain cross-class skills, and, again, that (to me) lackluster approach to DoH and DoL. (Granted, I've only taken either to 50, so take my disappointment with a grain of salt; I am however serious about the untapped potential for connecting these things, although that's quite the pipedream to get into.)
5. I feel like a flexible and manually challenge-additive system would be a good start. Work on the undermechanics, the reusable system first. Reap the benefits later. I can throw out a few spitball suggestions if you want.
6. My fear with trying to be more directly connected to the devs, and I understand just how elitist this may sound, is that not only might the opinions being heard not be representative—they might not even be thought out. I want more discussion with the devs, but I have no issue with the idea of "doing their work for them" first, because when we refuse to delve into that, we may well complain about symptoms whose removal only worsen the disease that caused them. I feel like aside from the CMs seeming to focus their interactions on just a narrow band of (typically UI/QoL/lore) concerns, we have almost a direct enough means to voice our concerns. What I don't feel we have are truly hashed out suggestions in most cases.