I don't see how there is a practical difference in these two things, despite the obvious philosophical difference.
<snip>
The only difference I can measure between the two is how likely it is that a DM deciding to rule differently than you expected from reading the rules yourself catches you by surprise.
At this point I'm not trying to interpret the OP, but just speak for myself.
I think the difference is a real one. It is about who is expected to have control over the fiction.
In 2nd ed AD&D, control over the fiction is expressly granted to the GM, and the capacity of the players to influence the fiction via their action declarations for their PCs is expressly mediated via the GM - who gets to decided whether it would be "good" or "bad" for the game to allow particular outcomes to occur within the fiction.
In 4e - to pick the version of D&D that I think is furthest from 2nd ed AD&D - the players are explicitly granted a high degree of control over the fiction. For instance, once the skill challenge is framed then if the players get N successes the fiction is changed in accordance with their desires (as expressed via the action declarations for their PCs). Or, to give a combat-related example, if a player declares that his/her PC uses Mighty Sprint to close the distance to the NPC bad guy, then uses a daily power to get some sort of buffed attack, and then spends an action point to follow up with another attack, the GM is not authorised by the rules of the game to interpose his/her "common sense" or his/her sense of what is "good for the game" to interfere with that sequence of action declarations.
(I think Gygaxian D&D is in many ways closer to 4e than 2nd ed AD&D, but uses a very different range of techniques - in particular, the very distinctive environment of
the dungeon, and conventions around that including GM notes and the turn sequence - to achieve that outcome.)
5e is not as clear as 2nd ed AD&D about who is expected to have ultimate control over the fiction, but I think at various points (eg its advice on DC setting, and its lack of a non-combat scene-resolution mechanic comparable to the skill challenge) it leans closer to the 2nd ed than the 4e direction.
is this merely an impression you get from reading the books, or is it something you've experienced in practice? Because I've played a fair amount of 5e and have yet to encounter the phenomenon of DMs denying player agency.
It's based on reading the rules and noting the resources available to players based on PC build. (Especially players whose PCs do not use spells.)
To step back: I don't think it can be true
that 4e is different from (say) 2nd ed AD&D and
that 5e harks back to 2nd ed AD&D and
that 5e is not noticeably different from 4e. I think the first two are true, and hence the third false. (In general. Maybe there are particular tables where things are done differently - most RPGs can be applied in pretty flexible ways.)
What is the noticeable difference between 5e and 4e? I think there are a few, but one of the main ones, it seems to me, is in the role the GM plays in framing and adjudicating action declarations.
None of this is a criticism of 5e. After all, a lot of people like 2nd ed AD&D better than 4e! It's an observation, and in the context of this thread is meant to show a degree of sympathy for the OP's post. (Even though I don't think we are coming from exactly the same place: for me the main issue is
finality, which is (3) in my post upthread; and the OP has endorsed that post, in which I conjectured that the OP's main issues were (1) and (2).)