Agreed.I think basically this view is, in part, outdated. Techniques and approaches to running RPGs, and thus the associated game designs, have evolved a huge amount in the last 20 years. In 1995, maybe even in 2005, I'd have agreed with you on a lot of this.
But the reality is a game like Dungeon World is simply not described by your view at all. The actual approach is super general and rarely gets in the way. It's not a set of rigid rules in the sense you all seem to mean. It's more of a supporting framework for how the participants interact and the game flows.
The actual mechanics are also extraordinarily general, and yet always applicable, unlike systems that try to handle endless specific situations, which I agree with you don't really work well.
There are other dimensions to this too.
For instance, look at Burning Wheel. To make a RPG work like BW, there need to be dice rolled at the moments of crisis/climax. This is what enables "say 'yes' or roll the dice", "intent and task", and "let it ride" to kick into action. Which, in turn, requires departures from some D&D approaches: every action declaration needs to have dice rolls to support it, including things like casting spells; and those action declarations need to be able to succeed based on a roll, including ones like "I look for a vessel"; and even the simplest actions, like that one, need to have some chance of failure (eg no +10 Perception vs DC 5).
It turns out that it's not completely trivial to come up with a design that fits these requirements. For instance, the way that 4e did it was to make all non-combat moments of crisis skill challenges, which achieve the chance of failure by distributing the checks across multiple participants on the premise that not all of the players will be able to auto-succeed on all of the checks that they have to make in the challenge. And that premise in turn sets up further design requirements.
If someone wants to run a RPG where the main focus is on the players resolving the problem/puzzle that the GM is presenting to them - "finish line"-oriented play, to re-use a term I used upthread - then these design challenges and the way that various RPG systems have met them won't be of much interest.
But for other sorts of play, where "GM decides" resolution is pretty anathema to the goals of play, these design challenges matter. And there has been a lot of development since the RM and RQ and similar sorts of systems were conceived of 40 to 50 years ago.