Yes, but if a GM is going to abuse things, the rules aren't going to stop them.
Agreed. The GM has too much authority for mere rules to stop a GM from being a bad GM.
"Let us take the really bad behaviors of users, and design our product to stop them," doesn't sound like a good design strategy.
Well, it's near the heart of every design strategy for every professional software product on the market. "Are you sure you want to do this?" is a very good validation. No one would suggest that a good design for a product is, "We'll just trust the users to only delete the things that they really want to delete."
As far as your complaint goes, I think the key to understanding what I'm proposing and thinking is in the sentence, "it quickly gets out of control
particularly if the game gives the GM no markers to represent how much resources/favor an NPC is receiving and doesn't encourage them to think about it or reflect on it." If the game begins with the assumption that ideally everything will work on a similar system - and if the system is really too complex for NPCs its probably too complex for PCs as well - but that the in the interest of time its often necessary and desirable to use simplified models for the NPCs, and then opens up its math and expectations to the DM and says - "If you must wing it, here is how to do so.", it's a very different mindset that the mindset that NPC's and PC's are supposed to be using different rules. While I'm a big critic of 4e in a lot of ways, the suggestions it makes for rapid templating of NPCs and for reskinning monsters are in and of themselves good approaches. The big problem of course is that to get the most out of the system, the NPCs not only are using a completely different system than the PCs but
are just as complicated to design and nearly as difficult to run as well. Any value in having the NPC/PC divide not share complexity is loss if the intention is to create these elaborate set piece combats with each enemy have specific flavorable powers, and the high level elites are very good examples of this. Conversely, on the other hand you have this huge stat block and while it doesn't involve nested text to the same extent as 1e-3e stat blocks, it lacks the out of combat refinement of the early editions as well because its gutted the "clutter" regarding what the monster can do in situations outside of set piece combat. This means out of set piece combat you're likely to get back to DMs relying on power of plot on one hand and not finding out of combat inspiration on the other.
Unless you have a really dim view of your customers, you are giving most of them reason to have to subvert your system to get what they want done.
Or looking at the otherway, a rules set is very much like open source software. You don't have to create a special privilege for the technical end users to modify the rules how they like (and accept the consequences). But you still want the system to work well for someone who isn't reprogramming it to suit his immediate needs all the time.
Good users will look at the system, and likely avoid it as being difficult to use (see the many, many comments of how 3e was difficult for high-level play, and how GMs avoided that range). Bad users may do likewise, or will simply go ahead and subvert the system (see the many discussions of house rules, and how to do quick NPC design without bothering with the full detail of PCs). So, using this as a design consideration seems futile.
This is a huge discussion that would end up in edition warring eventually, but while I fully agree that 3e had huge problems with high level play, I'd argue almost none of these problems had their real roots in NPCs sharing the rules of PCs. High level PC's were themselves too complicated, too reliant on absolute powers, too fiddly, too imbalanced, and too difficult to create (in that any noob following an intuitive character creation path in 3.5 would likely end up with an unplayable character given 3.5 assumptions). Without attempting to prove this for 3e, which would involve much discussion, I'll just point out that 4e threw out the NPC=PC assumption and still wound up with slow play, difficult to create high level encounters, and complicated PCs and NPCs.
So in short, the problem certainly exists, but it's not clear to me that the proposed solution is really a solution.
Designing some training for the user in why they shouldn't abuse the players, and how to avoid accidentally doing so, seems a far better approach than baking it into the rules.
If that's your thesis, then more or less we are in agreement.