Given responses in this very thread, I can guarantee that I would be considered a bad player by quite a few of the GMs that have posted in this thread.
Right. "Bad player" is relative to expectations.
In my Torchbearer game, one of the players deliberately initiated an interaction with a NPC anticipating that it would earn him an enemy. Is that good play, or bad play? I thought it was great play - it gave me a hook for hanging a twist to a later failed Resources check (the PC's enemy had had him black-balled at the markets), it prompted a later confrontation between the two characters (initiated by the players, and shifting the terrain from scholarship (where the PC is weak) to oratory (where the PC is strong), and it gives me material for future framing and complications.
At a turtling-type table I imagine it would be considered bad play.
Being a GM in D&D isn't really that hard. It's been made hard by the zeitgeist and certain cultural assumptions and playstyles that have gained traction both historically and recently. Much of this is the idea that the GM has to have some significantly worked on setting bible -- either borrowed from published or made themselves -- and have all of the responsibility to present this accurately to the players. Further, more "modern" developments (stemming from 4e and heavily embraced in 5e published material) is the idea that the GM should be presenting a plot (the alternative to this is an even more detailed "sandbox" of smaller plots and places) and then managing that plot to make sure it shines. These put a huge amount of extra work on the GM, but also create perverse incentives for the GM to feel they are more important because they've done this body of work and that they need the extra authority to make sure this body of work gets experienced.
I will query your timelines here - I saw the sorts of expectations you describe here, about both setting and plot, back in the 90s, and they seemed well-entrenched then.
I agree that it creates unnecessary hurdles to GMing and gameplay more generally.
"Trust the GM" means don't ask questions. This is the fundamental point -- you need to "trust" that whatever the GM is doing it for your own good as a player. This totally removes good questions about play. Let me give an example
When I was GMing 4e, I remember a player calling me on an aspect of framing that contradicted the players' success at a skill challenge in the previous session. It was a fair call.
If a GM is doing their best to frame scenes and narrate consequences, I'll let that go even if I can see flaws in what they're doing that (to my eye) seem like they could easily be remedied. I see that as an issue of
manners, rather than
trust.
Trust implies as-yet unrevealed consequences or competencies that can be relied upon. In the case of GMing, most of the time what you see is what you get, and so trust isn't a salient concept.
I mean, if the GM sets up a puzzle I guess I need to trust them not to have cocked up the solution, but even then there can be plenty of room for real-time back-and-forth between humans. (Eg one time I wrote a series of puzzles and codes for my daughter's birthday party. I did it late the evening before, and mucked up one of my Caesar shifts. So any trust the kids had in me was misplaced - but they worked it out anyway, noted the errors, and just made snide remarks about the incompetent dad!)