It comes back to my suspicion. A lot of homebrew is a few cursory notes and then winging it. Monster stats, NPC spells, room descriptions etc. I don’t want that from a DM and I won’t do it as a DM.
I have pretty high standards as a player and a DM regarding how a game is supposed to play. Every game involves improvisation and being good at improv is an important GMing skill. But if I find a GM is winging it I'm likely to politely excuse myself from the game and not ever return. When I'm doing improv, I'm doing because I'm reduced to a few cursory notes about the campaign setting. Most of the time, I have more than cursory notes for my homebrew adventures.
I have never met a GM that is as good as improv as they think that they are. There is a noticeable difference in the richness of a game when it is well prepared versus when it is improvised. Maps are more complicated. Scenes have more detail. Lore is richer and more important and more internally consistent. The story twists are more literary and better foreshadowed. Investigations yield more useful and logical clues. Even when you jump the rails, the game universe stays more grounded because the GM has prepared to improvise by establishing truths about the demographics, the setting, and what the resources the factions have and what they want.
For a while there in TV shows about 20 years ago the thing everyone was doing was improvising the plot of shows by introducing these "mystery boxes" where even the writer didn't have a clue what was going on. This technique was probably popularized by the X-Files where the shows generally disconnected episodic stories were implied to be part of some larger mysterious connected narrative through the dropping of scenes into the epilogue of the episode where mysterious figures said mysterious things. It was always implied that "The Truth was Out There". Of course, with no coherent idea what the truth was, that truth was never satisfyingly revealed because well, it never existed in the first place. The peak of this was the "Bad Robot" school of storytelling, formalized as an approach by J.J. Abrams, and seen in shows like Alias and Lost, and shows attempting to compete with viewership with those shows like The 4400 and Heroes.
What's notable about those shows is that they start well and then flop because the payoff they promise isn't there. They can't actually deliver on their promises using an entirely improved story. It's worth noting how good the first arc of Heroes is because it actually can deliver on "Save the Cheerleader; save the world" to a decent degree, and maybe it's unfair to Heroes because maybe the original writers knew the story before the writer's strike derailed it. But it's worth noting just how bad those shows end up being when you don't have that coherent story from the beginning, compared to things like Babylon 5 or the Harry Potter series where the author had worked out ahead of time what the story was and were improvising within that framework. In fact JMS often talks about how he'd plotted forks in the story for every character knowing that if he tried to film a coherent story for 4-5 years real life would cause actors to not be available for the whole period. Or you hear about Tolkien writing LotR like he was a GM, making maps and moving markers on them and making timelines with dates so that he can coherently track where everyone is at any point in the story.
And that sets my standard for an RPG. You can't get with improv alone where I want RPGs to get. Improv is necessary and GMs that are good at it are good at it because they are so deeply familiar with the material and so experienced telling stories in the setting. But you'll never get there with improv alone.
One of the reasons I so like Seth Skorkowsky as a reviewer is that I recognize in him a fellow GM that is putting in the work to run good games. He takes a 32 page module and says, "Ok, what do I need to run this that didn't make the page count", and then he makes it and tells you why he made it and why it improved the game from a working GMs perspective of running games that feel immersive and will tend to be exciting. That is the way.