The troubles with low prep:
a) It's impossible to be unbiased or fair when you are continually reacting to the meta.
b) It's very difficult to have Mystery Boxes with anything surprising in them if you don't plan ahead. If you don't know you needed clues until after you need them, it's hard to put them there. You'd have to work really hard to avoid Bad Robot style DMing, and as any writer knows Bad Robot is a big part of why the golden age of media collapsed.
c) It's very difficult to have anything large and coherent. Most of the time, you aren't ad libbing a castle sized dungeon (unless you've spent a lot of prep time studying castles and have several floor plans basically memorized). Good luck ad libbing something like Castle Ravenloft as original content. Most ad libbers I've met have no locations. You are on a stage and they change the drapes but there is rarely a lot to interact with or explore.
d) You are heavily reliant on your first instincts. You rarely have a chance to think about things and come up with better plans.
e) You are often heavily reliant on the players for content and ideas. But then you are kind of abandoning your fundamental duty to be a secret keeper if you don't make up some secrets. The last thing I want to do is set down at a table to play as a player and be hashing out what is actually going on in the setting like a collaborator on a script revision team, trying to come up with a script for an episode with a deadline.
A) are you focusing on the story , or what you are trying to do to players?
B) That is what out of game time is for. You gather your thoughts, think about what happens. Surprises come in many forms. For ex. Last week my players were in a necro lair, finding a copy of Necro weekly, a self updating magazine developed for the necromancer community. In this dungeon they came across a new type of undead which was extremely dangerous. After fighting one and making it out alive, my player asked, these are brand new yes? Are they featured in the magazine? I laughed and thought, that was a great idea and officially they were literally created for that weeks campaign...So I gave it to him, a full spread on this new type of undead, its strengths and weaknesses and the entire group laughed and thought it was the best.
C) This is false. Each year I sit my players down for a saga, go over the context of what is to be roughly expected, what is at stake, and try to get people to plan as a team. After which, I go for 1-2 years on a given campaign and they follow it through. I am just now getting the height of a campaign I planned over 20 years ago. Again, if you attach your GMing to the story and not what to do to the players you may find Ab libbing easier.
D) Again this is done outside of gaming where you think about the relationships of both the characters and the baddies. Some of us have a library of characters, motivations and backstories that can be called upon or created at a moments notice. In these cases those BGC's ( Background characters) are just as real as the players. Naturally there is a limit. If you want to get into the life and death of Jackwise the Smith, I am either going on a tangent or he is going to tell you ta mind yer own business!
E) I mean do secrets make the game? There is knowledge the characters have, and that which they do not. Again, I would base that on more real life to be played out. Revealing secrets often leads to story development but being secretive for secretive sake would be lost on me. That said having the characters in game, develop plans with the information they have is legit fine with me. But, keeping in mind that I run through a story, the motivations of the BGC's most likely are decided, or have been decided based on my story, making it flexible.
Lastly, I can see this breaks down to style of game play and preference, so in effect neither way is wrong. I have listed counters to your points not as an adversarial approach, rather to demonstrate a different approach. I am sure both work.
