Largely agreed - says he, whose first six DMed adventures back in the day were all homebrew.

(certainly a trial-by-fire on how not to design adventures!)
Well there's something to be said about necessity. Some kids learn to swim by falling in the pool.
If I did it this way I'd need to know who my players were going to be, and they'd need to know their PCs, weeks or even months before campaign start.
Do you not typically know who's gonna play? I know that's common with public play and with online games, but I always got the impression from your posts that you have a dedicated group. Is that not the case?
As for weeks or months....I don't think I'd agree you
need that much time. You might prefer it, sure, but I can't see how it's needed.
And given as my games tend to be rather lethal at low levels...
That can be an obstacle, I suppose. I always found that killing a first or second level character didn't pack that much of a punch since the player isn't as invested in them as they are after a couple more levels, at least.
However, we had a campaign we played where this was not an issue because we collectively created a roster of NPCs and possible PCs to populate the setting. We had 4 players, and once they selected their PCs, the others became additional NPCs and/or cohorts that helped the party a bit as needed. If a character died, there were still several replacements already at hand, and they had an actual place in the ongoing story already. It worked out really well.
If the new GM is also running new players, you're absolutely right.
But if the players are experienced you risk hitting a canon lawyer, which no GM needs be they new or otherwise.
That's more a case of annoying behavior than anything else, and can hopefully be curbed right at the start with something like "this is not THE Marvel Universe, this is OUR Marvel Universe" and then just periodically reminding everyone of that as you play.
Again agreed; but only to a point.
The pantheons at least need to be fully designed before roll-up so players know what their options are should one or more want to roll up a Cleric. Much of the rest can, as you say, be done as required; though I'd advise even a new GM to try and stay a step or two ahead of things if possible.
Again, I don't see this as a need so much as a preference.
Yes, having a pantheon gives players some options up front. But, not having it done up front may allow the player to add to the world by crafting their own deity for their cleric PC. This goes to both the point about allowing the players to craft world elements to make the more invested in play, and also on reducing some of the workload of the GM.
Yeah, a lot of these things have traditionally all been decided ahead of play. That's how most of us learned.....we looked at a list in the book, and we picked something that sounded cool, or that fit our character.
But that's not how it must be done.