Man, that's a great post and all, but every single one of your points has been discussed to death. Most of them in this thread.
The only one that I think is at all contentious right now is saying that early game hull sizes should become obsolete. I disagree and I think many others will too. You want each hull type to have its own place in the late game so that you're not forced to demolish half your fleet once it becomes time to upgrade. On top of that, straight upgrade from corvette through to battleship means a very linear and boring progression, with the added problem that as soon as someone researches destroyers they are ready and set to conquer everyone around them. It'd make Stellaris into just another Civ game where tech is king and you just want to rush for the next largest ship as soon as your current fleet becomes par for the course.
Good ideas don't self-propagate; they must be repeated in order to take hold.
Regarding your post, the reality is that under the current system, the hulls exist for the sake of having hulls. This is true of most of Stellaris' game mechanics near as I can tell -- someone went down a check list and added stuff they thought a space strategy game should have. None of the hulls have a unique place as things stand, and there will always be a strictly-superior hull choice. W
Regarding the current ship hull system, it has three problems as implemented:
(1) It encourages unit spam. Unit spam both severely degrades late-game performance and makes ship losses meaningless.
(2) It offers no meaningful choices. Hulls are HP pools for mounting weapons. There will always be some optimum hull which is most cost-effective for mounting weapons. Right now, corvettes are ideal because the cost of upgraded hulls doesn't justify the additional cost, missiles can't effectively After the next patch, I suspect we'll see a battleship meta or something similar.
(3) Unlocking new hulls feels pointless. When you unlock a new unit or hull in EU4, or Sword of the Stars, or other well-designed games, it feels like an actual upgrade. In Stellaris, it feels like one more pointless variable in a game filled with pointless variables. Tech research in a game should never feel pointless, like it's being done for the sake of itself. This doesn't mean newly-researched hulls should be overpowered, but there should be some benefit to being ahead in tech beyond merely the fact of being ahead in tech.
None of these issues are likely to be fixed any time soon. Paradox is already trying for over a year to balance six different types of weapons (missiles, fighters, lasers, railguns, artillery, disruptors) and three different types of defenses (shields, armor, evasion), without much apparent success. They're not going to balance five hull sizes on top of everything else.
My proposal is as follows:
1. Each larger hull is strictly superior to its smaller predecessor when in a 'generic' configuration.
2. Each hull, aside from corvettes, has a unique configuration that lets it do something special. Torpedoes for destroyers, XL weapons for cruisers, and other such.
This system will probably offer players more meaningful choices, because it will allow Paradox to balance around weapon systems and defenses, without also having to factor in hull sizes, and because hulls will, by the end game offer players clearly-defined choices rather than the current mess of assorted hulls that don't seem to have any clear role at all.
As far as hull upgrade cost, most strategy games don't have trouble requiring players to either pay for upgrading to new hulls (EU4, Civ), or requiring the player to demolish the old hulls and buy new ones (old EU4, Civ). The advantage of such a system is that it allows players who have fallen behind in fleet size to catch up, and reduces snowballing. Worst case, provide the player with a 'reprocess' button that returns 50% of the credits consumed. Admittedly, we currently must click through three screens to build ships, but this is shoddy UI design that should be solved by better UI.