Jumping into this thread quite late and admittedly haven't read all 18 pages. Overall I agree with the spirit of it and that more RAW elements of 5E needs to be brought into BG3.

I agree with #2 and #5 (stricter 5e rules, day & night cycle). Ritual Casting, cover system, reactions, proper short-rests (hit dice) and item attunement. A lot of these little things add tactical elements to gameplay which will ultimately make for a better experience. Night Cycle wouldn't only improve immersion - it'll also completely open a new avenue of gameplay for more stealth based characters. I.e. maybe instead of fighting or lying your way into the Goblin Camp, you can now sneak through. Homebrewing isn't necessarily a bad thing - but only when it enriches the tactical choices, as opposed to simplifying it. I.e. I'm quite happy with the weapon-based special attacks, just because it does help address one of the biggest issues Martials have in 5E.

I support #3 and #4 (larger parties and random encounters), but I don't think they're critical. 5E table top is "balanced" around 4 players, so I'm not strictly against that, but I do like offering player choice. Other than nostalgia for IE games, not sure why "6 max" is the chosen holy number. Honestly if they are opening the party size, I would not mind if they allowed parties of 7-8 if wanted (as long as it's properly balanced - i.e. level slower, encounter adjustments, etc). Let me get the entire fellowship of the ring together please. In terms of random encounters, they can add to immersion but also can become a slog if not well implemented. I think taking inspiration from the Fallout games would be best - i.e. there's a large list of hand-crafted, story-driven "random" encounters, but also some generic ones. So it's almost a treat each time you run into one.

I don't agree with #1 - the requirement for having "proper" stats for 5E monsters. Now, that doesn't mean Larian should change everything and lose the core mechanics behind iconic monsters. I.e. a Red Dragon should still breath fire, etc. However, I think strict adherence is detrimental. 5E monster stats/CR are already not very well balanced for table top (usually on the "too weak" side), over-simplified to make them easy for a DM to use on table-top, and that's before we factor in effects of the videogame conversion (just some high level thoughts):
- The Monster Manual is balanced against 4 characters (unoptimized, no magical gear, and 6-8 encounters a day) - which is quite different from what you'd expect from a video game party
- Mobility and extreme range (i.e. 100 ft+) isn't translated well to video games, but can be a huge part of monster tactics and defense in table top
- Per above, resting limits is hard to enforce without making the game too linear (like Solasta), which means you should always expect more player resources
- A single person managing a party is simply going to be far more coordinated and tactical than a group of 4 people (especially the very casual, role-playing focused D&D core of today)
- Combat loops in games are much faster than table top. 1 fight can last an entire session, whereas in 1 gaming session you might do 5-6 times that. This to accelerated player knowledge and loot. The latter especially makes a huge difference
- People in videogames (especially single player) are far likely going to be power-gaming/optimizing. Characters can vary dramatically in power-level depending on optimization
- The reload button changes everything

As someone who has DM'ed and built encounters, I've found stat adjustments to toughen encounter design is usually more precise than adding numbers of enemy, because action economy snowballs quickly and is much tougher to gauge. I.e. if you increase enemy hp by 20%, you can expect a slightly tougher encounter, whereas adding a fifth body to an enemy party of 4 can much more easily snowball an encounter - because there's so many more layers of RNG (initiative, attack rolls, etc).

My overall stance is, change is okay, as long as it's actually thought out and not something dismissed due to laziness. And realistically a lot of the current missing systems (reactions, attunement, etc) does feel like that a bit, as opposed to being well thought-out, deliberate system change.