1) I guess I see where he's coming from. However, at an actual table, it does seem kinda lame to be playing a class that doesn't have 'cool stuff' yet while somebody else is already getting it. Overall, I think I support his point of view, but it does seem to go against what I've come to expect from 5E.
2) So different classes will have different break points... I understand why, but I find myself curious how this will interact with multiclassing.
These are very good points, very "practical" questions!
Truth is,
narrative goes a long way, we can come up with nice narrative explanations for both domains at 1st or 3rd level, but why not 5th, 10th or 20th? I am sure we can find narrative excuses for each.
But the "feel" of the game is important, and they should not underestimate that subclass is a
very important choice for a lot of players, much more important than feats or proficiencies, and even more important than race (at least because race has a smaller impact on your abilities on the long term). Not all players will be bothered by subclasses occurring earlier or later for different classes, but part of the gamebase will, and this delivers a sense of unfairness, which is detrimental to the game "feel".
My opinion is still that they got stuck with a certain mechanical implementation of domains, and now they need too much work to find another (the game doesn't really need domains to grant bonus spells, that's just the idea of previous editions, but domains could have been used to grant
anything).
Plus, they probably thought that 3rd level is when spellcasters get 2nd-level spells, thus having both this and subclass might be too much of a boost: this is why IMHO they anticipated subclass to 2nd level for Wizards, Druids and Rangers.