The exceptional few who become really good in less than a year or eighteen months are all younger (below 40).
Few people here
start tango at over 40. I started at 45 and I was the "old man" of not just my class, but of many classes. There were a few people older than me in a whole week of multiple classes per day across six levels. I know this because I was an accompanier who practically lived at the studio. So, my comment here is, "Yes, we had some people who improved very quickly, but they were about the same age as the ones who did not."
Like I said before having danced in many other cities, close embrace is something that I call universal.
I am not doubting your experiences! However, I have several teacher friends who spend 100+ days of the year touring the world for dance, and they've come to call the
pseudo-close embrace (i.e., torso-to-torso contact only when convenient, separate axes, no weight sharing, turns are big orbits of follower around leader) the "international embrace." Genuine close embrace – torso contact at all times, dancing on three axes (leader's, follower's, shared), real lean and weight sharing rather than its illusion, turns are mostly mutual (around the shared axis, not the leader's axis) and tight – doesn't seem to be as common.
I am not sure why the younger followers in your scene are the way you described. My guess is they are learning tango from instructors who may not be teaching them right things.
You will get no disagreement from me there! My coach says that about twice an hour. I've been in the studio at the ends of lessons – group and private – before milongas and heard teachers talking to their students. In essence, they're creating a generation of show dancers for the social floor, people who think that a beautiful follower is light, floaty, and rarely standing on two legs.
And they are dancing with leaders they don’t challenge them.
Definitely. Leaders are in such profound shortage here that they have zero incentive to improve. A lot of them do weird stuff like visibly recoil back on their rearmost leg before each step to generate "push" (so it's like backwards-FORWARD, backwards-FORWARD all the time, and never forward-forward-forward-forward with drive and sustained momentum) and advance their forward leg before their torso in a sort of under-scooping movement (so their torso almost pitches backward and their weight arrives very late). I'm not sure where they're learning that stuff, or if they're even learning it at all rather than deciding on their own that's how tango works.
These skills come with both learning and experience. For example crossing instead of pivot is something that instructors should expose the followers too. Without that many followers won’t figure it out on their own. I haven’t experienced younger followers exhibit the symptoms mentioned more than the older followers.
My experience is that the younger the follower, the more flexible. This makes it more possible to engage in the extreme dissociation necessary to get a pivot in a situation that would block a pivot and naturally result in a crossing step in someone older and less flexible. Which is a case of "learning the wrong thing," because as a leader, when I lead a turn so tight that I'm expecting a cross, getting a pivot messes with the timing and footwork. Yes, of course I can read what I actually get in my partner's body, but the guys who dance as described in the previous paragraph cannot, so the results are bizarre.
I think the critical mass of social dancers only arose after mid-nineties.
That depends on where you live. I know people from communities where milongas were packed even in the 1990s. They just weren't where I lived then.

That's talking only of tango. Salsa is
also social dance, and salsa was a big deal down in NYC in the 1970s, and made it up here by the 1980s at latest – though in its Cuban form, not Puerto Rican. When I arrived in Montréal in 1990, there was a noisy salsa place near where I lived, always packed.