It's easy to explain. All answers that you put in the first category are people answering the wrong question, that nobody actually here asked. They are answering the question "what was the thing that lots of other people liked, that
I also liked", as if their opinion is somehow special here, on a semi-anonymous forum, where visitors don't recognize most other visitors. And since it isn't, their opinion is just one of the data-points in the "people, that liked product X" dataset. And people, who upvoted their posts also upvote something they like. And since things, that are liked and endorsed by a lot of people, but that are not worth the attention (in your opinion) you call "overhyped", it is pretty much by definition that the top answers to
that question (which nobody asked) would be a collection of overhyped things.
The answers from the second category are answers to the question that OP actually asked: what are the things that were hyped prior to their release (or shortly after) and didn't disappoint. Unlike the first question, this can be answered somewhat objectively, because it's not about the opinion of some semi-anonymous HN user anymore, it's about things that are hyped before the release and having impact after. You personally may not like the original iPhone, but you'll have to admit that the early hype didn't really fade — maybe it's a shitty product in your opinion, but it did conquer the market. And even if people are answering/upvoting that based on subjective feelings, in the end it's still turns out to be a somewhat fair assessment of the topic: because a product was hyped before the release a long time ago (well, not M1) and enough people still can vouch that it "lived up to the hype". That's impact.
Exceptions here are Apple Watch and Ruby on Rails, because they received some, uh, prenatal hype, but let's be honest: Apple Watch doesn't have nearly the amount of adoption iPhone had, and while Ruby on Rails was quite a significant thing at some point of time, it isn't anymore, other frameworks and languages filled the domain and now it's just "one of them" at best.