Recently ranked beatmaps are largely still good. But mapping is stagnating due to the modding environment. Cue rant about the bn test system:
There's a dearth of modders who know how to improve a mapper's concepts and visions for that specific map, a dearth of modders with differing opinions on flows, hitsounding, visual appeal, etc... and an abundance of modders who know how to improve the map according to the current meta's standard of quality. This results in modders pressuring mappers to "make high quality stuff" which mappers translate as "map using meta-driven concepts and patterns that are high quality".
Ranking maps isn't easy. The tier system has made it more difficult than before. New mappers and even some intermediate mappers don't bother experimenting or making more unique maps because the motivation for creating a "high quality meta-following" map leads to much easier qualifications. With new mappers especially, think of the logic this way: "Why map something experimental or innovative, and potentially have to completely remap, or never see the light of day, or map using the principles of the 2017 meta-mapping and produce something that can potentially become ranked?" Remember that especially for new mappers and intermediate mappers, even trying to map in the current meta-mapping isn't easy. "If I map like everyone else who gets their stuff ranked, I'll get my stuff ranked. And if I don't, I can easily find modders who can tell me how to map like everyone else so I can eventually get my map ranked." <--- thought process for many new mappers.
Modding is how a map goes from pending to ranked. So for me, modding is the root of the problem. The BN Tests were partly responsible for creating this because they only awarded people who modded a certain way and pointed out certain subjective issues. Modders quickly learned to mod this way in an effort to become BN's. We see a lot more focus on emphasis control, spacing, producing "good" flows, etc... Don't get me wrong though, stuff like this is worth pointing out in mods. I don't think the BN test failed in this aspect at all. Teaching modders to look for these things is good. The problem is the test only promoted these kinds of modders to BN, and removed BN's who didn't consider these aspects as thoroughly, even if they specialized in other areas or held different visions on flow choice / aesthetics etc...
Modders who specialized in hitsounding often were left out, as well as modders with different opinions on flow. There is no "good" or "bad" flow, only good and bad usage of movements (and even then, poor movement can complement the song too if the song is jarring). People who modded maps based on the mapper, ie "you usually do X in your maps, this is a good opportunity to do something like that instead of the usual jump spam" or people who try to mod with the mapper's vision in mind also suffered. When a map is purposefully created to be bad, you don't get many opportunities to "improve" the map as opposed to "fixing" it, and some modders are exceptionally good at improving good maps to become even better, but rarely deal with maps made by new mappers. I could go on about how the bn test system ended up stifling different mapping genres, and regions (knowing English for example, is essential to doing well in the Tests, and some modders who knew very little English suffered as a result). But you get the point now. A lack of modding diversity, and a lack of diversity in what BN's consider "high quality" is causing mapping to streamline.
As a result, we have a new generation of modders who define high quality as things that check the boxes of emphasis/spacing/flow that they received points in when taking the BN Test.
Mappers are mapping the 2017-meta style because it's currently the best definition of "high quality" whereas we previously had different and equally respected definitions, resulting in more diverse map pools.