I would like to advocate the use of tagging in some detail, as I have the impression that the power of tags is still massively underestimated by some of you.
Of course, in the end it always depends on what you intend to do with your notes - thus, every use case can only be an individual and subjective solution.
Hence, we first should clarify that tags can serve various purposes:
- They can establish categories. You can do this alternatively by sorting your notes into different subfolders organised by topics, sure. But tags are much more flexible as each note can thus be grouped under several topics without being copied into several folders. Another workaround is to establish MOCs, which is a nice idea, of course, because it helps you to gradually structure a topic you are working on. I am using MOCs (or whatever you may call it) a lot - but be aware that in this case you have to pigeon-hole your note immediately after writing it, otherwise it may become lost temporarily until you are explicitly searching for it. Tagging is much faster, especially when you just want to record a spontaneous thought and be sure it won’t become lost in the future.
- They can give you a quick overview of your note’s content. If you choose your tags in a well-directed way, you can immediately get an idea about what this specific note might be about, without reading it first.
- They can structure your work. This was already mentioned by several people here. Assigning the tag “elaborate”, for example, let’s you quickly find your notes that have to be further developed or supplemented.
These three purpuses are quite commonly used by most of us, I think. Yet, the real power of tags, in my opinion, can be found in a fourth argument:
- Tags let you discover new connections in an unintended and unforeseen way:
This is one of the major strengths of Luhmann’s approach to the zettelkasten. The above-mentioned use cases certainly facilitate your workflow a lot and relieve your brain - they are utilities that help you to make your work faster and more effective. But they all don’t change the way you are thinking about a topic from the very beginning - means they all support you in elaborating a structure of your content that is already in your mind when you start working on a topic but they don’t encourage you to consider a topic from a completely different angle.
I started realising the real power of tags while working for several years with Daniel Lüdecke’s zkn3. This is - for now - the programme that possibly best translates Luhmann’s method of working and connecting into the digital age. To be sure, tagging in an intelligent and proper way is a process of learning over time (tags should neither be too specific nor too general, not too many and not too few, etc.). But after some time your zettelkasten comes to life: it somehow develops its own personality and it starts to “speak to you” (that’s what Luhmann called it several times, if I remember right - he regarded his zettelkasten less as a simple tool for knowledge management, but rather as a good friend he could discuss his ideas with). It becomes your “second brain”, but one that can see much more connections than you ever can see by yourself (as you’re still limited by your routinised ways of thinking and structuring information into topics).
Let me give a brief example. I have, for instance, a note that originates from an article excerpt that defines conservatism as the “desire for structure”. Having tagged this note, amongst others, with “structure”, it suddenly shows you in another pane a link to a note that differentiates between two major forms of “learning”: a) Learning as adaptation in order to deal effectively with given structures as well as expectations and requirements from my environment; b) Learning as emancipation from existing structures that helps you overcome your environment’s expectations and self-confidently pursue your individual goals and dreams. This note, then, takes me both to several thoughts about how learning is organised in schools and to a note about the existentialist understanding of freedom. If I further follow the ideas my zettelkasten is sharing with me, I discover thoughts about such different topics as social organisation in the digital age, the organisation of work in global capitalism, the historically specific nature of desires, the process of identity formation in school and some thoughts on strategies of procrastination (!). In the end, I arrive at a strongly extended understanding of the different causes and mechanisms of conservatism I would have never discovered on my own. I suddenly see a bulk of unexpected and often surprising connections I couldn’t have thought about before because my way of thinking is significantly pre-structured and pre-defined by what I have learned about how things could (or should) be seen. Trying to find out something about conservatism, I probably would have searched for “conservatism”, maybe for “politics” or “parties” etc. - but not for “desire”, “school”, “procrastination”, “digitalisation” or “existentialism”. But the zettelkasten has triggered my thoughts in a completely unexpected way, so that I finally end up with a couple of new notes about new insights and connections. After some time, I could have written an article about the meaning of ‘conservatism’ that doesn’t just replicate and sum up anymore what others have said about this topic before but develops a completely new approach to this topic. (And no - I haven’t written an article about conservatism yet; I just took, for means of illustration, the first note my zettelkasten opened at random and followed its suggestions… 
)
Again, it all depends on what your main purpose is, of course. If you just want to use obsidian (or any other app) as a kind of wiki or structured repository of your knowledge, then tags might have a very different meaning for you, rather helping you to quickly find information you are looking for. But, if your main goal is to open up your mind, explore new ways of thinking and discover surprising and unintended connections between topics, then the simple structuring use of tags rather restrains than advances your work - just because all the links, hierarchies and categories you establish have always already been somewhere in your mind before and it is extremely difficult for all of us to overcome our habitualised ways of thinking.
The above thoughts basically form the background of a feature proposition I made here (cf. point 2.) - but maybe I should open a separate feature request for this, if only in order to see if others would be interested in such a feature as well.
Edit: here is my feature request