Tags are like pre-computed search terms that aim to figure out what podcast episodes are about. Aside from a few synonyms there isn't any magic going on as you can probably see from the results below.
However, there is a lot of room for improvement and there are a couple of changes on the horizon:
1. The actual tags aren't so great right now, for example an astute observer would notice the tags "ai" and "artificial intelligence" are both in the list. Worse, "ai" has a suspiciously high count because there are some low value matches being returned from other words that contain the letters "ai". This is easy to fix as I have both a white list and black list of terms that I can use to tweak the tags.
2. Tags really ought to have a hierarchy. For example, if an episode is about React, then it is also about JavaScript. Same with Kubernetes and DevOps.
3. There are a lot of "none" results. This is a combination of weak episode descriptions and topics that are hard to classify by search. Examples include things like "soft skills" or "career management". I think we'll need a bit of a human touch to handle these items.
Note: Because of some caching there may be a small discrepancy between the number you see below, and the count you see after clicking. The number over on QIT is closer to real-time.