As a result of thinking about my ongoing disagreement with Rick about what counts as PCT Science, I’m thinking that there are aleady too many categories, and only two or three about PCT itself should exist: Science, Real-World Applications, and Relationships to Other Work (I think here about such things as “Sociology”, “Deep Learning”, Friston’s “Predictive Coding”, “Reinforcement Learning”, and so forth). More specialized areas, such as those just mentioned, could be topics rather than top-level Categories. I think of Language, which actually belongs in all three of my suggested categories. Sociology does, too, if you think of it as incorporating Collective Control, Politics, Culture, etc.
Posters (or commentators) probably would need to use a lot of tags to allow coherent search of what otherwise is likely to become a rather tangled web of Categories and Topics.
Maybe it is too late to change, but I would think that any new category should at lease be subject to a vote (a “poll”, I believe, in Discourse) before being implemented.