I’m finding myself a little frustrated with how WK sometimes makes you learn the same combination of radicals multiple times, rather than teaching you the kanji (or hanzi or whatever) that those radicals make up and then using that kanji as a radical.
The most egregious example that I’ve noticed is that 識, 職, and 織 all have the same right half (戠), so I feel like it makes more sense to teach you that component as its own radical-- it’s “sound” and “drunkard”, so let’s call it, I dunno, “karaoke”-- and then teach 識, 職, and 織 as “say + karaoke”, “ear + karaoke”, and “thread + karaoke” respectively.
Off of the top of my head, other examples from among kanji that I’ve learned would be 要 and 腰 or 票 and 標. In both of these cases, rather than learning “moon + need” or “tree + ballot”, WaniKani has the right component kanji broken down into its “base” radicals (i.e. “moon + helicopter + woman” and “tree + helicopter + jackhammer”), which I feel like just creates extra memorization. Plus, WK already does this sometimes; from my last level, 障 is “building + chapter”, not “building + stand + early”. Why not standardize the approach?
Doing kanji-as-radicals consistently across the board is also helpful for the purposes of learning compound kanji as phonetic-semantic compositions-- to use one of the above examples, remembering the on’yomi for 標 is free if you know it as “tree + ballot”, but that association is harder to use if it’s broken down further into “tree + helicopter + woman”. I know WaniKani doesn’t directly teach phonetic compounds as such, which is fine, but it’s still a useful strategy that I think should be at least supported.