My understanding is that (at least on paper) there's a minor feedback loop where getting a lot of views on a video = video is recommended to more people = more engagement = more recommendations. And conversely, if too many people scroll past a video without watching it, that video gets de-recommended in the same fashion.
So now we throw text posts into the mix. Even if a lot of these text posts have music attached, I'd imagine people on TikTok are way more likely to stop scrolling and watch a video than stop scrolling and read a 1000-character post; unless TikTok's content recommendations specifically account for this discrepancy, I could see text posts getting drowned out pretty fast.