On-screen text and the caption box do different jobs
On-screen captions are for sound-off viewing and for the line you need them to read. The description or caption box is for the extra context, the keyword a search might use, and a question you actually want answered.
Write the way you talk. One idea. If you need a paragraph, the video was probably unclear. Hashtags are a seasoning, not the meal. Two or three specific ones beat a block of twenty.