@jordan-violet Appreciate the feedback! 
I’ve recategorized this as ux and added the onebox tag so it comes up as a data point as we explore improving the onebox UX in future.
Having the media content’s text description still present in the excerpt of the onebox is intentional and signals that there’s some additional media to be seen within that excerpt when it is expanded.
This aligns with the behaviour of oneboxing of external links (except that we do render those as thumbnails), but there’s some styling & performance considerations that we’ve had to trade off against when approaching whether to render it fully in local oneboxes’ excerpt; hence the use of the media content’s text description instead.