My moderation point is that with a video service you are forced to "watch" the content in the video in order to properly moderate (though you can just block people of course). You could have a bunch of filters like YouTube does to determine if your video should have ads and whatnot or you can rely on the community (or both).

The main issue with it is that we want to prevent "bad content" that being very poor quality content to skip being all detailed. To do that kind of filtering really requires some form of community review of the content as it's infeasible to have it all manually reviewed. If you have a community review process you open the door to mass reporting and the like so you cannot simply automatically remove content if it gets a lot of reports, it must be manually reviewed (by watching the content) to ensure it's fair to remove it. Lemmy, at least in my usage doesn't have this desired "bad quality" filter outside of up votes/down votes which notably don't remove the content (and so doesn't remove the immense storage requirement)