In the discussion about how to get more people watching more videos, and also the management of search, I'm surprised that nobody's mentioned tagging so far. (Actually, so surprised that the first two drafts of this post were a lot angrier than they really needed to be.) The infrastructure is in place on the current version on the .org to support this, and it definitely falls into the "well, as long as the hood's up" category of the other improvements mentioned by the dev team.
Right now, there is a halfassed, partial implementation of tagging. When you open up a video, the anime sources used are linked to a supersearch for that anime, and the creator's name(s) links back to their profile(s). The musical artists used for some reason don't have the same functionality that the anime sources do, and there's no provision for user tagging. If I watch a video using, say, Bela Fleck, I can't immediately check for any other Bela Fleck videos right from the video page, let alone follow a 'newgrass' tag to discover other videos using a similar style of music (with artists that I may not be aware of going in), or an 'awsmcore' tag to see what the hell that even means. I can see that a video participated in a given contest, but I can't click on the contest name and see the other videos that participated.
Users already have the ability to associate short strings with a video's db entry. If, when they're watching a preview or doing up a star rating, they have the opportunity to add or uprate up to five tags on the video, this adds trivially to the db footprint of the current 'quick comment' structure. Supporting tags will require additional coding to get the tags out as needed, keep the weights updated, and do the query behind clicking on the tag, but absent commentary from someone who is able to actually look at the site code, this shouldn't add too much burden in terms of computation over the queries that are currently run on the db.
Will people add stupid, useless, and insulting tags? Sure. But those that are less useful will get pushed to the bottom of the heap. If it helps sort through the thick middle of the database (approximately 86,000 videos score between 3.95 and 2.95 stars), make connections between videos that are related neither by artist, nor anime, nor creator, and gives people a way to add useful content to the site without being an editor or a programmer, it's worth at least looking into.