I think there are some tradeoffs for database vs plain text storage of community contributions. If we go with plain text like it is now people get to learn how to make pull requests on github but with a database it would be simple to make a form that makes it more accessible to the community as a whole.
The database model would require a moderation panel that is password protected adding to the scope of the project.
Adding a way to indicate the skill level the a video is directed at would be a great feature. Also being able to filter by that.
Frequently people seem to be annoyed by the search capabilities of YouTube but we don't have a great way to search the website. It would be nice to add this feature even if it's just a box that directs people to a google search of the site.
It would be awesome to have an integrated web editor window on video pages and the code is updated based on the video playing. My guess is this would require 2 forks of the web editor. One that Dan uses while recording a video that captures the current state of the code every few seconds or so. And another fork that displays these captures in an on page. It would require capturing time stamps that sync with the video and a way to sync those with the final edit.
Got this idea from khan academy: https://www.khanacademy.org/computing/computer-programming/programming/drawing-basics/pt/making-drawings-with-code. We could possibly reach out to them and see how they're doing it.
Also OpenProcessing has a tutorial mode that is interesting: https://openprocessing.org/sketch/672194.