I will 100% be taking some of your advice here, thanks! I just started a research project and haven't properly thought through what I want to accomplish. In my head, the scope is absolutely massive and I'm definitely going to burn out if I don't narrow it or at least set some intermediate goals. I'll start with a project proposal.

1. I'm treating the number of subscribers on 80,000 hours as the total pool of people at least somewhat interested in EA. I'm planning to do work on a task backlog because there's such a large difference between the total pool and those that are actively working on EA-adjacent projects. It seems like having a good onboarding process that could include some volunteer work for those that are just getting started would help drive more engagement down the line. 
2. I think it would probably start out as meta-focused, but I could imagine orgs from specific cause areas could post one-off projects to the backlog as well. That would be a good way for someone interested in AI safety, that isn't currently able to directly work on it with their career, to contribute.