I think Jukka's idea in the comments is great, and I expect Joe and I could figure out a good way to feature an unanswered question in September.
Regarding a "polytcs" project, suitable polymath problems are "parallelizable" in the sense that there are a lot of little pieces that can be handed out as modules to people if different skill levels. Like different people checking different special cases of a conjecture. I just had an issue with my phone when I tried to embed a link, so I will be lazy and say if you search "characteristics of a good polymath project" you will find quite a bit.
So I think someone interested would have to design a plan of attack on the question, and convince others the question is amenable to crowd sourcing. Also, especially for the first such question I think it is important that we guarantee success, by choosing something that will produce original research even if the complete answer remains out of reach.