There's at least two fronts, having good enough software to be compelling, and having the ops situation be radically easier. Honestly I think Social Web (a general idea not specific tech) is kind of in a doldrums, has some interesting projects, and a lot of enthusiasm, but there's a lack of key innovators. I'd probably get outright spat on by some but I truly think Kubernetes is the first diy-able enough platform to be interesting for ops, and I do think we're not far off from having some fairly good out-of-box autonomic systems, & that would change the game. Alas multi-tenancy is still extremely early days for nearly everyone, & crude, needs a lot of work, but that is kind of optional, but in the end is a key enabler for interesting home computing. Reciprocally Kubernetes really requires multi-cluster capabilities, clusters as cattle, & right now that's a cowpath we need paving.