We seem to be crawling out of the "uncanny valley" on the other side. Give it a couple more years, we've pretty much nailed the interpolation of human output (across conventional modalities).
I don't think this type of method can scale indefinitely, it's essentially just "better" sampling within dense areas of knowledge space. It cannot help with better exploration outside these dense areas, because these explorations won't have a consensus among agents almost by definition.
He described math more poetically and profoundly: "the art of giving the same name to different things"