Is it not possible? I'm not that familiar with the topic. Doing some sort of averaging over a large corpus of separate texts could be interesting and probably would also have a lot of applications. Let's say that you are gathering feedback from a large group of people and want to summarize it in an anonymized way. I imagine you'd need embeddings with a somewhat large dimensionality though?
Cool stuff. This will open up completely new ways for small / even one-man companies to build stuff with convincing graphics at least for MVPs. I imagine some lone coder with skills in gameplay mechanics, but none in graphics or a very small budget to spend on it, will get a lot of value from something like this.