@janellecshane I quite enjoy how it seems to maintain a few nouns in context from step-to-step (not sure if that’s coincidence or if the model is designed that way).
It would be awesome though if it could classify some of words in the text you’re selecting and structure it into an actual game mechanic with health and attacks and stuff. The generated text often lends itself really well to that.
@nbering @janellecshane It's plausible our choices have consequences, but the consequences cannot even be percieved because the choices cannot be understood.
@nbering I would definitely play that, though it sounds really hard to set up!