now that there's a working bot it's time to make a way to improve the quality of the bot with re-enforced learning.
the question that comes up in my mind: how do we tell the bot a post was "good"?