@timrichards
Indeed. It's a statistical model ... of word sequences scraped without permission from Reddit, Wikipedia, Xitter, and popular social media posts.
@plexus