GPT-4chan: This is the worst AI ever
In brief
Yannic Kilcher explains how he trained a language model on three years of 4chan's /pol/ board and let it post on the site. He evaluates how good the model was and discusses the ethics and fallout.