I have a hard time believing GPT4 would come out with anything that bad, which really makes me wonder which model they used...<p>Edit: answer is in the article, kind of...:<p>"In this case, the content was generated through a combination of algorithmic techniques with human review, not a large language model or AI system."