OpenAI Content Moderation Is Really, Really Bad

Author claims that these jokes are "clearly" inappropriate and violate policies but I'm not so sure. The main example is off colour, but does anyone really think it constitutes CASM?