I sincerely wonder what specific technical changes were made to Grok to cause it to start writing like this. Were a batch of new base prompts added? If so, what were they? https://newrepublic.com/post/197719/elon-musk-ai-chatbot-antisemitic-rhetoric