On Friday morning, Elon Musk declared, “We have improved @Grok significantly. You should notice a difference when you ask Grok questions.” While Musk didn’t say exactly what improvements to look for, he’d previously declared that xAI (which built Grok) would retrain the chatbot after it had been trained on “far too much garbage,” and he called on users at X (where Grok is heavily featured) to share “divisive facts” that are “politically incorrect, but nonetheless factually true.” (Musk recently merged the two companies.) One user subsequently asked Grok whether electing more Democrats would be a bad thing, to which it...