The author expresses frustration with the asymmetry of LLM use: while they appreciate AI for productivity tasks like code generation and content review, they resent receiving LLM-generated material from colleagues, which requires disproportionate effort to parse compared to the minimal effort needed to generate it. They question whether responding with AI-generated replies is a viable solution or merely a race to the bottom.