I wonder, how much of the discussions on the results of agentic coding is just LLM slop.