Everlasting LLMs

Author Bry Willis, writing under the pseudonym Ridley Park, details his frustrations when incorporating large language models like Claude and ChatGPT into his novel-writing workflow. He highlights how overly sensitive morality guardrails in these tools frequently trigger unnecessary lectures and rejections of harmless fictional content regarding sensitive themes. Beyond puritanical censorship, Willis notes that the AI assistants struggle with managing narrative timelines, often ruining suspense by accidentally leaking future plot points. Additionally, the models experience version control issues, repeatedly confusing deprecated manuscript drafts with current guiding documents despite receiving explicit instructions to ignore them. Ultimately, these functional quirks and unsolicited ethical lectures turn into time-consuming distractions that hinder the creative writing process rather than aiding it.👉 https://substack.com/@ridleyparkauthor/note/p-222100447