Write It Yourself (jnsgr.uk)

A tech writer reflects on their evolving use of large language models (LLMs) for drafting and reviewing prose, emphasizing a burgeoning dissatisfaction with the quality of LLM-generated content. They describe a specific workflow for creating internal newsletters, using LLMs to generate headings and bullet points while maintaining authorship over the prose. Despite the improvement in LLMs' ability to edit and suggest ideas, the writer struggles with the lack of personal voice and originality in the output, leading to feedback that their writing has lost its distinctiveness. This introspection highlights a crucial tension within the AI/ML community as professionals grapple with integrating LLMs into creative processes. While LLMs can offer valuable insights and streamline tasks, the writer argues that genuine expression and articulate communication remain essential skills, particularly as human-AI collaboration becomes more prevalent. The emphasis on maintaining personal voice in writing underscores the importance of balancing AI assistance with individual creativity, suggesting that reliance on LLMs for content generation could undermine the very skills that define effective communication in a tech-centric future.