Qiufan Chen, an award-winning science fiction author, discusses his creative process of integrating large language models into fiction writing with Amita Shukla. Chen explores how LLMs function across multiple stages of his writing practice—from initial research and structural development to drafting and editing—drawing on his experience as one of the earliest serious public experimenters with AI-assisted literary creation.
Topics
AI systems inherit the fundamental crookedness and contradictions inherent in human thought, making them unsuitable vehicles for creating perfectly rational or straight solutions. To responsibly develop and deploy AI, we must understand and work with this essential crookedness rather than pretend it can be engineered away, drawing philosophical insight from Isaiah Berlin's work on human limitations.
Venkatesh Rao
Sachin Benny launches a bio-thriller series set aboard the UET-1, the world's longest single rail line connecting Lisbon to Laos, following protagonist Rowan as she recruits unsuspecting friends for an invisible war while maintaining a meticulously optimized body through advanced cosmetic procedures. The narrative explores themes of bodily control, hidden agendas, and the blurred lines between self-care and weaponization in a speculative future.
Sachin Benny
Stanley Chen, a science fiction author who co-wrote AI 2041 with Google China's CEO, describes his hybrid writing process where he uses LLMs as essential tools but rewrites every line they generate, believing current models are insufficient as both writers and literary judges. The interview explores his practical experience with different LLM models for fiction writing, including their strengths with language-specific features like Chinese punctuation, and his philosophy of using AI as a collaborative partner rather than a substitute for human creativity.
Amita