Show HN: Storytelling for coding agents, using Pixar's story process (kashyab.com)

Karina Nguyen's recent Lux Summit talk explored the intersection of storytelling and AI, suggesting that the iterative process used by coding agents—planning, writing, testing, and debugging—can also apply to storytelling. By paralleling Pixar's Braintrust, a panel that critiques films at the storyboard stage, Nguyen introduced the “storygen-skill” for AI models like Cursor, Claude Code, and Codex. This innovative approach allowed the AI to craft two illustrated books, showcasing how structured self-review improves narrative quality. For instance, revisions to a bedtime story revealed crucial plot adjustments and character depth, demonstrating the potential of AI in creative writing. This development is significant for the AI/ML community as it not only pushes the boundaries of what AI can create but also highlights the importance of structured feedback in enhancing creative outputs. Future iterations aim to streamline the storytelling process into a single loop that encompasses planning, drafting, reviewing, illustrating, and publishing, potentially revolutionizing narrative creation. Moreover, the introduction of a “story bible” could enable the crafting of interactive narratives that maintain continuity, while meticulous logging of revisions provides valuable process data essential for improving AI's understanding of human storytelling. Nguyen’s work suggests that teaching AI to tell stories might be a crucial step toward fostering a deeper understanding of human experience.