I'd Rather Be Writing Podcast

This is a recording of our AI Book Club discussion of The MANIAC by Benjamin Labatut, held August 16, 2026. This was the club's first work of fiction, and much of the discussion is about what that mode makes possible: getting inside the madness of scientists like John von Neumann in a way a biography can't. We talk about the book's opening scene as a thesis in miniature, logic carried to irrational ends, the 'death of play' in the AlphaGo match against Lee Sedol, Move 78 as a case for human unpredictability, the garage door scene as a metaphor for...

In this podcast, co-host Fabrizio Ferri-Benedetti and I chat with Sarah Deaton, a technical writer at Anthropic who works on the Claude Code docs, about what documentation work looks like when much of it is automated. We discuss orchestrating docs through skills and agents rather than writing them, mining AI assistant conversations for doc gaps, running forensics on every false claim that reaches the docs, new metrics like time-to-correction and steers per PR, the fences that keep AI-generated docs honest, and how she closed more than a hundred pull requests in a single day.

This is a recording of the AI Book Club discussion of Sebastian Mallaby's The Infinity Machine: Demis Hassabis, DeepMind, and the Quest for Superintelligence. In the discussion, we talk about the complex character of Demis Hassabis (and comparisons to Ender's game), the idea of founders' personal morality acting as a final safeguard, and the relentless acceleration of AI development. We also talk about the historical rivalries between deep learning and reinforcement learning, the growing urgency around AI-driven job displacement, and what it means to train AI to automate our own work.

In this podcast, I chat with Fabrizio Ferri-Benedetti about a variety of topics related to AI and docs, such as applying Italo Calvino's literary principles of lightness and quickness to evaluate docs, the reality of AI review fatigue versus creator fatigue, whether vibe-coded tools are tools slop, developing internal skills for repeatable doc processes, and the utility of running local AI models.

This is a recording of the AI Book Club discussion of Max Tegmark's Life 3.0: Being Human in the Age of Artificial Intelligence. Our discussion explores Tegmark's visions of superintelligence, the unpredictability of AI goals and subgoals, and how the Asilomar principles fare against the relentless market pressures of today's AI arms race. We also contrast the book's theoretical focus on cosmic energy (and how superintelligence might more efficiently extract energy from matter) with the immediate, tangible threats AI poses to cybersecurity and global financial systems.