As I’ve been building this project I’m calling Better Political Conversations, I’ve stumbled across an AMAZING virus detector
for my writing. When I think a section of my content is complete and fully edited, I copy-paste it into NotebookLM and have it generate a podcast of the content. The way the podcast AIs talk about the content makes the flaws in my explanations very obvious, and I go back and edit. It also gives me a sense of how and where people will be likely to fold pre-existing knowledge and assumptions into their interpretations of my words, and I’m challenged to find a less-interpretable-but-still-accessible version. When it finally gives me a good podcast, I feel way more confident in what I’ve got.