Robots found my most depolarizing chapter
I've got two fun things to share with you:
First: I’ve answered your questions!
Ten of them, anyway, sent in by folks in Maryland, Indiana, Pennsylvania, and even Sweden.
Check out the answers on the latest episode of A Braver Way: Mailbox: Your questions on hope, betrayal and losing your temper.
Producer Jessica Jones joined me behind the mic and pulled out your questions from actual boxes. Really. It was quite fun and suspenseful. We could have gone on all day, but that darn time got in the way.
You’ll hear from me today if yours was one of the questions we talked about. Among them:
Involving whom or in what situation is your greatest challenge to keep your temper?
How do you handle fear?
What can our curiosity about nature teach us about being curious about people?
And! We’re going to keep the mailbox open. So if a new curiosity’s occurred to you, about any old thing you’d want to ask me given my work, you can drop your question here, and it’ll go in the boxes for next time we record a “Mailbox” episode.
I have a hunch this’ll become a habit.
Second: Robots found my most depolarizing chapter!
This is way cool.
Researchers Rémi Thériault, Dominic Packer, and Jay Van Bavel set up this big experiment using multiple AI models to ask how depolarizing 26 prosocial nonfiction books might be for real-life humans.
And lo and behold, I Never Thought of It That Way came in at a very respectable fourth place — just ahead of High Conflict by Amanda Ripley and The Righteous Mind by Jonathan Haidt, and behind Facing the Fracture and Beyond Your Bubble by Tania Israel, and Lovingkindness by Sharon Salzberg.
"The goal with these simulations was to narrow down which books to test in real human trials because there are just so many books out there we'd love to test,” Rémi told me. “But my hunch is that the best performing books in this LLM experiment are also the most likely to perform well when put to the test in humans — so it's good to see I Never Thought of It That Way up there.”
Bonus: I asked Rémi about one of the charts in his post detailing his results, and he sent along an even more detailed breakdown of something very neat — his AI models’ take on which chapter in the book they predict would have the most depolarizing effects.
The verdict? Chapter 8: Assumptions is the clear winner (it’s the twelfth section of the book, as charted below):
Which, frankly, I might have guessed! It’s the chapter that tells the story of a political bridge building trip that made a huge difference for the folks who were part of it, including yours truly.
As for his project’s hypothesis that books give us a boost, Rémi’s pretty encouraged:
“The main goal is to find effective ways to reduce polarization in a currently very politically divided America,” he said. “Based on my PhD findings, I really think that sustained intellectual engagement such as through books has the potential to help.”
Stay curious, all
