Quick one, but a fun one. Thinking in Bets is one of the first books to launch with Google’s new Expert Intelligence feature in Gemini Notebook.
Here’s why I’m excited about it.
Up until now, when you asked an AI a question about a book, the answer came from whatever the model happened to absorb during training. Sometimes that’s a faithful account of what the book actually says. Sometimes it isn’t.
The problem is you can’t always tell which one you’re looking at — the confident right answer or the confident wrong one.
Expert Intelligence works off the actual book. You add it as a source, and the answers come back grounded in the text, with citations to the passages they came from. So when it tells you what I said about resulting, you can go check whether I said it.
I asked it why probabilistic thinking helps. Here’s what came back: our minds form beliefs perceptually. We hear something, accept it on the spot, and then basically never go back and vet it. Saying “I’m 60% sure” instead of just asserting the thing represents what you actually know, protects you from the blind spots that smart people are especially prone to, and invites other people to help check you.
That’s not a model paraphrasing something it once read about my book. That’s the actual book.
The part I’m most curious about is what happens when you put your own material in alongside it. Upload your decision journal and ask where you’ve been grading decisions by how they turned out instead of by how you made them. Drop in the notes from a decision your team is still arguing about and ask what a pre-mortem would have caught.
I’m really excited to start getting feedback as people use it.
There is also a Featured Notebook to go with the book. Starter prompts, supplemental materials, and a Decision Quiz with five scenarios built to make you separate outcome quality from decision quality. If you’ve ever nodded along at that distinction while quietly suspecting you’d flunk a live test of it, start there.
And Google is giving away the book. Grab a free copy here. The reward gets applied at checkout, no code needed.
The whole thing is on the front page at notebook.google.com for about the next week, my book included.
If you try it, tell me in the comments what you asked it. I want to know what questions people bring to a book once the book can answer back.