Ethan Mollick is writing a follow-up to Co-Intelligence. The new book is called Co-Existence, and he announced it this week on his One Useful Thing newsletter. It arrives October 20, and you can pre-order it now.

I really enjoyed Co-Intelligence. That book was written for the chatbot era, when working with AI meant going back and forth with a prompt box and the human stayed firmly in charge.

That world is already fading. With agents now doing real work (and sometimes doing it better than we do), the question has shifted from “how do I prompt this thing?” to “how do I work alongside it?”

That’s the premise of the new book. Mollick describes it as a book about working with AI that is sometimes, but not always, better than you. That framing rings true for me. I watch these tools nail a hard problem in the morning and face-plant on an easy one in the afternoon.

He also admits his last book had 128 em-dashes, and he cut way back this time to prove a human wrote it. I know that feeling.

I’ve already pre-ordered. Mollick is full of good ideas on this subject, and I expect this book to give me plenty to chew on when it releases in October.