Book Review #1 Co-Intelligence by Ethan Mollick

I was reading Co - Intelligence by Ethan Mollick on the Tube in London and I missed my stop - ever the sign of a good read. I found myself reading it everywhere: with my kids at their Taekwondo lessons, at the hairdresser’s, waiting for delayed trains at Euston, and finally finishing it on a sunny day on the beach in Clacton-on-Sea. This book captivated me so much that I feel compelled to write a review. Allow me to explain why.

I read a comment in an online discussion about the book suggesting that it could be a great gift for your parents, and that those familiar with AI might not find much new information in it. I think I was this parent - I really benefited from a new level of understanding of some concepts and I hope, can now explain how LLMs work much better than before. 

The main reason I felt I wanted to write this review is because there are so many books about AI leaving you anxious about the future. They often portray AI trained on social media, inheriting human biases, stealing intellectual property, and depict a dystopian future where AI is inherently evil. I could make a list of these alarmist books, I read them too. 

Personally, I think it’s time to start listing a different kind of book - not about utopian worlds with AI solving all our problems, but books  stimulating critical thinking and uplift at the same time. Books that encourage you to experiment with AI, informed by a better understanding of AI. 

Number one in this list is Co-intelligence by Ethan Mollick.

First of all, this book is written in a reader-friendly tone.

There are many great metaphors helping to explain complex concepts, making them not only easier to digest but also stick with you. I don’t think this needs a spoiler alert, I hope you will have a similar feeling I had when I encountered a great metaphor explaining how LLMs work (watching the Bear on Disney+ added more to this) and on the phenomenon he calls Jagged Frontier.  

Ethan Mollick writes about  Centaurs and Cyborgs when discussing approaches to integrating the work of people and machines.

Being a Centaur means a clear separation of tasks delegated to AI and tasks for humans, with the choice based on the strengths of both and the nature of the tasks. There is a constant switch between human and AI tasks. 

Being a Cyborg means a blended model of co-working, with deep integration. You  can’t draw a line as everything is produced through the constant collaboration of humans and AI. 

My own story of working with AI is an evolution from Centaur to Cyborg, when I first started experimenting with AI it was important for me to be able to keep a line between non-synthetic texts and those created with the assistance of AI. However, working in tandem with AI is now the prevailing mode and the border between my own invention and that of AI has become blurry. I dictated parts of this book review to Chat GPT, with the prompt to improve the grammar and style without rewriting, keeping my voice. 

 There are 4 scenarios for our future with  AI - I wonder which one resonates with you? It will be fascinating to see which forecast comes true. 

Finally, is it just me, or if you’ve read Co-Intelligence have you noticed this too? We are now connected through the use of the Co-Intelligence code. For example, I was at an online conference where a designer discussed the use of AI tools, he mentioned treating AI like an intern but remaining the human in the loop. At that moment, I wanted to send him a Vulcan salute emoji. Too bad it’s not possible on Zoom! It felt like we were on the same page of the same book.  

There’s also a set of tools I got with the book, which I haven’t had chance to experiment with yet. 

That’s all for now.  Signing off, Inna (Cyborg).  

Check out Prof. Ethan Mollicks’s Substack One Useful Thing