I have a goal of reading one book a month. On weekends, I sit on the couch with a cup of coffee in another room and am on my phone. The purpose of this habit is to create space in my schedule where I am not tied down by time or what I have to do, but as my reading list grows, I try to get through more pages and more books.
We wanted to see if we could use artificial intelligence to summarize the key concepts, lessons, and wisdom of a book that we might not be able to read for months or years, leaving the beautiful writing on the physical page and using an AI to summarize, say, a non-fiction business book.
I chose Cal Newport's Deep Work to do a test run on ChatGPT. After trying out various AI tools, I thought ChatGPT would be the best fit, as it's one of the best-known text prompt chat tools. I also have a paid membership for $20/month, so I wanted to get the most out of it.
However, this mission was not very successful, partly because of the anti-plagiarism features built into the tools (obviously), and partly because it required a lot of quick engineering and original research to get anything worth using.
Thinking about parameters
The first thing I learned was that in order to avoid plagiarism and respect intellectual property rights, ChatGPT does not have access to the full manuscript, but only summarises it from existing information online.
If there's one thing I've learned about AI, it's that pre-prompt thinking is just as important as the initial prompt. I didn't just want a giant summary of the book. I wanted to learn Newport's big ideas, arguments, strategies, and frameworks for Deep Work and apply them to my own work.
So I started the chat by setting expectations.
First, ask ChatGPT if you can get access to the entire book. Tell ChatGPT that you want deep insights, not a superficial summary. Ask for suggestions on how to apply key strategies to your freelance business. See if you can glean more surprising and actionable tips from reviews and reader comments.
Prompt 1: “Have you got Cal Newport's book Deep Work?”
Screenshot by CNET
It wasn't ideal, but I had an idea: I found a YouTube clip of a six-hour audiobook and asked if I could use that to create a summary of the book.
But you're out of luck. You're asked to watch a video. This is also a plagiarism prevention measure. Watching a 6-hour video? That's not going to save you any time.
Since this book has over 32,000 reviews on Amazon, I thought there might be enough commentary on the book to write a detailed summary, so I started writing from this angle.
Next prompt: “I haven't read Cal Newport's book Deep Work. I want you to highlight the key ideas, concepts, strategies, and frameworks so that I can apply them to my business as if I had read it. A summary of the book is not enough.”
ChatGPT had a hard time interpreting “comprehensive” and spat out a ton of suggestions, and without even asking me about my work, started giving me advice like “Educate your team on the importance of deep work and provide training on time management and focus techniques” and “Allocate dedicated time slots on your calendar for deep work.” It also gave me general suggestions like “Focus on your highest value tasks and minimize time spent on low value activities.”
To me, this is all pretty mundane advice: apply pressure time for each question.
Screenshot by CNET
Still no breakthrough insight emerged, so I kept trying.
Screenshot by CNET
Alright, ChatGPT, let's head off to the Swiss countryside.
I hope.
I asked for examples of deep work for 2024 and we started working on something. I liked one suggestion: batching shallow work.
Screenshot by CNET
This is something I do instinctively, but it was helpful because it reminded me to batch tasks and be mindful of context switching.
Involving Google to get ChatGPT off the ground
At this point, I was using ChatGPT to ask random questions, like whether there was a limit on deep work time.
It said that deep work was limited to a maximum of 4 hours per day.
I remember a concept I like that details the two main schedule types, called the manager-maker schedule: As a writer, I follow a maker schedule, which means blocks of uninterrupted time are key.
I had to do a quick Google search of the key lessons from the book to know what insights to look to ChatGPT for. Apparently “productive meditation” was a key takeaway, so I asked about it in more detail.
Screenshot by CNET
Finally, a new concept: Deep Work is not just about having your head bowed at a desk.
This strategy was key: find the summary nuggets on Google, then come back to ChatGPT and extract them. I followed the instructions and it worked very well.
As I looked through the second abstract, I found another interesting concept: maintaining an engaging leaderboard. ChatGPT helped me unravel this concept.
Screenshot by CNET
What is the TL;DR?
If you want to use ChatGPT to learn about a book, you would have to spend at least 30 minutes reading the reader's summary yourself and creating prompts with the most interesting ideas. This will inevitably take up a large portion of your potential deep work time. You can't leave it to an AI tool.
If you make progress in this right direction, you might discover one concept that will change your productivity or perspective. But ChatGPT still doesn’t compare to reading a book on your own.
You should also consider that ChatGPT does not have access to the actual book, only summaries and reviews available online, so you might not get the exact gist.
It is still up for debate whether the use of AI summaries undermines the purpose of books on deep work.
