I’ve a aim to learn one e book each month. On the weekends, I cozy up on the sofa with my espresso, and I depart my cellphone in one other room. The purpose of this ritual is to create area that is unbound by time and to-do lists — however as my studying checklist grows, I discover myself making an attempt to get via extra pages and extra books.
I wished to see if I might use synthetic intelligence to summarize the primary ideas, classes and knowledge of a e book I may not get to for months and even years. Hold the attractive prose for the bodily web page, however use AI to summarize nonfiction enterprise books, for instance.
I picked Deep Work by Cal Newport to check run on ChatGPT. After utilizing numerous AI instruments, I assumed ChatGPT, one of many best-known textual content immediate chat instruments, would do the job finest. I even have the $20 a month membership, so I wished to benefit from it.
However the mission wasn’t very profitable — partly due to plagiarism protections in-built to the instruments (and rightly so), and partly as a result of it took a complete lot of immediate engineering and unbiased analysis to get something value utilizing.
Ponder on the parameters
The very very first thing I discovered was that ChatGPT does not have entry to full manuscripts — to keep away from plagiarism and respect mental property rights — and simply summarizes from current info on-line.
If I’ve discovered something about AI, it is that the preprompt pondering is simply as essential as the primary immediate. I did not simply need an enormous abstract of the e book. I wished to be taught Newport’s large concepts, arguments, methods and frameworks relating to deep work, so I might apply it in my work.
So, I began the chat with some expectation-setting.
- First, ask ChatGPT if it has entry to the entire e book.
- Inform ChatGPT that I would like deep insights, not a superficial abstract.
- Ask for solutions on how you can apply the important thing methods to my freelance enterprise.
- See if it could possibly pull from critiques and reader commentary for extra shocking and actionable takeaways.
Immediate 1: “Do you could have entry to Cal Newport’s e book, Deep Work?”
Not preferrred, however I had an concept. I discovered a 6-hour audiobook YouTube clip, so I requested if it might use that to summarize the e book for me.
However no such luck. It informed me to observe the video — one other plagiarism safety. Watch a video for six hours? That is hardly time saving.
There are over 32,000 critiques of the e book on Amazon, so I assumed perhaps there’s sufficient commentary on the e book to provide you with an in-depth abstract. So I began over from this angle.
Subsequent immediate: “I have never learn Cal Newport’s e book, Deep Work. Spotlight the important thing concepts, ideas, methods and frameworks, so I can apply to my enterprise as if I’ve learn it. I do not simply desire a abstract of the e book.”
ChatGPT had a tough time deciphering “complete,” spitting out a ton of solutions. It additionally began to provide recommendation with out asking me about my line of labor, like “educate your workforce on the significance of deep work and supply coaching on time administration and focus strategies” and “allocate devoted blocks of time for deep work in your calendar.” It additionally gave generic solutions like “concentrate on duties that present probably the most worth and reduce time spent on low-value actions.”
All fairly vanilla recommendation, in case you ask me. Time to use the strain, one query at a time.
I nonetheless wasn’t getting any breakthrough insights, so I stored pushing.
OK, ChatGPT, off to the Swiss countryside I am going.
I want.
I requested for an instance of deep work in 2024, and we began to get someplace. I favored one suggestion, about batching shallow work.
Whereas I do that instinctively, it was a useful reminder to batch duties and watch context switching.
Getting Google concerned to maintain ChatGPT on observe
At this level, I used to be utilizing ChatGPT to ask random questions, like if there is a restrict on deep work hours.
It stated 4 hours max of deep work per day.
I keep in mind an idea I like referred to as the manager-maker schedule, which particulars the 2 primary schedule varieties. As a author, I am on the maker schedule, which means blocks of uninterrupted time are important.
I needed to do a fast Google search of the important thing classes within the e book in order that I knew what insights to prod ChatGPT for. Apparently “productive meditation” was an essential takeaway, so I requested for extra particulars about that.
Lastly, a new idea. Deep work is not simply head down at desk time.
This technique was the unlock — wanting on Google for nuggets in summaries, then going again to ChatGPT to develop. When directed, it was nice.
I scanned a second abstract and located one other idea that sounded attention-grabbing: conserving a compelling scoreboard. ChatGPT helped unpack the idea.
The TL;DR?
If you wish to use ChatGPT to find out about a e book, you may need to spend no less than half-hour searching reader summaries your self and use probably the most attention-grabbing concepts to border your prompts. This inevitably takes a piece of your potential deep work time. You may’t simply ask the AI software to do it for you.
When you get heading in the right direction that method, you would possibly uncover one idea that adjustments your productiveness or perspective — however ChatGPT nonetheless is not pretty much as good as simply studying the e book your self.
You even have to think about that since ChatGPT does not have entry to the precise e book, and solely summaries and critiques which can be out there on-line, it’s possible you’ll not even be getting correct key factors.
And whether or not utilizing an AI abstract defeats the aim of a e book about deep work continues to be up for debate.
