I Tried to Summarize a E-book With ChatGPT, and the Plagiarism Protections Labored

On the weekends, I cozy up on the sofa with my espresso, and I go away my telephone in one other room, hoping to succeed in my aim of studying one guide each month. This ritual creates area that is unbound by time and to-do lists — however as my studying checklist grows, I discover myself attempting to get by means of extra pages and extra books.

I wished to see if I may use synthetic intelligence to summarize the principle ideas, classes and knowledge of a guide I won’t get to for months and even years. Hold the gorgeous 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 varied AI instruments, I believed 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 take advantage of it. 

However the mission wasn’t very profitable — partly due to plagiarism protections inbuilt to the instruments (and rightly so), and partly as a result of it took an entire lot of immediate engineering and unbiased analysis to get something price utilizing.

Ponder on the parameters 

The very very first thing I realized was that ChatGPT does not have entry to full manuscripts — to keep away from plagiarism and respect mental property rights — and simply summarizes from present info on-line.

If I’ve realized something about AI, it is that the preprompt pondering is simply as necessary as the primary immediate. I did not simply need an enormous abstract of the guide. I wished to study Newport’s huge concepts, arguments, methods and frameworks relating to deep work, so I may apply it in my work. 

So, I began the chat with some expectation-setting. 

• First, ask ChatGPT if it has entry to the entire guide. 
• Inform ChatGPT that I need deep insights, not a superficial abstract. 
• Ask for strategies on the way to apply the important thing methods to my freelance enterprise. 
• See if it may possibly pull from evaluations and reader commentary for extra shocking and actionable takeaways.

Immediate 1: “Do you’ve got entry to Cal Newport’s guide, Deep Work?” 

Not very best, however I had an concept. I discovered a 6-hour audiobook YouTube clip, so I requested if it may use that to summarize the guide for me. 

However no such luck. It advised me to look at the video — one other plagiarism safety. Watch a video for six hours? That is hardly time saving.

There are over 32,000 evaluations of the guide on Amazon, so I believed possibly there’s sufficient commentary on the guide to give you an in-depth abstract. So I began over from this angle. 

Subsequent immediate: “I have never learn Cal Newport’s guide, 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 guide.” 

ChatGPT had a tough time decoding “complete,” spitting out a ton of strategies. 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 strategies like “deal with duties that present essentially 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’m going. 

I want. 

I requested for an instance of deep work in 2024, and we began to get someplace. I appreciated 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 bear in mind an idea I like referred to as the manager-maker schedule, which particulars the 2 principal schedule sorts. 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 guide in order that I knew what insights to prod ChatGPT for. Apparently “productive meditation” was an necessary 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 increase. When directed, it was nice. 

I scanned a second abstract and located one other idea that sounded fascinating: preserving a compelling scoreboard. ChatGPT helped unpack the idea.

The TL;DR? 

If you wish to use ChatGPT to study a guide, you may need to spend not less than half-hour shopping reader summaries your self and use essentially the most fascinating concepts to border your prompts. This inevitably takes a bit of your potential deep work time. You’ll be able to’t simply ask the AI software to do it for you. 

When you get heading in the right direction that approach, you may uncover one idea that adjustments your productiveness or perspective — however ChatGPT nonetheless is not nearly as good as simply studying the guide your self. 

You even have to contemplate that since ChatGPT does not have entry to the precise guide, and solely summaries and evaluations which might be out there on-line, you might not even be getting correct key factors. 

And whether or not utilizing an AI abstract defeats the aim of a guide about deep work continues to be up for debate.