How I’m Using ChatGPT Projects to Learn From Books

I read a lot of books, take away a few ideas and then do something very human: forget most of them.

So I am trying a different approach. Instead of treating each book as something I finish, I am treating a group of related books as a learning space I can keep returning to.

The friction

I have books that belong together in my head. For example: Finish, Effortless, The Power of Small and Slow Productivity.

They are not identical, but they orbit similar questions: How do I do meaningful work without turning everything into a giant project? How do I stop perfection from becoming procrastination? How do I make progress with less unnecessary effort?

The normal reading workflow does not help me much after the book is done. Highlights sit somewhere. Notes sit somewhere else. The useful idea appears exactly when I need it least.

What I tried

I created a ChatGPT Project for this cluster of books and added the books as sources.

Now I can go back to that same project when I have a real problem and ask questions across the group rather than trying to remember which book said what.

That feels much closer to why I read in the first place.

What actually changed

The books stopped feeling like separate objects on a shelf and started feeling like a small advisory group around one problem.

I can ask for disagreement. I can ask which idea applies to my situation. I can ask for one tiny action instead of another summary.

Most importantly, I can come back weeks later without rebuilding the context from zero.

What I changed

I am organising by problem, not by book.

• A project about finishing and doing less can contain several books.
• A project about decision-making can contain another group.
• A project about influence or communication can become its own learning room.

The book is a source. The problem is the organising unit.

What I am watching for

This experiment can also become another form of collecting. I can imagine creating ten beautiful projects and never changing a single behaviour.

So I am adding one rule:

Every useful learning conversation should end with something I will try, decide or stop doing.

Steal this if it helps

Pick three or four books you already own that solve a similar problem. Put them together. Then do not ask for summaries.

Bring a real situation from your life or work and ask the books to help with that.

If the conversation produces one action you actually use, the books have already become more valuable than another folder full of highlights.