The Holon Project

Turning 52 books into computer code

This year, my goal is to read 52 books. However, this will be no ordinary reading process. For each book, I will do a standard readthrough, taking notes about the core concepts as I go along and then after finishing the normal reading, I will turn the core concepts into runnable computer code. Finally I will publish the code to github and review the book and my work through a YouTube video and a Medium.com/substack.com article. 

As far as I know, this is the first project of its kind. The reason I am doing this is because much of the information contained in books never really gets utilized by the reader. When Bill Gates was asked what sort of company he would start if he were just getting started today, his answer was that he would start a company that teaches computers how to read books so that the they could utilize all the world’s knowledge for the betterment of humanity. Bill thinks that such a company could be worth 10x what Microsoft is currently worth. 

My plan is to translate the most interesting books I can find into computer code BY HAND. Only then will I feel like I understand the process well enough to automate it. The plan that I am currently proposing goes strongly against the modern zeitgeist of machine intelligence. Surely, any company that heeds Bill’s advice about teaching computers to read would immediately fire up a deep learning suite and get to work throwing text at it. 

In 2013, before deep learning had been popularized, back when researchers at the Oak Ridge National Lab where I worked were still talking about support vector machines being superior to neural networks, back when I was studying biologically inspired computation under Dr. Bruce McLennan as an undergrad, I knew back then that I wanted to combine symbolic AI with small orthogonal neural networks to produce a hybrid system.

This project will be my attempt at producing that hybrid system, first by combining my own neural networks with book knowledge to produce computer code about said knowledge and then maybe, to clone my neural networks, piece by piece to do the work of encoding books automatically.

If you made it this far, I have a treat for you. This entire project will be programmed in JavaScript, the langua franca of the web. For my first book translation, I will compile the book “Eloquent Javascript” by Marijn Haverbeke into JavaScript itself. The book is suitable for beginner and expert alike. The goal is to review the technical fundamentals and best practices of JavaScript such that everything else we do in this project will have a strong foundation.

Call to Action: If you are interested in this project, consider purchasing “Eloquent Javascript” by Marijn Haverbeke and reading through it.