Being stuck in ASCII poverty, I’m too embarrassed to write Godel, Escher, Bach in the title. Here are my thoughts on it.

This book will take you a long time to read. I’m not a particularly slow reader, but it took me about 3 weeks to get through this, whereas a similarly sized novel would take one at the most. My main problem was being unable to get past certain paragraphs: every so often the book will hit you with something so beautiful or profound that you just have to put it down and read around the subject. At one point I took a break to design my own language, but eventually backed down and when I realised you need serious linguistics grokkage for something like that. That doesn’t mean I’ve given up; the project is simply postponed. You can expect to be speaking it in about 40 years time. You have been warned.

Critics have spent the last twenty years praising this masterpiece, so there’s not much left for me to say. Of course it’s brilliant, inspiring, deep, funny, clever and creative. Everybody knows that.

At the beginning of this year, my favourite book was The Hitch Hikers Guide to the Galaxy. Now it’s in third place, with A Fire Upon the Deep and GEB jostling for first place.

Some reviewers mark this book down for the massive self-indulgence on the part of the author, but he’s just about smart enough to get away with it. One of the book’s major themes is self reference, after all.

Like most timeless works, GEB is a book about problems. It offers few answers, but the questions asked are deep, difficult, and interesting enough to spend a normal lifetime pondering. I suspect a good long ponder is as close to an answer as most of the questions will admit. Like economics, it’s seems that intelligence is a subject which nobody truly understands in the slightest.

GEB glosses over the question of superintelligence, offering the superintelligent comment that “superintelligence would be weird.”

When online, people develop banner blindness, which filters out areas of the screen that 99.999% of the time contain useless information. Most books encourage a similar condition (or should it be an ability?): bibliography blindness. The endpages of your average book are just filler, adverts, or a pathetic attempt at a list of references. Soon you learn to just “stop reading when the text gets small”. GEB is (obviously) different. The bibliography is fully annotated and completes the book’s function as a pointer to fascinating things.

I suppose it’s not perfect - I read the 20th anniversary edition, which is just the exact original text along with an introduction. In some ways I hope he’s working on an expanded 25th anniversary edition. If you view the book as an object of literature - a singular “eternal golden braid” - then the book should rightly stay as it is, was, and always shall be. On the other hand, the book is immensely useful. It’s full of pointers to interesting topics, and revolves around an extremely intuitive explanation of Godel’s Incompleteness Theorem. There is a digression into the workings of “microcomputers” which is very late 70s, and from some of the chapters, a reader might be left thinking the AI scene is a thriving branch of computer science, instead of the unfunded dustbowl of a landscape it really is today. Since changing the text of the book would be like weaving a new thread into an already complete persian carpet, it would seem best to add a Post Scriptum to the end of each chapter, exactly as in Metamagical Themas. I would very much like to know Mr H’s thoughts 25 years down the line. Or maybe he really doesn’t want the page length to change from the perfect 777.

Tal Cohen has a notable review, and a page of nice book in-jokes from “otherwise serious books”. GEB mostly tells serious jokes.