Literary Theory for Robots: How Computers Learned to Write

I did not love this book. There’s a potential for a good book there, but it’s wrapped up in a very annoying writing style.

The book is a look at the history of how we have automated text generation. The author goes back into ancient times, and moves forward to Eliza and Claude Shannon, then eventually on to Markov Chains and modern AI.

But the writing style is just awful. It’s very, very obtuse. I bought this book because it was short (part of the Norton Shorts series), and I thought it would be concise.

It’s the absolute opposite of concise. It wanders around aimlessly, taking all sorts of tangents. The author seemed determined to demonstrate his well-rounded-ness more than he wanted to convey the source material. He tried to get lyrical and philosophic when all I wanted was a clear description of the subject.

The conclusion is entitled, “9 Big Ideas for a Conclusion” (so, so meta). I thought, “Awesome! A list! This will be concise!”

Spoiler: it was not. It made about as much sense as the rest of it.

It’s short, but I’m still annoyed that I read it.