A Few Books on AI for Wordcels
For most of us without advanced mathematics, how neural networks work will remain largely obscure, bordering on magical. Here are some books to at least give you some idea of what's going on.
I don’t think anyone is debating it anymore – we are the beginning of artificial intelligence fundamentally changing the way we work and live. This is a change at least as big as the development of the internet, maybe bigger. You don’t need anything all that much more powerful than chat GPT-4 to fundamentally upend the way we work and play, but I think in the near future, we are going to have tools much, much more powerful and disturbing. You can think of this as a good thing; you can think it’s a bad thing, either way, it’s a thing, and it’s best to have at least some understanding of it.
For most of us without advanced mathematics, how neural networks and machine learning algorithms actually work will remain largely obscure bordering on magical. This youtube video is about as technical as I can get and for truly technical people, this is a grade school level understanding. Still, that shouldn’t stop us from trying to understand at least the basics of the coming AI revolution. Below are three books I’ve found helpful for developing a basic frame work for understanding what the fuck is going on.
There are those who say recommending books on AI is a fool's errand – the space is changing far too quickly to be adequately chronicled through the long cycle of a standard book’s production. They’re at least partially right. The place to keep up with the latest in AI is the world of podcasts and twitter.* But to understand the scope of what we’re talking about, and some grounding to understanding the latest developments in the field, I think there are some books that help.** Here’s a couple.
If you’re going to read one book on AI, it should probably be this one. Yes, it’s outdated in the specifics (OpenAI is a nonprofit!) but I think the larger themes still hold true. Tegmark does an excellent job of using hypothetical futures to scare the shit out of you, and nontechnical writing to explain the development of the main branches within AI. Tegmark has the math chops to have a deep knowledge of how AI is developing, but this book is written for the common reader. Highly recommended.
An introduction to the various strands of work within AI told through the careers of some of the biggest players. Meaning, when you’re done with this book you’ll be able to use the big buzz words in AI (neural networks! Large language models! artificial general intelligence!) properly and with some understanding. Told narratively it covers the AI winter, the development of neural networks and the early stages of the AI arms race with enough character development to keep even the most hardcore english major interested.
No book exemplifies how quickly the world of AI is changing than this one. It isn’t so much that Roose was wrong in his rules here (you do need to be worried if you’re job is a switch point between machines) but this came out before the release of GPT-3 and as we all know now (including Roose, who had the unnerving conversation with Sydney) everything is different now.
Future proof is an easy read and I think the first popular book to really grapple with what AI will do to the workforce. Back in the old days of 2021, Roose’s book seemed almost alarmist. Now it’s pretty obvious he was actually downplaying how big a deal AI will be. That said, the basic theme of the book – lean in to being human – remains good advice.
* I’m resisting the urge to list specific podcasts and twitter accounts in this post, but they change so fucking fast. Check out my Friday round ups to see what I think is worth listening to and reading in a given week.
** Before you ask, no I haven’t read Superintellegence by Nick Bostrom. Maybe someday.