So I’ve been writing about artificial intelligence (AI) for more than two decades. AI and robotics were part of my dissertation at UC Santa Barbara (many thanks to my advisors for allowing a weird project about which they knew nothing and my former colleagues for hiring me to teach and research such weird things *way* before it made the New York Times).
Before Karen Hao’s book, Empires of AI, was published, I was super excited to read it. An insider’s look at OpenAI is the sort of thing I’m all in for. And I’m enjoying the book. I’m about 90% finished. It’s gracefully written and built on real effort to understand. I recommend the book thoroughly.
I definitely recommend the book. 🙂
That said, I feel compelled to write a couple of posts about the book because it engages things I, and others, have been working our way through for years.
Because of the nature of my research, I’ll start with the very obvious religious stuff that happens near the end. Hao points toward the religious commitments evidently espoused by Ilya Sutskever, one of OpenAI’s chief architects. Sutskever, Hao notes, believed that building an artificial general intelligence (AGI) would “bring about a rapture. Literally a rapture.” She quotes an OpenAI employee as describing Sutskever’s beliefs this way.
No doubt the OpenAI employee is correct, and no doubt the exact same perspective was widespread at OpenAI. I wrote about the religious desire to build transcendent AI two decades ago and one of the chief architects of that belief, Hans Moravec, has a book on the library shelves of OpenAI. As an aside, I should note that I think it would be a remarkable contribution for someone like Cade Metz (whose article I just linked) to write an entire book about the libraries in Silicon Valley companies. Anyway, at least one of Moravec’s books is on the suggestive shelves of OpenAI. Maybe another also. I’d wager almost any amount of money that at least one Ray Kurzweil book is there.
So, yeah, that Sutskever believes in a literal AI rapture? I mean, that was literally the whole point of my book, Apocalyptic AI: Visions of Heaven in Robotics, Artificial Intelligence, and Virtual Reality (Oxford 2010), not to mention preceding essays and suchlike.
And some folks even read that book, which I appreciate.
All I’m saying is that Sutskever’s religion isn’t new. It’s antecedents could get traced in different ways, and some scholars have. If you don’t believe me, grab one of my books and look at the references. 😉
Robert M Geraci 