I once met this man who had written a tome about the JFK assassination, a book available on Amazon (albeit in a self-published formed held together with twine—per a customer’s review). It’s premise was the FBI killed JFK, a far-reaching conspiracy hidden from the public all these years, ignored by the Warren Commission. When I met the man I had just returned from a trip to Dallas and had visited the book depository where Lee Harvey Oswald had perched, and where there is now a museum. So I had a fresh interest in the subject, even though the man had the disheveled look and twisted glint in his eyes of the crackpot. Not one prone to conspiracy theories, I asked him about his research, his visits to Dallas and the book depository.
He told me he’d never been to Dallas.
When it comes to AI, I’m a bit like that man.
If it weren’t for the help of a roommate who was a computer science major, I would have failed my college class in Fortran (or C++, or whatever program was in vogue many years ago). So how did I write a novel about AI? I sat in a room and thought about it. This is, of course, the purview of conspiracists.
For my novel, I did limited research, which is the best kind of research for a novel. You have to dig enough to create a foundation of believability, but not so much there’s no room for imagination. Have some inkling of the concepts of deep learning, neural networks, thought tokens, etc… You don’t need to understand the deeper underpinnings, vectors and linear algebra. Neither does the reader. Because a novel, a story, is not about such things. Too much detail bogs the narrative down; it’s like adding so many modifiers to a noun that the word you care about gets buried alive. The whole point of fiction is fiction, an imagined story. If your book is essentially a MadLibs (I’m dating myself) technical manual with a few blanks you fill in for your limited moments of creativity, then it will be limited as a story, however technically accurate.
The irony—or the essential nature—of AI is that you don’t need to understand anything about the inner-workings of AI (the programming, the deep neural networks) to understand what is most important about AI. The AI is doing all the programming, the AI is creating its own neural networks. This is extending beyond the comprehension of even the most elite computer scientists. What is important for all of us is the philosophical, ethical, and practical matters of how this will change us and the world.
Small Town Echoes of Metallic Minds had many jumping off points, one of which was empathy and the idea of how you might create artificial empathy, that is, the alignment problem. When I first began writing the story in 2019 I had no overriding interest in broader implications, but events have changed much since then. There are many journalists and thinkers currently sounding the alarm bells. I’m generally an optimist, but not when it comes to AI. Even if you think the present dangers of AI are hype, this is only the neonatal period, the infancy of a technology evolving in months, not eons. What will be our situation twenty, thirty, a hundred years from now? What has been most disturbing to me is the book I started seven years ago, a storyline that seemed incredulous at the time, seems not only plausible but real.
To put my mind at rest, wondering if I was overreacting, I recently listened to the Ezra Klein podcast “Jensen Huang Thinks A.I. Alarmism Has Gone Too Far.” I thought maybe this tech wizard, the chief executive of Nvidia, could talk me off the ledge. It was the least reassuring discussion I could have listened to: his central argument is that a US corporation would never release a potentially harmful product, as they would be worried about liability (see Purdue Pharma, Philip Morris, etc…).
And no, I don’t think the FBI conspired to assassinate JFK.
Some sources and recent releases relevant to Small Town Echoes of Metallic Minds:
1) For the basics of AI:
Godel, Escher, Bach: an Eternal Golden Braid Douglas R. Hofstadter
Why Machines Learn, The Elegant Math Behind Modern AI Anil Ananthaswamy
2) Empathy and alignment:
Looking for Spinoza, Joy, Sorrow, and the Feeling Brain Antonio Damasio
Animals, Robots, Gods, Adventures in the Moral Imagination Webb Keane
3) AI and writing:
Literary Theory for Robots Dennis Yi Tenen
The Artist in the Machine Arthur I. Miller
4) The political/worldwide implications:
The Rise and Fall of the Artificial State Jill Lepore
Ezra Klein podcast: “Bill Gates’s Blunt Warning on A.I.”
