Cory O’Brien is joining us today to talk about his novel, Two Truths and a Lie. Here’s the publisher’s description:
In a mostly underwater near-future Los Angeles, aging combat-drone veteran Orr Vue now lives a simple and small life, trading snippets of what’s become the most valuable currency: information. So when the cops show up at his door looking for data on a murder he’s not even aware has happened, things get interesting for the first time in 25 years.
At first, Orr is happy to exchange whatever he knows about the demise of InfoDrip’s top exec to buy booze and pay rent on his memory storage, but that plan goes to hell when Orr’s old boyfriend, Auggie Wolf, shows up as the number one suspect. Forced to stretch his atrophied spy skills and take his illegal horde of drones out of retirement alongside his busted knees, Orr finds himself in the crosshairs of the militarized police, a family of megarich corporate heirs, a clan of emancipated AIs, and a cult. Barely avoiding getting killed with every clue he collects, Orr realizes he’s uncovered not just a murder, but a conspiracy that threatens Auggie’s very existence. Ahh, the things we do for love…
But in a world where memories can be bought and sold, how can you truly know who anyone is—or what you yourself are capable of? Fast paced, funny, and shockingly romantic, Two Truths and a Lie is Raymond Chandler reinvented for the 22nd century.
What’s Cory’s favorite bit?
Writing a near-future science fiction novel can be tricky because the longer it takes, the nearer the future becomes. Things that seem like impossible inventions may be invented, or worse, rendered obsolete by other inventions. Given that, I’m really quite proud of how my new novel, Two Truths and A Lie, handles Artificial Intelligence.
Two Truths and a Lie is a sci-fi noir mystery that takes place in a future where information has become the primary form of currency. Since artificial intelligences are literally made out of information, it always made sense to me that they would be deeply involved in this surreal economy. What exactly that meant, though, changed over time.
When I first began writing, LLMs were barely an idea, and the most advanced algorithms making headlines were the ones used for high frequency trading. This is why most of the AIs that the protagonist (geriatric fact-checker Orr Vue) interacts with are named after stock-trading algorithms (Knife, Boston Shuffle, Carnival, etc.) The idea was that institutional greed had caused traders to graft more and more capabilities onto these algorithms, until they eventually became autonomous and went into business for themselves.
This idea survived into the final draft of the book, but it was complicated by the advent of Generative AI. ChatGPT started outputting startlingly lifelike dialogue, and there was a wave of panic/exuberance around “Artificial General Intelligence” emerging and taking over the world. It even slotted in nicely with my “trading algorithms become sentient” idea – just recently, an AI startup promised to simulate Warren Buffet’s mind to help it pick stocks! But at the same time, researchers such as Emily Bender et al warned that these apparently intelligent algorithms were in reality only “stochastic parrots,” machines regurgitating statistically likely strings of words, without any thought behind them. In spite of my technological optimism, I tend to side with this latter belief, and it’s made its way into Two Truths and a Lie.
There’s lots of science fiction out there in which AIs are, effectively, just really intelligent humans with poor social skills. Companies like OpenAI have capitalized on this popular portrayal to make their products seem more revolutionary, more genuinely intelligent than they actually are. They’ve also fueled apocalyptic theories about “The Singularity” and “Roko’s Basilisk.” Personally, though, I’m less interested in the question of “what if computers became sentient and took over the world?” and more interested in the question: “what if computers took over the world, without ever becoming sentient?”
The AIs in Two Truths and a Lie sound like they know what they’re talking about. They drive cars, run businesses, and make lots of money. In most situations, they function like really intelligent humans with decent social skills. But there are cracks in the facade that reveal there’s nothing underneath. An AI might demonstrate in conversation that it knows the word for chair, but not what a chair actually is. It might suddenly change its identity when its underlying dataset is altered. It might be tricked into spilling an unedited stream of private information, as in the notorious heartbleed exploit.
The common thread, the thing that I try to convey with all of these characters, is that they do not know what they’re saying. They aren’t “artificial intelligence” so much as “counterfeit intelligence.” Their behavior says nothing about them, and everything about the people who use and abuse them.
That doesn’t stop people from treating them like people, though, in real life or in the world of the story. LLMs are black boxes, with the appearance of consciousness arising out of complex mechanical interactions that no one – not even the people who built them – can fully understand or explain. This inscrutability makes them seem mysterious, powerful, and even dangerous. “Alien fucking motives,” Orr says, explaining his unease around the AIs of his world. A more accurate assessment might be “no motives.” But when dealing with something that seems so genuinely intelligent, what motives could be more alien than none at all?
LINKS:
WebsiteBlog: https://bettermyths.com
BIO:
Cory O’Brien is the author of Zeus Grants Stupid Wishes and George Washington Is Cash Money. He has written for numerous award-winning video games, including Monster Prom and Holovista, and designed multiple tabletop games, including Inhuman Conditions and Hand to Hand Wombat. He lives in Chicago.
