I wanted to start off by asking you about Dashiell Hammett. Right away I was excited by the hard-boiled style and references that firmly established your narrator as a modern interpretation of the Continental Op. Can you talk about how works from nearly a century ago influenced this story and why it paired well with the very contemporary ideas you are exploring?

The Continental Op is such an interesting phenomenon. Often the main character in a detective story is kind of a cypher—we know so little about how Nero Wolfe and Archie Goodwin met, we get drips at most of Phillip Marlowe’s life story or Sam Spade’s or Nick Charles’s—but there’s something mystical about the degree to which the Op is effaced—he doesn’t even have a name! The true Organization Man. Except he’s not—the organization defines his role but a glittering and furious humanity shines through. The Op playing things so close to the vest is of course also a clever storytelling decision of Hammett’s, but something about his sense of privacy bears thematic weight.

As for why this type of character right now: Well, everything 1930s is new again, isn’t it? In a moment of corruption and greed and impunity, when oligarchs are trying very hard to build a world where ordinary people (anyone with less than a few billion dollars, say) don’t matter, this angle really attracted me, particularly as a lens through which to tell a story about people trying to free themselves from the jaws of giant systems—even (especially?) ones they built themselves. The hard-boiled detective is such an evergreen American character—when there’s very little chance that the systems that define our lives will come through for us, what can we do but keep our eyes open, witness, and protect what we can?

And there was another reason, but it has more to do with one of your later questions, so I’ll save it for that.

As a writer known for novels—both in series and standalone—as well as your shorter work, when do you know the “size” of your story and how does your writing approach vary with different word counts?

Every story needs an ending. The question of length is a question of how close you start to the ending—and how much setup a particular ending requires. That doesn’t mean you have to know the ending before you start! But in short fiction I often work harder to find the shape. I have fewer words to work with, so each word has to carry weight. Sometimes that means precise outlining; more often these days, especially for new projects, I take a little while to find a voice around a concept, then give myself a little time to play in that voice before taking a long hard look at what lines I’ve drawn and determining where they converge. For this project I spent a while hammering at a real typewriter to get the voice right! For a novel, that playing-around process can take as much as 10k or 15k; for a 10k piece like this, maybe 1k or 2k.

This is a story that approaches debt from multiple angles, primarily financial and spiritual. Can you expand more on how the two kinds of debt are related?

Money and religion go way back: If you try to trace the origins of money, you find yourself in Sumerian temples, looking at cuneiform tax and granary records. Many religions discuss the question of “creation debt”—the question of what humans owe their divine creators (and whether this debt can ever be repaid). Sacrifices are often conceived as a kind of payment to God (which also has the benefit, from priests’ perspectives, of supporting the life of priests). Many Christian traditions analogize the crucifixion of Christ to a “repayment” for the sins of the world, or for the initial sin of eating the fruit of the knowledge of good and evil—in this case sin is a kind of debt. And all through the Bible we find regulations around Jubilee, around the forgiveness of debt. “Forgive us our debts, as we forgive those who are indebted to us.” So these senses of meaning have a lot to say to one another!

This story came out of a collaborative project at Dartmouth pairing scholars with science fiction writers; I had a few fascinating afternoons talking with Professor Devin Singh, who’s interested in debt as a theological concept, and debt in the Christian tradition in particular. Dr. Singh’s work teases apart these different senses of debt—which often lead us to regard, say, a contingent financial condition as having profound moral or spiritual weight, or a state of human obligation as a financial matter. We might say that a parent “owes” their child care, say, but what are we doing when we use the same word for that relationship as we’d use for my relationship to the holder of my mortgage? I thought it would be fun and helpful to approach this question in fiction by slamming the many senses of debt together, and seeing what contradictions and conflicts showed up.

Anyone interested in this question in greater depth should check out David Graeber’s Debt: The First 5000 Years and Michael Taussig’s The Devil and Commodity Fetishism in South America, as well as Devin Singh’s work.

I love how you’ve used the fallibility of AI to justify the old school gumshoe approach to detective work and also suggest a different kind of debt to technology. When the “old coats” were able to trust the answers they got from the box, that they did so at a price they did not fully realize. This brings to mind our current dilemma with LLMs. Do you think there is a way to change our approach to technology before we find ourselves in a world where the net’s “full of dreaming ghosts” or are we already there?

This is the more-fun answer to your Continental Op question up top: I had LLMs on the brain when writing this story, and thought, huh, let’s set aside the whole “AGI” / “self-improving AI” stuff for a moment and think about just what this technology is likely to do at scale if it gets, say, three times better. That’s just about it for the internet, right? At least, the internet as you and I know it. Scratch out “on the internet nobody knows you’re a dog”—on the internet, nobody knows you ARE, and you don’t know that anybody else is either. So why waste time there? Why go there to persuade, to chat, to sell, to confess, to communicate? Bots will be extremely good and cheap and have no loyalty to anything like a “truth.” No online video any more real than a drug trip. Digital documentary evidence will be borderline useless, because of how easy it will be to fake. So much for “social” media. And if enough bots start seeding hallucinated data into “official” channels—if, say, a bunch of teenagers start giving an LLM access to social security data—well, that’s a big challenge for the modern state.

Among other things, this sounds like a boom market for professional eyewitnesses! “PI”s if you will. And: I had a nice chuckle thinking about how LLM text poisons LLM training data, and how the models’ hunger for new actual human natural-language input about the actual world would lead, quite naturally, to a justification for the most artificial element in any hardboiled PI story, that is: the story itself! The carefully written, exhaustively detailed, hard-boiled PI narrative of the case, including all the irrelevant-for-a-police-report detail, all the broken bottles in all the alleys, all the sex and all the exhaustion and all the sitting alone at a table and all the wisecracks and metaphors. It’s not a foreground element of the story, but one element of our genderfluid Op’s revenue stream is that this whole adventure gets written down and fed into a model. Good times.

Now, I don’t think this is at all inevitable. Who knows what the future holds? I think we’re in a massive bubble and LLM tokens are being sold at something like one-tenth cost, so we may be in for another long “AI” winter unless the current clowns screw up badly enough to upside-down the entire world economy and end up with some of that sweet, sweet government bailout money. Which might well be the whole idea, but then I wonder how they’ll cool off the mark (the mark being, in this case, “everyone”).

I don’t think the internet’s dead yet or likely to be DEAD-dead, the way I describe here, in the next year or so. It’s still pretty good for Some Things. There’s still blood in that thar stone! But I do wonder if the real is where the juice is going to be moving forward, after a prolonged fascination with magic mirrors. I hope so! After all, the real is where I live. And it’s where we have a lot of important work to do in the next hundred years or so.

Where can our readers look for more of you and your work?

I have a website at maxgladstone.com and a newsletter at https://buttondown.com/MaxGladstone. My most recent book, Wicked Problems, came out last year, and is available wherever fine books are sold! And less fine ones too.