Hell on $5 A Day: Getting Some Mojo Back

As some of you may know, Hell on $5 A Day is not a “standalone” book… per se. It was originally conceived to stand on its own (as in you could read it and have a complete experience without there ever being another book), but as I dove into the world-building and got to know the characters better…

I mean, when Kurt fell through the portal, I didn’t know he was as special as he’d turn out to be. Honestly, sometimes I didn’t know if I was telling the story or just transcribing it as it told itself. Still, the deeper down that fictional universe rabbit hole I went, the more I wanted to explore. Currently, Hell on $5 A Day is now the first in a decology (10 novels).

There are currently, in various stages of planning/writing:

  • Sodom All Over Again: Alain, Marie, and a new group of friends must stop a brewing battle between angels that threatens to destroy Los Angeles in 1946. Marie takes on a much bigger role here and really comes into her own, plus I get to have fun writing a story in post-war Los Angeles, the stomping grounds of film noir’s greatest detectives. I expect to be done with this in late 2025.
  • New York, by Midnight: Alain, Marie, some vampire friends, and a handful of Beat era luminaries must stop a plot to bring about the end of the world against a backdrop of the Cuban Missile Crisis.
  • The New Defenders of Old (series working title): A three-volume YA series spun off in a way from Sodom All Over Again, but set in a more modern day Los Angeles. Still trying to decide if Kurt stays in NYC and lives his life, makes a cameo, or takes a more significant role as a mentor to the kids.
  • Kevin Be Praised (series working title): Hard to explain. While Hell on $5 A Day started with the thought of a “coffee klatch” becoming a “coffin klatch,” this started with the thought, “long ago, when the world was still young, the spirit of darkness killed the gods and stole their magic.” I’m still early in fleshing out what that means, but it’s fun, will have 3 novels and will flesh out the multiverse.
  • Multiversal Healthcare: All the favorites from the prior three series unite for a final battle against a cancer trying to wipe out the multiverse.

With all that in the hopper, I started wondering if I really wanted to really write New York, by Midnight. It probably has fewer notes and less character and plot development than New Defenders and isn’t as new and sexy as KBP. I started talking myself into dumping it, even though it’s the novel that explains why Vinnie says “this one’s for ‘sixty-two” during the battle in Pandaemonium.

So I went to review what I’d written so far, looking at my notes, and I found myself inspired, outlining the origin story of one of the primary characters, Midnight herself. It gave her so much more dimension than my brief intro and notes, made her backstory way more compelling. And when I started diving deeper into her relationship with Avery Stark (who confronts Kurt at the end of $5), I really got into their love story. I mean he loves her more than Alain loves Marie. And now I am not at all interested in skipping this book.

And my AI assistant

You’ve probably seen my article testing the reading comprehension of 15 LLMs. I did this because I wanted to use an LLM to save me from having to do a LOT of background reading, but instead give me just the information I needed on everything from the Paris sewers to how to create a convincing stage psychic’s act. I thought I’d have to pump one full of a bunch of books and maps to get that, so when so many did so poorly, I was a bit forlorn.

But do you know how much knowledge a 70 billion parameter LLM has? Neither do I, it seems, but it’s a LOT of information. Would it possibly know about the sewers of Paris already? Would it know how to perform a cold reading? Yes and yes. Many of the larger models gave fairly credible answers.

This is sort of frightening, though, because cold readings are what many stage psychics use to convince you they have a message from your dead uncle. If an LLM could explain one to me, could I create an AI agent that convinces people it can talk to the dead? Talk about a ghost in the machine.

But that was a good thing overall. I’d been feeling disheartened about whether an LLM could help me as the research assistant I’d always wanted, but it looks like (for the most part), it can. If I need it to absorb really specialized knowledge, maybe not, but it’s got enough information in that 40-gigabyte model to answer many of the questions I’ll come up with as I further plot and write the next books.

So, no guarantees, but I think I might be able to write one a year despite all the other fun projects I have planned. I’m feeling kind of stoked. Hope you will be too.

Add a Comment

Your email address will not be published. Required fields are marked *