Tales & Trails of a Traveling Wiseguy

I've written recently that I'm getting a D&D adventure started. Sometimes, though, getting started is hard. Like, I have an idea of the theme or setting upon which I want to base the game, but I'm not sure what the story should actually be. Other times I've got the kernel of an idea, and it's elaborating it into a storyline with plot points and multiple encounters that's difficult. I figured generative AI could give me a hand at these challenges.

I used Google Gemini to assist with fleshing out two adventures. In one I described the basic setting and prompted "It should include undead among the monsters" and asked the AI to elaborate the major plot points and encounters of the adventure, and to detail the villain. In the other I described an initial encounter I imagined and asked what it might lead to.

In both cases AI was very helpful. It came up with creative ideas for encounters and summarized them as key points in a storyline. The AI even prompted me to ask it followup questions, like "What might be the villain's motivations?", "What help could a key NPC provide?", and "What are some unique magic items involved in the story?"

While the AI was helpful it also made mistakes. When I described this to a few friends recently, one jumped in with, "It's important to proofread what AI gives you!" That's true but it's not the problem I had. While we've probably all seen fails reposted online where a student copy-pasted an AI answer including the prompts, thus revealing that they were so lazy in using AI they didn't even read what they copied, there are failure modes in AI that go well beyond what can be solved with basic proofreading. These projects demonstrated that using AI requires you have significant domain knowledge to check its output.

The errors I caught were ones where the AI cited D&D rules and had them wrong. For example, it listed the wrong Challenge Ratings (CRs) for about half the monsters it put in the adventures. CRs are simple data lookups from monster stat blocks. It shouldn't be hard for AI to get them right. But they were wrong— and deadly wrong in at least one case. If I didn't know so many CRs by heart I might have taken an encounter with a recommended monster way too tough for the party.

In another instance, the AI assured me that the party of the 4th level characters (a detail I specified) would have key spells like Fireball and Cure Disease to overcome specific challenges. Well, those spells are both too high level for 4th level characters to get. When I challenged the AI on how 4th level characters would get such spells, it initially offered me a spirited— and completely bullshit— defense of its creation. When I challenged it a second time it admitted that it made a mistake.

"Okay, now go back and revise the encounters to correct this mistake," I prompted it. And, to its credit, it did! But the problem remains that I had to have significant domain expertise to fact-check what the AI was giving me.

I began a journey down memory lane yesterday when I wrote a journal entry about how my parents never liked attending the church of their faith that was right in our neighborhood. Instead of a short walk up the hill behind our house to the local church where we might see our own neighbors, we piled in the car and drove 20 minutes each way to a church in the next town over.

As I wrote in that blog, my parents were evasive about why they preferred the one far away. My parents, especially my father, gave only vague non-answers whenever I wondered. After a while I stopped asking.

The truth about the church up the hill came out decades later, not long before my father passed away. He knew he was in his last few months of life. He told me one of his goals then was to square things with relatives who were estranged from him. I wasn't estranged by any stretch of the imagination. I was traveling coast-to-coast every few weeks to visit and support him. During one of our bedside chats he told me the story. Well, not the whole story. He gave me just the one or two missing pieces that allowed me connect up the puzzle from other things he'd told me over the years and from things I remember from as far back as my own early childhood.

The story goes back to the mid 1970s when my dad lost his job as a store manger in a retail chain.



The mid 1970s were a tough time in the US. The country was just coming out of a deep economic recession spurred by the first oil embargo. The recession was probably why his employer folded. And even though the recession was over by the way economists define it, it wasn't over by the way ordinary people might define it. Companies were failing. Those that weren't failing still weren't hiring. The unemployment rate was above 7%. So when my dad's employer shut down and sent everyone to the unemployment line, finding new work wasn't easy. It took my dad months... maybe even a year or more.

By the way, yes, I'm using AI image generation to help illustrate this story. No, I don't have real photos to share from that time. I was too young even to hold a camera then. I mean, I was still filling diapers when this shit went down. And my parents never snapped many photos during my childhood. That always struck me as weird when I was older, because my dad had been a semi-pro photographer when he was in high school and college.

I saw some of his 1960s era work decades later. It was in a box from his mother, who'd just passed away at age 101. It looked good. He could have made it a career. Why did he put his cameras down and then not pick up another one for, like, 40 years? And also, his mom kept copies his vintage work as mementos; he never did. I might've asked him "why?" about either of those facts, but as I already explained early in this story, my dad was famously loath to answer such questions. In that respect he was like a perpetual pouty teenager giving guttural one-word answers.

Anyway, AI image generation. I'm using it here because I think telling the story with some pictures improves if, even if the pics are not authentic. For one, having pictures beats walls of text. Two, I've iterated on the prompts for these pictures to have them reflect, accurately, particular elements of the story. Of course it's impossible to have them accurately reflect everything, even the spotty parts I remember in snapshot memories from my early childhood. I've got a funny story to share about some of the prompting I had to do while creating an image I'll use later in the story. I'll share that anecdote when we get there.

To be continued....

There's a water leak in our condo complex. These happen frequently with the landscape irrigation system; squirrels and other critters chew chew the half-buried plastic pipes. This leak seemed a bit more persistent than a landscaping pipe, though. Water was leaking steadily, not just for the 15 or 30 minutes a day that the irrigation system runs. Concern about the problem led to a robust discussion in our neighborhood email forum.

"Supergirl has looked at the water leak", the HOA president assured us.

That was certainly an autocorrect mistake. 🤣 Our landscaper's name is somewhat similar, at least in terms of how autocorrect works, to "Supergirl". But I couldn't resist picturing this...



...with the help of Google's Gemini AI.

That's right, AI. The thing that's going to take all of our jobs in a few years. We'll be sitting at home, surviving off our unemployment checks— at least for the 13 weeks those last— but we'll be able to entertain ourselves by prompting AI to draw pictures making light of our woes!

I made that first picture with a simple prompt like, "Draw a comic book style picture of Supergirl as a plumber." I then refined it a bit to include cues about where the leak is in our neighborhood and got this: