I recently discovered the concept of a journaling RPG, which can be a tool for creative writing. In Act I of these posts I narrate this discovery in more detail. In Act II I narrate how I finally got around to playing one, called Borealis, by overengineering a Lean implementation. As promised, the third post in this series is a transcription of the journaling I wrote in response to the game prompts. (You can read just this post standalone if you want.)

I’ve added the intro text from the game, since that’s essential for understanding the story. I won’t explain all the rules of the game, but they determined whether a given story beat went well or poorly. During gameplay, you flip over cards and reference their corresponding prompt from the game booklet. You start out with two safe cards, the Three of Spades and Six of Spades, and I’ve copied their text here. I also copied the relevant prompts for the later cards I pulled, but I wrote the journal such that those are actually not essential to understand the story. In some cases I kinda ignored part of what the prompt said, to better fit how I felt the story was going — this is considered a good thing to do in journaling RPGs.

Normally I would do a bunch of editing, major or minor, to get a short story into a state that I was very happy with. But I think it’s fun to see how this one turned out exactly as it happened from the gameplay. So the below is not edited at all. I did a tiny bit of retro-editing during gameplay.

The transcript

“This day you become one with the land,” said Elder Mishan. Her wrinkled face looked stern as she repeated the words as she had for all those of the People around us both, just as had elders before her back to before even our oldest writings. “This day you walk with winter, and winter shall learn your worth. May Boreas watch over you and his wind bring you home safe to us.”

I glanced at my mother and father, standing proudly at their place in the circle, the brand of their own walks clear against the skin of their clasped hands. For a sudden awful moment my mind was empty, and I had no idea what I was supposed to say.

“I shall walk with winter,” I said, the words coming back to me, and I remembered my brother standing here on the altar stones reciting the same. “I shall walk in Khione’s care. I shall harm none, neither plant nor tree, bird nor animal. And by this sign shall the winter know me and learn my name and see my worth. All this I promise to you.”

There was a brief intake of breath from everyone. Promise made. And now I had to keep it. The circle parted and my boots crunched against the snow as I set out to find the Borealis.

[Three of Spades]

Your maelnra, or walking stave, is the key tool of the People. Sharpened at one end and carved with runes of knowledge and protection, it can be used to drive through ice, as a pole for a bivouac or skis, or to hold up a weary traveller as well as a myriad of other things.

[Six of Spades]

Your sealskin coat is lined with fur and treated with Elder Imelion’s finest unguents to keep the winter out and you inside, as warm and dry as possible.

As soon as my family faded into the distance I began to feel an immense sense of regret. Not regret in leaving, but regret in not having asked more about what I will experience. Somehow I suddenly feel like I have no idea what will happen, or when I will even know when I’m supposed to do something different. What if I just walk and walk forever, until I freeze? I already know that I won’t go back unless I succeed. It is who I am, but that doesn’t stop it from filling me with adrenaline at the thought.

With only my coat and maelnra, I try to remain calm and walk on.

[Jack of Hearts]

A roaring sound chills you to your bones as you realise with horror that ahead is an ice bear, and it is protecting cubs. The bears are formidable opponents themselves and can act in frenzied rage to defend a family: find a way to proceed, if you can!

My fretting over whether anything will even happen was immediately shown foolish. Though the icy world looks barren, we know it is not. It holds great danger, and there is little natural danger greater than that of an ice bear.

I know not whether she knew the People were close, or whether it is bad luck. But my journey got very desperate very quickly. She does not seem calm or healthy herself. Most animals know that men are a great threat to them, and will run away. But if they are starving, they may attack themselves. And attacking she is.

She runs toward me at a full sprint. My heart sinks deep into my stomach. I can run down a glazelle through persistence, but almost all animals are faster at a sprint than men. And so I know there is no running from her. I must fight.

I meet her challenge, and scream at full volume, bolting directly at her. If I’m lucky, she will be more scared than I, and stop her charge. I am not lucky.

She saw me from a moderate distance, and continuing to run at the bear requires me to renew my commitment over in my heart. Finally, after agonizing seconds, she is within spearing range. Mercifully a bear will not throw it back at you, so it is a safe try. I throw with all my strength, and foolishly, at the cost of control. My maelnra plunges into the snow beside her.

She banks left but keeps barreling toward me. In desperation I try to evade but it’s clear that I must make it past her to get to my maelnra. The hopelessness of my situation is so simple, there is little else to do. I run at her, and as we are about to collide, she jumps to land on me, and something in my heart tells me to dive. I jump under her and roll. The great bear, body emanating heat and musk, flies just over me. As I get back on my feet and bolt toward my maelnra, I realize her back claw has caught on my coat, and it pulls off me swiftly.

As I reach my maelnra and yank it out, I turn around to see the bear starting to trot away with my coat. She has dug her strong teeth into it. I throw again, injuring her, but the maelnra falls out, and she starts running away. It would be foolish to pursue. But I will not last overnight without warmth. A hundred yards away, I see her stop and chew up the coat. It is then that I realize she can smell the sealskin and oils of it. She must truly be starving to try to gain nourishment from it. And when she is done trying, the scraps will be of little use to me. I grab my maelnra again — now the only tool I have against the world — and trot in the opposite direction. I must test my luck again, for the other option is self-sacrifice.

[Ace of Hearts]

The snow grows deep and cold. Draw upon your experience of the ice lands to proceed.

Some of the People find rune-working the most natural thing, more natural than nature. For myself, it’s easier for me to forget that they even exist. I feel that my strength is my body, my strength is my cunning, and the trust that I have built with my family. If I am honest with myself (and I must be, here in the land of Boreas) I distrust the runes. I have seen the runes fail those I love. We have gone hungry because my Uncle trusted the runes, and we lost our third Brother because the runes deceived us. I cannot trust them as I strive to be trusted.

But I am also a fool. I am out here, essentially naked with just a sealskin and stave. Good workings, made for me with love by my trusted ones — but they are not enough. I forget the runes that make my stave a maelnra. I forget the most powerful tool I have.

“Remember,” my mother said, “the runes work not on belief, but on possibility. Do not convince yourself of something unknown to you. And do not forget to acknowledge the possible.”

She said this to me just before I became a man in body. (For I have not yet become a man in spirit.) It was an unusual thing for her to say; though she applies runework daily, she rarely speaks of it. Instead, her work speaks for her, in the way that it keeps her family warm and protected. But that day she expressed that wisdom which I will never forget. Or at least, there will always be a time when I recall it. That time will not always be the fortunate one.

So with the feeling of the rune lines under my freezing palm, I continue walking forward, and open my mind to the possibilities. I search the horizon, I look at the snow at my feet. And I wait.

[Seven of Diamonds]

Huddled in the snow is the frozen form of another traveller. You hurry to their side but this person is long, long gone; another walker with winter, perhaps, but less successful? And then as the frigid form turns its face to you and its eyes flood with a chill blue light you realise that this is how it hunts…

And for my watching, the runes do not reward me. Over the next drift is a huddled body in a tattered coat, not like mine, but one much older. Out of instinct I run toward the Person. No doubt they are long dead. I see in my mind the possibility — is this from the runes? — of taking the coat. At least then it could serve the living. As I approach the body, I place a hand on its shoulder. Suddenly it falls over and my heart skips a beat. No harm done, but I silently chastise myself.

Then I see the eyes. The eyes of the body are blue. Not the blue of eyes, but the blue of the glaciers, vibrant and threatening. I scream, jump back, and run away. No good can come from this disfigured body, if it was even a Person to start with. I run away until I am over the next snowdrift, out of sight of the pale blue eyes. I cannot deny that no harm was done to me, but I do not feel more inclined to trust the runes.

The runes under my palm. Except they’re not. I realize in this moment, in what must be my lowest point, that I have left my maelnra where the body lies. How could I have forgotten my only remaining tool? If I ever see my family again, it will not be due to my worth. I must go back.

[Ten of Diamonds]

Upright in a pillar of ice you find a sword, a weapon of war, designed for killing. Tentatively you touch its hilt, untrained in weapons, when a voice sounds inside your mind. “Pick me up! Wield me! Kill with me!” and you try to thrust the chill thing away as it batters into your thoughts, seeking to make you its unthinking warlike slave. Fight it off to continue!

I make my way to the top of the snowdrift and stop. I see a body. But it is not the body I saw before. I see a weapon. But it is not my maelnra.

What I see is the ice bear. Her coat is my sealskin. That which was crafted by my trusted, I have left in tatters. And into her body is plunged the maelnra, my maelnra, but it has turned to solid glistening ice from my betrayal.

This is not my lowest point. That has already happened, when I forgot the promise I made, less than one day ago. As the sun now sets behind me, sealing my fate, the words echo in my mind. “I shall harm none, neither plant nor tree, bird nor animal.” My first action was to harm the first living thing I saw. On that first encounter, I had forgotten my promise, and I had forgotten that there is always another possibility. I had forgotten the runes.

The winter now knows my worth, and my tears freeze to my face. I fall to my knees, and my wet face binds with the snow, the snow I have tainted with blood. This day I became one with the land.

FIN

Postmortem

During this game, I got quite a bad deck shuffle and pulled only challenge cards. This makes the story kind of brutal, but still pretty coherent. I would almost never write a tragic story like this, so being forced to do so was a good exercise for me.

On my first prompt, I legitimately forgot about the no-harm promise when I chose to attack the ice bear. The fact that this incepted an original sin that could explain my ultimate demise felt so lucky.

On the last card, I had lost the game according to its rules, so I had to end the story there. Instead of introducing all the new details from the prompt into this story that I knew was about to end, I mapped them onto what had already happened in the story.

Overall this was super fun. I won’t likely play this exact game again, but am eyeing several more whose setting aesthetics are more intrinsically interesting to me.

See Act I for an explanation of journaling RPGs.

I wanted to play a journaling RPG but didn’t have a deck of cards with me. Should I ask Claude for help?

It had certainly already occurred to me that I could get an LLM to just make up a cool custom-themed journaling RPG for me. But I didn’t really want to do that. I wanted to have an experience that was created for me by another human. Now though, it was occurring to me that I could have Claude write a simple program that simulated tracking the state of the cards as they would be arranged and manipulated on the table. Then I would still be playing the game that was created for me by another human, just aided by the tool that is the LLM.

It further occurred to me that I could wrap in another one of my current interests, which is Lean, the programming language. Lean was designed for formal verification and is used mostly, at least as of the time of this writing, for proving theorems in mathematics. But it is a fully-fledged programming language, and it can do anything any other programming language can do. In theory you could build a phone game or the next social media platform in Lean.

I am currently extremely enthusiastic about learning Lean, except that I haven’t really done that much of it yet. The reader might see parallels between this and my relationship history with journaling RPGs. But I thought if I can have Claude write this game simulator for me in Lean, then I can use that tiny, self-contained codebase as an object of study for better learning Lean. And also, for finally playing a journaling RPG.

After a few backs-and-forths with Claude, he produced a minimal spec. I then spent several hours reading through and trying to understand each line.

And this worked! I got a bunch more experience reading and understanding Lean syntax for a project where I knew exactly what it should do and where I was motivated to make sure it worked correctly. I learned about lots of little quirks of Lean, like the fact that subtracting natural numbers truncates at zero, or that the set of numbers from 0 to 5 is its own type, Fin 6. (The 6 isn’t special; every such finite set is its own type.)

But more than that I got a little bit more of an understanding of what formal verification looks like. I had asked Claude not just to implement the game but first to write out a behavioral spec for the game’s rules. This is a necessary stage of formal verification and there are parts of it that can feel counterintuitive on first exposure. The specs can sometimes feel like a copy of the implementation, especially with the simple examples one must pass through while learning.

A common class of thing one proves in formal verification is called an invariant, which is some property that holds on every step. For Borealis, the game I chose from the first act, an invariant would be that no card appears in the deck twice, or that the pawn is always on a face-up card. If you prove these properties for the initial deal, and then prove that every legal game action preserves these properties, then by induction you’ve proved that they will always hold true during the session.

After I read through the whole spec and was satisfied that I understood everything, I asked Claude to actually implement the game. This request was, as often seems to happen, over-eagerly responded to, and Claude pumped out another thousand lines of Lean code. Having worked my brain pretty hard to read the first several hundred lines and it being close to midnight, I just let out a big sigh, closed my computer, and went to bed.

The next morning I got up and asked Claude to redo the implementation work, without the features I didn’t ask for, and keeping the Git commit history more fine-grained.

After that came the play testing. I was pretty sure Claude got all the game behavior right on the first try so I wasn’t really testing the Lean implementation. Instead I was back in the mindset of being excited to finally try a journaling RPG and see how I liked it. The structured procrastination had done its job beautifully, and it was time to get down to the real work: playing.

It went so well that my journaling of the session came out as a basically fully coherent story. This has… never happened to me before? To be clear it’s not a story that I love. It does not, for example, treat themes of rationality, cultural heritage, or anything else that I would deliberately write a story about. But I like it, and the fact that it came out so easily on the first try was a pretty strong effect size.

In fact, I liked it enough that my play session journal is the third post of this series.

I don’t remember where I first heard about journaling RPGs but the concept was immediately appealing to me. Nonetheless, it took me over a year before I actually played one. This is my story. (It’s mostly procrastinating.)

RPGs (Role Playing Games, like Dungeons & Dragons) are also appealing to me, but they are a very intensive time commitment. And not just you, but for a whole group of friends. First you have to find a group of people willing to commit who will work well together, and then you have to schedule a recurring event for many weeks. You usually need full attendance for the game to move forward.

I personally am also not particularly attracted to a bunch of common features in RPGs, such as the grungy dungeon aesthetic, or frequent combat. I have played one-shots (or three-shots) with friends before and always enjoyed it, but it’s not something I would go out of my way to make happen.

Journaling is also very appealing to me and yet somehow I’ve never been able to do it. Fiction writing, too is very appealing, and despite the fact that I can generate detailed plot outlines and sophisticated world-building, I have never managed to get down more than a handful of very short stories.

What is a journaling RPG?

Journaling RPGs are a subtype of solo RPGs. So let’s back up a step and define those. A solo RPG is any game very similar to an RPG, but which is designed to be played by one person alone. When RPG players first hear this, it often sounds oxymoronic or pointless. But it turns out they work!

Solo RPGs seem to have had an explosion in popularity over the last 10 years or so, especially in online spaces. Ironically, though (group) RPGs are ostensibly more social by virtue of being a group game, solo RPGs manage to be very social via the hobby community spaces dedicated to solo RPGs, like Discord or Reddit.

Now for the journaling part; because all RPGs require a degree of creativity and informal gameplay (and not just following a rigid set of rules as in many games) they naturally have some overlap with creative writing. A journaling RPG is just a solo RPG in which writing down the fictional content or ideas produced during the game is a large part of the activity and purpose. They can span the full spectrum from what is little more than a collection of writing prompts to what is essentially a solo RPG where you write down what happened.

Window shopping

Many solo RPGs are free, have easy-to-understand rules, and can be played through in just a couple hours. So naturally, I proceeded to research them for several hours.

I googled things like “best journaling RPGs” and “which journaling RPG to start with”, and opened up three dozen tabs. I watched videos with advice, reviews and opinions. I fawned over the beautiful cover art. I hovered my cursor over the “buy” button on some gorgeous physical book editions. I learned about the different subtypes, the award-winning games that kicked off new rule sets, and the games everyone disagrees about. I determined the optimal list of games to start with, given my personal preferences and prior experience with RPGs.

Finally, at long last, I chose Borealis, a game with minimal rule complexity, and which promised a playtime of one to three hours. I bought the PDF.

And then I felt satisfied.

Second wind

Over a year later, I’m traveling away from home for a while. Apropos of nothing a friend mentions they’re playing a journaling game that reminds them of me. I remembered that journaling RPG that I’d downloaded and decided to try it.

The only problem was I couldn’t remember what it was called or what folder I had put it in. I searched in my file manager. I searched in Dropbox. I searched my Gmail. I asked Claude if he had any idea what game I was talking about. Alas he did not.

Eventually I gave up and do a second round of window shopping.

Amusingly it had been long enough that I had forgotten much of what I learned the first time, and this round of window shopping was just as fun. I even went to a brick-and-mortar game store (they only had one). Eventually I found two journaling RPGs that looked really pretty and had themes that appealed to me: Artefact and Tales from the Gods. But I didn’t start either one. They seemed like they would take a while or something. Instead I bought a hard copy of Tales from the Gods, and watched some more review videos.

Eventually, I realized I could search my email for the name of the stores that people use to sell their solo RPGs. Then I was able to find and re-download a copy of Borealis.

Freed from my indecision, I eagerly opened the PDF and began reading the rules. After reading the pretty introductory flavor text for the story, I got to the “Materials” section, which told me that I needed a deck of cards. Given that I was traveling, I did not have a deck of cards nor did I have particularly easy access to one. I kept reading through the setup and play rules, hoping that I could maybe just use paper or use some kind of online card deck simulator. But it quickly became clear that the cards were being used in a very physical way that would not be practical to simulate.

Or would it? For we now live in the great age of LLMs, and they are supremely good at coding.

Act II: Structured Procrastination

From a civilization I’ve never heard of

I was immediately drawn to the stele, in a way that I always am but most people aren’t. It’s a flat rectangular slab, totally blank except for some random text on it you can’t read. Why would that be attention drawing, when the rest of the room is filled with a variety of colors, 3D shapes, depictions of animals and human expression? This particular slab is not even decorated. Not the slightest bit. The only thing that saves it from being the monolith from 2001: Space Odyssey is a very slight taper.

But exactly what excites me about it is the nature of writing itself, the intense abstraction it involves, and the motivations people had for inventing and learning such a craft.

When I approach to read the object label, the one part I can understand, it says “Marble stele with a Lydian inscription”. If I’ve ever seen “Lydian” before, I don’t recall it. Sometimes a category will have few enough things in it that one could reasonably learn what they all are, like countries in the world or taxonomic phyla (not that anyone would…). Sometimes a category has unreasonably many, like the number of cities in the world or taxonomic genera. And sometimes it will have just enough that you always feel tantalizingly close to knowing them all but never quite get it. Writing systems are in that third kind of category, for me.

While the Lydian script is understood to be an alphabet similar to Greek, the label goes on to tell me that the Lydian language itself has not been fully deciphered. So not only do I not know that this stele says, but in fact no living person on earth knows.

(It is in this moment that I wonder whether Large Language Models have been trained on this particular inscription, and the Lydian corpus as a whole, and whether, in the vast latent space of the model’s parameters, it has perhaps already arrived at a representation for the text that lies close to its correct semantic meaning in English. I could not myself investigate this question, but those working at frontier AI companies could. Thus it may be that some non-human on earth knows what this stele says.)

I stare at the inscription for a while. It is very cleanly rendered and preserved, such that if I wanted, I could easily identify the character set and transcribe it. I see a few instances of a character that looks like the Chinese character for king, 王. (Their similarity is of course entirely coincidental.) In this inscription it appears doubled, 王王, most of the time, but not always. I see another distinctive character that looks like if a tilde was a three-leaf clover instead of two. And I see another that is likely the ancestor of the letter A. I start searching for repeated sub-strings, to identify words repeated in the text. I find both a repeated word, and words which share a prefix. Maybe they’re different verb tenses or noun declensions. In a short inscription, repeated words are either extremely common like “the”, or they’re what the inscription is about, like a person’s name.

This is all very pointless of course, as academics will certainly have identified all such frequency patterns long ago. But I get the same joy out of it that one does from doodling a frog, even though humanity has mastered the art of depicting frogs. It’s also quite satisfying to know that, if I kept going, and tried with patterns that weren’t just sub-string frequencies, I would in theory converge on a full understanding of the content of the inscription — at least as far as remains objectively present in the corpus.

Almost all writing in the world is incredibly boring, repetitive, distasteful, or inane. This also goes for stelai, which are quite fond of telling you who owns what or how great the current king is. But I still find myself intimately in contact with a level of appreciation beyond what the authors of the inscriptions choose to write. Inventory tablets are brilliant in part because the sheep could never count themselves. And it’s better for the king to declare his greatness by word, thus dissuading potential pretenders, than to be forced to prove his greatness by blows.

This week I am reading the much-anticipated new novel by YouTuber, unpaid intern at Good.Store, and best-selling author, John Green.

Although John is one of my favorite people, he’s not actually one of my favorite authors. I adore The Anthropocene Reviewed, but I’ve felt only lukewarm about his young adult fiction. To be fair, I read them in my 30s. I don’t read his young adult novels to enjoy them, but instead as an act of love for a person who I deeply respect. But Hollywood, Ending is not a young adult novel.

I’m only going to talk about the first chapter here, which is 4 quick pages. If you want, you can listen to John read them to you, as is his pre-release tradition. I did so, but when I got the book I reread them anyway, and somehow noticed several more things.

I’m in a phase of my life where I’m annotating books by default. It’s not very natural for me to annotate — indeed, to write things down. (If I’ve already had a thought, why do I need to write it down? If it was true or useful I’ll probably remember it, or regenerate it next time I read this.) So usually my books end up with just a handful of annotations scattered across three hundred or so pages.

But with Hollywood, Ending (or H,E as the Nerdfighters are calling it) I seem to be doing something closer to Bible study. “If this sentence had a secret message for me, what would it be?” or “Didn’t so-and-so say something that contradicted that earlier?” or “Why did John choose to tell us this detail, in this place?”

Here are some things I noticed in the first chapter. The book is about some actors during the filming of a movie about Andy Warhol. So while the book isn’t about Warhol, he’s a major presence.

• The very first line gives the very first name in the book, Matthew Malone, which is the actor playing Warhol. I noticed that “Matthew Malone” sounds a lot like “Marylin Monroe”, the subject of one of Warhol’s most famous works.
• On the second line, we get the name of the movie; Andy Warhol Never Gets Old. Presumably this refers to the fact that Warhol has remained iconic. But (I see from reading Warhol’s wikipedia page) Warhol died at the age of 58. This is arguably before “old age”, so he arguably never got old.
• The introduction of a character on the second page sprinkles the paragraph with color. Andy Warhol’s work is known for being very colorful.
• Two pages in, and I notice that there have been a lot of brand names: Uber, Altoids, Moleskine, AirPods. Warhol’s work is known for bringing branding and consumerism into the realm of art, and causing us to question their presence in our lives. Is Green doing the same?

This is a pretty fun way to read a book. I crank up the pattern-matching, puzzle-solving part of my brain.

I’ll just give you one more; the first-person narrator, Juniper, is not unambiguously gendered. There are plenty of indicators of Juniper’s gender, but no one says a pronoun, nor “girl”, “boy”, etc. But we also don’t get any indication that this book is about transness, or even gender. John tends to do some kind of easter egg thing in each of his books, so I can’t help but wonder if the easter egg for H,E is that Juniper is never gendered.

I’m thrilled to read this book, and I hope you will too. It’s been a long time since I’ve gotten to participate in a book release as a cultural event.

On the way to the Met, you get a good slice of NYC atmosphere.

Let’s say you take the 4, 5 or 6 train to the 86th street station. Once you find your way up to street level, you’ll be greeted by some sidewalk sheds, and then you’ll have to figure out which way is west. If you can see the super skinny-tall 432 Park Avenue, that’s south. Walk one block west, and you’ll cross Park Avenue itself. It’s a two-way street, and it has a big median down the middle that’s filled with trees. I’ve always loved the way this double-wide canyon opens up to show more of the city in one view.

You’ll know when to turn left, because suddenly in front of you will be all green, with blue sky above. As you walk down 5th Avenue, you have a charcuterie board of Manhattan architecture on your left, and the vast greenery of Central Park on your right. If it struck your fancy, you could step into that park, head down one of the many footpaths, and spend a lovely evening lost among the trees and rocks and fields nestled inside one of the highest population density areas in the world.

But we’re here on a mission, so you keep straight. The sidewalk below your feet is oddly something resembling cobblestone. It’s pretty, and not quite made for walking, and it does a good job indicating that you’re not on just any sidewalk.

As you get closer, the food trucks get louder and more numerous. You’ll find the New York classics like hot dogs and gyros, and plenty of variety besides. You’ll also encounter the rows of people selling their own art. It’s never been clear to me how many of these stalls are genuine hand-made crafts made with love, and how many are tourist traps designed to feel like hand-made art. But I have a fondness for them nonetheless. Long ago, when I lived in New York, I considered spinning up a little art stall myself. I had some clever ideas for making art from wires, combining an aesthetic of industrialization with a feeling of optimism. But I’m not a doer, so I never did.

When you’re almost there, you’ll see a little playground. And wouldn’t you know, it’s a Met-themed playground, complete with pyramid and obelisk. Whenever I see it I think, “I should go check out that playground” before I realize that solo adults aren’t really supposed to do that.

Then you cross the next street. It’s kind of two streets, actually. The first one is just a street into Central Park. But the second one is 30 feet of road that then dives, understatedly, half a floor under ground level, into a dark little cave that is clearly not meant for you. I like to imagine that all the wonders of the museum have passed through this threshold, tucked away in an average-looking van or truck. In reality there are probably several loading areas into the museum. This is probably one of them, but I’ve never investigated; it feels like you’re not supposed to look directly at it. Just mind your business, and cross the street. Nothing to see here, just city stuff.

And now you’re reached the eastern-most cornerstone of the museum. Just on the other side of that limestone are some truly ancient treasures — the Egyptian wing. To get in there, you’ll have to walk a ways further, to the entrance at the top of the steps. But you wanted to do that anyway; they’re a worthy sight of their own.

This is my post for day 30 of the Inkhaven writing retreat.

If you’ve ever played video games, you might be familiar with the concept of a mini game. Mini games are a separate, smaller game within a larger game, with different rules and a well-defined context. They can add more novelty to the player’s experience while not over-complicating the game as a whole. But they also reflect how life works sometimes.

Our daily lives just so happen to be well-modeled by using lots of different mini-models that fire off in the right context. You only need to use your driving skills when you’re in the car, and you only need to consult your model of shoe laces when you’re tying your shoes. Another example is the whole phenomenon of buildings being made up of rooms. When you’re in a room, you mostly only need to think about what’s in that room. You mostly only need to know how the rooms relate to each other by means of walking between them. Only rarely do you need to know exactly what’s on the other side of a wall, or what room happens to be above your kitchen.

One of the miracles of modernity has been the slow but steady process of discovering and understanding how all of reality in its multitudinous richness emerges from one simple model, one set of equations that turned out to be so dominant and unwavering that we decided to call them laws.

Though these laws are in some sense the most true model we have, they are not the most useful in any practical sense. We don’t use the Standard Model to fix our cars or mow the lawn, nor do committees use them when choosing zoning or fire safety policies. We don’t even do engineering with software that runs quantum field theory, except in the cases where the purpose of the engineering at hand is to further probe quantum field theory. Our best modeling software is highly optimized for the context, using simplified laws like Kirchoff’s rules or classical Hamiltonians.

The more “unified” models of our world have higher prediction accuracy if you are willing and able to trace out all their implications, and they are the only models that have a chance of getting the answers right at the edges between contexts, or at the extreme values. But that “if” is a huge one, and I think it’s essentially never true. We have a lot of spare compute in the world, and if we could fix cars better using quantum mechanics I think we just would.

People are constantly doing predictive modeling in their brains in order to navigate the world, and we are essentially stuck using the highly contextualized and simplified models. An enthusiastic foodie might use some domain knowledge from biochemistry to design a recipe, but once they do, they’ll essentially cache their calculations into a “if I’m cooking X, set the oven to Y” type of rule. If you ask them why they’re doing that, they might be able to recover their original modeling as the justification, but they also might not.

That said, I believe that most people could move their minds substantially more toward unification than they do. I mostly do so because I value having true beliefs for its own sake, but I have noticed that over time I tend to get better results in predicting related outcomes than my more contextually-modeling friends. We are making different trade-offs, and they are better than me at many more specific skills.

I do think that a unified mindset is especially important in epistemics, since this isn’t domain modelling so much as meta-modelling. I think having a deep understanding of the Bayesian worldview can help someone in all areas of life, personal, interpersonal, and impersonal.

All that is to say I am the kind of person who knows what room is above the kitchen, because I subconsciously made sure to model that when I moved into the house.

I think that if you want to be someone who is highly skilled and adaptive in a wide range of contexts, if you want to go around the world navigating it like a video game and solving lots of problems, then you probably shouldn’t prioritize this kind of mental unification. Being this kind of person is extremely valuable, and I salute you.

If instead you want to try really, really hard to get the right answers to the world’s hardest and most important problems, then I think you need to be the kind of person who focuses on unifying your models. I see existential risk from AI as one of these hardest problems.

Without a doubt these types of people need to work together. The contextual modelers need to be willing to trust the unified modelers’ claims about things like overarching principles and limitations. And the unified modelers need to be willing to trust the contextual modeler’s claims about how specific domains work. The contextual models have lower leverage, and the unified models have a higher probability of being catastrophically wrong. I’m not sure who should be in charge of when exactly we trust who exactly, but it’s somewhere in between.