☞ Microcosm Industries
On simulation toys and software microcosms
Longtime readers know that I am obsessed with simulation toys: the kind of software that allows you to play with a miniature digital world. Think SimCity or SimEarth or the like. I’ve written before about the software company Maxis and its creations, the construction of emergent microcosms, the weirdness of Dwarf Fortress, and even explored some of these ideas in my recent book The Magic of Code. Given my love of this genre and my love of lists, it shouldn’t be that surprising that I now have a new website for this: Microcosm Industries. From the about page:
Microcosm Industries is devoted to fostering simulation toys: software that allows you to play with a complex miniature world. These are not quite games and not quite computational models. They are simplified but still complicated enough to yield surprise and excitement, with an aesthetic that leans more towards delight and construction than slick visuals and overwhelming details. These are open-ended building toys built in a computational medium, often drawing magpie-like from fields as broad as biology, urban design, and physics.
Microcosm Industries is currently just a list and clearinghouse for these simulation toys (you can even filter by those that are immediately playable in the browser). But I want it to be more! I think this aesthetic needs to be supported, so if you are working on something in this space, please let me know . And if you are interested in connecting to like-minded fellow travelers or want to be otherwise involved, please also contact me. I want Microcosm Industries to make this space of simulation toys into one that is as strong and vibrant as possible. If you are interested in following along with this journey, you can also sign up for its newsletter.
To quote again from the about page:
Microcosm Industries is devoted to the notion that this type of software was not just the province of a specific gaming moment in the 90’s. In fact, in many ways, these simulation toys are easier to build than ever, due to advances in processing power, AI coding advances, and even better datasets. We are entering a golden age not only for simulation and modeling, but for the simulation toy.
Let’s make this happen. ■
Speaking of lists, my Tech of the Rings list keeps on growing, from Gamgee and Iluvatar to even Mordor(!) as names of tech companies that originate in the world of Tolkien. Check out the complete list.
The Enchanted Systems Roundup
Here are some links worth checking out that touch on the complex systems of our world (both built and natural):
🜸 The AI people have been right a lot: “Try to keep an open mind as the world gets increasingly wild.”
🝤 Your Backpack Got Worse On Purpose: “In 1986, a corporation that made women’s lingerie bought every backpack brand you’ve ever trusted.”
🜚 Recreating the Canon Cat document interface
🝤 Extrapolated Futures Archive: “Mapping real-world scenarios to the science fiction stories that explored them first.”
🜹 The Ever-changing Art of the Screensaver: “Once central to our digital experiences, screensavers now hold a powerful place in our nostalgia for early computing”
🝊 I put Sefaria on a Gameboy: “Digital Torah needs better form factors. AI can help.”
🜸 Are “Vintage LLMs” the start of a new humanistic field? “Thoughts on Historical Language Models and Talkie-1930”
Until next time.
Lists shall cure listlessness!
Puns aside, I've found resonant inspirations from e.g. David Ha's interactive papers & experiments https://blog.otoro.net/archive.html
I'm intrigued, I want to believe in this kind of game that's like a tool for exploring or experimenting on practical/intellectually fruitful models, and I'm guessing that deep down that might be where your interest in this stuff is coming from as well, the hope that computer models can tell us useful things. But you've gathered a list of things that only superficially resemble that, and decidedly aren't really that. I'm *achingly* interested in cities but Cities Skylines bored me to sleep, it can't model really any of the questions I have about cities, it hard-codes too many assumptions, and I've heard others say this too. Factorio can teach you about industrial systems, but afaik, not through simulation, more through training. The wisdom contained within the system seems to be a subset of the wisdom of its author? Which is not what we want.
I think Golly has driven intellectual progress, but even there, few people had to play with Golly for the rest of us to learn the computational metaphysics insights it contained. It's not even totally clear that the conway rules needed to ever be implemented on a computer, the whole thing might have worked as a thought experiment.
So I suspect the thing we're looking for is mostly just going to be... a kind of real, but lightweight, simulation software, that's simple enough that people who aren't experts in the field they model for can use them through an LLM, and use them exploratively (fast), and ideally plug them together in surprising new ways (so, interoperability), plus a suite of *culture* for using simulation software in highly explorative ways, including a name for the demographic who use it, and public knowledge that this demographic reliably turns up whenever authors put in the effort to optimise their stuff and support the relevant standards.
That's what I think we should build. Maybe. It would certainly be neat.
Right now, intellectually interesting simulation software is more often spectated than played. Would you agree our proposal is to tilt the ratio for serious simulation software towards play, because there's a lot of valuable insights buried in them that requires a whole lot of fucking around to find? And right now not enough fucking around is happening?