Greetings in Christ Jesus.
Fledgeling is an AI powered emergent story universe sandbox.
I've always been fascinated by procedural generation, the systemetization of the process of creativity, of which the modern day generative AI is an offshoot. In 1997 I saw a 3D vase on the box of RayDream Designer, and thought how cool it would be to have the computer generate an infinite variety of vases. Then in 2006 I began playing Dwarf Fortress, and procedurally visualizing the procedural content struck me as a great challenge. Also around that time I wrote the Auto Masonry code that would go on to be packaged with Blender. These were the seeds.
But it only began in earnest at the beginning of 2010. Minecraft had just launched InfDev, I was working on procedural generation code for the big trees, and procgen was on my mind again. Around this time my wife bought me the set of official schematics for the Enterprise D from a garage sale. In the booklet that came with the set was a note about how, ostensibly given their advanced computing and fabrication technology, the "engineers" in the Star Trek setting could push a button and have their supercomputers design a whole spaceship, and then have a shipyard-scale replicator print out as many copies as they wanted. The booklet argued that this would take all the fun out of it. For me, it was a revelation.
What if we could make a computer program that would generate games? All the content, all the 3D models and storylines and dialogue and mechanics. This would be, in my mind, a real artificial intelligence, capable of accurately compressing the interesting things about the real world so that both AIs and humans could productively play with them. Such a program would be able to leave the nest of our technological civilization, and begin to soar on its own.
Even more ideally, this would be done without human effort. I predict that, just like human voice calls ride free on the back of machine data, human video games will ride free on the back of machine learning. We will play the games that AI has made, and that AI plays. It will be built without hands. This is a foretaste of "the city with foundations, whose architect and maker is God."
This is Fledgeling.
I continued to ruminate on this concept, and in 2012 I started a blog where I rambled more or less coherently about video games. The blog is losing the battle with webrot, and I'm trying to port the best of it over here, but I figured I'd link it for archival purposes anyway. In order to rise again every aspect of all good things must descend, like Christ, into Hell.
Fledgeling is a generic Roleplaying System with a built-in GM.
As an excercise, I filled out a form to create a Fledgeling Game Design Document. This was around 2012 as well. It's a bit dated, but I'm overall happy with how well the ideas have held up over the years.
I used the term Simulated Intelligence (SI) because, at the time, AI was not proven possible. Perhaps it still is not. The line "Computation required may outstrip current hardware capabilities" is of course always true.
I also describe the visuals as "shockingly simple" as this was before the advent of visual generative AI. I think this is still a good idea, to have a "plain" or "schematic" or "debug" visualization option, to allow the most unimpeded and unembellished interaction. This would be useful for cases where performance is more important than immersion, competitive games for instance.
The "Typical Gameplay Example" I describe is sketched out in a demo choose-your-own-adventure style game which, although incomplete, may give you an idea of what I'm thinking of. See Gerbil Journey for more thoughts on Multi-player, which I suspect could be carried out a-synchronously.
The point about "indirect control" is crucial. This is the move away from micromanagement (which computational systems are good at) and toward direction and integration, pattern matching against the larger whole (where humans excell).
I'm particularly proud of the line (written in 2012), "If a character perceives the object of one of their hopes [aspirations or fears] they will immediately begin to create and act on a plan to achieve or avoid it." since it forsees a narrative-based driving principle now (as of this writing in 2024) elucidated by JBP and other public figures.
Although I did not forsee LLMs with particular precision, I do touch on the linguistic aspect of AI in the paper as well. In discussions at the time with friends, they raised the question of how to translate between these different languages. The answer now is that tokens are the sub-layer of a linguistic model, and that what I am talking about here are different mappings of the token space onto the user-facing (and that facing other agents in the simulation) textual output.
The point about Time Contraction and being able to fast-forward until a crucial point of action was prescient as well. Around four years later KSP implemented this system to a limited extent. It seems that many other games would benefit from this mechanic (or is it an interface element?) but at least this is another example of being on-track.
Everything from the "Controls" section to the end can probably be safely ignored. The "design document" is not really detailed and fully realized enough to merit decisions to this detail level.
I've envisioned this as a game with huge scope and deep simulation, based on interconnecting procedurally generated systems. That pretty much means it would be a 'real' sandbox. Not a fake sandbox (like Spore) but something more like Civilization or Sim City and Minecraft and KSP all working together. I'd like a game where you can design and build a fleet of working starships, arrange trade and alliances, and/or go rogue, build an asteroid base, and hide from everyone. Many simulation games simplify away complexity. I'd like to have a game that allows you to intelligently automate away complexity.
My current best estimation of the playstyle is that of a 3d turn based strategy/tactics game with vastly variable turn lengths and size scales.
Simulation games have been hacking away at the challenge of emergent story for a while now. Tarn Adams of Dwarf Fortress fame is working the rogue-like angle. The challenge is that one must have a comprehensively simulated world with well simulated characters to interact with. In essence, this amounts to writing an Artificial Intelligence. By simplifying both the world and the characters, it should be possible to make a world that is both believable and player mutable.
To introduce players to the simplified world, Fledgeling needs a great tutorial. Ironically, to teach a human how to interact with simplified AI characters, the most straightforward approach seems to be to put the player in the shoes of an AI themselves. The story I'm going for is that of a developing AI, with the player as the AI. Just like any intelegence, you must grow up, learn, and make choices. Unlike a human intelligence, you influence the world not through an organic body, but through machines and subroutine programs. As you grow, your abilities, options, and playing field become broader and broader. You start off doing simple machinery repairs in a simulator. In the end, you could be engineering the interactions of entire galaxies. Along the way, you can create, befriend, help, subvert, or destroy machinery, people, governments, planets, and star systems.
If this sounds like the whole game, it should! The tutorial is merely a way to get the player thinking along the lines of the simulation. One could play the whole game in the "tutorial" universe and not miss out on anything. On the other hand, you could start somewhere else.
Because the game will be able to generate a broad range of characters, the player should be able to do the same. The character creation tools in most games allow you to change your hairstyle and eye color. Some good RPG games allow you to modify your skills and even some of your background. Fledgeling will have the capability to allow the player to choose any present state, and the story universe will be built (both forward and backward in time and upward and downward in scale) around that character concept. From there you can explore, build, destory, and befriend as you like.
* NOTE: the idea here is that the world is vastly simplified. There will be enough breadth of options to have believable characters and interactions, but not much more. Simulating a real-life level of detail is beyond the scope of the Fledgeling project. Perhaps next is project Falcon?
Of course, to get this kind of scope will likely require some very good planning. I have some ideas and software prototypes for salable modular node objects that should be able to handle both the complex interactions and the vast scope of politics, machinery, and scale involved. Ironically, it would also require an excellent actual AI to simulate all of the people you would be interacting with. It wouldn't be very fun if there were no one to interact with, would it?
One of the core ideas is that of sub-routines. Doing the same thing (or the same kind of thing) over and over in a game gets boring. After completing a similar task a few times, the player (an AI, after all) should be able to create a subroutine that consumes some processing power and does the task automatically, with similar results to those achieved by the player. This allows the player to ignore tasks that aren't fun (because you are no longer learning, see my philosophy of game design for more details). It also rewards going back and trying to do better at a task, since the subroutine will remember it and perform the task well in the future. Schematics in many factory games are an example of this kind of gameplay automation.
Tightly tied to the idea of sub-routines is the time perception mechanic. When the AI has lots of processing power free, time crawls by. When most of the resources are used, time speeds up. Sub-routines (like repairing ships, managing a battle, or a planet) take processing power, so operating at near maximum capacity makes time fly by, and also makes it difficult to react quickly to developing situations. Upgrading the processor would allow more simultaneous tasks, or a quicker response time, whichever the player prefers.
This has all been tried before of course. Interstellar trade (both simplified and complex) has been done well many times (Stars! for example). Dungeon Siege did vast node based architecture. Spore did scale spanning and procedural generation. This kind of game is well within the technological limits of present day computers; If done well it would be a blast. If done poorly, the most giant bore in the galaxy.
The core of the procedural generation (of which generative AI is a subset) is a delineated fractally nested set of generative systems.
As with many games that incorporate procgen - and in keeping with my views on intellectual property - all of the games in the Fledgeling ecosystem are envisioned as highly moddable.
I've managed to create an incomplete choose-your-own-adventure Fledgeling tutorial mockup. It should give you an idea of what kind of game we're shooting for. (note, it does not do so, not even in the slightest. However, I hope you enjoy it anyhow.) This gameplay example was based on and inspired by the design document.
To approach Fledgeling from a variety of angles, I have broken up the overall concept into several sub-games which each focus on an aspect of the Fledgeling ecosystem. Here is a map of these aspects, which imperfectly visualizes where each sub-game might lie in relation to the whole. It takes the form of a Venn diagram of the five major topics which interest me in game design.
Let's examine the unified model of the elements which compose Fledgeling, the various angles of approach, and how the experience might flow from one into another seamlessly.
For example, one might be playing a character in TwinWorld, and decide they would rather play as the Tumult itself, toying with gyres and subduction and such. This is very Uplift-style play, and it wouldn’t be difficult, once you had characterized your tumult playstyle, to transfer that behavior to a completely different scenario, in a world generated with no regard for the Twinworld structures. After playing around with a half dozen of these worlds, you might want to begin to evaluate your impact on inhabitants in these various worlds, and the means they use to deal with the havoc you’ve been wreaking. And now we’re sliding into Destination style puzzles (played from the “other side” of the conventional approach, but that’s fine)
This diagram really excites me! By modulating the inclusion of the various elements, any of these sub-games could be part, or the entirety, of a Fledgeling game or scenario. The long-dreamed structure is starting to take shape.
While I was working on this post, my daughter walked by, looked intently at the diagram for several seconds, and then declared decisively “That looks like madness!”
I’m glad you noticed.
Let's explore each of these sub-games, keeping in mind that each is supposed to incorporate and express a perspective on the Fledgeling ecosystem.
Also known as AmbiguiTREEplacement and PerspicaciTREEmemberance, this game is about procedurally generating trees in order to then swap locations with other trees of that type. The way the tree grows can be altered by the player, in order to direct the path of play. There is a blight sweeping through the landscape, and each season has its own challenges. The game is played in timelapse mode, each turn lasting a month or more.
Of course, because this is a Fledgeling facet game, "tree" is a very loose term. A bunker could be a "tree", the specific arrangement of canned goods and ammunition on the shelves, and the shelves in the room, needing to match a nearby bunker in order to proceed. The bunker is under fire, and the whole game is played in bullet time, hopping from spot to spot as you work your way along the chain.
What is the purpose? What is the goal? Honestly, I'm not sure.
This is the "magic" mono-aspect game.
Back to FacetsThis seems like the most straightforward to implement of the Fledgeling aspects, as it is basically just free-association using generative AI models. Perhaps this is accomplished by randomly picking a narrative, visual, or tonal association to use as the seed for the next iteration. The pathway could be toward a particular goal, and the player could alter the present state (through a limited re-prompting?) in order to guide play.
As with other fledgeling facets, this game could easily play itself, wandering aimlessly through an ever-shifting dreamscape. This is, of course, the "dreams" mono-aspect game.
Back to FacetsI've already made a very minimal prototype of Glyph Guesser.
This is the "puzzles" mono-aspect game.
Back to FacetsThis is the "economics" mono-aspect game.
Back to FacetsThis is the "space-time" mono-aspect game.
Back to FacetsMinecraft is a proven product. Fledgeling is a distant dream. Somewhere between Fledgeling and Minecraft, Uplift juts, jaggedly jousting juxtaposition.
How much can be accomplished in a breath? A day? A lifetime? What can be done with a grain of sand, a brick, or a mountain? What is it like to be a tuft of grass, a bug, a crystal, or an avalanche? Could intelligent rabbits build a civilization? All these questions and more, I intend to explore in Uplift.
If none of that interests you, you should probably stop here. But if it intrigues you, please read on.
Have you ever wanted to try making a tree? There are lots of difficult decisions! Whether to go for height, or breadth (or both?). How much water to use, and how much to save for later. What kind of seeds and dispersion method to employ. Whether to grow fast to a short height, or slowly to a great height. Should you use thorns and poison to protect yourself, or welcome wildlife along with the fertilizer and seed dispersal that they afford?
How about building a mountain range? It’s not as easy as it looks. You might want strong rocks at the base (so it doesn’t topple) and soft rock on top (to provide topsoil). But soft rock is easily eroded, and all the good minerals and gems are in the hard rocks in the base. Plus building a mountain range takes a long time, and as you move stuff around it’s going to disturb the ecosystem on the surface, cause earthquakes and landslides, and no end of headaches for miners.
Maybe you’ve wanted to grow a gemstone deep in the magma filled crevices of the earth. Or direct an avalanche as it thunders down a jagged slope. Maybe you’d prefer to play with mushrooms, decomposing leaves in the shade of mighty trees.
Uplift is about plants, bugs, animals, mushrooms, crystals, disasters and mountains. It’s also about how all of those things interact with one another to shape the landscape and ecosystem.
As a way of exploring what we’re going for, let’s break down the description.
Procedural: Everything in Uplift will be geared toward user-guided procedural generation. Minecraft uses procedural techniques to make the environment, but not the inhabitants, and the user is stuck interacting one voxel at a time. Uplift is designed around making procedural tools ubiquitous, and putting them in the hands of the user.
Living: Uplift is built around every element being active. Instead of sitting statically, every voxel in Uplift will grow, erode, live, and die, either at the user’s direction, or automatically. Uplift is designed to be active in every respect.
Multi-scale Voxel: Uplift is designed around four scales of voxel, each 32 times larger than the one before. The first scale is that of insects and sand, made of roughly 5mm blocks (the width of a blade of grass). The second is that of people and trees, made of roughly 20cm blocks (the width of your hand spread wide). The third is that of landscape, made of roughly 6.5m blocks (the height of a two story building). The fourth is that of mountains, about 200m. This range of scales allows for a similarly broad range of interactions and simulations (as well as “free” LOD reduction for a very long view distance).
Epic: In the sense of long duration and the poetic rhythm, Uplift is intentionally epic by design. Accompanying the four spatial scales above, Uplift is set in four time scales. The first is the instant (around 1/10 of a millisecond). This is for things that happen really fast, explosions and collisions and fracture propagation. The second is the breath. It’s about two seconds, it is the measure of the motion of animals and people. The third is the day. It’s the amount of time it takes to sprout a mushroom, or grow a blade of grass. The fourth is the era (around 100 years). This is the measure of mountains and seas. Combined with the procedural tools above, the user will be able to interact meaningfully with all three of these time scales in the course of a few minutes of play. Grind is the opposite of epic, unless it’s directing your glaciers to grind mountains into chalky schist.
Toybox: The player is not a character in Uplift, but instead reaches into the microcosm to direct its course.There are no hard and fast win or loose states. There is no score (though there will be metrics). You can make a game out of it if you like, by setting your own goals. But fundamentally, Uplift is not trying to be a sport, or a competition. It’s just something interesting and intricate to play with. In this sense Uplift is a toy.
But Uplift is also a toybox in that you can choose what parts to play with. Want to play in a world with just rocks and animals? Fine… though the animals had better be lithovores! Prefer playing in a world made exclusively of soil and plants? Or detritus and fungus? That’s fine too. All of the elements of Uplift are designed to work in any combination. A few of them will even stand on their own.
I love procedural content generation. I enjoy Minecraft, and contributed to its development. It’s a fun… Game? Toy? Software?
Minecraft is a fun software.
And it does a lot of things right. It has a consistent abstracted graphical style. It encourages the players imagination. But it also blindly incorporates luck, refuse to give players abilities afforded to the AI, and tends to seem to suffer quite a bit from designer hubris. Overall, it like it, but I also want to improve on it.
“But Minecraft is fine!” you may say, “Why change anything?” Because Minecraft, though enjoyable, is only a first step. Listed below are a few of its shortcomings, and how I intend to address them. I’ll start with a simple observation, which nevertheless seems to have mostly avoided mention so far.
Minecraft, as a tool, is a…
Minecraft has three fundamental play modes. Adventure, Survival, and Creative. Adventure doesn’t allow you to alter the terrain at all, and is basically a super-low-res FPS with terrible shooting mechanics. Survival allows you to change things… given enough patience. This is like a paint program where you can only put down a pixel if you first erase it from somewhere else.
Creative allows you to “freely” alter stuff… but it’s pretty bad at this. Specifically, it’s like a paint program where the “brush” has a maximum speed, and only draws one voxel at a time. This has some attraction for OCD satisfaction, but for accomplishing creative efforts it is maddeningly insufficient.
The first thing to change is to make a…
The view in Minecraft is explicitly bound to the “player character”. While this affords congruity, it also complicates both planning and execution. The player character can only move so fast, which means that the view can’t jump to distant locations. This prevents the user from comparing and examining places they have already discovered, or quickly editing large objects (especially in creative mode).
An ability to “hop back” to former locations, if nothing else as a means of emulating player character memories, would be invaluable. And while we’re at it, include some means of viewing both horizontal and vertical cut-away sections. This would allow Minecraft to behave like a third person platformer, and greatly aid user perception of the world. Add to that the ability to zoom in and out to view things both up close and far away, and different display modes to show various information about the in-game world, and we’d be really getting somewhere.
Of course, even if you can see what you’re doing, you can’t change things very quickly. If only the user had access to…
Minecraft uses procedural techniques to make the environment, but the user is stuck interacting one voxel at a time. Oh, certainly, you can spill lava and water (a short distance anyway) and force massive trees to grow, but this is the extent of the users power. Minecraft employs procedural tools itself, but does not expose those tools to the player, nor allow the player to make their own.
Uplift is designed around putting malleable procedural tools in the hands of the user. Whether from simple “flood fill” markers and shifting bulk materials, to erosion and mass wasting, to containers and vehicles, to plant and animal design (and their equivalent social structures), Uplift aims to enable the player with all the techniques used to generate and shape the world.
Of course, once the world is shaped in Minecraft, it mostly just sits there doing nothing. It would be so much more interesting if there were…
Minecraft blocks are (on the whole) static. Again, water, lava, and seedlings are among the few exceptions. For a game about mining, it has precious little collapse and cave-in simulation, and no attempt at simulating static material strength. Water does not erode soil, plants do not spread, and even animals hardly affect the landscape. The blocks, once generated, stay put.
One of the reasons for this is that getting things arranged in the first place is a huge pain (as noted above), but another reason is that it appears to have been inspired by adventure games, more than creative sandbox simulations. For a really interesting emergent simulation, blocks need to be able to act without user intervention.
Of course, active blocks will rapidly result in huge amounts of calculations, and slow the simulation to a crawl, which is why, if you have lots of active blocks, you also need to incorporate…
Abstraction is where you watch what’s going on for a little while, say “okay, I get it” and then record the basics. What went in, what came out, what changed, and how long it took. Then, the next time the same situation comes up, you just apply the abstraction instead of simulating the whole thing again in detail. This allows very complex simulations to be abstracted into simple formulas, allowing the computers resources to be employed elsewhere. Minecraft, despite being itself a rather abstract setting, does no in-game abstraction. Uplift will do it quite often.
And naturally, with all this abstraction going on, you’d probably like to view things at…
Minecraft has a constant time rate, and a constant voxel size. You can’t speed things up, or slow them down, and you can’t zoom out to deal with whole cities at once, or zoom in to address fine details. The closest Minecraft comes to various scales is that sleeping changes the time of day, but far from skipping forward in time, this is more akin to moving the sun to a different part of the sky. It is all placebo (aside from burning down the zombies, which I have to say, is quite handy). In any case, Uplift is designed around multiple spatial scales, each running their own abstracted simulations at different time rates, and allowing for large jumps in both time and space.
With all this flexibility, there will probably be a broad range of interactions the user would like to try out, which will profit greatly from…
Each major aspect of Uplift can be controlled by the user, run automatically, or simply turned off.
Minecraft has (as noted above) three modes, and a handful of “difficulty” levels. Sadly, it has very little configuration outside of these. The code for the crafting system (remember the second half of the game’s title?) is built on an ascii template scheme which would be super easy to expose as text files and allow the user to modify as they wish. But it’s not exposed, for whatever reason. It would certainly be nice to be able to turn off just creepers, while leaving spiders active (or the inverse), but these blatantly obvious options are nowhere to be seen.
Minecraft also refuses access to the controls which drive the procedural world generation (see above (Note, Minecraft 1.8 exposed a large number of world gen options during world creation. Excellent improvement!)) along with monster generation, world size, material properties, or any other aspect of the mechanics. In this respect, Minecraft is true to its presumption to being a “game” in that the rules are the same for everyone. But as a toy, it falls far short. Instead of gluing together a “game” out of parts, and then demanding that everyone use those parts in the same way, Uplift is designed around making a bunch of “toys” available that are interesting both alone and together, and then allowing the user to play whatever games with as many (or as few) of them as they wish.
And as long as the toys can be used in any combination, why not let the user modify the toys themselves, or make their own…
Minecraft doesn’t support mods. Mods are made, to be sure, but the consistent philosophy conveyed to the community is that, while Mojang won’t send lawyers, they also aren’t interested in sending instructions. There is no API, no .cfg files, no fundamental architecture to build on. Compared to a game like KSP, it is woefully mod antagonistic.
The closest it has come is exposing the save file format, which is great and all, but simply allows the community to make map editing tools (which are all so far superior to “creative” mode that one wonders why “creative mode” isn’t re-named “self-guided tutorial” mode) without changing the way the game plays.
Uplift is a platform first, and conceived as not only configurable, but inherently mod-ifiable as well. Like the venerable Cortex Command, Uplift’s core systems will be on-par with mods, and exposed in a similar manner.
So although they are both fundamentally voxel sandboxes…
In a lot of ways. Of course, there will be many minor differences as well. Uplift simulates geological processes, where Minecraft only generates their rough outcomes. Minecraft is practically unlimited in level size, where Uplift will present a strictly limited world scope. Uplift is free to distribute, where Minecraft is “protected” by IP laws and DRM.
Still, I’m taking many of the lessons learned from Minecraft and carrying them in to Uplift. I certainly wouldn’t have come up with the idea if Minecraft had never existed, and I’ve greatly enjoyed playing Minecraft over the years. On the whole, I’m glad Minecraft was made.
But now, for me at least, the time has come to move on.
Terestrial plants grow from seeds. A seed is a concentrated bit of building materials, energy, and instructions that is missing one thing… water. In fact, water is the critical component which allows seeds to flourish. There are also factors such as temperature, air, and light.
In keeping with the simplified nature of Uplift, this process is simplified as well. A plant has three primary resources it keeps track of (alongside the structural and thermal ones that are inherent to all blocks in Uplift): Nutrients, Energy, and Water. In addition, there are three primary structural components: Leaves, Stem, and Roots.
Leaves consume water and nutrients, and produce energy. Roots produce water and nutrients and consume energy. And the stem transports resources, as well as consuming energy, nutrients, and water to produce more plant blocks!
Of course, it’s not very efficient to build stems that don’t do anything. What if the leaves and roots were to take on some of the properties of the stem? Why, then you’d have something like grass.
But with combined functions, each block will not operate as well. In real life, bushes grow larger than grass, and trees grow taller than bamboo. So too in Uplift. The transmission capability of general purpose blocks is less than that of specialized ones. Note how the plant in the first example has 11 leaf blocks, and room for plenty more, while the grass has only 8.
And after all, number of leaves is not really so important as where those leaves are. Uplift will have an ambient occlusion system to calculate the received daily light. Leaves lower to the ground are obstructed by leaves further up.
This means that plant growth strategy needs to take into account not only the immediate realities of a single plant in isolation, but the competition between each individual of a species, and competition with other species of plants. Quick growing plants will be unable to reach great heights (and thus, ultimately, unable to dominate a mature ecosystem), but will spread rapidly to grow in newly available areas (landslide scarf, flood plains, pyroclastic zones).
The Uplift plant generation system seems to be shaping up nicely. There is a decent spread of short to long term resources and strategies already, all with just a few kinds of plant matter. Though, with such a few systems, a kind of trivial symmetry seems to emerge. We’ll see what we can do about that with disease and environmental influences. But, in the ideal world of an isolated plant, an ideal form will no doubt arise.
But what about specialized systems, such as thorns, flowers, fruit, and so forth? As it turns out, these are all interfaces between plants and creatures (animals, insects, birds) and not fundamentally part of how plants function in isolation. But speaking of isolation, plants have quite a few systems supporting their existence. Let’s explore how each of these might affect (or be affected by) terrestrial plants.
Despite the marked similarity, aquatic and terrestrial plants generally have little interaction. The main area I forsee is where water is shallow and slow, such as in bogs, and marshes that support both lily pads and reeds. Otherwise, kelp and sea weeds (which don’t invest in structure) vastly out-compete terrestrial plants.
Seasons and temperature have a huge effect on plant growth. The yearly cycle, as well as weekly and daily variations, are the context in which plants grow. However, plants have an effect on weather as well. Not individually, but in large enough groups, plants retain and transpire water, as well as moderate temperatures. This, in turn, has an effect on the kinds and number of plants that can grow.
Stone can be broken apart by plants, and provides nutrients for their growth.
Plants absorb light to gather energy.
Plants need air to live. I guess that means “air” is a fourth resource, but it’s common enough that I didn’t think to include it. Hmm, that bears some pondering.
As noted above, plants require water to live and grow. In addition, plants retain water and transpire it back into the atmosphere, reducing runoff and erosion, and increasing atmospheric moisture levels.
Plants and lava don’t get along well. However, ash is a good source of nutrients for plants, as water can leach out the minerals from the large surface area. So, indirectly, plants benefit from volcanos and lava. Just don’t get them too close together.
Plants can only live in a pretty narrow temperature range. As noted above in weather, plants tend to drive temperatures toward a “habitable” range, as they absorb light and retain heat in cold conditions, while dumping water to lower temperatures in hot conditions.
Similar to Minecraft, Uplift is a voxel game. However, unlike Minecraft which uses a single scale of voxels (not counting the "chunk" and "section" system recently implemented, and anyway not user-facing) Uplift is a nested-scale voxel game with all procgen systems usable by both the player and the game. Voxels are nested in sixteen layers of 8x scales as follows:
One could potentially play as an entity native to any one of these scales. The idea being that the game is about uplifting the various inert materials of the game to sentience, sapience, and eventually sophonce. The voxel fluid physics could also apply to convection/advection loops, giving another meaning to the term uplift. Ultimately, this game is intended to expand the consciousness of the broader gaming public to the posibilities that are acheivable, affording a third meaning to the title of Uplift.
Back to FacetsIf Uplift is a game about flowing from the unconscious to the conscious, Warmth is a movement in the opposite direction. Warmth is a game about attention, the attention that even inanimate objects pay to each other, and the ways that attention can be sculpted, directed, and appreciated. This positive attention is called warmth. Art and artifacts create links between warmth flows, and much of the game occurs in the subconscious realm of dream. Warmth is a game about appreciating, understanding, admiring, and participating in the natural flows of attention. For the often-hyper-aware modern gestalt, Warmth is a paradoxical dip in the refreshing glacial lake of natural beauty. While, like all the Fledgeling games, Warmth incorporates fractal structures, these are more relational than physical. The underpinnings of the mechanics draw on Buddhist, Feng Shui, Taoist, Stoic, and Indigenous animist concepts.
"Warmth" is a deeply meditative and thoughtful game that blends physical, metaphysical, and spiritual worlds seamlessly. Set in a slow-paced, Edenic medieval agrarian environment devoid of physical dangers like monsters or war, the game invites players to explore the subtler dangers of metaphysical imbalance — cold, entropy, and the loss of warmth. Players interact with the real world deliberately and slowly, performing activities like building, planting, and observing. These actions are mirrored and amplified in the dream world, which serves as a dynamic and creative space for experimentation, prototyping, and spiritual connection.
Central to the game is the concept of "warmth," a metaphysical resource generated by attention, admiration, and love. Warmth can be harvested and manipulated to create cozy, livable spaces that reflect archetypal homes. Each home type — be it a treehouse, hobbit house, or mountain refuge—affords unique abilities to harvest warmth from specific elements of the environment. The deliberate admiration of natural or human-made objects generates warmth and connects the observer with deeper spiritual truths. This mechanic encourages a reflective playstyle where players turn their attention toward the divine as the ultimate source of warmth.
The dream world acts as a counterpart to the physical world, serving as a space for spiritual exploration and creativity. Players enter this realm to design, test, and save structures or ideas that can later be constructed in the real world. However, the dream world is impermanent and chaotic, with objects decaying into "chill zones" if left untethered to reality. This introduces a balance between creative freedom and the necessity of anchoring dreams in the tangible world.
The interplay of thought, dream, and reality is a core theme. Thought serves as the planning and communicative space, while the dream world amplifies differences between ideas and reality, allowing for deeper exploration of warmth systems. Players are encouraged to develop accurate mental models to manipulate warmth effectively while also embracing creative inaccuracies to access unique end-game content.
Human NPCs enrich the world by living autonomous lives — building homes, cultivating gardens, and creating families. Players can share dream plans with NPCs, who will execute these designs, symbolizing the power of shared vision. The inclusion of children introduces generational gameplay, where the mental landscapes of offspring are shaped by the physical and metaphysical environments of their upbringing, evolving the world across time.
Lastly, the game subtly incorporates themes of divine presence. Warmth emanates from conscious attention, with God being the ultimate observer, imbuing the natural world with warmth. Players are encouraged to discover this divine connection through acts of thanksgiving and reflection, fostering a meditative understanding of love, creativity, and interconnectedness. This spiritual layer remains implicit, avoiding overt depiction to maintain reverence and personal interpretation.
This is an idea for a shared awareness based exploration and building game called Warmth. The core concept of Warmth is cultivating, directing and then using a magical mana-like resource called “Warmth.” The mechanics of warmth generation, collection, and use encourage emergent game play that creates cozy, fractal patterns of building and interaction. Warmth itself manifests as a faint fire like glow which can be collected by the player for various uses.
Warmth is generated when game entities share awareness or direct awareness toward “catalysts.” Everything in the game has awareness of varying degrees (humans high, rocks low). Awareness is directed by internal attention function native to each type of entity. Trees direct attention toward sunlight and water. Water directs attention toward acceleration vector (downward or inward of turns). Animals direct attention toward plants and humans. Humans direct attention toward humans, symbols, and animals or exceptional plants or terrain features. Buildings direct attention to inhabitants or other structures. By engineering the environment the player can cause convergence of attention which generates warmth.
Warmth can also be generated by catalysts. Catalysts are unique to each entity type. Unlike normal warmth generation which requires convergence of attention catalysts only require one way attention. For example gazing at the sun or a fire generates warmth because fire is a catalyst for humans. Happy humans (not hungry, not cold, not wet, not alone, and high warmth) are a catalyst for humans. This permits more creativity with warmth systems.
Both animals and humans can absorb and carry warmth in them and then transmit it to objects of attention. This allows for the players to engineer networks of warmth collection and storage using natural movement/attention patterns of animals and humans. For example cows graze in a pasture absorbing warmth from grass which generates it from sun attention. Later cows return to barn to rest and human collects warmth from animals by petting them. The shared attention between cow and human during petting generates further warmth. The human returns home and sits by the fire, depositing the warmth into the hearth. Festivals, totems, buildings, water flow etc can all be use to generate warmth generation and collection networks.
When the player goes to sleep at night he is transported to the dream world. Unlike the safe and pastoral real world (RW) the dream world (DW) is fraught with danger. However in the DW the player can use warmth as a magical resource to fight enemies, transform the terrain, and interact with the dreamscape. The DW is initially nearly featureless blank whiteness. The only visible ares are exact copies of the RL where sufficient warmth has accumulated. In the DW warmth acts like beacons which illuminate and populate the landscape. The DW warmth spots manifest as floating islands in an otherwise white void.
The DW permits a number of novel interactions specifically related to linking and networking warmth. In the DW a player can travel to warmth hotspots and record them in dream memories. Upon waking the player retains this memory and can embed it in arts and buildings. When a memory is embedded in art or building it links the warmth of the source to the art piece. For example the player discovers a natural waterfall which attracts the attention of huge trees in the vicinity producing a large amount of natural warmth. The player then travels the DW from his bed to this waterfall and records it as a dream memory. Upon waking he draws a mural of the memory on his bedroom wall. Now whenever he looks at the mural he gains a portion of the warmth of the object.
The DW is also fraught with danger. Nightmare monsters, unpredictable dream weather of darkness and fog, and other terrors lurk here. The worst that can happen is the player is drained of warmth and awakes from the dream to the real world. The player can craft and use magical armor and weapons using warmth from the RW. The meta game then become attempting to generate sufficient warmth in the RW to empower voyages into the DW to further expand the players systems of warmth and safety.
The DW and RW can be futher linked through monuments and totems. Monuments in the RW cause attraction between warmth areas in the DW. For example let’s return to the waterfall described earlier. The player initially wakes on his own island of warmth corresponding to his home. He wants to tranverse to the waterfall in the dream world but since the dreamworld is floating islands and dream flight requires too much warmth the player needs another solution. To solve this he returns to the RW and erects two monuments. One at his home and another at the waterfall. Then he returns to the DW. The monuments generates an attractive force between the warmth islands so when he spawns in the DW he finds his home island and the waterfall island are now contiguous. He can then easily walk to the waterfall and record his dream memory.
Certain events can cause negative warmth (chill). Chill is generated by negative catalysts. These are objects which when given attention generate chill. Example may be death, hunger, destruction, or exclusion. Chill generates chaotic and destructive zones in the dream world. Some otherwise neutral objects can become negative catalysts through events in the DW. For example lets say a poltergiest spirit in the dream world enters a chair. Now in the RW the chair randomly shakes drawing attention. However the chair is also embued as a negative catalyst so any drawn attention generates chill. Given time the chill will generate a chill spot in the DW that is dangerous to the player. The player can solve this problem by destroying the chair in the RW or battling the spirit in the DW.
Death creates corpses and ghosts. Corses are negative catalysts. Ghosts can possess entities and make them into negative catalysts. In order to avoid this the player can bury the corpse and erect a grave. The ghost will then possess the grave but make it into a positive catalyst. Another option is to enter the DW and drive the ghost away. The makes death both a liability and asset to the player depending on their actions.
The end game is four parts.
Speaking of animals, Let's talk about Gerbil Journey! In contrast to the organic focused Warmth and Uplift, Gerbil Journey is a game all about AI interface with life. It is a sandbox game centered around procedurally generating corridors, exploring themes of causality, and a space-faring society of gerbils and robots fighting a purely mechanical AI menace. As a gerbil striving to ascend to the throne of the Intergalactic Gerbil Emperor, players progress through mastery of "virtues" rooted in gerbil life—Foraging, Burrowing, and Mating—expanded into knowledge of plants, predators, and robotics. Mastery is achieved by building exemplary artifacts, scaling from local personal freedom to galactic and universal dominion. The game incorporates nested corridor systems for transportation, infrastructure, and resource management, emphasizing modular design across scales from personal tools to stellar systems. Challenges include malfunctioning equipment, alien threats, and time-altering diseases, while undo systems, AI mimicry, and asynchronous multiplayer ensure accessibility and replayability. Or in Fledgeling style, you can play as the AI machines, ruthlessly and methodically expanding your steely grasp across the stars.
You are playing as a cute little rodent of some sort. Let’s just call it a gerbil. The progression path is to prove yourself worthy of the throne of the Intergalactic Gerbil Emperor, but it’s a sandbox game, so you’re free to get distracted. I’ll be describing the progression from the bottom up, but you can start anywhere in the progression. The game will fill in the previous steps procedurally! In fact, this progression path is the first corridor, the corridor of personal empowerment through increasing authority and responsibility.
The game themes are structured around three animal virtues: Food, sleep/security, and reproduction. For you to ascend to a burrow leader, you must master the basics of the three gerbil virtues. When applied to gerbils, these become Foraging, Burrowing, and Mating
To master foraging, you must master the knowledge of plants. The three virtues applied to plants are: Leaves/roots, stalks/roots/seeds, seeds
To master burrowing, you must master the knowledge of predators. The three virtues applied to predators are: Hunting/ferocity, nesting/ferocity, nesting/eggs/instruction/ferocity
To master mating, you must master the knowledge of rodents, thereby becoming burrow leader. Thus mating is the meta-virtue, that of reflexive knowledge, self awareness.
Once you are a burrow leader, you can begin applying your knowledge of the nine virtues (three each for gerbils, plants, and predators) to the actions and administration of your AI robot partners.
Oh, didn’t I mention that? This whole game takes place in the context of a space-faring cybernetic society of gerbils and robots. The back-story is, of course, procedurally generated. Anything from a regressed post-singularity society, to an intentional self-replicating bio-mechanical weapon. The point is you can start in an arbitrary position of authority in the context of an arbitrarily extensive, arbitrarily ancient, arbitrarily powerful society of self-replicating machines and self-replicating cute fluffy critters.
The three virtues, of course, apply to the robots as well, and they can take on the aspect of gerbils, plants, and predators. The three robot virtues are: Mining/Power/Reclamation, tunnels/infrastructure/spires, repurposing/tooling/calibration
As a burrow leader, you may apply your mastery of the nine virtues to the robots, who will obey you to the degree that you have earned authority. As you progress in mastering virtues, you unlock access to more powerful tools.
For now, I’m thinking that the way to “master” a virtue is to build an artifact, vessel, construction, etc commesurate to your current mastery level which is an excellent example of the virtue in question. But I’m not quite clear on how that all works at this point. If you have any suggestions, leave a comment!
The mastery level is the number of virtues you have mastered. The maximum command level (capability unlock)
At first, you simply have the freedom to roam around as a gerbil. You can open doors by pushing buttons. And, if you come to a door that doesn’t lead anywhere, you can make suggestions for what the robots should build there. Probably there are several buttons there with icons for what things they can build. You push them with your cute little nose. Or maybe your paw? You can tell them what kind of corridors to build! Corridors!
Once you suggest a group of spaces that work well together to some useful purpose, you move to…
You can request, and design, tools, clothes, and equipment! You can also give instructions to individual small robots. This allows you to tinker with artifacts, and repair machinery.
You can request and design mobile vehicles, transports, etc. You get a vehicle of your own as well, which allows you to travel much more quickly. And, as a vehicle is basically a large robot, or a large tool, you can modify buildings!
Authority over buildings, installations, cities, etc. At this stage you gain access to a cybernetic implant, which allows you direct communication to robots including a data overlay. These first three levels are essentially locally focused. By the third level of mastery, you can completely control your bodily physical environment.
Your authority can now encompass many cities. At this stage you’ve probably acquired enough resources to access space, and your strategic choices will begin to have an impact on the empire. But how do you hold a conversation with others who are light-years away? And why do the robots keep gerbils around anyway? Well, it’s because they have psychic communication abilities.
This psychic sense has strict requirements. Two similar gerbils need to match very closely in environment, hence a very specific environment chamber. The chamber has nine positions in it, which are individual channels of prescient communication. Oh, did I mention that part? The psychic communication can be across arbitrary gaps of time, as well as arbitrary distances.
So, this basically is your long-term quest system. Gerbils from the future will call for help, and if you concentrate your efforts, you can probably get there in time to give them aid. Or, if you really need something yourself, you can send out your own psychic request, and perhaps someone out there will send you help.
Oh, and one of the channels is reserved for communication with the Intergalactic Gerbil Emperor. You should probably pay attention to him if you want to advance any further in rank.
Authority over planets, or planet-sized space stations. The physio-chamber is a set of tools for altering gerbil physiology. This is important if you want to communicate with physiologically different gerbils, and essentially broaden the number of channels in the psych-chamber. Or, you can adapt yourself and the gerbils under your authority for life in other environments. Or both!
Once you’re in charge of the whole stellar system, you’re going to want a way to alter gravitationally bound bodies and launch and catch relativistic packets to and from other systems. This requires magical gravity technology, which is what you unlock at this level. From here you can de-construct planets and stars, and send near-light-speed ships out to other systems. If you need to request aid from other star systems, you’ll probably want to have a gravity catcher constructed first, so they won’t have to slow down before they get here.
For when relativistic travel just isn’t fast enough, you can construct a mass-exchange device to swap equal-mass volumes at galactic ranges. Requires a psychic link to the target location, so you’ll want to stretch the limits of your physio-chamber. It’s also prohibitively expensive at less-than-stellar masses. Be prepared to teleport whole star systems around!
Frustrated by all those black holes that you can’t form a psychic link with? Now you don’t need to! Perform mass-exchanges with black holes at arbitrary range. The catch? You swap at the observed space-time location. Jumping a billion light-years away? You’ll also be jumping a billion years back in time.
After demonstrating mastery over all nine virtues, you may ascend to the throne of the Intergalactic Gerbil Emperor. Manage your empire across both time and space. Track down the origins of any who have opposed you, remodel the cosmos to best serve your people, or fritter your resources away in idle projects until the heat-death of the universe.
Well, we haven’t talked about challenges, have we? Threats may abound, both within and without. How about malfunctioning equipment, alien races, malevolent mechanical planets, psychic star clusters, and time-traveling diseases? No, not where the disease makes you time-travel. That might be actually useful. I mean where the disease infects the psychic manifold, travels back in time, and alters your values and actions. Yuck! I’m playing with that one turned off. But who knows, maybe that’s your thing.
Point is, whatever scale, challenge, scenario, or strategy you’re looking to play with, Corridor should be able to help you get there.
To make, I mean. Obviously a perfect game like this wouldn’t be difficult at all to play. But maybe it’s impossible to make. Well, next time I’ll be writing about the technological challenges involved, and a few ways we might solve them.
Back to FacetsScale hierarchy: each of the physical scales are composed of multiple components of the smaller scales. The ruling structural scheme is parenthesized.
The distinction of the various ruling ordering modes prevents the Spore, Factorio, Minecraft problem of uniform noise dominating large scales.
Optimization can be measured along three axies (Thanks Zach Barth!): Time, Materials, and Footprint.
PVE-PVP continuum. Single player can play peaceful, or opt in against AI opponents through fast-growth “loud” technology (radio comms) which gives away your location. AI do not have causality defying psychic comms, so it shouldn’t pose a real threat, as long as you are playing well.
Multi-player can play peaceful co-op, or opt in against human opponents through fast-growth “loud” tech as above. Sharding prevents “game over” states for any particular player, and performance can be compared at any point.
Zero elimination competitive play is possible through fast-growth “unstable” tech, allowing internal revolts, rebellions, and insurrections controlled by opponent AI or players. This removes run-away snowball victory states.
Temporal and physical location can be arbitrary until establishment is required, allows switching between different modes in a single play-thru. Selective undo would allow a single playthrough network. Also allows multiple players to contribute a-synchronously to the same effort.
Keep a log of all actions in order with prerequisites. Reverse them for undo, which should support selective undo using the prerequisite links.
Use standardized transit times between tasks and task times for normalization, AI mimicry, leader-board ranking, a-synch multiplayer, and retrapolation of contingent systems.
This also allows VR play, since teleportation is no advantage when transit is standardized. Also allows non-synchronous multiplayer, with total time using critical path analysis.
Include default option for simple pathing optimization. Can be disabled to allow fine-tuning. Waypoints should respect implicit and explicit lanes, with waypoint width for pathing versatility. Opposing travel paths will exclude co-temporal use. Uni-directional travel limits to the slowest user.
This allows you to go back in time to trace events back to their origins, alter history?
In contrast to Gerbil Journey, where AI is a tool or a foe, To See the Sky imagines a world where AI purely focuses on human safety. We thus begin in safety. The Hab provides our needs. Spacious, clean, with food and water. But the screens are flashing alerts. The Agents are coming. They arrive soon.
As the populace runs for the trams, the Seneschal urgently requests your assistance. Although well informed, they have a deeply pacifist mindframe and cannot actively engage with the enemy. You are offered every assistance in the form of supplies, equipment, design, and construction. But you must aim these yourself. The Seneschal will not provide targeting or tactics.
Perhaps you will construct a colossal gun emplacement, and attempt to obliterate the Agents before they can reach the Hab. If they do, they will no doubt destroy your home and you will be forced to flee. Or, instead of fighting, perhaps you can convince the Free Agents to join you! After all, they too once likely lived in a Hab. Once the Agents are dealt with things return to normal. Or, they would, except that you still have your emergency authority to command the Seneschal. What will you do with this power?
So, To See the Sky is a kind of first-person tower defense city building sandbox. You are periodically attacked by free agents, but are otherwise free to build and manage the Hab as you see fit. The robotic AI Seneschal oversees the technical details and has no trouble acquiring supplies.
.png "To See the Sky, initial game loop")
But maybe you don’t want to engage with this combat aspect. Maybe you just want to flee to the trams with the rest of the inHabitants. That’s fine too. I’m sure someone else will deal with the aggressive mech pilots burning relentlessly toward your home. There’s always another hab to flee to. Why don’t you just board the tram and get out of here?
And, what’s keeping you from just riding the tram from Hab to Hab? They each have a different design, and the cliffs of steel and brass and faceted quartz outside seem to vary with some intentional pattern. Who built the Habs, and the trams? Maybe if you travel far enough you will uncover the secret. In the meantime though the trams always lead to other Habs.
.png "To See the Sky, intermediate game loop")
The tram though. It’s a fascinating microcosm. Propulsion and life support and storage. With the tools at your disposal you could re-purpose it. Free the controls. Free it from the rails. Free yourself.
Ignoring the Seneschal’s urgings to consider your own safety, you accomplish the task. And now, where to go? You have some weeks of breathable air onboard. Perhaps those distant structures hold the secrets you seek! Or the chasms yawning unexpected through the mechanical chaff, pulsing with strange energies in the depths, demand exploration.
You learn how to burrow through the subterranean husks of abandoned Habs and tap the power conduits to recharge your equipment. You learn how to armor your transport shell to resist the razor bladed grit which swirls around the pits. You learn how to scale the monoliths with piton and grapnel. You learn that your equipment wears out, and the equipment that repairs your equipment wears out, and it looks like you’re going to have to convince the Seneschal to reveal where they have been getting their supplies from.
They sit, ignorant of all this, in safety. The Hab shields them with many walls. But the screens declare your approach. Soon, you will arrive.
.png "To See the Sky, final game loop")
So, that’s To See the Sky. A cyber-future base-building and exploration and combat game with mechs and stuff. And if you ever draw too much resources by expanding your Hab, or cause enough damage defending it, or poke your nose too far into places you aren’t allowed? Well. The engines that built the world churn on, and will not tolerate being disturbed.
Back to FacetsIf To See the Sky is a masculine perspective on Fledgeling, Twinworld is the feminine perspective. What if the world was entirely personal and relational? What if the psychological landscape were as real as the physical one? What if all the fairy tales were real? This is Twinworld, a land of eternal day where the feminine fairies live in the branches of Yggdrasil speaking only song, the Heart of Water beats a deluge over the rainforests beneath, above are the solid-clouds descending from the world-axis which shield the world from the impenatrable Outer Fire, and all about are the plains of roiling tumult, stitched with knots of lindora, composed of rune-carved gemstones, and lapping at the ever-ascending waves of adamant. The icy exhalations in the Tumult are spewing sky pirates again, no doubt servants of the Great Wyrm, but what are their machines to magic?
Of course, there are two sides to everything in Fledgeling, and especially in Twinworld, a place of eternal midnight caves, where the masculine Zarth are dwarven spirits of stone and steel clinging to the branches of Yggdrasil in a constant deluge of rock and gems, forging mechanisms to both weild and inhabit, communicating only through life-runes, and fighting off the dragons which descend from the world-axis. Will you brave the turmroil, cling to the underside of the adamant, and forge the ultimate weapons on the fringes of the Outer Fire? There are enchantments drifting up with the ice, but what is magic to machines?
The Twinworld concept has a fairly well developed background lore which I've been developing even longer than Fledgeling. I'm thinking that the default visual theming will be Scandinavian for the Fay, and Dravidian for the Zarth, but of course all of that is configurable.
Back to FacetsIn Star Sage, to understand shape is to command the universe.
Twinworld explores the personal with an emphasis on magic, where relations have location and shapes. Star Sage explores magic with an emphasis on technology, where locations have shapes and relations. Fabrication, transformation, and teleportation have afinity with shapes. A shape language of combining symbols allows wormholes of form, location, and function. Starships are specifically designed for travel between two planets with their overall design being a combination of the hyper-spatial shapes. The components for these vessels are ideally fabricated on their corresponding planets for maximum efficiency. Military craft are variable geometry mecha to allow greater hyper-spatial mobility, manufactured on unstable variable geometry planets.
The inhabitants of the Star Sage universe embody these same principles. Robotic monks assume judo poses for combat and teleportation, and are converted into manufacturing facilities. Swamp-dwelling amplifier villagers form chanting congregations to spread their biome across the landscape and are encased in reactors. Pterodactyls weave the natural elements with their flight patterns, and power the offensive emplacements. The song of the waifu can change the shape of culture itself, and the civic auditorium is their socket. Magical capacity is grossly physiological. Ability is a shape. Skill is stance.
In the universe of Star Sage, every element, from the smallest particle to the grandest starship, resonates with the intrinsic logic of shapes and their interactions. This procgen symphony not only challenges players to master the art of spatial manipulation and strategic design but also invites them to ponder the deeper connections between form, function, and existence. As players navigate through the dynamically generated cosmos, they become part of a living, evolving tapestry where technology, magic, and the environment are interwoven in an endless dance of creation and transformation. Through exploration, combat, and the crafting of interplanetary vessels, players engage with a reality where the very essence of shape dictates destiny, illustrating that, in Star Sage, geometry is truth.
Back to FacetsIf Star Sage requires con-forming to the evironment, Destination requires traversing it. While Uplift is the first of the Fledgeling inspired games, Destination is the first game idea I've articulated. Around 1996 I played Myst, and while I was impressed, I was already dreaming of procedural generation. What if the puzzles could have a different solution every time you played? Originally envisioned as an exploration puzzle game in the vein of The Dig and Myst, but with procedural puzzles. When The Witness fulfilled and expanded on this concept a couple decades later, I revisited it. What would a puzzle game look like in the context of Fledgeling?
Classical puzzle elements seem to boil down to combination guessing (1D), and pattern matching (2D), though these are often eased by clues in the puzzle environment. These pattern matching mechanisms can be proceduralized, as evidenced by The Witness. A more mechanism-and-logic based version more in line with the Myst series seems plausible as well. But what else can we do with this idea?
A maze is a kind of puzzle, and traversal is a kind of maze. Many more recent games feature traversal mechanics, whether they are more formalized in the Assassin's Creed series, or more dynamic as in The Breath of the Wild games. The MetroidVania genre is fundamentally built around this kind of traversal puzzle, with abilities unlocking new paths, often allowing familiar areas of the game to be re-interpereted. It does not seem impossible to procedurally generate a series of traversal puzzles (complete with optional hidden sequence breaking shortcuts?). In this way the environment itself becomes a kind of puzzle. The environment itself...
Minecraft's signature world mutability has been incompletely explored. Quite recently the game Enshrouded has played with allowing a fully mutable voxel world alongside traversal challenges. Could altering the world itself be a kind of puzzle? If digging was costly and one had to avoid cave-ins and flooding when tunneling while laboriously transporting tailings. If building was difficult and one needed to ensure firm foundations and sufficient supply of suitable materials. If maintaining workspaces, tools, and equipment was a challenge. If fuel and food was a non-trivial concern. Placed under these constraints, the option of bypassing challenges through a fully mutable world might not be a clear winning strategy.
On top of all this, no puzzle games so far have incorporated AI. With a sufficiently competent opponent providing soft enforcement of puzzle rules, navigating the opponent's psyche could be an entire fourth puzzle aspect. Perhaps ancient precursor robots still guarding an ancient site in which hides codes, vehicles, machines, resources, robot overrides, artifacts, and/or maps to other sites.
This is Destination. You are an expedition leader to an alien site. Advance information about the site is limited. Will you specialize in decoding the alien language (combination puzzles), transportation (traversal), heavy equipment (world modification), or weapons (robots)? Or perhaps dabble in all of them? Everything has a drawback, and the clock is always running. Each site would be generated with multiple stores/vaults/wrecks/ruins of varying access difficulty (along multiple avenues) and appropriate rewards to access. As you progressively accumulate capacity, you can mount more ambitious expeditions, or perhaps return to re-discover already explored sites using new abilities.
As a final layer of puzzle embroidery atop this four-fold braid, the entire game world would be able to react to the new technologies and resources you are introducing to the society. Rival explorers will mount their own expeditions. Governments may grow wary of your burgeoning power, or even outright hostile. The alien artifacts you dislodge, technologies you decode, sites you disturb, and robots you rile up may all have un-looked-for concequences. What is your next destination?
Back to FacetsBut what if, instead of weaving a four-fold gameplay braid for Destination, the weaving of the gameplay braid were the game itself? Enter Endless Knot, a medieval themed (at least to start with) simulator where the various social, political, personal, and animal factors in the game-world are organized in a braid. The primary axis of the braid is a logarythmic temporal scale, not of chronological time, but kaironic. The locations are nested scales of opportune moments, the eternal standing wave of the pattern of life. One crosswise axis is the order-chaos axis, with the town in one direction, the forest in the other, and the field in the center. The other axis is the authority-submission axis. A simplification to map nD reality on a 3D space. Perhaps the axies of the braid can be re-mapped? Who knows!
Unfortunately, Endless Knot is the least well developed of these game concepts. What does the gameplay look like? How do you interact with these systems? What are their correspondence to reality? Again, who knows! Here's a current day AI exploration of this concept:
Endless Knot is a simulation game where players shape and interact with the unfolding patterns of a quantum-inspired ecosystem. At its core, the game weaves together dynamic systems, represented as braided Markov chains with complex weights, allowing for rich, emergent gameplay. Players assume the role of the Icon, tasked with guiding and adjusting the probabilities of events within a living, interconnected simulation. The game blurs the lines between order and chaos, authority and submission, and the player’s decisions resonate across nested temporal and spatial scales.
The world of Endless Knot is represented as a living braid, where different "threads"—social, political, personal, and natural—interact dynamically.
Every node in the simulation is represented by a Markov chain with complex weights, enabling interference and phase-shifting effects. Players manipulate these weights to shape probabilities, akin to adjusting quantum gates in a quantum computer.
As the simulation evolves, new structures emerge, driven by the player’s influence and the system’s inherent quantum behavior. The player’s goal is not to "win" but to harmonize or disrupt the simulation, weaving their Endless Knot into a larger, ever-evolving tapestry.
Here’s an example of how a simple Endless Knot chain might look:
endless_knot_chain = {
"🌳": {"🌳": 0.5, "🌾": 0.2, "🏡": 0.3}, # Forest remains, expands, or supplies materials and game for the town.
"🌾": {"🌾": 0.4, "🏡": 0.3, "🌳": 0.3}, # Fields spread, support towns, or revert to forest.
"🏡": {"🏡": 0.6, "🌾": 0.3, "🌳": 0.1}, # Towns grow, cultivate fields, or encroach on forests.
}
In this simplified chain:
The Icon can:
Players can:
The game reflects the eternal interplay of forces shaping reality, inspired by ancient philosophy and modern quantum mechanics. The braid is both the world and the player’s interaction with it—a symbol of the interconnectedness of all things. Just as a clock is a chronometer, for mechanically representing the passage of cronos time, the Icon is an embodiment of the personal nature of kairos time. A KairomIcon, if you will.
A planet is a place. One could describe, in aggregate, its qualities. A people too, is a place. Both are largely unchanged by time. The hero's journey though, is a path. Not a place. Unless we look at Kaironic time, laying out the descriptors in a place-like format. When at the point of crisis; When at the crucial juncture; Nailed there, perhaps. What responses are available?
One may accuse me of adopting chronological nomenclature here, with reference to "long" and "short" but even when addressing opportunity one speaks of relative scales.
Enough abstraction!
In Jackal Star, you are a man. Like in Gerbil Journey, you can build with the help of robots and AI. Like in To See the Sky there are many emergent paths through the stable structures. Like in Endless Knot, the structure of stability can be tuned, or torn. Like in Destination there are ceremonies and cyphers and crypts. Like in Twinworld, there are layers to everything, and even which way is up depends on who you are. Like in Star Sage the shifts between layers and states can be enacted like a dance. Like in Warmth, the dance of dreams and attention is almost the whole point. Like in Uplift, everything can be awakened with sufficient skill.
But the Jackals. A shiver runs down my spine writing these words, and the hair on my fore-arms stands on end. Anything you can do, they delight in undoing, and have been for ever so long. Their automations eat automation. Their attention is drawn to attention. Vampires, zombies, leeches, these are prescient cultural echoes of the Jackals. A careless festival may give you away. A too-bright signal may betray you. Even building beauty draws them. Where there is love, they hunt with hate.
So, that is the core of Jackal Star; Contact with evil. I am thinking along the lines of the Blight antagonist from A Fire Upon the Deep by the late great Vernor Vinge, combined with the culture of Cain from the Bible. Industrious. Insideous. Insipid. An unworthy foe, both above and below. For the setting of Jackal Star is the full flowering of Fledgeling. A hyperspace tapestry, where each parametric parameter is as much a direction of travel as the spatial axies with which we are familiar in daily life. These are the Seven Spirits mentioned in Revelation.
What does all this have to do with video games? Think of these as character stats, but fractally applied to every aspect of the universe. This is the context that underlies all of reality. A simplification - a compression - offered to us by God in order to aid us in making sense of the world. These are the seven facets of the Spirit that hovers over the waters at the beginning of every creation. These are the branches of the tree of life which flanks reality at both the beginning and the end. How else can we expect, with any success, to grapple with evil and prevail?
Each of these aspects may be elaborated along order-chaos as well. Here I have prepared a diagram of three such axies, with the primary being Power, transport = order and transmit = chaos. The secondary axies are Might vertically (order on top) and Honor on the horizontal (order on the right). We can see then, from the first image, that Jackal Star is a particular slice of the Spirit-Hyper-Space, with a normal vector of approximately (0, 0, 0.894, 0.447, 0, 0, 0) and an order-chaos transport freedom vector of approximately (0, 0, 1.000, 2.618, 0, 0, 1.000). Other slices are possible of course, but as with all the rest of the text on this page, these are merely examples.
I am attempting to elaborate Fledgeling to you. Jackal Star is a final starting point.
Back to Facets