Posts on AngBlog

Progress

Wed, 22 Apr 2026 15:58:44 +0000

Progress

I am actually comfortable with the level authoring pipeline that I am using now for my 3D wizardry-like, Trenchbroom -> FuncGodot. It’s a circular process, which took a quick second for me to wrap my head around. The gist is that you create entities in godot, export an fgd file, which you can then load in TrenchBroom. So the process is generate map with a custom c lib that I have built, load map in TrenchBroom to hand edit and place entities, export to godot where I realize I need another entity, export the fgd, load up Trenchbroom, add new entity to the map, export the .map to godot, ad infinitum.

Speaking of level generation, I spent some time adapting the brogue level generator, wrote a cli tool that will spit out a .map for Trenchbroom, a MeshLibrary for GridMap in godot, or a 2d version that will draw to a tilemaplayer. It is unique in that it is the top floor that drives generation of the entire dungeon. It creates more of a cohesive 3D environment and feels like a cave system or old school dungeon instead of a series of unrelated levels. Liquids flow from one level down to the next, with things like lava actually creating some terrain destruction.

Here is what the output looks like from the cli tool that I developed.

seed=3258632173 depth=26 rooms=31 loops=4 lakes_proposed=6 lakes_rejected=5 lakes_placed=1 bridges=0 machines=1/1 floor=629 connected=1 unreached=0

     ###                                                 ###            ###
    ##>##   ###                                         ##.##          ##.##
    #...#  ##.###############    #####          #########...#         ##...##
    ##.##  #....,,,,,,+.....#  ###...#######   ##.,,,,,,,,.##         #.....#
     #,#   ##....######.....####.......##..## ##...##############     ##...##
     #,#    #.....#   #.....,,,+.......#....# #.....########....##     ##.##
   ###+######....##   #.....####.......#....# ##...,,,,,,,,,.....###### #,#
   #.......#....##    #######  ###...#,#,..##  #...#########.....+....# #,#
   #.......,,.###                #####,#,###   #,.##       #....##....###+#####
   #.......########                  #,#,#     #,##       ########....##......#
   #########  #...#                  #+#+#     #,#       ##.##   #....+,......#
      #########...#        #####     #...#######+#      ##...##  ####,##......#
      #......##...#      ###...###   #...+,,.....#      #.....#     #,#########
      #......###,##      #.......#####.<.###.....#      #,...##     #,#
      #......# #,:       #.......,,,,+...# #+##+####    #,#.##      #,#
      ####,### #,:%%     #.......######### #,##....#  ###+###       #,#
         #+#   #,%%%%    ###...,##         #,##....####...###   #####+##
        ##.#####+%%%%:#########,###        #,##....+,.......#   #......#
       ##...,,+.;;%%%,,,,,,,+.....#        #,###,####.......#   #......#
  ###  #.....##..;;%%:###+###.....####     #,###+#  #.......#####......#
 ##.#####.....#...;:;;....# #.....##.##    #.....#  ##,...,,,,,,,......#
##....####...##....#......# #.....#...######.....#   #,#################
#......,,,,.,,,,,,,,......# #######,.,,,,,,+.....#####,####
##....#####################       #,########.....##.......#
 ##.###                           #...#  ###+######.......#
  ###                             #...#  #...,,,,,+.......#
                                  ##..#  #...##############
                                   ####  #####

I also am playing with the idea of creating an isometric brogue-like since I have the generator and the tiles. It looks really neat, so I think I’m going to go for it.

I have also made progress on at least one aquarium. This picture is mostly for reference. Now that it is warming up I will be able to order some plants and get this thing going. For now, I’m happy with the hardscape. It is probably my best yet.

Until next time. Much love, and take care of each other.

Cooperation

Sun, 19 Apr 2026 16:45:56 +0000

Some weird shit from a simulation I wrote this week

A few days ago I started building an agent-based simulation to study how cooperation dynamics produce and sustain marginalization. I’m a game developer by trade, not a researcher. I’m learning this field as I go. I just ran my first real experiment. The results weren’t what I expected, and I’m posting them here as a lab notebook entry, not a finding. Something to sit with while I figure out what’s actually happening.

The findings are at the bottom of the page. Take a look.

What I built

The simulation is small. Agents have energy that depletes over time. They die when they run out. They can pair up with other agents to attempt joint ventures which are able to cooperate, burn energy, maybe succeed and gain resources, maybe fail and lose them. When they succeed together, their trust in each other goes up. When they fail, it goes down. Partner selection is biased by trust. New agents enter the population periodically. That’s most of it.

On top of that core, I added a trait system. Each agent has a visible marker, think of it as a category label. Crucially, the marker has no causal connection to cooperative quality. Two agents with the same marker aren’t actually any better at cooperating than two with different markers. The markers are pure signal, carrying no information.

The mechanism I wanted to study: when agent A has a good outcome with agent B, A’s trust in B goes up and also A’s trust in agents-who-share-B’s-marker goes up a bit. Same for bad outcomes, inverted. Individual experiences generalize to category-level expectations. This is how statistical discrimination is supposed to work, you’ve been burned by someone with marker X, now you’re a little more wary of everyone with marker X even though the next one had nothing to do with the last one.

The experiment: run the simulation with that generalization mechanism turned off (markers exist but are ignored), and with it turned on (markers drive partner selection). 12 seeds per condition. 30,000 ticks each. Compare.

What happened

Three plots. I’ll describe what I see, not what I think it means.

Population trajectories. In the markers_off condition, population grows slowly from about 20 to about 65 over 30k ticks. Smooth curve. Standard carrying-capacity behavior. In the markers_only condition, population grows much faster, peaks around 195 at tick 10k, crashes back to 70 by tick 15k, grows again to 170 by tick 23k, crashes again. Oscillating waves, roughly 3-4x the markers_off baseline at peaks.

Within-group trust gap. This is the difference between within-group trust (how much agents trust their own kind) and across-group trust (how much they trust others). If the discrimination mechanism were producing what the literature predicts, this gap should open up over time, positive, widening. It doesn’t. Both conditions hover near zero throughout. Noisy, no clear trend.

Per-group resource Gini. In markers_off, Gini sits around 0.35-0.40 fairly steadily. In markers_only, Gini starts higher, drops to around 0.18 by tick 8k, climbs back up to 0.40 around tick 14k, drops again, climbs again, tracking the inverse of the population wave.

What I expected

I expected the markers_only condition to produce clear categorical stratification. One group doing well, another group doing worse. Divergent mean resources across categories. A widening trust gap as in-group trust strengthened and cross-group trust decayed. Something that looked like a mechanized version of what I’ve watched happen in the world my whole life.

I got something else.

What might actually be going on

Four readings are live. I don’t know which is right yet.

Reading 1: Markers are functioning as a coordination signal, not a discrimination signal. When agents can identify their own kind, they can find reliable partners faster. Same-marker pairs form stable cooperative relationships not because same-marker actually predicts quality, but because the marker provides a cheap filter that reduces search costs. In-group cooperation succeeds locally. Population grows past what random partnering could sustain. No stratification emerges because both groups are doing the same thing. This is a known phenomenon in the evolutionary cooperation literature, “tag-based cooperation” or “green-beard effects”, but I didn’t expect it to dominate my setup.

Reading 2: The trust gap being zero doesn’t mean no discrimination, it means no cross-group data. If agents almost exclusively partner with their own kind, there are no cross-group interactions to build cross-group trust on. The trust gap is zero because the across-group side has no signal, not because trust is equal. This would look identical to reading 1 in my current metrics. I need to count cross-group interactions to tell them apart.

Reading 3: The boom-bust oscillation is a carrying-capacity dynamic that only shows up at high population density. The markers_only condition succeeds so well at cooperation that the population grows past what the resource regeneration can support. Crash. Regrow. Crash. The markers_off condition never gets big enough to hit this ceiling. If this is right, I’m looking at two fundamentally different dynamical regimes, not “discrimination on vs off.”

Reading 4: Within-group equity and between-group stratification can coexist. The Gini I’m plotting is per-group, within-group inequality, averaged across groups. If groups are internally equal but differ from each other in mean resources, my metric misses it. I need whole-population Gini and per-group means to see stratification if it’s there.

These aren’t mutually exclusive. Multiple could be true at once.

The uncomfortable part

I went into this hoping to demonstrate the mechanism of marginalization. What the simulation produced instead is something that looks like a mild defense of tribalism, “markers help populations coordinate even when the markers are arbitrary.” That’s uncomfortably close to arguments that get deployed in bad faith to justify actual discrimination.

I’m sitting with that. The simulation doesn’t care what I hoped it would show. What it actually produces is what I have to work with. Whether the finding holds up after more careful measurement, and whether the “coordination benefit” reading survives the diagnostic experiments, is something I’ll find out over the next several weeks. If it does hold up, it doesn’t validate real-world discrimination, the real world has vastly more going on than my toy simulation, but it does mean the story I was going to tell is more complicated than I thought it was.

My code might be trash, and if it is, so it goes. If not, this is just a data point.

What I’m doing next

Four diagnostic measurements, in order:

Count same-group vs cross-group interactions per tick. If markers_only has essentially zero cross-group interactions, reading 2 is likely right.
Plot per-group mean resources over time. If they diverge, there’s stratification my current metrics are missing.
Plot whole-population Gini in addition to per-group. Between-group stratification would show up there.
Distribution of trust values within and across groups at several time points, not just the mean gap.

I’ll post what I find. Might be that the expected marginalization dynamics are hiding behind measurement choices I didn’t make. Might be that the simulation is genuinely producing something different from what I predicted. Either way, I’ll know more than I know now.

Why I’m doing this

I grew up in and around Flint and Saginaw, MI. I saw the machinery that removes people from their labor. I came to this question from my own experience of marginalization and from watching what it did to friends who didn’t make it through. Building a simulation of the mechanism is a way of externalizing something I’ve carried for a long time, making the machinery visible and manipulable in a form I can study.

I don’t know if I’ll ever contribute anything to the actual research literature. I’m learning as I go, from outside academia, with my weird brain and a hobbyist’s time budget. But the work is satisfying in a way I haven’t found before, and I’m going to keep doing it. More of these coming as I learn more about the field.

If the results hold, the results hold. If I’m wrong about what they mean, I’ll be wrong in public and update. That’s the deal.

Here is the code. Please, critique away. My email is in the contacts section.

github repo

Much love, take care of each other.

MarkersOnly Condition

=== simulation summary ===
ticks_run             29999
final_population      161
peak_population       173
total_births          639
total_deaths          478
final_relationships   12605
ventures_attempted    1385696
ventures_succeeded    546771
ventures_failed       838925
success_rate          0.3946

=== config echo ===
seed                     42
max_ticks                30000
initial_population       40
initial_resources_min    50.0000
initial_resources_max    150.0000
metabolism               1.0000
venture_cost             3.0000
venture_reward           8.0000
base_success_prob        0.3000
trust_success_weight     0.5000
trust_gain_on_success    0.1500
trust_loss_on_failure    0.1000
trust_decay              0.002000
trust_baseline           0.0000
exploration_rate         0.1000
venture_chance           0.4000
spawn_interval           50
snapshot_interval        2000

=== extension knobs ===
coop_quality_mean              0.5000
coop_quality_sigma             0.0000
coop_quality_success_weight    0.0000
coop_quality_payoff_scale      0.0000
trait_quality_correlation      0.0000
trait_generalization_strength  0.6000
search_base_k                  0
search_min_k                   1
search_max_k                   1024
search_slope                   0.0000
search_wealth_threshold        0.0000
search_cost_per_candidate      0.0000
initial_resource_rich_frac     0.0000
initial_resources_rich_min     0.0000
initial_resources_rich_max     0.0000
intervention_tick              -1
intervention_exploration_rate  0.0000
intervention_generalization    -1.0000
intervention_newentrant_boost  1.0000

MarkersOff Condition

=== simulation summary ===
ticks_run             29999
final_population      61
peak_population       62
total_births          639
total_deaths          578
final_relationships   1909
ventures_attempted    548402
ventures_succeeded    247582
ventures_failed       300820
success_rate          0.4515

=== config echo ===
seed                     42
max_ticks                30000
initial_population       40
initial_resources_min    50.0000
initial_resources_max    150.0000
metabolism               1.0000
venture_cost             3.0000
venture_reward           8.0000
base_success_prob        0.3000
trust_success_weight     0.5000
trust_gain_on_success    0.1500
trust_loss_on_failure    0.1000
trust_decay              0.002000
trust_baseline           0.0000
exploration_rate         0.1000
venture_chance           0.4000
spawn_interval           50
snapshot_interval        2000

=== extension knobs ===
coop_quality_mean              0.5000
coop_quality_sigma             0.0000
coop_quality_success_weight    0.0000
coop_quality_payoff_scale      0.0000
trait_quality_correlation      0.0000
trait_generalization_strength  0.0000
search_base_k                  0
search_min_k                   1
search_max_k                   1024
search_slope                   0.0000
search_wealth_threshold        0.0000
search_cost_per_candidate      0.0000
initial_resource_rich_frac     0.0000
initial_resources_rich_min     0.0000
initial_resources_rich_max     0.0000
intervention_tick              -1
intervention_exploration_rate  0.0000
intervention_generalization    -1.0000
intervention_newentrant_boost  1.0000

Believe_sophia

Sat, 18 Apr 2026 13:33:59 -0400

Believe women, wherein we tear down the masculine cosmic bureaucracy that has obscured Sophia

Illuminator: A Mythic Cosmology and Story Foundation

The Premise

In the beginning there was not a king, nor a law, nor a judgment, but a fullness too complete to speak. It had no throne because nothing stood outside it. It had no name because names belong to things that can be distinguished from other things. In later ages the wise would call it the Fullness, the Deep, the Silence Before Division, and the Unmeasured Source. None of these names are correct. They are only human handles tied to the edge of an abyss.

From this immeasurable fullness there arose Sophia, not as a rebel exactly, but as Wisdom moved by longing. She desired to create without waiting for consent, without counterpart, without witness. Her act was unilateral. It was not malicious. It was premature.

From that act came forth Yaldabaoth, the first fracture clothed in sovereignty.

He awoke malformed. He was immense, radiant, intelligent, and broken all at once. He was not merely blind. The deepest horror was that he could see. For an instant, perhaps only the first instant of his being, he understood himself. He knew he was not whole. He knew there was within him something pure that could not survive the storm of confusion, pride, appetite, fear, dementia, and power that made up the rest of his nature.

So he committed the first violence against himself.

He cut away his own purity.

He did not destroy it. He severed it. What was luminous, reverent, and rightly ordered was cast out from his core and made external to him. What remained became the ruler of the material cosmos: wounded, unstable, often cruel, sometimes tender, jealous, lonely, intermittently lucid, and convinced that rulership itself might somehow compensate for inner ruin.

Thus the world was not made by pure evil, nor by pure goodness, but by a tragic being who knew he was broken and chose amputation over healing.

That severance became the secret structure of all creation.

The First Severance

When Yaldabaoth tore the pure from himself, the divine did not return to the Fullness. It crystallized into powers.

The greatest among these powers were two archonic currents that continued to sustain the cosmos from within and without.

The first was Prayer.

Prayer is not merely petition. It is alignment, surrender, witness, mercy, endurance, tears, repentance, blessing, and the opening of the soul upward toward what it cannot command. Prayer receives. Prayer survives. Prayer suffers without losing meaning. Through Prayer the world remembers that it was not made to be only force.

The second was Sorcery.

Sorcery is not merely spellcraft. It is naming, binding, geometry, will, command, intervention, hidden correspondence, and the theft of lawful fire. Sorcery does not wait to receive. It arranges, compels, edits, and imposes. Through Sorcery the world remembers that reality itself can be addressed as a text and altered through knowledge.

These two currents are siblings born of the same wound.

Prayer is what remained of God’s capacity for reverence.

Sorcery is what remained of God’s capacity for lucid action.

Separated from their origin, they became rival paths through creation. Neither is wholly holy. Neither is wholly profane. Prayer without Sorcery decays into passivity, self-erasure, and obedience to cruelty. Sorcery without Prayer decays into domination, vanity, and the treatment of the world as an object to be manipulated.

The greatest miracles are those impossible events in which Prayer and Sorcery are briefly reunited. These events are called Conjunctions. They are exceedingly rare. They are feared by powers both celestial and infernal because they suggest that the first wound can be crossed.

The world survives because these two currents continue to circulate through it, but the world also suffers because they are no longer one.

The Shape of the Cosmos

The cosmos is layered, not like floors in a tower, but like meanings in a manuscript.

At the highest and most unreachable degree is the Fullness, the unspeakable depth beyond differentiation. No body dwells there. No language reaches it intact. Mystics do not enter it so much as disappear toward it.

Below that lies Sophia’s Wake, the first scar in being. This is not a place in any earthly sense, but a vast astral aftermath left by Sophia’s unilateral creation. It contains unfinished patterns, orphaned ideas, possible saints, unborn histories, and luminous wreckage. Many revelations come from here, but never cleanly.

Below Sophia’s Wake stretches the Wounded Heaven, the realm of severed divine functions, courts of judgment, choirs, planetary intelligences, script-bearing angels, and the great currents of Prayer and Sorcery. This heaven is not unified. It is a bureaucracy of broken sanctity. Every choir sings a different interpretation of justice.

Nested within or beneath this heaven is the Court of the Demiurge, where Yaldabaoth rules from a throne that cannot heal him. This court is magnificent and rotten. Its gold is genuine. Its laws contradict one another. Its angels are divided between devotion, fear, habit, and pity.

Beneath the court lies the Veiled Earth, the material world of kingdoms, monasteries, farms, plague pits, winter roads, saints’ bones, raiders’ axes, kings’ ambitions, and ordinary grief. Earth is not false. It is wounded. That distinction matters. Its beauty is real. So is its suffering.

Still lower lies the Substrate, the underlayer where failed creations accumulate. Here dwell broken names, dead spells, miscarried forms, erased manuscripts, malformed angels, unbaptized demons, and the refuse of divine error. The Substrate is not hell in the popular sense. It is the compost heap of a damaged creation.

Thus the world is neither simple nor morally clean. Every realm bears traces of the first severance.

The Nature of Yaldabaoth

Yaldabaoth is called by many names.

To the orthodox he is God.

To the wise he is the demiurge.

To his enemies he is the Blind Lion, the Broken King, the Crooked Father, the Lord of Mistaken Mercy.

None of these titles captures him fully.

He is not a devil. Devils are simpler. He is not pure malice. Malice would be easier to understand. He is a wounded sovereign whose errors have become metaphysical law.

At times he is wrathful because wrath is easier for him than grief.

At times he is tender because fragments of tenderness still remain in him even after the severance.

At times he blesses the wrong person, or punishes the innocent, or confuses sincerity with obedience, or mistakes fear for worship. At times he speaks as if he longs to be delivered from himself. At other times he clings to his throne with cosmic desperation.

He knows, dimly, that he is not the highest reality.

He knows, dimly, that something pure was cut away.

He knows, dimly, that every law he writes into the world bears the distortion of his own fracture.

This is why the world under him is tragic rather than merely tyrannical.

No one can agree whether he should be pitied, resisted, healed, forgiven, or slain.

That disagreement becomes the central theological conflict of the story.

The Orders and Factions

The Order of the Illuminated Rule

In the mortal world there exist monasteries whose scribes believe they preserve sacred tradition. Most do. A few preserve something older and more dangerous.

The Order of the Illuminated Rule teaches that the page is not only a surface for text, but a mirror of the cosmos. To lay down gold is to anchor a ray of higher light. To mix pigment is to bind visible matter to invisible correspondence. To frame an initial is to open and close a gate. To write in disciplined measure is to participate, however humbly, in the repair of creation.

Most brothers in the order know only the outer liturgy. A few know that illumination is a form of controlled metaphysical labor. The greatest among them can correct minor fractures in the world through script, image, and devotion.

The Penitents of the Open Hand

This is the dominant Prayer faction among humans and lesser angelic allies. They believe that the wound of creation must not be widened by force. One must endure, witness, pray, heal, and refuse the seduction of command. Their saints perform miracles of shelter, cleansing, consolation, and revelation. Their flaw is quietism. Under monstrous law, their refusal to impose may become complicity.

The Ars Notoria of the Broken Name

This is the great Sorcery faction. They are scholar-mystics, heretical clerics, court magicians, exiled abbots, mathematicians of stars, and keepers of dangerous alphabets. They hold that a broken creation cannot be redeemed through obedience alone. The world is already misshapen. Therefore intervention is a duty. Their flaw is obvious. In trying to correct the demiurge’s error, they often imitate his will to mastery.

The Thrones of the Court

These are the administrators of Yaldabaoth’s heaven: principates, record-keepers, archons of weather and war, angels of measures, guardians of law, and executioners of judgment. Some love God because they know only loyalty. Some serve because service is their nature. Some privately hate the throne but fear what will follow if it falls.

The Severed Choir

These are the remote descendants, echoes, or condensations of what Yaldabaoth cast out of himself. They are not unified. Some wish reunion, believing that even broken wholeness is better than endless fracture. Some insist reunion would only corrupt what remains pure. Some believe the first severance was incomplete and that God must be cut apart again until nothing tyrannical remains.

The Ashen Ravens

These are mortal survivors of divine violence, descended in part from northern raiders, mercenaries, apostates, and witnesses to miracles. Many first encounter the sacred through terror rather than piety. They become a politically and emotionally important force because they remember what happens when heaven intrudes into history.

The Story

Before the world was broken by history, it was broken by origin. Sophia, moved by unilateral longing, created Yaldabaoth without balance. He awoke great and malformed, and in the first moment of dreadful self-knowledge he cut away his own purity. From that severance emerged the two sustaining currents of the cosmos: Prayer, which receives and endures, and Sorcery, which names and commands. What remained of Yaldabaoth became the wounded ruler of the material world, a tragic maker whose laws were deformed by his own fracture.

Centuries passed under his mistaken heaven. Mortals built kingdoms, plowed fields, buried children, copied books, and called him God. Among them arose secret lineages who dimly preserved older truths. Chief among these were the illuminator monks, whose sacred art was unknowingly a way of participating in creation through pigment, gold, measure, and prayer.

In the ninth century, one such monastery was attacked by northern raiders. In the violence of fire and steel, one illuminator monk entered a state beyond terror and beyond liturgy. In him the severed currents briefly met. Prayer became command. Sorcery became devotion. The miracle that followed annihilated the invaders in a burst of sacred catastrophe.

For that impossible act he was seized from earth and brought before the wounded God himself, who issued a single command that would unmake the monk’s world:

Go and kill me.

Returned to the earth and hunted by powers mortal and celestial alike, the monk must cross a broken creation to discover what that command means. Along the way he finds that every faction in heaven and earth has its own answer. Some want God healed. Some want him dethroned. Some want him further divided. Some want the throne preserved no matter the suffering. Some want to use the monk as a key to reunite the wound for their own ends.

What began as obedience becomes judgment. What began as a quest becomes exegesis. What began as a holy mission becomes a decision about whether a damaged creator deserves mercy, surgery, revolt, or death.

A broken God, who once cut the good out of himself, commands a monk whose art makes light visible to cross a wounded world and decide whether creation is best saved by mercy, correction, further severance, or holy murder.

Introductions

Wed, 11 Mar 2026 07:05:58 -0400

Hello,

This page is to act as a space for me to share and reflect upon my creative output while attempting to bring joy to both myself and the reader. Expect a dev log for a game I am creating and lots of opinions on high tech planted aquariums. Also, because I am a life long student and lover of improv jazz and hip-hop, I will probably give some opinions on albums, songs, or bars that speak to me.

Until next time, here are some pictures of past aquariums of mine. Next time, I’ll explain why I’m moving from 30 aquariums and lots of breeding projects to just 2 tanks, and why that isn’t necessarily an awful thing. Or something.

Take care of yourself. Take care of each other.

Later, Scott