Getting Back into Magic: The Gathering

Pompous pontification on AI and re-mixing genre tropes

2025 update by ChatGPT of the 2004 MTG Legends art by Richard Kane-Ferguson

Return of the Lone Wolf

I was living in Columbia, Missouri, attending grad school in the mid-1990s when I first discovered Magic: The Gathering. It was a brilliant card game that combined the mania for collecting with strategic gameplay. It scratched the pay-to-win itch, sure—but it also rewarded ingenuity. Budget players could build competitive decks using cheap, plentiful cards.

I was never a particularly good player. I was more interested in getting Kasimir the Lone Wolf into play than I was in winning games. I identified with Kasimir—not because of his stats or special abilities (he didn’t have any), but because he looked like Ezferan, my first and favorite Dungeons & Dragons character.

Kasimir cost five mana, including one blue and one white, and had no special abilities. Just a 5/3 body and some evocative flavor. I often held him in my hand, unplayed, waiting for the right combination of lands to bring him onto the battlefield. And when I did bring him out, he didn’t turn the tide in my favor. Focusing on Kasimir wasn’t an effective strategy, but I favored flavor over victory.

I don’t remember building themed decks back then, or thinking much about the game when I wasn’t playing it. I was in grad school and had a lot of demands on my attention. MTG was just something fun to do with friends.

Thirty years later, I’ve returned to Magic—but the landscape has changed. There’s a robust secondhand market, searchable databases, and phone apps to scan and catalog cards. And there’s ChatGPT.

I’ve been using Chaffo¹ not just to build decks, but to flesh out worlds behind them—fictional scenarios that anchor card choices in narrative logic. The AI has been more than happy to assist. I’ve been repurposing elements from other genre properties like Dungeons and Dragons and Warhammer 40K as the starting point for these exercises in generative mythology. Remixing scraps of lore, myth, and cultural debris is what Chaffo does best.

One monstrous faction I repurposed for my personal MTG myth-making was the githyanki, a race from the Dungeons & Dragons universe. The name githyanki comes from a novel by George R. R. Martin, but their backstory—the astral-plane-dwelling, mind-flayer-hating psychic warrior culture—was the invention of science fiction author Charles Stross.

The githyanki entry from the Fiend Folio (1981), a classic D&D bestiary.

Appropriation and Transformation

When I told ChatGPT I wanted to build a world—something loosely inspired by the Githyanki and the Warhammer 40K Eldar—it didn’t hesitate. Within seconds it offered names, factions, motivations, visual motifs, and thematic hooks. Most of it I discarded or reworked. But I wasn’t starting from scratch. I was no longer just building decks—I was modeling an ecosystem.

In Charles Stross’s version of the Githyanki backstory, they were former slaves of the mind flayers, a cruel and powerful psionic species. They won their freedom, but in doing so became tyrants themselves. A rival faction, the Githzerai, split off in an effort to reclaim a moral center and became their sworn enemies.²

Meanwhile, in the Warhammer 40,000 universe, the Eldar³—later rebranded as the Aeldari—were once the dominant civilization in the galaxy. Their decadence and depravity led to the birth of a Chaos god called Slaanesh, an entity formed from excess and desire. In the psychic cataclysm that followed, nearly the entire Eldar race was consumed. The sliver of the remaining Eldar included the Exodites—voluntary exiles who had already turned their backs on decadence, embracing low-tech, hardscrabble lives on Maiden Worlds.

These were the raw materials. Two estranged peoples. Two histories of pride, fall, exile, and return. ChatGPT eagerly began to fuse them.

What emerged was a loosely sketched mythos centered on the Verdani—a forest-adapted culture descended from voluntary exiles from a vast and decadent empire—and the Aethari, the condemned splinter faction from the slave revolt that destroyed that same empire.

Both groups originated in the Old Empire. The Verdani were once citizens who broke away and inadvertantly sealed themselves into a pocket universe, believing they were escaping into another realm. The Aethari were a slave race—shaped by oppression, refined by resistance. They participated in the slave revolt that consumed the Old Empire but then turned against the revolutionary leader turned tyrant. The dominant group crushed the conscientious splinter faction, a few of whom managed to flee and find refuge with the Verdani.

Together, Verdani and Aethari became the seed of a hybrid faction; the Emergenti. When Verdani and Aethari work together, their latent powers amplify: Verdani regain suppressed psionic strength, while Aethari gain concealment from the forces that hunt them. This emergent synergy—not shared history or belief—is what binds them.

Their great adversary is Rha, and the faction that follows her—the Edictari—embodies the logic of domination cloaked as liberation. Rha led the rebellion that shattered the Old Empire, but in doing so became something more than a revolutionary. The scale of her wrath transformed her into a god of judgment, fury, and orthodoxy. The Edictari are her executors: radiant, ruthless, doctrinal.

Chaffo was already familiar with my tendency to frame ideas in terms of complex adaptive systems. Et didn’t need to be told what kind of mythology I was building. Et shaped everything through that lens: power emerging from feedback loops, factions defined not by essence but by interaction, survival depending not on strength or purity but on flexibility and emergent capabilities.

That scaffolding held even as we pulled elements from Stross, Tolkien, Games Workshop, and Dungeons & Dragons. The result wasn’t a pastiche. It was a structure—one tuned to the kind of thinking I’ve been doing for decades, now externalized and accelerated by a machine that doesn’t understand any of it but models it faithfully.

Learning Before the Machine

The Emergenti didn’t emerge from the numinous ether by some mystical process. They’re a variation on a familiar genre trope as viewed through the lens of my particular interests. They’re a specific instance of a familiar archetype—a token of a type. I invoked the type, and Chaffo returned plausible variations. I selected from its offerings, modified, and elaborated.

ChatGPT didn’t originate the ideas. Neither did I, in the strict sense. We both operate within a shared symbolic and genre framework laid down by generations of writers, illustrators, game designers, and fantasists. I didn’t create the language I use to think. I didn’t invent the tropes I’m recombining. But I’m the one imposing structure, making selections, choosing what fits and what doesn’t. The model offers possibilities. I create coherence.

This is exactly what companies like Games Workshop do. They repackage mythic and science fictional archetypes under a proprietary skin, then wrap them in copyright and send lawyers after anyone who gets too close to the aesthetic territory they’ve claimed as their exclusive property. They didn’t create the symbolic inheritance they rely on, but they assert legal ownership over the transformed output—and by extension, over the source material itself. That kind of enclosure of the cultural commons deserves explicit condemnation.

The model doesn’t enclose anything. It returns what I ask for, based on its training data. It doesn’t understand, and it doesn’t care. But it’s good at what it does—offering recognizable, resonant fragments that I can use or discard.

The danger isn’t that it replaces my thinking. The danger is that it might replace the part of the process that’s hardest to learn: getting started. When a project is still formless—when the ideas haven’t locked into place—it’s easiest to let the model fill the void. But that early friction is necessary. It’s how ideas are forced into shape. Skip that stage too often, and you don’t learn how to do it yourself.

I’m from the analog age. I learned to write and think long before the debut of the first web browser, decades before the first large language models. I grew up doing the tedious parts myself—reading, outlining essays, assembling meaning from fragments without a real-time assistant. By the time AI came on the scene, I was already fully formed. I never had the option to let it do everything for me.

Younger people do. They’re growing up with the constant temptation to let the model do the hard parts for them—generate the structure, supply the variation, simulate the voice. If they lean on it too soon, too often, they may never develop the internal capacity to create independently. Their output will be faster, but shallower—less organic, more algorithmic.

I use the model as a collaborator, but I know how to write without one. I’ve spent decades consuming, transforming, and propagating culture. Using AI with that prior formation produces something different than letting it do everything from the outset. Without that foundation, I suspect it becomes much harder to avoid outsourcing the parts that actually matter.

Young people who grow up with AI will develop fluencies I never had. I didn’t build my cognitive habits from scratch—I re-mixed what I inherited, same as they will. But I did it at pre-digital speed. I didn’t have frictionless tools for ideation or output. The limitations of the medium shaped the work—and shaped me. That’s not virtue. I’m not taking credit for the time of my birth. That’s just what it meant to be Gen X.

Freddie deBoer’s recent post, No, I Mean It – AI Maximalists in the Media Should Really, Actually Take the Shitting-in-the-Yard Challenge, cuts through the noise. He distinguishes between AI doomers and genuine skeptics—those who remember how many overhyped revolutions never arrived. The Human Genome Project didn’t give us tailored medicine. Crypto didn’t give us sovereignty.

I feel the pull of AI maximalism, both utopian and dark. Doom and hype are mirror images: both assume inevitable, epochal transformation. Skepticism, by contrast, entertains the possibility that we’re overfitting the narrative. That the hype may land closer to the Human Genome Project than to fire, electricity, or municipal water systems and indoor plumbing.

I’m playing Magic: The Gathering again after a 30-year break, but I’m not doing it as an exercise in slow cognition or enact some idealized rejection of AI. I got back into it because my brother had a stash of cards, and because my oldest son—who doesn’t play Magic but thinks like a shape rotator—is coming to visit. In fact, once I hit “publish,” I’ll head to the airport to pick him up. First stop: the game store, to buy some basic lands.

That’s not a philosophy. It’s just life. And in a time when so much is abstract, optimized, or mediated, that might be the best I can reasonably aspire to.

1. ChatGPT-4o

2. Mind flayers (also called illithids) are iconic D&D villains: squid-faced psychic entities who enslave other beings and consume their brains.

3. Eldar (singular Elda) comes directly from Tolkien. It was the name given to the Elves by the Vala Oromë when he first found them wandering in the starlight of Cuiviénen. Games Workshop, the company behind Warhammer 40K, is notorious for lifting ideas wholesale from broader speculative fiction—from Tolkien’s Eldar, to Herbert’s God Emperor to the Necrons (Egyptian undead meets Terminator)—and then aggressively pursuing copyright infringement cases against anyone who wanders too close to their terrain.