Sunday, August 23, 2026

The Medium That Was: A Eulogy for Interactive Art (1947-2017)






"Every single audience in history has traded permanence for convenience, every single time it was offered the choice, in every medium, without exception, and has then complained about the consequences from within the arrangement it chose."

An Obituary, Written Late

There is a particular cruelty in being the one who notices. 

Everyone else is still talking about the thing as though it were alive. They are making plans around it, buying tickets to its conventions, arguing about its future. And you are standing at the back of the room with the terrible knowledge sitting in your chest like a stone, and you cannot say it out loud because to say it out loud is to be the crank, the reactionary, the man who complains that music was better before, that films were better before, that everything was better before.

But I am going to say it out loud anyway (to a different tune than a few years ago, where I still had hope), because I have carried it for long enough and it has not gotten lighter.

Video games were a medium. Past tense. 

They were a medium for roughly sixty years, from the middle of the twentieth century to somewhere in the middle of the second decade of the twenty-first, and then they stopped being one, and what we have now goes by the same name and occupies the same shelf space and employs many of the same people, but it is not the same thing. It is a different thing wearing the skin of the old thing. It has the vocabulary but not the grammar. It has the iconography but not the faith.

I want to be precise about this, because imprecision is how these arguments get dismissed. I am not saying that nothing good has been released since 2016. I am not saying that the people making things today are less talented or less sincere than the people who came before. I am not saying that I personally have stopped having fun, which would be a small and boring claim about my own tastes. I am saying something structural. I am saying that the conditions that allowed video games to function as an art form — the material, economic, technical, and cultural conditions — were dismantled, deliberately and profitably, and that when you dismantle the conditions for an art form you do not get a worse version of that art form. You get something else entirely

You get a product category.

That is what we have. A product category. A very large, very lucrative, extremely well-lit product category, with an enormous press apparatus devoted to reviewing the products and an enormous influencer apparatus devoted to demonstrating the products and an enormous discourse apparatus devoted to litigating the products, and none of it, not one single part of it, is about a medium anymore, because there is no longer a medium there to be about.

The corpse is still warm and it is still moving and this confuses people. Bodies do that. There are electrical impulses that continue after the fact. There are twitches. What I am going to try to do in the following pages is walk back through the whole thing, from the room-sized machines to the present ruin, and show you where the wound was and when it opened and why it could not close.

I am not going to offer you a silver lining at the end. I want to be honest about that upfront, so you can stop reading now if that is what you came for. There is no silver lining. There is no scrappy underdog movement that is going to restore what was lost. There are a handful of ghosts, and I will talk about the ghosts, and they are beautiful, and they are ghosts, and a ghost is not a person.

The Machines That Filled Rooms

It began, as most things in the twentieth century began, in the gap between what an expensive machine was purchased to do and what a bored engineer wanted to do with it after hours.

In 1947, Thomas T. Goldsmith Jr. and Estle Ray Mann filed a patent for what they called a cathode-ray tube amusement device. It was, functionally, a missile simulator: you turned knobs, a beam of light moved across the face of a tube, and you tried to make it intersect with a target that was not even rendered by the machine — the targets were drawn on a transparent overlay and taped to the glass. This is worth sitting with for a moment. The very first thing that resembles a video game required a piece of paper stuck to the screen with adhesive, because the machine could not yet imagine the world it was supposed to be depicting. The world was outside the machine. The machine only knew how to move a dot.

Everything that follows is the story of the world moving inside.

In 1950, at the Canadian National Exhibition, an engineer named Josef Kates built a machine called Bertie the Brain that played tic-tac-toe against fairgoers on a fourteen-foot-tall display of light bulbs. It was a demonstration of vacuum tube technology, a technology showcase, an advertisement. It was dismantled and forgotten almost immediately. In 1951, at the Festival of Britain, Ferranti exhibited the Nimrod, a computer built for the sole purpose of playing the matchstick game Nim, and the exhibition literature made a point of explaining to the public that the machine was doing serious mathematics and merely condescending to play. In 1952, at Cambridge, a doctoral student named A. S. Douglas programmed noughts and crosses into the EDSAC as part of a thesis on human-computer interaction. In 1958, William Higinbotham, a physicist at Brookhaven National Laboratory who had worked on the timing circuits for the first atomic bomb, connected an analog computer to an oscilloscope so that visitors to the lab's annual open house would have something to do besides look at static displays about nuclear physics. He called it Tennis for Two. Visitors lined up around the building. He did not patent it. He did not think it was important. He thought of it, by his own later account, as a way of making a very dull afternoon slightly less dull, and he spent the rest of his life somewhat bemused that anyone cared

And in 1962, in the Kluge Room at MIT, a group of young men with access to a brand-new PDP-1 wrote Spacewar!.

I want to dwell on Spacewar! because everything is already there. Two ships. A star in the middle with real gravity that pulls you in and kills you. A hyperspace button that will save you from an incoming torpedo but might, at random, destroy your ship for no reason at all — a mechanic they included because they thought a game should have an element of desperation in it. Steve Russell wrote the core of it. Peter Samson, dissatisfied with the randomly scattered dots that served as the starfield, wrote a subroutine called Expensive Planetarium that plotted the actual night sky, accurate to the fifth magnitude, scrolling correctly as the field of view moved. Nobody asked him to do that. Nobody was going to pay him for it. It did not improve the gameplay in any measurable way. He did it because the sky was wrong and it bothered him.

That impulse — the sky is wrong and it bothers me — is the entire art form in embryo. It is not a commercial impulse. It cannot be reverse-engineered from a spreadsheet. It is the thing a person does when they are making something they believe in and they have decided, privately, that it will be correct even in the places nobody will look.

Spacewar! spread across the country on paper tape, hand to hand, for free, because there was no mechanism to sell it and nobody had thought to try. DEC eventually started shipping it with the PDP-1 as a diagnostic program. That was the whole economy: a thing so good that it became infrastructure.

The medium was born in an environment of total non-commerciality, on machines that cost as much as houses, made by people who had nothing to gain. It would spend the next fifty years trying to preserve some fraction of that condition inside a commercial container, and it would succeed, remarkably, for a long time. Longer than it had any right to. And then it would fail completely.

The Coin Slot and the Cabinet

The 1970s did something to the medium that we do not talk about enough: it gave it a body.

Nolan Bushnell and Ted Dabney, having played Spacewar! on university hardware, understood that the thing had commercial potential and that the only way to deliver it to the public was to put it in furniture. Computer Space, in 1971, was a beautiful fibreglass cabinet in a swooping, molded shape that looked like something from a spaceport lounge, and it was a commercial disappointment because the controls were incomprehensible to anyone who had not grown up around a terminal. The following year Pong appeared, and Pong could be explained in a single sentence printed on the cabinet — avoid missing ball for high score — and it made a fortune.

The same year, Magnavox released the Odyssey, built from Ralph Baer's Brown Box prototypes, and it too came with plastic overlays that you taped to your television screen, because the machine still could not draw the world. Twenty-five years after Goldsmith and Mann, we were still sticking paper to the glass. The medium's entire technical history is the story of removing those overlays one at a time.

What the arcade gave the medium was physicality and, crucially, *finality*. A cabinet was a manufactured object. It was welded and bolted and shipped on a pallet. It sat in a bowling alley in Ohio and it did exactly what it did, forever, without revision. If Tomohiro Nishikado's Space Invaders sped up as you destroyed the aliens because the hardware could render fewer sprites faster — an accident, a limitation, a bug that became the most famous difficulty curve in the history of the form — then it sped up, permanently, everywhere, in every cabinet on earth. Nobody was going to patch it. Nobody could. The accident became canonical. The accident became the *point*.

Think about how much of what we love in this medium came from that arrangement. Pac-Man's kill screen at level 256, an integer overflow that renders the right half of the maze into garbage, is not a flaw that was tolerated; it is a place. It has a geography. People made pilgrimages to it. Billy Mitchell built a reputation on reaching it. It exists in the cultural memory as a location, a genuine frontier, the edge of the map where the world runs out — and it exists that way *only because it could not be fixed*. A modern developer would have shipped a hotfix within a week and issued an apology on social media, and the edge of the world would have been quietly paved over, and something irreplaceable would have been lost without anyone noticing it was there

This is the first principle I want to establish and it will run through everything that follows: finality is not a limitation on the art form. Finality is the art form.

A film is a fixed sequence of frames. A novel is a fixed sequence of words. A painting is a fixed arrangement of pigment. Every art form we take seriously is defined by the fact that at some point the artist stopped, and what they stopped at is the work, and the work includes the mistakes, and the mistakes are frequently the best part. The moment you introduce infinite revisability you have not improved the art form. You have destroyed the possibility of a work existing at all. There is only ever a current build.

We will come back to this.

1983, and the Sealed Object

In 1983 the North American market collapsed. Too many consoles, too many publishers, too much shovelware, retailers drowning in unsellable inventory, Atari burying its unsold cartridges in a landfill in Alamogordo, New Mexico — a story that everyone assumed was apocryphal until archaeologists dug them up in 2014 and found that the industry had, in fact, literally buried its failures in the desert.

The recovery was engineered by Nintendo, and the mechanism of the recovery deserves attention because it set the terms for the next thirty years.

Nintendo's insight was that the crash had been a crisis of trust, and that trust could be manufactured through control of the physical object. The Famicom cartridge, and after 1985 the NES cartridge, was a sealed unit. It contained a lockout chip. It bore a Seal of Quality. It could not be manufactured without Nintendo's permission and it could not be altered after manufacture by anyone, including Nintendo. When a game shipped, it shipped. The developer's last chance to be right about anything was the moment the ROM was burned.

We now regard this as a constraint that developers laboured under. I want to suggest that it was closer to a covenant.

Because here is what that constraint produced. It produced a generation of developers who understood, in their bones, that they were making an object that would outlive their ability to touch it. It produced testing regimes of genuine seriousness. It produced designers who thought about the whole shape of a thing before they built it, because you could not discover the shape through post-launch iteration. It produced the phrase "going gold," which is one of the most beautiful pieces of industry vocabulary ever coined — a reference to the gold master disc, the final authoritative copy, the version from which all others would be pressed. Going gold meant: it is finished. It has left us. It belongs to the world now

Nobody goes gold anymore. The phrase survives vestigially in press releases but it means nothing, because the disc that gets pressed is not the game; it is a coaster, an installation stub, an object that requests the real game from a server. The gold master is dead. We killed it and we did not hold a service.

And the object itself — the cartridge, and later the jewel case, and later the big cardboard PC box — was not incidental packaging. It was part of the work. Consider what shipped in a PC box in 1993: the disks, yes, but also a two-hundred-page manual with lore and illustrations, a cloth map, a code wheel, a novella, a set of blueprints, a fake newspaper. Infocom shipped feelies — physical props from the fiction, a matchbook, a piece of stone, a vial of liquid. Ultima shipped cloth maps and metal coins and pewter figures. These were partly anti-piracy measures, sure. They were also worldbuilding delivered through the postal system. The game began before you turned the computer on.

Nothing today is an object. Everything today is an entry in a list on a server that belongs to a corporation, and the corporation can remove the entry, and when it does, the thing is simply gone, and there is nothing on your shelf to prove it ever existed. We will come to that. I am trying to keep it in order.

The Long Golden Afternoon: 1985 to 2000

I do not want to write a nostalgia piece. Nostalgia is a cheap drug and it makes bad arguments. So let me be careful here and say what was actually good about this period in terms that would survive a hostile reading.

What was good was that the medium had, briefly, all of the following at once:

Sufficient technical capability to express nearly any idea. By the mid-nineties the hardware had escaped the era where every design decision was a negotiation with the sprite limit. Not entirely — never entirely — but enough that a designer could conceive of a thing and then build approximately that thing.

Insufficient budget to require mass appeal. This is the crucial one and it is almost never stated properly. A Super Nintendo game could be made by a team of fifteen people in eighteen months. A PlayStation game could be made by a team of thirty. At those budgets, a game did not need to sell to everyone. It needed to sell to *someone*. And when a work only needs to find its audience rather than the entire audience, you get specificity, and specificity is the precondition of every good thing that has ever been made in any medium at any time.

A distribution system that rewarded completed objects. Retail. Shelf space. A thing in a box that had to be worth the money on the day it was purchased, because there was no second chance, no roadmap, no season one.

A critical culture in its infancy that had not yet been captured. Magazines, mostly. Frequently corrupt, frequently stupid, but operating in a market where the review arrived after the object was already fixed, which meant reviews described a thing that would still be that thing when you got home.

Look at what came out of that configuration. 

Look at 1998 alone: Ocarina of Time, Half-Life, Metal Gear Solid, Grim Fandango, Baldur's Gate, Thief, Resident Evil 2, Panzer Dragoon Saga, Banjo-Kazooie, StarCraft, Xenogears. That is not a good year. That is an eruption. That is what a medium looks like when it has found its footing and has not yet been financialised, when a form is mature enough to do anything and not yet valuable enough to be worth destroying.

And note the shape of those things. They are *complete*. Grim Fandango has an ending, and the ending is the ending, and Tim Schafer did not return two years later with a narrative expansion pack that recontextualised Manny's fate for $19.99. Metal Gear Solid ends. Thief ends. Half-Life ends, and its ending is a strange, hostile, deliberately unsatisfying conversation on a train, and Valve shipped it that way and let it stand, because in 1998 you shipped the thing you meant and then you lived with it.

That is what an artist does. An artist makes a decision and then does not get to take it back. Everything about the modern arrangement exists to spare people from that experience, and the sparing is precisely the damage.

The Tether

I have thought for a long time about where to date the wound, and I keep arriving at the same unglamorous place: not a game, not a scandal, not a single betrayal, but an infrastructure decision.

In November 2002, Microsoft launched Xbox Live.

Everyone remembers this as the arrival of console multiplayer, headsets, voice chat, teenagers screaming slurs into the void of Halo 2. That is the surface. What actually happened is that a console was permanently connected to a server owned by its manufacturer, and once that pipe existed, everything that could travel down it eventually did. Patches. Content. Advertising. Telemetry. Revocation. Every subsequent development in this story is just water finding the channel that was dug in 2002.

In September 2003, Valve launched Steam. It was, at the time, universally loathed. It was slow, it was buggy, it broke games that had previously worked, and its principal function from the user's perspective was to prevent you from playing a product you had purchased until it had finished talking to Bellevue, Washington. When Half-Life 2 shipped in 2004 and required Steam activation even for the retail boxed copy, the discourse was ferocious. People understood exactly what was happening. They said so, loudly, in the plain language of people who have not yet been trained out of it: *I bought this and I cannot play it and a company I did not agree to deal with is now standing between me and my property.*

They were right. They were completely right. And they lost, and they lost so thoroughly that within a decade the same population would be defending the arrangement, evangelising it, mocking anyone who raised the objection they themselves had raised, and treating a seasonal discount on a licence that can be revoked at will as a form of consumer liberation.

That is how these things go. 

Then, in April 2006, Bethesda put a set of decorative armour for a horse in The Elder Scrolls IV: Oblivion on the Xbox Live Marketplace for two dollars and fifty cents.

The horse armour is a joke now, a meme, a punchline so worn down by repetition that it has lost the capacity to mean anything. I want to try to make it mean something again. What the horse armour established was not that publishers were greedy — everyone already knew publishers were greedy. What it established was a *category*: content that exists, that is finished, that is on the disc or could have been on the disc, and that is nonetheless not part of the work, because the publisher has drawn a line around it and priced it separately.

Once that category exists, the work has no boundary. And a thing with no boundary is not a work. It is a supply.

Follow the logic forward and you get everything. If content can be carved out and sold separately, then the correct commercial strategy is to design games that maximise the carvable surface area. If the game can be updated after release, then the correct strategy is to release before it is finished and let the audience do the quality assurance, unpaid, as a community. If the game can be tethered to a server, then the correct strategy is to tether it, because tethering permits measurement, and measurement permits optimisation, and optimisation of a game means optimising the extraction of money and time from a person, which is a different activity from making something good and is in most cases directly opposed to it.

None of this required villains. That is what people misunderstand when they want to name a bad executive. It required only a set of incentives and about ten years, and by the end of that decade the incentives had eaten the medium down to the bone with the polite, patient thoroughness of a fungus.

The Decade of Enclosure: 2007 to 2015

Watch the enclosure happen. It is almost tidal in its regularity.

2007. The Orange Box ships and Team Fortress 2 begins receiving free content updates, which everyone correctly perceives as generous. Two years later it will receive a hat economy and an item drop system, and the generosity will have been the loss leader

2008. Development budgets for major releases cross the eight-figure line as standard. The number of publishers capable of financing a top-tier release contracts sharply. Studios that had made three idiosyncratic games in five years begin making one enormous game in five years, and the one enormous game must sell four million copies to break even, and a thing that must sell four million copies cannot afford to be strange.

2009. Fumito Ueda's Team Ico shows a boy and an enormous feathered animal at the Tokyo Game Show. Hold that thought. We are coming back to it. It is the hinge of this entire essay.

2011. Sony's PlayStation Network is compromised and taken offline for twenty-three days. Seventy-seven million accounts. For over three weeks, people who had purchased digital games discovered the terms of the arrangement they had entered into: the games were not theirs, the library was not a library, and when the server was down the shelf was empty. There was outrage. There were free games as compensation. Nothing changed. Everyone went back.

2013. Two events in the same year, four months apart, which together constitute the last moment the audience had any leverage. In March, EA and Maxis released SimCity with a mandatory always-online requirement, justified by claims about server-side simulation that were shortly demonstrated to be false when players discovered the game ran perfectly well disconnected. The launch was a catastrophe; nobody could play; the servers buckled. In May, Microsoft revealed the Xbox One with a licensing regime that would have required a periodic online check-in and restricted the resale and lending of physical discs. The backlash was violent enough that Microsoft reversed the policy within six weeks.

That reversal is the last victory. I want to state that plainly. June 2013 is the last time the audience for this medium collectively refused something and the industry retreated. Everything the Xbox One was pilloried for in 2013 is now simply the ambient condition of the medium, adopted voluntarily, universally, without a fight, because it arrived slowly rather than all at once and because the generation that would have objected had by then aged into irrelevance.

2014. Assassin's Creed Unity ships in a state that produced screenshots of faceless characters — eyes and teeth floating in the void where a head should be — and the images went around the world as comedy. But look at what the reaction actually was. Not this should not have been sold. The reaction was they will patch it. And they did patch it, and the patches were themselves enormous, and the story ended with a shrug. The Overton window had moved. The unfinished product was no longer a scandal; it was a phase.

The same year, Bungie released Destiny, and the phrase "live service" entered the general vocabulary, and a game became a place you were expected to return to daily, forever, on a schedule set by someone else, in exchange for numbers that increase.

2015. The year everything ends, and it ends four separate times.

In February, Ready at Dawn released The Order: 1886 — an original property, no sequel, no franchise, a self-contained ten-hour story about a Victorian London of half-breeds and arc-lightning rifles, presented in a fixed cinematic aspect ratio with black bars its creators refused to remove. It was destroyed. Not criticised — destroyed, culturally, on the grounds that it was too short and too linear and did not offer sufficient value per dollar. There would be no sequel. The studio moved to contract work and virtual reality. The lesson was legible to every publisher on earth: a bold, complete, finite, authored object is a financial liability, and the audience will punish you for making one.

In April, Konami removed P.T. from the PlayStation Store. Not delisted-but-redownloadable — removed, unobtainable, gone. A playable teaser, one of the most formally inventive pieces of interactive horror ever constructed, and it ceased to exist for anyone who had not already installed it. Consoles with it still on the drive began selling for thousands. That is not a market anomaly. That is a medium discovering that its works can be *unmade*, retroactively, by an accountant, and that there is no library, no archive, no vault, no Criterion, nothing standing between any of it and deletion.

In September, Metal Gear Solid V shipped visibly incomplete, its third act existing as unfinished assets on the disc, while its director's name was being systematically scrubbed from its own marketing materials by a publisher that had decided the prestige game business was less profitable than pachinko.

And in November, Fallout 4 arrived to enormous sales and a season pass, and the arrangement was complete, and no one was angry anymore.

Four events. An original work punished for being finished. A work erased. A work amputated. And a work absorbed. That is the whole disease presented in a single year, in sequence, like a demonstration.

The Last Guardian, or: The Boy and the Animal at the End of the World

On the sixth of December, 2016, Sony released The Last Guardian.

It had been in development, in one form or another, since roughly 2007. It had been announced in 2009. It had changed hardware generations. It had been declared dead, publicly, more than once. Its trademark had lapsed. Its director had left the employ of the company making it and had come back as a contractor to finish it, because he would not let it go and they would not let him have it. It arrived seven years late into a market that had reorganised itself completely in the interim, and it arrived exactly as it had been conceived, which is to say: as an object from a dead world, carried forward intact, like something recovered from a shipwreck.

I want to say why I have chosen this as the terminus, because I know that a lot of people did not like it and I am not making an argument about quality.

Here is what The Last Guardian is. It is a game in which you are a small boy, and there is an enormous animal, and the animal is not a mechanic. That is the whole thing. The animal does not obey you reliably. You issue a command and Trico ignores it, or misinterprets it, or does it forty seconds later after becoming distracted by a butterfly. Players complained about this constantly. Reviewers complained about it. There were long, patient discussions about whether the "AI was broken."

The AI was not broken. The AI was the argument. Fumito Ueda had spent a decade building a system whose entire purpose was to deny you the sensation of control, because the subject of the work was the difference between commanding a creature and being trusted by one, and you cannot make a game about earning trust if the creature does what you say. The friction was the thesis. The frustration was load-bearing. Every complaint about the controls was a complaint about the presence of an idea.

And it was finished. It ended. It had an ending, and the ending was devastating and complete and did not set up anything, and there was no season pass, no roadmap, no live-service layer, no cosmetic store selling alternative feather patterns for Trico, no post-launch content cadence, no engagement loop, no daily login. It was a story that a man had wanted to tell for ten years, and he told it, and then it stopped, and that was all.

Nothing like it has been made at that scale by that kind of company since, and nothing like it will be, and I do not say that rhetorically. The financial conditions that permitted a first-party platform holder to fund a decade-long personal project with no franchise potential and no monetisation surface no longer exist. They did not degrade. They were removed. Sony's own subsequent output tells you exactly what replaced them, and it is not shameful work, and it is not the same species of thing.

So that is where I am putting the stone. December 2016. The boy and the animal.

You could argue for other dates. You could put it in 2015, with The Order buried and P.T. deleted. You could put it in 2017, when Star Wars Battlefront II shipped with a progression system so nakedly predatory that governments began writing gambling legislation about it, and the response of the industry was not to stop but to get subtler. You could put it in 2019, when Google announced that games would be streamed from a data centre and you would own nothing at all, an experiment that failed only because Google is bad at finishing things and not because anyone objected to the principle.

But I like the boy and the animal, because a medium should be allowed a proper funeral, because of the 'Last' in the name evoking the last of something, and because a proper funeral has a body. That is the last body we got. Everything after is exhumation.

What Occupies the Grave

Let me now describe, without euphemism, what the thing called "video games" actually consists of in the present day.

You do not own anything. This is not a philosophical provocation; it is the plain text of the licence agreement you accepted, and in the last two years it has stopped being subtext. In 2023 a publisher announced it would terminate the servers for a racing game and render it permanently unplayable — a single-player-capable game, purchased, installed, on people's machines, revoked. In 2024 an executive at that same publisher said publicly that players would need to get comfortable with not owning their games. He was not being sinister. He was describing the business he works in accurately, and the only surprising thing about the statement was the frankness. Storefronts for two entire console generations have been closed. Thousands of works are now unpurchasable by any legal means and exist only because of criminal activity by unpaid volunteers who will themselves eventually die or lose interest.

Imagine a film medium in which the negatives are destroyed by the studio on a rolling basis for tax reasons and the only surviving copies are bootlegs, and in which this is not a scandal but the ordinary operation of the industry. That is the situation. There is no Library of Congress here. There is no BFI. There is a handful of exhausted archivists and a legal regime that classifies their work as piracy.

Nothing is finished. The release date is now the beginning of the development cycle rather than its end. Day-one patches larger than the entire product they are patching. Roadmaps published in advance, so that the audience knows what the work will eventually become and can therefore never experience it as a thing that *is*. Early access as a permanent condition. Games that have been "in development" and simultaneously on sale for eight, nine, ten years. Games whose ending is added in a subsequent chapter. Games that are silently rebalanced in the night, so that the version you played and the version your friend played and the version you will replay in five years are three different objects with the same name, and no one of them is authoritative, and criticism therefore becomes impossible because there is no fixed text to criticise.

That last point is more serious than it sounds. Criticism requires an object. When the object mutates continuously, the critical apparatus collapses into consumer advisory — is it worth your money *right now* — which is exactly what happened, and it is why almost nothing written about games today is criticism in the sense that a person writing about a film or a novel would recognise. It cannot be. There is nothing to hold still.

The work is designed around extraction rather than expression. The battle pass. The daily. The weekly. The seasonal reset. The engagement metric. The retention curve. There are people employed in this industry whose job is to determine the optimal interval at which a reward should be granted to sustain a behavioural loop, and they are drawing on the same literature as slot machine designers, because it is the same literature and the same loop. This is not a fringe practice. It is the dominant revenue model of the largest products in the medium, and the largest products in the medium constitute the overwhelming majority of the hours played.

A form in which the most-consumed works are optimised for compulsion is not an art form having a bad period. It is a gambling industry with a rendering pipeline.

The economics have eaten the people. Between 2023 and 2025 something on the order of thirty thousand people were made redundant across this industry, including entire studios shut down in the same week they shipped critically acclaimed work, including studios shut down *by owners who had bought them eighteen months earlier*. Budgets of two hundred million, three hundred million dollars for single releases, which means that a release must reach an audience of a size that forbids specificity, which means every one of them is the same, which is why they are the same. It is not a failure of imagination. It is arithmetic. You cannot spend three hundred million dollars on a strange idea. You can only spend it on a familiar one, executed at a level of surface fidelity that nobody asked for and that has never once been the reason anybody loved anything.

And...there are no more new things. Look at the release calendar of any given year now and count the works that are not sequels, remakes, remasters, adaptations of existing properties, live-service entries in an existing ecosystem, or franchise extensions. In the golden period, a major publisher would launch several genuinely new properties a year as ordinary business. Now a new property from a major publisher is an *event*, discussed in the press as an act of institutional courage, and it usually fails, and its failure is then cited internally as evidence for never doing it again.

The Order: 1886 is the ghost that haunts every greenlight meeting in this business. Not because it was good — it was uneven, and its critics were not simply wrong — but because it was the last major publisher-funded attempt at a self-contained original work with a defined ending and no monetisation apparatus, and the market taught everyone the lesson it was always going to teach.

On the Objection That I Am Simply Old

I want to deal with the counterargument honestly, because I have made it to myself many times at three in the morning and it deserves better than to be strawmanned.

The counterargument goes: every medium has a period that its adherents believe was the peak, and that period is always coincident with the adherent's adolescence, and therefore the claim is not about the medium but about the adherent's own diminishing capacity for wonder. You are not mourning video games. You are mourning being fourteen.

I have interrogated this and I do not think it holds, for one specific reason.

The nostalgic complaint is a complaint about *quality*: they don't make them like they used to, the new ones are worse, the writing is lazy. My complaint is not about quality. It is about *ontology*. I am not saying the products are worse. Many of them are astonishing. The craft on display in a major release today would have been science fiction in 1998. The animation systems, the lighting, the density, the sheer number of extraordinarily skilled people bringing their whole lives to bear — it is genuinely magnificent and I do not sneer at it.

I am saying that the thing they are making is not the same *kind of thing* anymore.

If you told me that novels still existed but that they were now sold by the chapter, that the chapters could be revised by the publisher after purchase without notice, that the ending was released eighteen months after the beginning, that your copy would be recalled if you stopped paying a monthly fee, that the author's name was removed for legal reasons, that the text contained periodic prompts to purchase additional descriptive adjectives, and that all previous novels had been destroyed except for illegal photocopies — you would not say "the novel is going through a rough patch." You would say the novel, as a form, had ended, and that a different commercial product had taken its distribution channel.

That is my claim. It is not about taste. You can test it structurally. Take any work from the medium's mature period and ask whether it could be greenlit, funded, completed, and released today in its actual form. Not a version of it. It. The answer, for very nearly all of them, is no — and the reasons are never artistic. The reasons are always that it is too short, too finite, too specific, too uncommercial, too unmonetisable, too unlikely to sustain a two-year content roadmap. Which means the form has lost the ability to produce its own canon. A medium that could not make its own masterpieces if it tried is not a living medium. It is a genre that used to be one.

And the second objection: *but there is more good stuff released now than ever before.* True in raw volume, and irrelevant. The medium did not die of scarcity. It died of the conditions under which things are made. A thousand small good things do not add up to a form. They add up to a scene, and a scene is what remains after a form collapses.

Which brings me to the ghosts.

The Ghosts

There are still people making complete objects, and I want to be careful here because this is the place where an essay like this usually turns and offers you the consolation, and I am not going to do that. I want to describe the ghosts accurately, which means describing both what they are and what they are not.

In 2024, a man named Billy Basso released Animal Well. He had worked on it alone for approximately seven years. It is a metroidvania of extraordinary density, built on a custom engine, in which almost nothing is explained and the world is layered with secrets nested inside secrets to a depth that the community required weeks of collective effort to reach, and there is a final tier that was designed to be unreachable by any individual player working alone. It shipped finished. It fits on a floppy disk's worth of ambition and a lifetime's worth of craft. It is one of the best things this medium has produced in twenty years and I say that without qualification.

And it is a ghost. Here is why.

It is a ghost because it is *one man*, working for seven years, in isolation, on a thing that no institution would have funded and that no structure exists to reproduce. There is no pipeline behind it. There is no studio that will now make five more of these. When Basso is finished, that is the end of the lineage; there is no apprentice, no house style, no accumulating institutional knowledge, none of the things that make a form rather than a series of miracles. The golden period was not golden because individual geniuses existed — individual geniuses always exist. It was golden because there were *institutions* that could take a talented twenty-four-year-old, put them on a team, ship three games in five years, and produce a designer at thirty who knew how to build a whole world. Those institutions are gone. What remains is people teaching themselves in the dark, alone, for a fraction of a wage, and occasionally one of them emerges holding something perfect, and we all crowd around it like villagers around a saint's relic, and then it is over.

That is not a medium. That is folk art after the fall of the empire that supported the cathedrals.

And there are others. There will keep being others (I myself have designs, through The Quantumyth, to this end: 100 individual self-created, one-man-audience titles, all developed through this blog). 

Every year or two something arrives that is finished, and specific, and made by people who believed in it — a game with no roadmap, no season, no store. And each time, the reaction is instructive: the entire discourse treats it as a *miracle*. As a rebuke. As proof that it can still be done. And that reaction is itself the evidence for everything I have written here, because you do not call a thing a miracle when it is the normal operation of the world. In 1998 nobody wrote three thousand words about how remarkable it was that Grim Fandango was finished when you bought it. It was finished when you bought it because everything was. The miracle framing is an autopsy report.

So: the ghosts are real, and they are beautiful, and they are ghosts. A ghost is not a person. A ghost is the shape a person left in the air. You can love it. You should love it. But you should not mistake the loving of it for the presence of the thing that made it, because that thing is not coming back, and pretending otherwise is how you spend the rest of your life waiting at a station for a train that was decommissioned.

Why It Cannot Come Back

I am asked, when I say these things aloud to people who are kind enough to indulge me, what would have to happen for it to return. As though the question were open. So let me close it.

For the medium to return, the following would have to occur simultaneously:

Ownership would have to be restored, which would require legislation against the licensing model in every major market at once, against the unified opposition of every software company on earth, an industry with vastly more lobbying capacity than the audience has organisational capacity. Efforts exist. Petitions exist. They are the political equivalent of the ghosts: sincere, admirable, and structurally incapable of reversing a settled arrangement that generates hundreds of billions of dollars annually.

Post-release revisability would have to become socially unacceptable, which is impossible, because it is convenient, and every single audience in history has traded permanence for convenience every single time it was offered the choice, in every medium, without exception, and has then complained about the consequences from within the arrangement it chose.

Budgets would have to collapse back to a level where specificity is survivable, which would require the audience to stop rewarding surface fidelity, which it will not do, because surface fidelity is legible in a thirty-second clip and specificity is not.

The archive would have to be reconstructed from a position where most of the material is already lost and the legal framework criminalises the reconstruction.

And an entire generation would have to develop a taste for finitude — for works that end, that do not want anything further from them, that offer no progression outside themselves — having been raised from childhood inside systems specifically engineered to make finitude feel like deprivation. You cannot un-train that. It is not a preference. It is a nervous system.

Every one of those is individually impossible. They would all have to happen together.

It is over. Not "over unless." Over.

Coda

I keep thinking about Peter Samson and the starfield.

A young man in 1962, working on a machine the size of a wall, in a room he did not own, on a program with no commercial purpose whatsoever, decides that the fake sky is not good enough. So he sits down and encodes the real one — the actual constellations, magnitudes correct, scrolling properly — into a toy that will be given away for free to anyone who asks. There is no money in it. There is no credit in it; his name is a footnote and most people who have played the descendants of that program have never heard it. He did it because the sky was wrong.

That is the beginning of this medium. A man correcting a sky that nobody was going to check.

And the end of it, sixty-odd years later, is a boy and an animal, made by another man who spent a decade on a creature that would not obey, released seven years late into a market that had stopped having any use for such a thing, and then silence.

Between those two points there was an art form. It was young and vulgar and frequently stupid and it produced perhaps twenty or thirty works that will stand as long as anything from the twentieth century stands, and it did this in about forty years, which is an astonishing rate, and it did it under conditions that were themselves an accident — a brief window in which the technology was capable enough to say anything and the money was small enough to let you.

The window closed. It closed the way windows close: because someone found it more profitable shut.

What remains is not nothing. It is worse than nothing, in a way, because it is *almost* the thing. It has the same names on it. The same studios, some of them, the same franchises, the same faces at the same conferences saying the same words. It is a very good imitation and it is getting better at imitating every year, and there are children being born right now who will grow up inside it and will never know that there was ever anything else, and they will be perfectly happy, and their happiness will not be a refutation of anything I have said. They will simply have been born after.

I do not have advice. I am not going to tell you to support the indies, though you should, or to buy physical, though there is barely any such thing anymore, or to keep hope alive. Hope is a discipline appropriate to situations with uncertain outcomes and this outcome is not uncertain. This one is settled. The right response to a settled loss is not hope; it is grief, and then, eventually, accurate memory.

So that is what this is. Accurate memory.

There was a medium. It began with a dot on an oscilloscope in a physics lab because a man wanted the open house to be less boring. It ended with a bird-dog the size of a house who would not come when called. In between, for about half a century, people put things in boxes and sold them and could never take them back, and because they could never take them back, they meant them.

That is all a work of art has ever been. Something a person meant and could not retract.

We do not make those anymore.

Thursday, August 20, 2026

Black Myth: Zhong Kui (A Culture and Mythology Reading)

 

The Clam, the Paper Man, the Water Ghost, and the Man Who May Become The Ghost Hunter

Game Science has finally shown us what Black Myth: Zhong Kui actually is, or at least what it is beginning to become, and the fifteen-minute gameplay demonstration released on 20 August 2026 is frankly astonishing. This is the first substantial playable footage from the second Black Myth game: a long, strange procession involving an unnamed swordsman, an elderly traveller, a wooden raft, an overwhelming wall of water, human-shaped enemies, a man who becomes a sheet of yellow paper, hound-like things, long-haired ghosts, Daoist-looking combat iconography, a gigantic humanoid clam called Dà Bàng Zhēnjūn — 大蚌真君, a towering black-draped ghost who can apparently be summoned and inhabited, an extraordinary woman erupting from the water, skeleton musicians, malformed bodies, bound creatures, figures sitting in pots, mysterious ritual specialists, and finally the unmistakable red-bearded Zhong Kui riding his tiger. Game Science describes the footage as an early work-in-progress build running at 21:9 and specifically calls it our first look at “the protagonist” in combat and at fragments of the story. That wording is already interesting: they do not simply call the playable swordsman Zhong Kui. Art director Yang Qi has subsequently stressed that the area shown is essentially a test map, and even said the protagonist currently lacks dialogue because they have not finished selecting his voice actor. The newly updated official Q&A is almost comically resistant to explanations, saying that fan interpretations are already more elaborate than the development team's own present understanding of the project. So everything below needs a clear hierarchy: there are things the trailer unambiguously shows, things Chinese folklore gives us strong reasons to associate with those images, and theories which Game Science may deliberately encourage without having committed to them yet. 

And yet I think the trailer already reveals something far more fundamental than plot. Game Science seems to have found the conceptual engine for its second mythology. Black Myth: Wukong drew much of its playable identity from the Monkey King's body: extension, agility, duplication, transformation, the staff, the extraordinary physical possibilities inherent in Sun Wukong himself. Zhong Kui appears to be approaching the supernatural from another direction entirely. Here the operative ideas are judgement, invocation, command, embodiment, paper, spirits and uncertain categories. What exactly is a ghost? When is a ghost an enemy? Can a ghost become a companion? Can a human become a ghost? Can a god appear as an old man? Can a man become paper? Can a monster carry an honorific normally associated with exalted beings? Can one spirit attack another? Can a dead thing sit under a flowering tree and play the erhu for pleasure? The original official text accompanying Zhong Kui already asked precisely these kinds of questions. Game Science's 2025 introduction begins with the idea that demons arise from the human heart and ends with Zhong Kui drawing his sword to distinguish the “clear” from the “turbid”; elsewhere the developers have said the central subject of the game will be life and death, deliberately choosing a theme that can function without the player having read a single foundational Chinese novel beforehand. Zhong Kui is uniquely suited to this because there is no single Journey to the West-equivalent text that owns him. He accumulated across ritual, popular religion, painting, theatre, storytelling, New Year imagery, Daoist practice and later fiction. Game Science therefore has centuries of overlapping Zhong Kui traditions rather than one authoritative narrative to adapt. 

The First Mystery: Is the Swordsman Zhong Kui Before He Becomes Zhong Kui?

My current read is yes, quite possibly — and I think this is presently the strongest single theory — but it is absolutely not confirmed. More precisely, I suspect Game Science wants us to consider whether the playable swordsman is a mortal or pre-apotheosis Zhong Kui, before death turns him into the enormous red-bearded demon-queller shown riding the tiger at the end. The reason this theory has accelerated in China today is not simply that the game's title says Zhong Kui. The protagonist has a prominent injury or scar around the head, carries a long sword whose broad visual identity recalls the weapon associated with the fully formed Zhong Kui, and is deliberately presented separately from the monstrous Zhong Kui we see at the trailer's conclusion. Chinese players immediately connected that head wound to one of the most famous later versions of Zhong Kui's mortal story: the gifted but physically unattractive scholar succeeds in the examinations, is rejected or humiliated because of his appearance, and kills himself by striking his head against the palace architecture. CCTV's descriptions of the traditional opera Zhong Kui Marries His Sister preserve variants in which he “碰死” or “撞阶而死” — strikes himself against the palace gate or steps and dies — after the court refuses him because of his appearance. In other tellings he kills himself differently, which itself demonstrates how unstable the biography has always been. Today's Chinese discussion is therefore not inventing the scar theory from nothing: it is matching a conspicuous visual wound with one of the best-known narrative events in Zhong Kui's post-medieval biography. 

There is another reason I find the theory persuasive: life and death is exactly the theme Game Science says it is pursuing, and Zhong Kui possesses an almost perfect character arc for dramatising the passage between them. He is one of those unusual supernatural figures whose identity depends on death. The scholar dies; the ghost becomes an officer, king, judge, general or demon-queller depending on the telling; the rejected man gains in the other world an authority denied to him by the human one. If Game Science wants to make life and death experientially comprehensible rather than merely discuss them in dialogue, letting us spend the first part of the game inside the mortal Zhong Kui before crossing the border ourselves would be extraordinarily elegant. The final tiger shot could consequently be a glimpse of where this scarred swordsman is going rather than proof that the swordsman and Zhong Kui must be separate people. The radical change in his appearance would also be thematically appropriate. Zhong Kui's famous frightening face is deeply intertwined with stories about ugliness, social judgement and his posthumous identity; a game about his becoming could literally make bodily transformation part of the tragedy. Chinese folklore never supplies Game Science with one mandatory sequence of events here, which gives the studio enormous latitude. CCTV's older survey of the tradition even notes that historically there may have been no single real Zhong Kui at all and discusses the theory that the name developed from zhongkui/zhongkui as an older exorcistic implement before the figure became increasingly personified. 

The strongest alternative being discussed is Du Ping, Zhong Kui's friend and eventual brother-in-law in the enormously popular Zhong Kui Marries His Sister tradition. Du Ping finances or assists Zhong Kui on his examination journey, later gives him an honourable burial, and is rewarded after Zhong Kui becomes a supernatural authority when Zhong Kui returns with his ghostly entourage and marries his younger sister to him. This gives Du Ping an obvious reason to stand adjacent to Zhong Kui's story without being Zhong Kui himself, and fans understandably reach for him once the trailer shows the playable swordsman and the tiger-mounted Zhong Kui as separate images. But there are problems. The traditional Du Ping is defined primarily by generosity, friendship and filial/social virtue; he is not conventionally a demon-hunting swordsman with supernatural possession abilities. Game Science could reinvent him, certainly, but at present the playable character's sword, scar, centrality and potential death-transition strike me as a tighter configuration. There are also theories that he is an earthly vessel, disciple or partner of Zhong Kui rather than either man. Game Science has intentionally preserved all these options. Even Chinese same-day coverage is careful to report the mortal-Zhong-Kui, Du-Ping and “human vessel” ideas as speculation. 

There is a wonderful additional possibility hidden in Zhong Kui scholarship itself: there may never have been one historical man waiting for Game Science to reproduce accurately. One long-standing antiquarian explanation treats “Zhong Kui” as the personification of an older exorcistic object or concept, while other traditions connect his development with ritual expulsion, door guardianship and demon-quelling imagery. The character gathered biography over time. By the Tang and Song periods we have the great dream story in which an enormous ghost destroys the smaller spirit troubling the emperor, followed by the celebrated image of Zhong Kui supposedly painted by Wu Daozi; later storytelling gives him examination injustice, suicide, Du Ping, a sister, ghost armies and adventures cataloguing the moral deformities of the human world. The Black Myth version therefore has an unusual licence: Game Science can create a human “origin” which feels ancient without pretending there has always been a single canonical one. 

The Old Man, the Raft, the Wall of Water — and Why He Becomes Paper

The beginning might be my favourite part of the entire trailer because it instantly refuses to explain itself. The protagonist encounters an elderly man in a mountainous waterside region; the pair work together around a crude wooden raft; a huge surge of water overwhelms the scene; after the protagonist gets the old man settled and fights through the hostile group nearby, the old fellow eventually ceases being an elderly human body at all. He becomes a yellow sheet, talisman or paper image which escapes through the air, and the protagonist follows it. This sequence is confirmed in detailed Chinese recaps of the footage, while the unnamed man's performance has already been identified as using the facial/performance work of veteran Beijing People's Art Theatre actor Li Guangfu. His character has not been officially named. 

The temptation will be to immediately declare him some named god — a local earth god, an immortal, Zhong Kui in disguise, a Daoist master — but there is nowhere near enough evidence yet. What is much safer, and much richer, is to recognise the cultural technology of the transformation. Chinese ritual traditions have a deep history of treating paper images as interfaces between visible human life and invisible beings. Shenma, jiama and zhima — divine images, “paper horses” and related printed ritual sheets — could represent gods, spirits, supernatural officials and their means of conveyance. The Daoist Culture Database describes such paper images as symbolic representations of divinities and records an old explanation in which the paper itself serves as a material support or locus upon which the divine being may depend. Museum and arts scholarship similarly describes paper-horse images as ritual carriers: figures are printed, invoked, displayed, burned or otherwise dispatched in acts linking the worshipper with divine and spirit domains. In other words, Game Science showing a seemingly flesh-and-blood old man suddenly resolving into yellow paper is drawing on a visual idea that Chinese audiences can read immediately: the body we were looking at may have been an emanation, effigy, temporary embodiment, ritual proxy or summoned presence, rather than an ordinary biological man. 

The yellow material also naturally evokes fu and fulu, Daoist talismanic writing and diagrammatic forms associated in historical sources and popular practice with commanding, invoking, restricting or expelling spirits. I would not say the old man “is a Daoist talisman” as a factual identification, because we need to see the marks on the paper clearly before saying precisely what Game Science is quoting. But the trailer deliberately places paper metamorphosis inside a game whose user-interface symbols are already being compared by Chinese researchers and creators with components of Daoist talisman designs. A same-day analysis highlighted by GamerSky identifies forms resembling a fire-wheel design, the head and central structure of Sanqing talismans, and yunzhuan or “cloud-seal” forms. That reading is currently a fan/research comparison rather than a Game Science annotation, but the convergence is striking: the protagonist's powers, the old man's body and the visual interface all appear to be speaking related ritual languages. 

Then there is the raft and the water itself. I would be very cautious about searching for one named myth to decode a raft, because in this case the broader structure is stronger than any premature identification. Shores, crossings, flooded ground and unstable boundaries are natural spaces for a game about life and death: land and water interpenetrate; solid ground gives way; apparent human beings become spirit-media; submerged entities later erupt into the fight. The sequence is almost a declaration that physical categories cannot be trusted. The old man is introduced as a mundane travelling companion, the sort of NPC one might escort through an action RPG, and then Game Science simply takes his body away from him. From that moment onward, every person in the footage deserves suspicion.

There is even an intriguing possibility that the water imagery connects indirectly to the clam boss itself. This is speculative, but it has a real mythological foundation. In ancient Chinese imagination the shèn (蜃) could be conceived as an enormous clam or related marine creature whose exhalations produced phantom buildings and cities over the sea — one of the traditional explanations behind haishi shenlou, the mirage. Shiji's Treatise on Celestial Offices includes the famous idea that “shèn vapour” beside the sea resembles towers, and later accounts elaborate the creature as an immense bivalve capable of generating deceptive visions. The game's boss is called Dà Bàng Zhēnjūn, literally something close to “Great Clam Perfected Lord,” rather than Shèn, so there is no basis for saying these are identical. But putting a giant supernatural clam into a waterside chapter already involving unstable appearances makes the old association worth holding in mind. If later material reveals illusions, false landscapes, dream architecture or deceptive manifestations around this creature, the shèn tradition will become much more significant. For now it is resonance, not identification. 

The Human-Shaped Attackers and the New Sword Combat

Once the water sequence breaks, the protagonist fights a group of largely human-proportioned enemies wearing white coverings and broad-brimmed hats and wielding knives, hammers and spears. They are especially interesting because they prevent the trailer from establishing an easy “human hero versus obvious monsters” schema. Some look like men, some may be spirits, some may be members of a ritual or martial organisation, and Game Science gives us no labels. The current footage shows blocking or deflection, evasion, recovery, multiple attackers, sword-energy effects and the ability to set the blade alight from an external flame. The great mechanical separation from Wukong is obvious: the staff is gone, just as Game Science said it would be in its July interview, and the new protagonist is a dedicated swordsman. The developers had also suggested before this trailer that Zhong Kui could broaden the weapon vocabulary further, although the exact final arsenal remains unresolved. 

The sword is particularly appropriate for Zhong Kui. Across painting, theatre and popular iconography he is frequently represented carrying or drawing a blade as the instrument of demonic judgement. Even when he is rendered comically — drunk, escorting his sister, surrounded by mischievous ghosts — the sword keeps him legible as the demon-queller. That connection makes the playable man's matching choice of weapon one more point in favour of the mortal-Zhong-Kui reading. If this were simply an unrelated wandering Daoist protagonist, Game Science has nevertheless chosen to make him fight using the object most immediately associated with the title deity.

Yet I think the more radical departure is occurring in the supernatural abilities surrounding the sword. Wukong gave us transformations because transformation belongs to Sun Wukong's mythology. This trailer suggests Zhong Kui may convert the old series idea into something closer to spirit jurisdiction. You encounter ghostly beings, summon them, fight beside them and perhaps enter them. The difference could eventually become one of the defining distinctions between the two games.

Dà Bàng Zhēnjūn: The Clam Man Is Much Stranger Than “Big Clam Boss”

The clam creature is the trailer's most completely demonstrated monster and probably the most immediately unforgettable new design. Its name shown in the fight is 大蚌真君 — Dà Bàng Zhēnjūn. Dà bàng is straightforwardly “great/large clam.” Zhēnjūn, however, is a far grander piece of language: “Perfected Lord” or “True Lord,” a religious honorific associated with revered transcendents and divinities. Game Science has therefore produced a revolting, low-crawling humanoid mollusc encrusted with shell and marine growth while giving him a title with spiritual dignity. Whether the name is comic self-aggrandisement, a sign that the creature really is some fallen or local numinous power, or a deliberate challenge to the player's assumptions about what counts as god and what counts as monster is unknown. The title alone makes him more interesting than an anonymous yaoguai. 

Mechanically, his enormous paired shell functions as armour and as part of the encounter's evolving physical state. He charges and rolls, exposes vulnerable flesh when opening himself, and eventually loses or has part of that shell broken, pushing the fight into a second phase. Game Science has not confirmed a generalised monster-part destruction system, so the shell break should currently be treated as a property of this encounter rather than evidence that every large enemy will have severable components. The basic design is still revealing: the enemy is built around a natural biological defence, and overcoming him changes his silhouette rather than merely deleting a health-bar segment. 

The deeper folklore connection is the giant-clam tradition mentioned above. Chinese thought contains an unexpectedly rich imaginative relationship between huge shellfish and illusion. The shèn becomes an explanation for impossible structures hovering beyond the coast; the creature sits between zoology and monstrosity, something apparently mundane enlarged into a source of ontological deception. Dà Bàng Zhēnjūn need not literally be the shèn for Game Science to be drawing from the same cultural reservoir. This is precisely how I hope they approach the entire bestiary: not turning a classical encyclopedia into a checklist, but absorbing older visual associations and generating creatures which feel like they could have grown out of Chinese storytelling themselves.

And this is where the title Zhēnjūn becomes delicious. Zhong Kui's world should contain bad humans, decent ghosts, fraudulent gods, legitimate local spirits, dangerous animals, supernatural bureaucrats and beings whose legal or cosmic status we do not yet understand. A clam calling itself a Perfected Lord might be ridiculous, sacred, corrupt, tragic or all four at once. The 2025 official verse already sets up this moral disorder by contrasting monsters and immortals as two faces and asking who possesses the authority to judge life and death. The clam is the first boss we have seen whose very name embodies that ambiguity. citeturn306843search5

The Wrapped, Black-Draped Companion — and the Mechanic I Think Defines the Game

The biggest gameplay revelation arrives once the clam's shell has been smashed. The protagonist calls forth a gigantic humanoid figure, substantially taller and heavier than himself, wrapped or draped in dark material. Chinese reports have been cautious and simply describe him as a mysterious tall man in a black covering. Game Science's own current character roundup material does not provide him a name. What happens next is far more useful than a name anyway: the entity remains in the battle and fights alongside the protagonist, and then the protagonist can shift directly into or through this being and attack using its body. Chinese coverage describes the mechanic with language equivalent to possession or body-shifting. 

This immediately suggests a whole potential architecture for Zhong Kui. Imagine that the supernatural progression system is built around ghosts rather than conventional transformations. A defeated, judged, recruited, bound or otherwise acquired spirit could become a temporary ally; the player could then invoke that being beside the protagonist, and some spirits could additionally function as bodies the protagonist temporarily inhabits, supplying radically different ranges, masses, weapons and movesets. Game Science has shown only enough to establish that one summoned ghost-like companion has both collaborative and embodiment functions. A collectible ghost roster remains my extrapolation, not an announced system. But if that is where the mechanic goes, it would be superbly appropriate to Zhong Kui.

The reason is historical. Zhong Kui is frequently represented with ghosts, rather than eternally alone against them. The Zhong Kui Marries His Sister theatre tradition turns the ghost entourage into a procession of distinct personalities: productions include large ghosts, lamp ghosts, burden-carrying ghosts, umbrella ghosts, donkey-leading ghosts and other supernatural attendants. CCTV's preserved cast list for one production gives us exactly this kind of retinue. The Qing novel Zhan Gui Zhuan goes even further. There Zhong Kui is supplied with named assistants and three hundred yin soldiers, and the story's supernatural bureaucracy explicitly tells him that ghosts cannot all be treated identically: some are to be killed, some pacified, with judgement depending upon severity and circumstance. That is extraordinary material for a game system. Zhong Kui's power is not merely “I hit ghosts with sword.” It is authority over a society of ghosts.

That old novel also offers one of the most useful lenses through which to read Game Science's creature designs. Its “ghosts” include satirical personifications of human failings, habits and social types. They are a supernatural taxonomy of corrupted behaviour as much as a zoological bestiary. The king of the underworld effectively tells Zhong Kui that some of the hardest ghosts are difficult to prosecute precisely because their wrongdoing is habitual, social and morally slippery rather than a simple criminal offence. This sits beautifully beside Game Science's line that demons arise from the human heart. It makes me suspect the game's great question will eventually be what deserves the sword. A summoned monster fighting beside us is already evidence that “ghost” cannot simply equal “enemy.” 

The wrapped companion may therefore be our first proof of the game's central mechanical idea: Zhong Kui's mythology becomes playable through relationships with spirits. You may kill them. You may command them. You may inhabit them. You may perhaps judge them. Some may already serve you. That is an extremely fertile replacement for Wukong's transformations.

The Woman in the Water

Then Game Science does something wonderfully strange. Dà Bàng Zhēnjūn attempts to flee towards what appears to be its refuge or nest and a woman suddenly erupts from the water, seizes it and violently beats it. Current Chinese coverage describes her only as a mysterious female ghost or water ghost; Game Science has not named her. GamerSky's same-day character inventory similarly labels her simply “神秘女” — mysterious woman. 

There is an obvious general folklore association with shui gui, water ghosts, drowned spirits and other beings attached to rivers, pools and deaths by water, but I would resist assigning her a canonical water-spirit identity this early. Chinese supernatural literature has far too many female aquatic beings, drowned revenants, river spirits, dragon-associated women and local numina for “woman emerges from water” to lead automatically to one name. More interesting is what she does in the scene. She does not attack the protagonist. She intercepts another monster. We therefore see, within the very first gameplay demonstration, another breach in the simplistic division between player and monster. The supernatural world has its own alliances and grievances.

Some fans are already asking whether she might be Zhong Kui's sister, particularly because Du Ping and “Zhong Kui's brother-in-law” theories have circulated since the earlier New Year footage. I presently consider that identification weak. The traditional younger sister from Zhong Kui Marries His Sister is usually the living woman Zhong Kui returns to escort into marriage with Du Ping. She is not conventionally an aquatic revenant who bursts from a marsh to pummel shellfish. Game Science can of course rewrite the tradition, and the February 2026 in-engine special may eventually prove to have reused designs that become important in the actual game, but the studio explicitly positioned that celebratory short as separate from the final story. So “water woman = sister” is possible in the broadest sense, yet currently unsupported. 

There is another interpretation I find more interesting: perhaps she demonstrates the same summon or adjudication system as the dark companion but in a more cinematic form. Maybe the protagonist has invoked her. Maybe she is a local spirit who hates Dà Bàng Zhēnjūn. Maybe she claims the boss after defeat. Maybe she belongs to an ecosystem in which supernatural beings have obligations and territories. The footage cuts before it settles the question.

Her emergence also transforms the preceding boss fight retrospectively. The protagonist did not simply travel to a boss arena and kill the large thing occupying it. He broke the clam's shell, drove it from the battlefield, and then something already present in the local supernatural world intercepted it. The environment contains agents. That is a promising direction for a game in which folklore can feel inhabited rather than arranged exclusively for combat.

The Creature Montage: What Is Actually There

Same-day Chinese coverage has done something very useful: GamerSky went through the trailer and assembled an image catalogue of the presently visible characters without pretending it knows their identities. Its labels amount to our best provisional register. Beyond the protagonist, mysterious old man, Dà Bàng Zhēnjūn and mysterious water woman, the footage contains the protagonist's summoned/possessed ghost, long-haired ghost creatures, a mysterious ghost-faced person, a grotesque person in a pot, an apparent boss-sized unknown being, a mysterious eyebrow-painting or face-painting figure, a bound ghost-creature, a mysterious Daoist who fights the protagonist, the fully realised Zhong Kui, and several categories of ordinary smaller enemies. That source explicitly warns that most identities are currently unknown.

This is precisely where I would caution against the inevitable online tendency to identify every unusual creature as “from the Classic of Mountains and Seas.” One Reddit reaction today already made that association, and visually I completely understand it: the bodies have the startling taxonomic strangeness people associate with Shanhaijing. But Chinese strange-creature culture is vastly larger than that one book. Game Science can pull from zhiguai tales, local cults, Daoist ritual images, Buddhist infernal imagery, exorcistic theatre, opera, New Year prints, ghost novels, paintings, oral traditions, occupational folklore, temple sculpture and wholly original invention. Black Myth: Wukong already demonstrated that Game Science enjoys discovering obscure textual and artistic fragments and then combining them into a new organism. Treating every deformed body as a direct classical-creature adaptation would reduce what is actually exciting about this design language.

The long-haired creatures seen after the paper-old-man sequence appear much closer to the broad visual language of revenants and wild spirits than to anything we can responsibly name today. The hound-like animals could be ordinary yaoguai, corrupted dogs, infernal beasts or simply an original enemy family. The ghost-faced person may draw on masks, ritual face painting or bodily supernatural mutation. Chinese nuo traditions alone contain an enormous vocabulary of exaggerated divine, demonic and exorcistic masks — including figures with fire-shaped brows, tusks, horns and other features designed to render supernatural type through the face — so “strange face = named monster” is a poor inference. 

The person in the pot is immediately suggestive of punishment imagery, cooking, sacrifice and the grotesque domestic humour that Game Science clearly enjoys, but no current source gives the figure a mythological name. The bound creature is similarly tantalising: bondage could indicate punishment, capture, transport, restraint of dangerous spiritual power, or a creature literally using another body as part of itself. The half-human rider or partly human mount image in the montage that immediately jumps out in the footage currently has no reliable official identification that I can find. I would leave it unnamed rather than forcing it into a source text. Visually, one of Game Science's recurring pleasures is turning social relations into anatomy: carrying, riding, binding, eating and inhabiting other bodies become creature design. If this being is indeed one body bearing another, it might represent parasitism, punishment, hierarchy, symbiosis or simple grotesque locomotion. A later bestiary entry will probably transform what now looks like visual madness into some sad, comic or vicious little story.

The mysterious Daoist-looking opponent is particularly intriguing. If the playable man himself uses a sword and ritual powers, a second ritual specialist fighting him immediately opens the possibility of competing exorcists, sects or systems of supernatural authority. Zhong Kui's world should benefit enormously from human institutions trying to regulate the same territory as the ghosts. A supernatural game becomes much richer once “exorcist” ceases automatically to mean “good person.”

The Skeleton Band May Be the Most Revealing Image in the Trailer

Then there is the image I keep returning to: the protagonist calmly seated beneath a flowering old tree playing guqin, while three skeletons accompany him as an ensemble. According to the close breakdown circulating in Chinese coverage, the skeleton on the left wears a broad hat and plays erhu, another behind plays a flute, and the one on the right beats a drum. Petals fall around them. Nobody is being attacked. Nobody is screaming. They are making music. 

This single image tells me more about Game Science's conception of the world than another ten bosses could. Death has not abolished personality. A skeleton can have an occupation, rhythm, social role and perhaps friendship. The afterlife is a community. Ghosts are capable of art. Zhong Kui's relationship with the dead can consequently contain companionship alongside violence. This is fully compatible with the old iconographic and theatrical Zhong Kui tradition, where his attendant ghosts are often comic, industrious, celebratory and domestic. Zhong Kui Marries His Sister literally sends them through the night carrying lamps, belongings, umbrellas and musical objects as part of a wedding procession. The demon-queller is surrounded by the beings he supposedly exists to “quell.” 

That contradiction is central to Zhong Kui's charm. He himself is a ghost who hunts ghosts. He is frightening yet protective. He belongs to the underworld but guards human houses. He can be represented with savage violence, then appear in paintings drinking, travelling, marrying off his sister or surrounded by comic assistants. His monstrousness protects people from monstrosity. A game based upon him should therefore resist a clean human/monster split at every opportunity, and this trailer repeatedly does exactly that.

The Tiger Is Not a Random Mount

The ending finally gives us the figure everyone expected to control from the title announcement: the enormous red-bearded Zhong Kui riding a tiger. It is tempting to treat this simply as Game Science's trademark spectacular mount image, or even as evidence that we will eventually ride the tiger through the game. The latter is possible, but there is no official confirmation of a mount system, and Yang Qi's reminder that we have seen a test map should make us extremely cautious about extrapolating world structure from one montage. 

The image itself, however, has real historical precedent. A surviving Ming-period painting traditionally titled Zhong Kui Riding a Tiger depicts Zhong Kui carrying his sword on the back of a tiger, looking towards bats whose presence also invokes auspicious symbolism through the fu / “fortune” homophone. The tiger-mounted Zhong Kui is therefore part of an existing visual tradition, and Game Science has chosen an image with historical ancestry rather than simply sticking their character on the coolest animal available. 

The tiger is also perfectly suited to Zhong Kui's exorcistic persona. It is one of the great dangerous animals of Chinese visual culture and can itself carry protective, military and spirit-subduing meanings. Zhong Kui sitting above it converts ferocity into controlled authority. The 2025 announcement trailer already established this mature Zhong Kui and the tiger as the game's defining icon; the 2026 gameplay trailer now withholds him until the final instant after spending almost its entire running time with somebody else. That editorial choice is deliberate even if its exact narrative explanation is not yet fixed.

If the playable swordsman is mortal Zhong Kui, the final image is essentially prophecy: this is what he becomes after death. If the swordsman is Du Ping or another figure, it functions instead as the arrival of the being around whom his story will eventually orbit. If the protagonist is some earthly manifestation of Zhong Kui, the tiger image could be another state of the same consciousness. All three explanations remain alive.

My strongest bet is still the first. The head scar is difficult for me to dismiss, especially when paired with a narrative explicitly organised around life and death. There is something dramatically powerful about Game Science titling the game Black Myth: Zhong Kui and then showing us fifteen minutes of a man whom everybody insists does not look like Zhong Kui. The obvious answer may be: because he isn't Zhong Kui yet.

The Deeper Zhong Kui Tradition Game Science Can Use

What fascinates me about this project now is the sheer volume of cultural territory available to it. Black Myth: Wukongcould excavate endlessly within Journey to the West and its surrounding traditions, but it remained tethered to a massive literary gravitational centre. Zhong Kui gives Game Science something stranger. His mythology exists almost as a network. One branch is the emperor's dream and the grotesque ghost who destroys another ghost. One branch is the gifted scholar rejected because of his appearance. One is the suicide. One is the posthumous appointment. One is the loyal friendship with Du Ping. One is the wedding of the sister. One is the household protector printed and hung against evil. One is the sword-bearing figure in paintings. One is the tiger rider. One is the comic ghost procession. One is the satirical supernatural official of Zhan Gui Zhuan, travelling through human society and discovering that the worst “ghosts” are recognisable human behaviours. One is the historical possibility that the personal biography was itself built over a much older exorcistic name or symbol. Game Science can move through all of these without violating a single dominant source because no single source owns the whole man. 

That Zhan Gui Zhuan connection deserves particular attention over the coming years because its moral structure is almost uncannily aligned with Game Science's announced text. In the novel, Zhong Kui is shown a catalogue of ghosts embodying forms of deceit, greed, shamelessness, pretension, drunkenness and other human failings. The supernatural bureaucracy explains that some are difficult to punish through ordinary legal logic because their wickedness is habitual and socially embedded. Zhong Kui is instructed to distinguish between those who should be destroyed and those who can be pacified. This is exactly the kind of material which could turn the phrase “distinguish the clear from the turbid” into game design. Imagine bosses where victory produces judgement rather than automatic execution; ghosts recruited after defeat; humans revealed as spiritually worse than monsters; local gods whose titles conceal corruption; innocent revenants trapped by ritual rules; supernatural officials who disagree over jurisdiction. Nothing in the current footage confirms those systems, but the underlying Zhong Kui literature offers Game Science a remarkable foundation for them.

The current footage already behaves as though moral taxonomy interests the team. An old man is apparently a piece of spirit paper. A “Perfected Lord” is a clam monstrosity. A gigantic ghost fights on our side. Another ghost's body becomes temporarily playable. A woman emerging from haunted water attacks our enemy. Skeletons form a band. Human-shaped ritual figures attack the hero. The titled demon-queller himself appears only at the end and is far more monstrous-looking than the man we have spent fifteen minutes controlling. Every category leaks into another.

A Provisional Bestiary After Fifteen Minutes

So, as of 20 August 2026, the safest complete reading of what Game Science has actually shown is this. We have the unnamed white-robed or pale-clothed swordsman protagonist, probably human at this point and possibly mortal Zhong Kui; the Li Guangfu-faced old man, apparently able to collapse into or reveal himself as a yellow ritual sheet; several groups of human-shaped armed enemies in pale cloaks and wide hats; hound-like supernatural beasts; long-haired ghostly creatures; the boss Dà Bàng Zhēnjūn, a gigantic humanoid clam whose shell breaks during combat; the protagonist's towering black-draped summon, which can fight independently and apparently become a temporary possessed body; the mysterious water woman/ghost who attacks the fleeing clam; the three skeleton musicians playing erhu, flute and drum; a ghost-faced humanoid; a grotesque person in a pot; at least one bound creature; a mysterious face- or eyebrow-painting figure; an unidentified large probable boss; a Daoist or ritual swordsman who appears to oppose the protagonist; the strange part-human rider/carrier configuration which remains unnamed; additional unclassified small yaoguai; and finally Zhong Kui himself riding the tiger. Anything more specific than those names currently begins moving from identification into theory. GamerSky's exhaustive same-day visual roundup uses almost exactly this cautious nomenclature because the official bestiary names simply have not been supplied yet. 

And I love that we do not know. I hope Game Science resists explaining these things for a long time. Part of what made the earliest Black Myth: Wukong footage so electrifying was seeing beings which looked culturally specific before knowing their context. Zhong Kui has already gone further because its source field is considerably stranger. A monkey king brings expectations. Zhong Kui brings a profession, an image and a pile of contradictory stories. Behind him is effectively the entire Chinese ghost world.

What I Think the Trailer Is Telling Us About the Finished Game

My working theory after going through the footage, the Chinese same-day analysis, Game Science's current statements and the older Zhong Kui material is that Black Myth: Zhong Kui may eventually be structured around a human journey into supernatural jurisdiction. The scarred swordsman begins relatively mortal. He deals with apparitions whose physical reality is already unstable. He learns ritual abilities. He encounters and perhaps captures or binds spirits. Those spirits become allies and alternate bodies. The division between human, ghost, god and monster progressively collapses. Death is not a failure state in the story but an ontological threshold. At some point — perhaps early, perhaps halfway through — the protagonist dies or undergoes an equivalent crossing and becomes the fully realised Zhong Kui: grotesque, empowered, accompanied by ghosts and associated with the tiger. The game's announced “life and death” subject then becomes the actual arc of the playable body.

I would put mortal Zhong Kui ahead of Du Ping as the identity theory today, although only provisionally. The scar tied to the suicide tradition is the biggest clue; the sword is another; the shared visual weapon language between protagonist and mature Zhong Kui is another; the life/death theme is another; the deliberate withholding of the name “Zhong Kui” for the playable protagonist while presenting the iconic Zhong Kui separately is exactly the kind of misdirection an origin story would exploit. The chief reason for restraint is Game Science itself. The company is openly telling us that development is early, that this is effectively a test map with incomplete story slices, that fan interpretation is currently outrunning the project, and that even the voice cast is unfinished. The mystery may therefore contain genuine undecided space, not simply secrets sitting inside a completed script. 

What looks far more solid is the design philosophy. Game Science has found a way to make Zhong Kui mechanically different from Wukong while preserving the Black Myth fascination with embodied myth. The staff has become the sword. Transformation appears to be mutating into invocation and possession. A bestiary derived from one famous novel is becoming a much broader ecology of Chinese strange tales and folk religion. Ghosts appear to have social lives rather than existing exclusively as combat targets. Daoist visual components are entering the interface. Paper is becoming supernatural matter. Boss bodies can physically change. Human enemies coexist with monsters. Musicians can already be dead. A clam can be a “Perfected Lord.” Zhong Kui can sit on a tiger while looking more demonic than half the entities he exists to destroy.

The tiger itself feels like the final seal on the whole fifteen minutes. We begin with an ordinary-looking man helping an ordinary-looking old man move a raft. By the time the montage finishes, ordinary humanity has almost ceased to exist as a stable category. The old man is paper. The landscape contains ghosts. The swordsman's powers reach into other bodies. The clam has a divine title. The woman in the water belongs to some unknown supernatural order. Skeletons have music. Bound and hybrid bodies move through the world according to rules we cannot yet read. Then Zhong Kui appears: enormous, red-bearded, sword-bearing, riding the tiger in an image with centuries of visual precedent. He is simultaneously human memory, ghost, god, official, monster and protector.

That is why this trailer has impressed me so much. Game Science could have made the obvious sequel: another extraordinarily expensive Chinese myth action game with a famous hero, spectacular bosses and a different weapon. Instead, the first playable footage suggests they are reaching for a stranger question. Who gets to decide which thing is the monster? Zhong Kui is perhaps China's perfect protagonist for asking it. He dies because humans judge his outward appearance, then becomes the supernatural authority charged with judging beings whose appearances are themselves horrifying. He is a ghost assigned to deal with ghosts. He resembles the things ordinary people fear and becomes the image they hang on their doors for protection. The ancient stories make him both victim and judge.

And if the scarred swordsman truly is Zhong Kui before all of that happens, Game Science may have found an extraordinary way into the myth: let us meet the man first, let us cross the boundary with him, and only then let us become the thing riding the tiger.

I can keep tracking Chinese-language frame analysis and official comments as people identify the unnamed montage creatures.