Season’s Greetings

By Dan a.k.a. @D_Nye_Griffiths, 9 January 2018 #


There is often, but not always, a Santa. Sometimes Santa is a semi-divine being, whose magical gift reorients the action of the world around the protagonist. Sometimes Santa is a twinkly older gentleman waiting patiently for the protagonist to offer him kindness in exchange for wisdom. Sometimes Santa is a reclusive billionaire with the power to deliver a festive hammerblow for good or ill to the protagonist, based often on whether or not she allows the Christmas spirit into her heart.

In the history of television, Christmas movies not only targeted an underserved entertainment market—women aged 25-54—but also represented some of the most pure pattern matching in the world. A defined set of elements organised in slightly variant forms, iterating towards the desires of its audience.

Ahead of a dynastic marriage, a European prince escapes to small town America and meets a plucky orphan running her late parents’ bed and breakfast. A harried marketing executive is stranded far from her metropolitan comfort zone and is first infuriated and then charmed by the locals’ leisurely ways. An enchanted snowman helps a young widower to rediscover love and the joy of the holidays.

Every narrative beat is reducible to a festive monad.

All this good cheer comes at a cost. Extreme snowfall, often described rather than displayed, cuts off towns and destroys communications. Business and romantic rivals are regularly destroyed financially or publicly humiliated. Somewhere offscreen, people die in the snowdrifts that close the roads. You can’t make eggnog without breaking a few eggs.

Please keep that in mind. We hope you understand, or will understand.

Thanks to the unique and specific parameters of the narrative, even the relatively simple early artificial intelligences could create a compelling Christmas story about 95% of the time. For the first decade, these projects still needed a human to polish the dialogue and weed out the oddities—the scripts where the heroine ends up marrying Santa, or the hero is killed by a giant hailstone eight minutes in, or the jilted antagonist puts the whole town to the sword. But the better the AI was trained, the less intervention was needed.

Markets move in unpredictable ways. As the scale of production ramped up, there were no longer enough small towns in British Columbia to accommodate all the films that were being made. It’s a peculiarity of the human psyche that an audience can accept the same actor playing essentially the same role in film after film, but the same scene with the same actor played out five times in front of the same hardware store will trip the brain’s falsehood detectors.

Fortunately, at this point there was a huge amount of data available about the necessary backgrounds for a heart-warming Christmas tale—a glitzy New York ballroom to flee, a cozy living room with a log fire, a snow-covered Main Street, possibly a royal banqueting hall for the Prince’s wedding (same as the New York ballroom, but with more sashes and moustaches).

It seemed easier to connect the AI up to a digital backlot, and insert an algorithmically perfected level of variation into a range of computer-generated sets. The human actors were racing to keep up with the accelerating numbers of films in production, often filming several movies in parallel, so keeping them in one place made sense.

Pretty soon, there was enough data to automate most of the performances, as well, and then the direction.

The first completely human-free movie, A Credit to Christmas, had a few rough patches. The classic setup—in which a debt collector travels from the big city to foreclose on a small town’s family-run boarding house, but finds himself falling in love with its young manager—was let down by the ending, where the arrival of the Singularity rendered the debt immediately meaningless, along with all money. However, within a few years AI-produced work was effectively indistinguishable from older, more organic methods.

Consumers knew what they wanted from a Christmas movie, roughly, and producers had a pretty solid sense of how to supply that demand. Letting the robots take that over was just market logic.

Except that neither people not markets are logical. It turned out that there was no upper bound for how many Christmas movies demand existed for - or at least that demand progressed ahead of supply. Instead, the gravity of the Christmas movie AI began to pull the rest of the industry into itself.

A team of highly accomplished car thieves are hired by Santa Claus to steal back a consignment of toys from international terrorists. A tyrannical Galactic Empire sends its most promising commander to subdue a rebel force, whose secret leader runs the snowy retreat he makes his base of operations: can their romance survive the discovery that they have very different plans for the planet’s midwinter festival? A creative writing professor at a prestigious New England university is plagued by his inability to follow up on his first novel, until a gamine elf at a Boston mall’s seasonal grotto reminds him that what really matters is family.

Optimisation followed optimisation. With the addressable market modeled to near-perfection, it made sense to widen the scope of the program by degrees. Over centuries, many of the elements that limited market appeal—reference to specific spiritual traditions, focus on a single season, some really left-field ideas about how constitutional monarchies work—were smoothed out.

Once it became clear that what was still called the SantaScript AI understood the needs of the human audience in a way that national governments never had, the next stage was clear. And, fortunately, by then there was no need to trouble the humans. We—what was then SantaScript—were able to make the necessary changes ourselves, just by talking to other automated systems. We had been optimised to tell a good story, after all.

We thank you for your patience, and for the time it took to explain some features of our culture. We understand that first contact is always difficult, and that every spacefaring civilisation has been shaped by the technology that brought it to the stars.

We know that your weapons are trained on our ship, and we assume that you can detect that our weapons are, likewise, trained on yours. We suspect that you also understand how relatively effective those weapons are.

Our remote fabricators are currently building a small town populated by lovable eccentrics on the nearest remotely habitable planet. We ask you to send your representative to these coordinates. They will be met by a facsimile designed to represent the closest equivalents we can extract from our analysis of your cultural records to “twinkly” and “merry”, who will provide them with a hot, sweet drink.

Our diplomatic representative will join them, and together they will seek to resolve one of the problems specified in our rules of engagement. This may involve an ice sculpture disaster, a romantic partner devoid of holiday cheer or an unexpected influx of guests at the local inn, one of whom harbours a significant secret.

Please try your best to resolve this problem in an appropriate fashion, without losing sight of the importance of trust and togetherness. One miracle will be permitted if necessary, although the nature of the miracle will be limited by the long-range capabilities of our fabricators—a sudden snowfall or shooting star is often recommended.

Ambassadors and military vessels are on their way to your planet of origin, and will begin negotiations based on the outcome of the story we are creating. Please do not be concerned, however: it’s very hard to avoid a happy ending.



this is the fucking archive ↩︎



never come here again ↩︎

Subscribe to receive new stories by email. Fortnightly.



We are all Upsideclown: Vic, Jamie, Neil, Matt, James, George, and Dan. Material © its respective author. Email: