My next book is The Reverse Centaur's Guide to Life After AI, out next month. Pre-order it now, including as a DRM-free audiobook or ebook, at my Kickstarter, and help me continue to prove that DRM-free isn't just the right way to reach an audience, it's also the best way to reach them.
Yesterday, I attended a Brian Eno talk about the nature of creativity and art based on What Art Does, the short book he published with Bette Adriaanse last year:
I haven't read the book (yet – I just ordered a copy), but the talk really got me fizzing. The subject matter (not just what art does, but also what art is) is one I've given a lot of thought to, and Eno's characteristic mix of gnomic koans and deceptively plainspoken assertions brought me along to some realizations of my own.
For Eno, art is "everything you don't have to do." You have to wear clothes to protect yourself from the elements, but you don't need to adorn those clothes. You need to speak to make yourself understood by the people around you, but you don't have to sing or write poetry or make up stories.
This is a really critical point, and I think it can be further refined by this: "Art is intended to make other people feel something." This distinguishes "art" from "beauty." A sunset can be beautiful, but no one intends anything by it. An artist who takes a photo or paints a picture of a sunset does so in the hopes that it will make you feel something, but the sun and the atmosphere and the Earth's curvature and rotation don't hope anything, because they are inanimate.
This distinction has lately become far more significant, thanks to the rise of images and words that have the seeming of intent, but who don't have an intender. When you paint a painting, every brushstroke conveys an intent, even if you can't point at an individual brushstroke and articulate its purpose. The same is true of prose: every word and punctuation mark is there for a reason, and "being good at writing" (like "being good at painting") is how we describe someone who has practiced so much that these reasons can be infused into each micro-decision on a near-totally subconscious level.
Contrast this with AI: when you prompt an AI to generate words or pixels, you are conveying some intent about the feeling you want the people who experience the model's output to experience. The problem is that the AI doesn't have any intent of its own – it just has statistical predictions, based on other people's intent, which it has analyzed through its training data.
So when the AI expands the three sentences in your prompt into 100,000 words or 1,000,000 pixels, it isn't adding any of its intention to the finished work, it's diluting the intention you fed to it. Three sentences divided by one million pixels yields an image that has an average intentionality that's so low that it's practically homeopathic.
Until recently, we weren't accustomed to encountering coherent strings of words or polished images that had no intender, so we imputed the existence of that intender to them, and we did what we always do when we encounter a work of art: we tried to mentally materialize a facsimile of the feeling the artist experienced while creating the work.
Because the intention of these works was so dilute, we ended up hallucinating an intent. We made up an imaginary artist who meant something by every choice in the work, and experienced an emotional affect that we ourselves had created out of (nearly) whole cloth.
As a species, we've been through this before. Think back to those sunsets. There was a time when we all thought of sunsets as being explicitly created by another being, who was in communication with us through the natural environment (some people still believe this). Looking at a sunset was an exercise in asking yourself, "If I were God, what would I be trying to say to me with this sunset?" just as looking at one of my photos of a sunset would be an exercise in asking yourself, "If I were Cory, what would I be trying to say to me with this photo of a sunset?"
The rise of materialism and scientific rationalism is sometimes called a "disenchantment" and indeed, there's a sense in which a sunset that we know to have no intender is no longer "enchanted." The experience of a sunset becomes something like, "Those colors and their interplay with the physical world is very beautiful." It might even be, "How could I capture that beauty in a painting or a photo or a description so that I could communicate it to someone else?" But it's not, "I wonder what God wants me to feel when I look at this sunset?"
So for many of us, the experience of AI "art" went from, "Wow, there's a person in the machine that's trying to tell me something," to "Wow, that is an impressive feat of software design, but it doesn't say anything to me." Maybe some of us think, "Huh, I could take some element of this, refine it with my own brushstrokes or words, and make something out of it." That's like thinking about turning a sunset into a painting: the sunset is striking and maybe beautiful, but it doesn't become art until you work at it, in order to make it communicate something:
Mark Fisher describes the "seeming of an intent without an intender" as "eerie." It's true: when the door slams in the night and there's no one else in the house, it's eerie. But eeriness is easily dispelled: once you locate the open window that's creating the draft that's blowing the door closed, the eeriness regresses swiftly to the mean:
Banishing eeriness may be straightforward, but preventing eeriness is much harder. We are prone to imputing intent to the things we see in the world. In "Genesis," an essay from EL Doctorow's (no relation) collection The Creationists, Doctorow describes the origins of the Babylonian creation story (which the Hebrews ripped off for Genesis 1:1-29 – Genesis is Babylonian fanfic). The Babylonians made up this story about how God created the heavens and Earth and so forth, and this story was so cool that they couldn't believe that they had just made it up, so they concluded that God must have put it in their minds:
Back to Eno: central to his talk was the "theory of mind." To have a theory of mind is to be able to impute someone else's intent. It's when you ask yourself, "What does that person mean by the thing they just said or did?" Because art is a process by which an artist tries to get you to feel something, it requires that the artist have a theory about your mind. And because experiencing art is a process of trying to figure out what the artist wanted you to feel when you experienced their work, experiencing art also requires a theory of mind.
From time to time, I teach fiction writing workshops, and one of the lectures I always give is about how stories are a "fuggly hack":
It's very weird that storytellers can trick our brains into experiencing emotions based on empathy for "people" whom we know to be imaginary. Romeo and Juliet are made up, they never lived, they never died, and so, objectively speaking, their deaths are less tragic than the death of the yogurt you ate for breakfast. That yogurt was alive and now it's dead, after all. And yet, we weep for Romeo and Juliet.
Our automatic "theory of mind" processes create empathy for stuff even when we know that stuff is inanimate. But the purpose of narrative isn't getting you to experience empathy with an imaginary person. The purpose of narrative is to get you to experience that empathy so that you will feel something. In other words, the storyteller who describes a character who is swept away by the beauty of a sunset is trying to get you to feel "swept away" not "empathy for someone who is swept away."
There's lots of art that skips the step in which you are asked to first experience empathy for an imaginary person in order to arrive at some feeling. A lot of music, visual art, dance, and poetry seeks to evince that feeling in you directly.
When this works, it's profound. I think about this a lot in terms of built environments, specifically Disney themepark rides. When I started hanging around with Imagineers (the multidisciplinary artists who design and execute these rides), I noticed that they made frequent reference to the role of narrative storytelling in their ride designs, which was weird, because the very best Disney rides do not use narrative to evince a feeling.
Think of two Disney rides: Snow White's Enchanted Wish (1955); and The Little Mermaid: Ariel's Undersea Adventure (2011). In Snow White, riders follow a track through a series of animated vignettes with UV-fluorescing painted backdrops and an orchestral soundtrack. There are almost no words spoken in the soundtrack. The ride's vignettes recreate scenes from the 1937 animated film, but they don't make any attempt to explain the plot of the movie.
A rider who'd never seen Snow White and the Seven Dwarfs could not recount the plot of the movie to you. However, that rider could absolutely convey the emotional affect of every scene in the film. It is a near-perfect transmission of the feelings evinced by the movie, notwithstanding that it bypasses recounting the film's narrative.
By contrast, The Little Mermaid ride is what's sometimes pejoratively called a "book report ride." The scenes are full of dialog, and they explicitly re-create the storyline of the 1989 film. These scenes are well-executed, with lots of clever mechanical effects and skillfully painted and sculpted scenes and robots. A rider who never saw the film could give you a scene-by-scene breakdown of it – but they could not tell you about any of the emotional beats of the film. For all that the ride faithfully recreates the story of the film, it does so at the expense of the purpose of the film, the feeling the film is designed to evince from its audience.
As a novelist, I find it natural that someone trying to build a Little Mermaid ride would start from the premise that it should explicitly retell the story of the film. If you want an audience member to experience a feeling, narrative gives you the opportunity to explicitly describe the feeling you want the audience member to experience. You can situate a character on a lonely beach at sunset and tell the reader how that character feels.
The problem is that while this has an increased likelihood of being high-fidelity way of transmitting a feeling, it also has an increased likelihood of being a low-intensity way of conveying that feeling. When you tell someone about what's going on in another person's mind (including an imaginary person's mind), it doesn't fire up the theory-of-mind machine in the way that asking someone to infer the state of someone else's mind from implicit cues does.
This is why fiction writers are exhorted to "show, not tell." Dramatic, implicit evocations of an emotion are intrinsically more interesting than explicit statements about emotions. That's not to say that exposition can't evince an emotion – it can and does. It's just harder to do this with exposition than it is to do it with dramatization:
In his talk yesterday, Eno discussed abstract art, and the way that it evinces feelings in the viewer directly, without ever telling you what to feel. This is in keeping with much of Eno's own art (he recently told me that when he writes lyrics, he never uses the words "I," "me," "you," or "love").
In this theory I'm developing here, we could say that the more abstract a work is, the harder it is to evince a specific feeling with high fidelity, but the more likely it is that the feelings it does evince will be intensely felt. When your aesthetic sense resonates with a Henry Moore bronze or an Eno ambient track, the thrum is deep and strong.
Key to this theory is that it's about how hard it is for an artist to evince a feeling and how hard it is for the artist to make that feeling intense. Abstract art is more likely to be misunderstood (or not understood) than explicit narratives, but lots of abstract art is very well understood by people for whom it resonates. Explicit narratives are more likely to have a flatter affect than work that attempts to skewer your emotions directly, but plenty of explicit narratives make you feel the most profound emotions you're capable of feeling.
Imagine a 2×2 grid with "intensity" on one axis and "fidelity" on the other. It's easier to evince an intense feeling when you are more abstract, but it's harder to control what that feeling will be. These are works that operate on an implicit theory of mind ("I think I know what you'll feel when you see this"). It's easier to control the feeling you're evincing when you are more concrete, but it's harder to make that feeling an intense one ("I will tell you what someone else is feeling using this work").
None of this is to establish a hierarchy of art. As Eno says, the value of art is in whether it makes you feel something and what it makes you feel – not how that feeling is drawn forth. In What Art Does, Eno describes both art and science as an extension of our natural, in-born tendency to play. The difference is that we judge the success of science based on whether we can validate its conclusions, while we judge the success of art based on whether it excites us:
'Excitement' is to art as 'falsifiability' is to science.
(With thanks to Brian Eno.)
If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
sometimes I think about how far we still have to go with consent
my worst relatives try to sneak meat or meat products into my food despite the fact that I'm a vegetarian
my ex's brother gave his mother an edible without her knowledge and when she got freaked out and paranoid they laughed, and people I've told that go "yeah that's shitty but it's just weed"
when I go to the doctor and ask them to describe what they are going to do before touching me they get frustrated
when I ask a friends of a friend who is a small influencer to keep me out of frame in videos they film for social media in public they look at me like I've pissed in their cereal
Genuinely, I think the most publicly visible AI advance so far this year is how thoroughly and abruptly the "image generators can't do legible text" problem has beem fixed.
For those who haven't been paying attention, this is a recent ChatGPT output. It can do this sort of dense sensible text in images totally one shot, so don't rely on bad text to identify AI images any more.
Here is the entire interaction that produced this image:
I've seen other people do much more sinister things with this! You can produce 100% realistic fake screenshots of news articles, tweets, that sort of thing, if you can convince the bot it's kosher to do so.
We're in the middle of the school. Guys, we're done building the classrooms. School thanks to your donations, this is a very wonderful thing Thank you to everyone who donated and participated But we still have a lot of lacking
Hey guys, let's be snug. Why don't we complete donations quickly? Why don't you help us spread this out for fundraising? This is considered an institution. Why doesn't anyone care about her? I'm overloading my energy. Please create posts everywhere. Everything is available in Gaza, but education is still not available please help with posting We're doing a good job, please share and donate now.
The Coastal Initiative in Gaza urgently needs your support to laun… Asma Yunis needs your support for Help the children of Gaza get a safe a
Games Workshop has turned a corner after years of silence allowed hate to fester in its community
Despite its recent association with the right, Games Workshop’s games from the 1980s adopt a broadly anti-authoritarian if not leftist stance. Zhu, who has written extensively about the politics of games set in the Warhammer 40,000 and Warhammer Fantasy universes, notes that the company frequently poked fun at the right.
An early Games Workshop scenario invites players to side with dwarven characters patterned on the workers in the 1984-1985 U.K. miners’ strike, a labor action that was crushed by conservative prime minister Margaret Thatcher’s government. One orc figurine appearing in White Dwarf #81 (Sept. 1986) even hoists a banner with Thatcher’s face on it.
Warhammer 40,000: Rogue Trader takes on a similar political bent. The Imperium of Mankind often borrows from 2000 AD’s Judge Dredd comics, a British science fiction satire that mocked Dirty Harry-style tough cops celebrated in much of American popular culture for their excessive vigilante justice.
As Zhu says, the early Space Marines were painted as petty tyrants rather than superheroes: “A lot of the artwork emphasizes their dehumanized nature. They’re presented as kind of monstrous, horrible. […] More often than not, they are depicted getting shot by space goblins, or carrying out trivial duties like arresting graffiti writers, or doing obviously bad things like taking slaves.”
Zhu finds other radical political elements in early Warhammer 40,000. For example, the alien Genestealer cultists in the 1980s appeared as limousine-driving industrialists who exploited rural planets in what seemed like a warning about the dangers of capitalist and colonial oppression.
Games Workshop has downplayed these more satirical elements in later decades. Some of this tonal shift was due to personnel changeover, Williams argues, but much of it was driven by the company’s growth. As Games Workshop expanded into overseas markets such as the U.S. in the 1990s, it dropped its working-class critique of British politics along with characters with names like “Obiwan Sherlock Clousseau” in favor of a more corporate style that would translate better for young American gamers.
As the most popular faction in the game, the Space Marines and their allies in the Imperium began to get the good-guy treatment in marketing. Warhammer 40,000 fiction in the Black Library series has been better about emphasizing the morally problematic nature of the Imperium, but advertisements and fluff as well as tie-ins such as video games undermine this by asking players to pledge loyalty to the game’s totalitarian empire.
Williams sees Games Workshop’s Primaris redesign of the Space Marines as reflecting a broader context of creeping authoritarianism. While the old Space Marine lines looked like “bulky space knights,” he says, the new models wear armor influenced by the “tacticool” aesthetic favored by SWAT teams and right-wing paramilitaries.
Games Workshop did not invent the glorification of violence, Williams argues, but it has been shaped by a global culture that glamorizes deadly shows of force.
The Imperium of Mankind often borrows from 2000 AD’s Judge Dredd comics, a British science fiction satire that mocked Dirty Harry-style tough cops celebrated in much of American popular culture for their excessive vigilante justice.
I know nothing at all about Warhammer so no opinion there, but I once asked a congregation of comic nerds (so, a forum) what’s the deal with Judge Dredd. I knew the concept (a cop / judge / jury / executioner dripping with machismo in a fucked up future), but what’s the angle? Is it played straight? Is it satire? Is Dredd the hero of the story, the anti-hero, the villain? Is it a device to say something else entirely? And pretty much everyone replied “YES”. It’s all of these things, occasionally at the same time.
So I tried it, I read a few volumes before I lost interest (it’s not bad there’s just a lot of it), and I came to the same conclusion. Depending on the story, Dredd can mean “fuck Dirty Harry” or “fuck yeah Dirty Harry!”, and sometimes it’s straightforward and other times it’s very complicated.
But here’s the thing. When an author with clear anti-authoritarian intent uses an authoritarian POV and narrative framework in order to ridicule it, and goes at it for too damn long, the framework takes over. The POV wins. The medium is the message.
I firmly believe that Verhoefen’s Starship Troopers would be the best satire ever made (that’s a big award, I mean it 100%) if it was a short film. It started very strong, and it was hilarious throughout, but there was one problem: it went on for too long, and by the end you were kinda glad that the drill sergeant made it out of the giant slug alive, because that’s just how film works.
hooray!
It isn’t an intellectual failure, it’s only a natural emotional response. You just gotta refrain from rationalising it (“actually the drill sergeant was a good egg!”). Of course this doesn’t negate the satire, but I think it does weaken it.
Think of nature documentaries. If the camera follows the wolf, you root for the wolf. If it follows the caribou, you root for the caribou. And if it’s complicated, it follows both the young frail caribou that got left behind and is getting eaten AND the wolf cub that starves to death because daddy didn’t find any caribous and didn’t bring home any food.
Now, if the camera follows a ridiculously exaggerated caricature of a wolf, you are at first conscious it’s a caricature. But keep that up for an hour and the gimmick fades to the background, the caricature becomes just a wolf, and you root for the wolf.
No idea if this applies to Warhammer (again, I know nothing about it, no opinion there), I’m just noting it as a general phenomenon.
they gotta put all their eggs in the "male socialization" basket cause some trans women don't even have the body parts they fearmonger about but they still want to control them
like at this point in the like... 10+ years trans women have been a go-to wedge issue, everyone knows how chromosomes don't necessarily determine phenotype, everyone knows bottom surgery exists, everyone knows intersex people exist, everyone knows some people don't even produce gametes at all, everyone knows we don't stop calling cis women women after a hysterectomy or sterilization
the well of bioessentialist arguments is starting to run dry cause everyone knows the facts that disprove them. so they've had to greatly exaggerate how socialization works and how immutable it is cause what the hell else are they gonna do, lol
still thinking about the time my bf asked about favourite dinosaurs and my brother said quetzalcoatlus (pterosaur (not a dinosaur)) and my bf said dimetrodon ((synapsid (stem mammal) (went extinct 50 million years before dinosaurs evolved)) and i said crows (bird (dinosaur)) and yet i was judged to be the incorrect one in this scenario
