Most guides on detecting AI images are from 2022 and have aged like dinosaur milk.

“AI can’t draw hands.”

“AI can’t draw straight lines.”

“AI can’t spell words.”

This is fantastic advice…assuming you still live in the year 2022. However, studies show that 46% of people have upgraded to a later, worse year called “2024”, and if you are one of them (condolences), this no longer applies. You cannot rely on AI images having such egregious errors. So what can you rely on?

There are still some “tells” of AI imagery—many are strangely getting worse as the technology advances—but often they’re not errors so much as they’re “conceptual tension.” Basically, an image generated by (for example) Midjourney has several different priorities:

1) fulfill the user’s prompt

2) look coherent

3) look attractive

4) satisfy a moderation policy

5, 6, 7 etc) do a bunch of crap the user isn’t even aware of.

This is a lot of plates to juggle, and if the goals clash (ie, the user prompts for something ugly or incoherent), the image can get subtly pulled in different directions. I’ll show examples of what this looks like soon, but this sense of confusion—and not extra fingers—is arguably the best way to spot AI images now. It’s not a smoking gun. It can be subtle, and requires a careful eye. But sadly, this is an age of photorealistic fake media, and we must learn to be careful.

Keep an eye on what your older relatives are doing on social media, by the way. They might not be safe.

PSA: Delete Facebook

Facebook is the internet’s fecal tract. A viscerally unpleasant place, but you can learn a lot about what’s making the online world sick from studying it. And right now, it’s inundated by AI image-spam.

As we speak, millions of retirees are blissfully thumbing their way down a black sewer of nonexistent houses, fake people, and assorted nightmares.

These pictures are bizarre. Many don’t even try to look convincing. Some appear to be the work of aliens with only the faintest understanding of our world.

What’s “peach cream”? Who makes their own birthday cake? Why does history’s second oldest woman have prison tattoos? Why does the cake say “386”? Why is there a gun in my mouth?

(Without getting distracted by this, it’s probable that these accounts are run by the same few bad actors. The repetitive phrasing and subject matter is telling—I’ve seen dozens of cakes made with “peach cream and filling”—which makes me wonder about their sky-high engagement numbers. If you have the technical know-how to manipulate dozens or hundreds of fake accounts, you might also be viewbotting.)

Should we care about this? Even if you have a “let people enjoy things!” outlook on life…

…It’s never a good feeling to see an older relative losing touch with reality. And suppose this is all the front for a scam? I’m not sure this alien race of 396-year-old cake-baking prison-tattooed grandmas has our best interests at heart.

But when you tell a boomer that an image is fake, they might ask “How do you know?” And that’s a fair question: how do you know? Vibes? Because that’s the world we live in? Some AI images are incredibly realistic. It’s easy to get stuck in a Twilight Zone of “I know this is fake but I can’t explain why.”

This is an AI generated image. Can you find any mistakes?^[1]Tricky, isn’t it? Here’s my stab at what’s wrong with it. First: it’s convenient that the photo is too blurry for us to see fine details. I don’t say … Continue reading

In an age of photorealistic imagery, here are what I see as the signs of AI fakery.

AI Images Are Excessively Archetypal

When you prompt AI for a cow, it’s never content to just give you a “cow”. It gives you the cowiest cow that ever cowed. The God-emperor of cows. A divine bovine. A cow with every “cow” trait cranked to the max on its stat sheet. A cow-icature.

You’ll always get a Holstein or a Jersey, because those are the most famous cow breeds. It will be standing in a green field on a farm, because that’s where cows live. There will be other cow-related details thrown in (like a red barn in the background, or a bell around its neck). Everything will be cow cow cow.

The image above is by Dall-E3. Midjourney V6 is more realistic^[2]I have long suspected that Dall-E3 images look shiny and fake on purpose. Sadly, there will be a major tragedy or crisis triggered by deepfakes at some point, and OpenAI doesn’t want one of … Continue reading but has similar issues.

Contrast these perfect AI-generated cows with an actual photograph.

A real cow does not exist as a perfect Platonic archetype. It has a certain amount of not-cowness mixed in: flies, a crooked horn, a limp, asymmetry, visible scars, or whatever. AI seldom (if ever) depicts these details, because they constitute deviance from the prompt. The user did not ask for a cow with a crooked horn. They asked for a cow.

Synthetic images have little of what Roland Barthes called “The Reality Effect”—the thousands of small imperfections that make an object seem “real”. AI apparently needs to see a lot of something in its pretraining data before spontaneously including it in an image (note how infrequently Midjourney remembers to put a tag in the cow’s ear, even though nearly every real-life cow has one). When you see an image that’s just a perfect depiction of something—and nothing but that thing—get suspicious. Real life is never that perfect.

AI image models struggle with subtlety. They can mimic almost any art style, but it’s usually a loud, witless parody of that style. The concept that a picture might have depths—that its theme might require thought and interpretation, instead of being blared at the cheap seats—is foreign to them.

I find a lot of AI art really tiresome, to be honest. It’s the visual equivalent of being shouted at through a bullhorn. Human-created artwork is often subtle and seductive: the longer you stare at “Too Late” by Windus, the deeper you fall into its world and story. AI artwork often just feels like disposable junk: a loud surface with nothing substantial underneath.

Going in a darker direction, on 12 Oct 2023, the office of Israeli Prime Minister Benjamin Netanyahu shared a photo of what’s believed to be a burned baby.

(Content warning: burned baby.)

This is a horrifying image, and certain ideologically motivated individuals claimed it was an AI fake. They even got a Ben Shapiro retweet tagged with a community note, stating that this had been “proven”.

I took one glance at that image and thought “real”.

The reality effect is like a slap in the face. You see those sad little smudges of ash on the rails? Those random dust bunnies (or whatever) against the skirtingboard? AI never puts things like that in an image. When AI makes something dirty, it does so in a stylized, aesthetic way. The dirt will look like it was brushed into place by a professional set designer. Those random smudges in Netanyahu’s photo are too anticlimactic and “incidental”.

Also, if you could prompt an AI image generator for a “burned baby”, you would get something that unmistakably looks like a baby. Not a hunk of charcoal that barely looks like anything.

Other people cropped a few pixels from the image, uploaded it to AIorNot, and got a verdict of authentic, thus “proving” not a damned thing, because AI detectors are trash. They do not work.

I am not saying Netanyahu’s photo is definitely real. It could have been faked through some other method (not that there’s any reason to believe it was). But anyone calling this image “AI generated” is either lying, or being a liar’s useful idiot.

Nothing is Ugly. Everything is Beautiful.

In moments of doubt, ask: “Is there anything about this image that should be ugly, but isn’t?”

Every real photo has some ugliness. Zoom in on the nose of the real cow above, and it’s packed with dried snot.

Compare with Midjourney’s bovines, who thoughtfully blew their noses before being photographed.

I have mentioned the lack of flies. Also, the grass isn’t littered with piles of shit. Anyone who’s seen a cow knows how unrealistic this is.

AI models have a strong aversion to gross or nasty things. They struggle to capture the mundane grubbiness of real life, except in a stylized way. Prompt for an ugly person, and you either get a fantasy goblin… (Midjourney V6)^[3]Midjourney also has a “–style raw” flag, which supposedly omits the model’s “beautifying” process. It didn’t help much. To Midjourney, ugly person = DnD … Continue reading

…a “gritty” character actor in a Coen Brothers film that women would probably thirst over if he was real… (SDXL)

…or a refusal. (Dalle-E3)

Genuine, untrammeled ugliness is so rare that it sometimes goes viral. I remember when Loab became a thing. Someone claimed that by prompting SD a certain way you could uncover a paranormal cryptid in latent space.

As was noted by feminists, in most pictures of “Loab” there’s nothing supernatural about her. She’s not a demon or Lovecraftian monster. She’s simply an unattractive, unsmiling woman with a skin condition! People like Loab are everywhere in real life—I see uglier faces every time I go to a hardware store—but they’re strikingly rare in AI-generated images. So rare that they can literally be passed off as paranormal monsters.

Look For “Concept Merge”

This is awkward to explain, but you’ve seen it before.

Basically, it describes the AI tendency to get stuck between two contradictory ideas. Rather than choosing one or the other, it mashes them together. Here’s an image I swiped off Facebook. (I’m not crediting any of these losers.)

Looks good, but what are those objects on the wall? Brightening the image…

Are they frying pans? No, they have shield bosses. But if they’re shields, why do they have frying-pan handles?

This is concept merge. The prompt was clearly something like “a rustic cabin with a Viking boat in it”, but when it came time to add objects to the wall, Midjourney had a choice between “frying pans” (appropriate for a rustic cabin, inappropriate for a Viking boat), or “shields” (appropriate for a Viking boat, inappropriate for a rustic cabin). It got stuck, couldn’t make a choice, and eventually created weird frying pan/shield combinations.

Concept merge is a frequent feature of AI images, like the one below. It was shared on Twitter by someone who was convinced it was perfect, with no AI mistakes.

Obviously it has many: the girl has three nostrils, weird small gerbil-teeth, hair made of transparent plastic (note the left eyebrow visible under her bangs), deformed egg-yolk irises, and hairstrands that combine, multiply, and flow against physics in dozens of places. She has the generic “sameface anime girl” look endemic to a thousand LORAs such as Wowifier. The image makes incongruent stylistic choices: her exaggerated facial design (huge eyes + tiny nose + pointy chin) looks odd next to the realistic-looking skin and fabric.

But most striking are her clothes. What is she meant to be wearing?

1. A studded leather jacket?
2. A crop top?
3. A schoolgirl sailor fuku?

StableDiffusion couldn’t make up its mind…so it gave her all three at once!

Human artists do not make mistakes like this. Even the shittiest artist does not draw two totally different things at the same time. AI images split the difference and compromise. Faced with two choices, they dive right down the middle.

Concept merge is most obvious in small details. Look for least important thing in the image that’s still a distinct object. Often, it will be a couple of different objects smushed together.

Bricks. Tiles. Planks.

This will be fixed some day, but right now, AI cannot do tesselated patterns. Even the task of depicting a chessboard or a brick wall defeats every model.

It’s a classic case of attacker’s advantage. Tesselated patterns are defined by their uniformity, and every brick the AI renders must be perfect. A 95% success rate isn’t good enough: one malformed brick gives away the illusion. The AI has many chances to fail and only one chance to succeed.

So whenever you’re on the fence about an image, look for a fence in the image. Or a brick wall. Or roof tiles.

“Honey, can you go to Home Depot and pick up a single brick that’s about 4 feet long and bendy in the middle? You know the kind.”

“We’ll also need a perfectly square wooden plank for the floor.”

(This is more “concept merge”, by the way. The prompt was probably “a kitchen in a rustic cabin”. Rustic cabins have wood floors, but kitchens are usually tiled, so the planks assume a weird tile-like character.)

What other things can you look for?

Buttons. Is a row of buttons on a jacket spaced uniformly? Or are some further apart than others? In the grandma cake image, note that the top two buttons on her cardigan are spaced closer together than the bottom two. Likewise, the knobs on the thing in the background are unevenly spaced.

Electrical cables and ropes and strings—particularly if there are multiple overlapping ones. Are they sensible? Can you trace them from place to place with your finger, or do they become scribbled tangled-up garbage?

Tree branches are a big one. They’re so complex that they often look like random scribbled nonsense, and AI depicts them as such. Zoom in and you’ll see something’s not right.

Some beautiful concept merge here too. Is the cabin made of logs or stones?

(The comments on that picture were depressing. “Ha! You’re gonna feel like an idiot when the river floods and your cabin gets washed out!” If Mount Stupid was real, Facebook’s users would be Tenzing fucking Norgay.)

Also, zoom in on chain links.

Again, this might be fixed next year, or tomorrow for all I know. But for now, AI struggles with thin, complicated objects that exhibit a coherent structure.

The wonderful thing about the real world is that physical laws apply everywhere, equally, to all objects. A speck of dust is no less subject to the law than the titanic mass of Jupiter. As is written in the gospels, the regard of God is upon even the smallest sparrow.

AI does not care for sparrows. It has a limited bucket of competence, and lavishes it upon the central, important subject of the image. The small things get forgotten about, so zoom in on them.

Learn Basic Media Literacy

I hope I do not sound condescending or sarcastic. But whenever something happens in the world of synthetic media—whether it’s the GAN deepfakes of 2017, or Sora in 2024—we see the same histrionic reactions, over and over.

“Facts are over. Truth is over. We can’t trust photos anymore!”

Why were you trusting photos before? Does Photoshop not exist? Can’t a 100% real picture still be used to trick and deceive? How does a naïf like you have any money left in his bank account?

As Randall Munroe once observed, every anxiety about deepfakes could be applied to words. “You mean people can just…lie to you? Say that the sky is green when it’s blue? Say that the moon’s made of cheese? Oh my God, nobody can trust anything! We’re in a post truth era!”

Yes, deception is a problem. But it is not insurmountable. It hasn’t spelled the end of truth as we know it. We can defeat liars, whether they use words or photos or (soon) videos. We need to take the natural skepticism we have toward words (“talk is cheap!” / “money talks, bullshit walks!”) and allow it to flow over to images as well. This is an easy thing to do. So easy that there’s no excuse not to be doing it already.

When you see a photo of something alarming, ask questions like:

• Where did this image come from? What is its source?
• If it’s a photograph, who took it?
• Has the person sharing it built up any social capital that would inspire me to trust them?
• How easy would this image be to fake? Would anyone benefit from doing so? (if unsure, default to “very easy” and “yes”)
• People never take just one photo of really cool thing. Can I find multiple photos, from different angles? Is it plausible that the Pope would go out in public wearing a puffer jacket and exactly four photos of it would exist?
• Is this photo even physically possible to take?

Someone shared this photo.

The pelican’s body looks AI generated (“fur or feathers” concept merge). I don’t know about the head. It looks far more detailed than any I can get out of Midjourney—did someone crop out a pelican’s head from a photograph, and paste it in using Photoshop? It doesn’t seem to connect to the body.

But there’s zero chance that it’s real. It’s implausible that a drone would snap such a high-quality photo, perfectly framed and lit. We’ve seen pictures of birds flying into drones. They look like this.

This is another case of AI getting too good. It creates a photorealistic image even when we wouldn’t expect one.

Postscript: How To Recognize AI Generated Text

Models Have a “Voice”

We all know “ChatGPT tone” by now. People mistakenly call it “bland” and “generic”. It’s actually the most distinctive writing style on earth. I can recognize ChatGPT-created text within the first 1-2 sentences.

“Dive into the fascinating world of “skub“, a mesmerizing concept that echoes through time like a haunting symphony, unraveling a captivating portrait of the human imagination. “Skub” is more than just a word. It’s a philosophy. A lifestyle. A poignant reminder that divisive concepts can also unite us. As we delve further into “skub”, it is important to be mindful of diverse perspectives, and to recognize that “skub” may be controversial among certain groups. Together, “pro-skub” and “anti-skub” activists weave a synergistic tapestry, an intimate dance that highlights the positives of “skub” while remaining mindful of its potential impacts. So, in conclusion, whether you wear a “pro-skub” or “anti-skub” T-shirt, let us stand together in solidarity and mutual respect, allowing our diverse viewpoints to forge a better, more inclusive tomorrow.

(I didn’t generate that, I typed it from memory.)

Nostalgebraiest once described this as a HR managerial fantasy of how humans talk. Peppy and upbeat, with a hint of supercilious smugness. Supremely confident, while being frantically neurotic that it could offend or confuse someone. It over-qualifies and undermines itself with weasel phrases and ambiguity. ChatGPT is so terrified of saying something wrong that it doesn’t care whether it says anything right.

ChatGPT has the most exaggerated writing style of any chatbot, but they all have distinctive voices. Gemini loves contrastive “it’s more than just [thing x], it’s [thing y]” sentences. When you ask Grok to diss something, it always says “you’re like a walking talking [x]”

Many other models are trained on synthetic data from ChatGPT (whether they know it or not), and echo ChatGPT’s pablum.

Lots of Filler Adjectives

AI models are addicted to pretty but empty adjectives. Everything is captivating or enchanting or inspiring.

These are filler adjectives. They are vague, nearly meaningless, and add virtually nothing to the text. Any writing coach would tell you to cut them.

Take “captivating”. Its definition is “capable of attracting and holding interest”…but that could be anything. Sunsets are captivating. Car wrecks are captivating. Crossword puzzles are captivating. If an adjective could equally be applied to love’s first kiss and to a weird bug smashed on a car windshield, then it’s probably better to drill down to the metal and explain exactly what makes the thing captivating.

AI-generated text abhors such specificity. It casts a broad net over the reader’s mind, encouraging them to add their own meaning. You might have your own idea of what “haunting” means, so when you read “he played a haunting melody on the violin” this can easily engage your senses in a way that seems profound. It seems like the AI is reading your mind! In reality, it’s something of a parlor trick. There were no magic slippers, and the profundity was supplied by you.

Sarah Constantin wrote about a thing called “Ra”.

Ra […] can roughly be summarized as the drive to idealize vagueness and despise clarity. […] You know how universal gods are praised with formulas that call them glorious, mighty, exalted, holy, righteous, and other suchlike adjectives, all of which are perfectly generic and involve no specific characteristics except wonderfulness? That’s what Ra is all about. […] Ra is evident in marketing that is smooth, featureless, full of unspecified potential goodness, “all things to all people,” like Obama’s 2008 campaign.

“All things to all people” is a good way of putting ChatGPT’s writing style. It appears to fluently speak dozens of languages. In truth, it speaks only one: Ra.

AI Text Is Tonally Inappropriate

If someone is using AI to write their social media posts, this is usually obvious, because the text will not match the tone expected of the conversation.

When a human enters a social space, they subconsciously “read the room”, and adjust their voice (or writing style) to fit. If everyone is using slangy contractions, they will too. If everyone’s writing in a stiff, formal way, they will too. Linguists call this a “code switch“. AI usually has no idea where its text will end up. Thus, it cannot read the room, cannot code switch, and its output will be very “general”, disconnected from the discussion going on around it.

This is easy to recognize in practice. @fchollet tweeted this:

“Accelerating progress” should mean increasing the breadth and depth of our search for meaning & solutions. Increasing the diversity and creativity of our cultural production and consumption. To a large extent, I feel like generative AI achieves the opposite. It collapses everything into a narrower, standardized, derivative space. It might even result in cultural stagnation if left unchecked. Not for too long though, since people will simply grow out of it, eventually. Humans have a built-in drive towards interestingness. They get bored quickly.”

He received a number of replies. One was written by an AI. Can you spot it? Read them carefully.

1. Depends on the input, bland input=bland output.
2. Yes. Several years ago I was impressed… even now I’m impressed, but I’m not interested anymore: meaningless texts, generic art
3. We get bored too quickly
4. Very well said.
5. There’s a parallelism with sports betting: the most probable outcome is the less interesting, thus the smaller prize.
6. And/or the AI may get better at producing interesting results
7. Generative AI is a pattern extender. That’s it. This is a very good thing though.
8. Certainly, generative AI can seem limiting in terms of truncating our creative scope. However, shifting perspective, we may see it as a tool that can enhance our own creativity. Much like a kaleidoscope, it’s our interpretation of AI’s output that truly shines.
9. This is 100% about not allowing a better approach and a 1a trojan
10. “people will simply grow out of it, eventually.” That sounds like an implausible prediction. (Did people grow out of their propensity to get sucked into addicting tech in other cases?). Is there a concrete outcome in, say, a decade, that would tell us if it was or was not true?
11. Companies will pay people to do nothing but create so that they can give data to the ai

#8 isn’t a sore thumb, it’s an amputation. Aside from being generic feelgood blather (not that the human replies are necessarily any more insightful), it simply doesn’t look like a Twitter reply. Nobody says “however, shifting perspectives…” on the Bird Site.

(Plus, kaleidoscopes are hollow tubes containing angled mirrors and chips of glass. They don’t “shine”. Mixed-up metaphors are very common in AI writing.)

You can tell AI to imitate the a style, but that often leads to an unconvincing “hello, fellow kids!” affect (“Not gunna lie, some of those AI art pieces be lookin fire 🔥 tho lol”), burdened with excessive amounts of “style”. Conversational registers can get incredibly specific, and a failed mimickry attempt can be even more revealing than no attempt at all (I asked Gemini to write replies in the style of a /r/redscarepod poster. It came up with stuff like “The hot take I didn’t know I needed. Generative AI is like the cultural equivalent of the McRib.” Talk like that on /r/RSP and you’d get called a cop.)

Yes, some of those replies are innocuous, and wouldn’t attract notice. But if a bot constantly talked like that to an adversary dedicated to unmasking it, it would be exposed.

It’s easy for AI to fool a careless human. They were doing it in the 1960s. An observant one is a different matter. It’s the brick problem again: AI can get it right some of the time, and maybe even most of the time, but if it even slips up once…

“All of your clues can be defeated with about 5 seconds of prompt engineering. You know that, right?”

What do you want from me? If there was a bulletproof way to spot an AI image, I’d tell you. There simply isn’t one.

All of these hints are vague and “vibes” based. All require a close eye. All produce false positives (many photographers employ a staged, archetypal style. And ChatGPT had to learn its writing style from somewhere…) All can be overcome by prompting, inpainting, and Photoshop.

But there’s good news: most AI scammers are lazy too lazy to do that.

Whenever an AI-related scam hits the public’s eye, what’s striking is how flagrantly incompetent they are. They refuse to do any work to make their fraud believable. Remember that fake Willy Wonka event in Ireland? Remember how it had gibberish AI text on the flyers? This is your outreach! The one thing people see! It boggles the mind!

This is a level of laziness that I didn’t know existed. He could have easily fixed the text in Canva, or, hell, MS Paint. Even if he had zero graphic design skills, Fiver’s right there. It would have cost virtually nothing to have his goddamn ads look professional. Yet for the person running the event, “virtually nothing” was still a price too high.

And what about those repugnant Facebook images from pages called JESUS LUV U and the like?

This is not the work of a Machiavellian criminal genius. This is the work of an impoverished Malaysian contract worker, slaving away in a 45C cubical, creating 600 pieces of content per day so he doesn’t get fired or deported. He doesn’t care about quality. Why would he? It just has to be good enough to satisfy his boss (who also doesn’t care about quality, because he, too, is but a cog in a larger system. Hey, you look like you need to read Zvi on Moral Mazes. Understand its lesson, and the world will make about 17% more sense.)

There is no reason to do a good job in that scenario. It’s not as if he’ll win a trophy at the Scammy Awards this year if he makes an image look realistic. (“And the award for best deepfake goes to…Muhammad Awang! Keep it up, and we’ll print an extra visa for your sex-trafficked sister!”). Literally nobody cares.

There’s also the Nigerian Prince effect. Scam emails don’t actually benefit from being clever. They target the mentally ill, or people who don’t have a skeptical bone in their body. If you are capable of asking critical questions, you will never fall for the scam anyway.

Obviously really stupid con artists get caught. But are there smarter ones that aren’t being detected?

I don’t know. My honest impression, based on a glance at the rubbish flowing across my own Facebook feed, that >90% of these people are lazy, careless bottom-feeders. Click on their pages, and everything is bad. Look at the grandma baking a cake at the start of the article. The person prompting didn’t even take ten seconds to prompt a plausible number onto the birthday cake. Expecting them to fix fine mistakes in Photoshop is like expecting them to sail to the moon in a teacup or empty the ocean with a spoon. They are seemingly all like this.

Tech evangelist types are fond of telling us that AI has democratized software engineering, allowing non-programmers to implement their ideas. Well, that’s also true for frauds: specifically, lazy wannabes who would like to make money unethically but can’t quite figure out how to do so. AI—which can generate limitless text and imagery on command—is their Wonka’s golden ticket to a land of “ENCHERINING ENTERTAINMENT”. But they cannot escape who and what they are—incompetents—and give themselves away at every turn.

Horace once wrote “Caelum non animum mutant qui trans mare currunt”. Translation: if you suck, you suck. That’s my hope: that most criminals are just too lazy or dumb to do much with AI. Capabilities advance, but human ineptitude is forever.

You are not supposed to accuse people of being insane on the internet. It is obvious that Kanye West suffers from bipolar disorder. His swings between wild creativity and nihilistic self-destructiveness are indicative. But pointing that out would, of course, be stigmatizing people with bipolar disorder. I’m not his therapist. I have not made an informed diagnosis. So I’m not allowed to speak the obvious, no matter how much sense it makes, or how nonjudgmentally I say it.

Yet occasionally someone is obviously, deeply, flagrantly mad in such an expressive way that there’s no way to escape a reckoning with it. But our reactions are counterproductive, and often mad themselves. Often, we valorize the person, and regard their madness as a form of “specialness”.

Here are two pieces of writing on this topic I find profound. They both basically typed out my beliefs better than I could. I have all these little half-excavated thought-fossils in my mind, buried in sand, and I can’t quite get up the energy to rip them out. It’s nice when I don’t have to, when someone else has captured those same thoughts as huge, roaring, living dinosaurs, far better than anything I could write.

The first piece is Matt Lakeman’s 2020 blog post about The Room, and the Disaster Artist, and the toxic cult around “auteur” Tommy Wiseau. He has since taken it down, and turned his site into a travel blog. But it’s really good.

I read Disaster Artist on a whim when the movie came out. I’ve since gone through the audiobook 3.5 times and can confidently say it’s one of my favorite books of all time. I expected just to hear funny anecdotes about the making of a famously awful movie and the man behind it, but I found so much more depth. In my eyes, Disaster Artist is an examination of insanity (which I am defining as “the inability to perceive reality to the degree of low or non-functionality in regular life”). The book is a pushback against a subtle cultural norm that sees crazy people as having some sort of gift or potential or insight that everyone else doesn’t.

The next is Tom Ewing’s 60,000 word exploration of Cerebus, which is a 300 issue British comic about an aardvark whose creator went insane. Apparently the back half of the comic devolves into bristling thickets of gynophobic rants and mystical Torah analysis.

Here is a comment I left on the last entry, Aard Labour Epilogue: Dance Of The Aardvark Catchers.

Thanks for writing this series. I enjoyed it immensely. It actually inspired me (contra doctor’s orders) to start reading Cerebus.

I read issues #1-#3, then skipped ahead to somewhere in the high 200s. The difference was amazing—going from a bud sprouting on a stem, to a huge rotting flower, petals dissolving into muck. Obviously I didn’t expect to understand the story, but the sense of aesthetic collapse was stark. The start and end of Cerebrus barely look like they were created by the same species, let alone the same human!

Even in the early issues, Dave Sim isn’t an amazing writer. I noticed this in the letters of introduction: Deni Loubert’s will be fun and witty and engaging…but then I turn the page, and crash at high velocity into an ENDLESS BLATHERING TEXTWALL that I kind of bounce off. I’m sure Dave sharpens up after a few volumes, but I’m not looking forward to his Torah exegesis.

You’ve mentioned anime several times. Cerebus is also seen as an influential early work in the furry fandom (along with other culturejamming “comix” like Fritz the Cat and Omaha the Cat Dancer). Ironic, given Dave’s stance on gay people, that furries later became possibly the gayest subculture of all. (“By and large, furries are bi and large”—Eric Blumrich).

“The world, or the further-right parts of it, have moved closer to Sim, and one of the questions I had sitting down to re-read Cerebus is “how come this guy hasn’t become a cult figure on the alt-right?”. Reading it sorted that one out: Sim’s views on gender and politics are ordinary enough in those circles, but his religious convictions are intense and bizarre and inseparable from anything else he thinks, worlds away from the convenient surface performances that pass for faith on much of the right, and grossly heretical to anyone who does believe.”

Well, there’s also the paradox at the heart of extremist movements. The worst thing you can do is ACTUALLY BELIEVE.

The ones who rise to the top tend to be grifters and carpetbaggers. People with no real ideological attachment to the cause, but who have glommed onto it as a way to raise their own status. (Example: one of the main figures of the GamerGate movement was Milo Yiannopolis, who a year previously had tweeted that men who play videogames are losers.)

The grifter’s strength lies in knowing when to backpedal, when to apologize, when to moderate their words and behavior. They have rabbit ears to how they (and their movement) are perceived by outsiders, and are willing to pull things back to the center (at least on a shallow rhetorical level). They play the game.

Dave Sim does not play the game. He doesn’t think it’s a game at all. He’s on a holy quest to share the truth which supersedes all politics and optics. He either doesn’t know how repellant he looks or doesn’t care.

Hardcore culture-warriors of every stripe—Ayn Rand, Andrea Dworkin, Kellie-Jay Keen—often end up as pariahs in their own movements. Their charisma and force of will gets bums onto pews, but eventually, they bog the movement down, jamming it up with their unwillingness to compromise. Not that Dave Sim is necessarily on a level with those people (he’s not charismatic), but the same principal applies.

I mostly stand by that, although Milo Yiannopolis did not call gamers losers, he said “Few things are more embarrassing than grown men getting over-excited about video games”. I have unalived myself in Minecraft for this error.

But I also deleted the latter half of my comment, because I wasn’t sure how it would be received.

One of Ewing’s recurrent points is that Sim’s “madness”, when verbalized, sounds like standard cant for parts of the extreme right. There is some truth to that. But that doesn’t mean Sim also isn’t mad. It is not true that being mad gives you a unique and interesting perspective on life. Often the opposite is the case.

We have this cultural idea of mad people that isn’t true to reality. Louis Wain drew cute cats. As his schizophrenia advanced, the cats become distorted and twisted, resembling owls and hyenas and demons, before finally becoming abstract detonations of light. (Or so goes the story. Much of Wain’s work is undated and often nobody knows when a certain illustration is from.) Madness, like what happened to Bowie in 1976, gets seen as a kind of kind of gnostic initiation. Sane people have a locked door at the back of their minds, leading to worlds undreamed. Madness turns the key.

But that’s not how to works. In reality, mad people are usually the dullest people you will ever meet. They waste their lives retweeting political slogans on Twitter. They tumble down into predictable conspiracist obsessions. They end up as hollow spaces, filled with wind, chanted slogans, and fragmenting memories. They are tragic, but also boring and tedious.

Obviously it’s not the job of mentally ill people to be entertaining. That’s offensive—they’re not zoo animals. But there’s a stark difference between madness as popularly portrayed and madness as it truly exists. I dislike the pop cultural idea that mad people are “gifted” or have some creative spark denied to sane people. Madness isn’t creative freedom, it’s chains.

My father suffered a slow and hideous decline. He spent his final days nearly catatonic, watching Yes Minister DVDs. Occasionally he’d enter a manic state and “clean the house” (frantically strew rubbish everywhere, and then leave it like that until someone else tidied the mess.) When he stopped taking his medicine he did things that were embarrassing and out of character. At one point, he asked me to copyedit his autobiography. I agreed, and he he emailed me a few pages of rambling nonsense that trailed off mid-sentence. I found the document heartbreaking, and although I have kept it, I still wonder why I did. There is nothing of my father in that document. I am preserving some scrambled words that could have been generated by a Markov chain. In any meaningful sense, my dad died before he wrote it.

Sometimes crazy people are interesting (I know some), but they are not the plurality. Madness (on the whole), is “an empty head thinking as hard as it can.”

Today I watched the legendary 1940s Fleischer Superman cartoons. I thought “they’re in the public domain. What do I have to lose?”

Everything, as it turned out. I had everything to lose.

But let’s not think about that. This is my safe space, and I will fill it with happy things. Such as how Paramount acquired the film rights to Superman in 1940 and then contracted Fleischer Studios to animate a series of shorts based on the property and these came out between 1941 and 1943 and are regarded to this day as milestones of Golden Age animation and there were seventeen of them not sixteen not eighteen happy things happy things happy things

Was This The True Birth of Superman?

Superman changed greatly as he slid out of the cultural birth canal. At the very start, he wasn’t even a hero.

In 1933, writer Jerry Siegel and artist Joe Shuster self-published a short story called The Reign of the Superman (read it here). Using a fragment of a meteorite from “the Dark Planet”, a mad scientist turns a homeless vagrant into a “Superman”, a megalomaniac villain who seeks to rule the world.

This character, of course, wasn’t Superman.

In June 1938, the Siegel/Shuster duo reinvented the character for Action Comics #1. This new Superman was a heroic figure: an intrepid crime-fighter who lives under the undercover identity of Clark Kent.

But this character also wasn’t Superman.

Yes, the hero of Action Comics #1 has many similarities with Superman as we currently recognize him, and could fairly be called the genesis of the character. But he’s not all there yet. He’s not (explicitly) from the planet Krypton. He cannot fly (he jumps really high instead). He doesn’t live in Metropolis but Cleveland, apparently (in Action Comics #2, it’s mentioned that Clark Kent’s a reporter for the Cleveland Evening News). He apparently enjoys killing people. It’s OK, though—they’re gangsters or Irish or something.

The 1941 cartoons reflect another evolution of the Superman conceit. The Fleischers, faced with the task of animating the character, felt it would look stupid to have him jumping around everywhere (as opposed to every other part of Superman’s design, which is far from stupid) and gave him the power of flight. This is now Superman’s marquee ability. Additionally, many famous Superman phrases (“It’s a bird, it’s a plane!”, “Faster than a speeding bullet!”, and “I wish I had a third example!”) hail from the cartoon, as does the trope where he changes into his costume in a phone booth.

So do these shorts depict the “final form” of Superman?

No. The Fleischer Superman doesn’t live in Cleveland or Metropolis, but in Manhattan. There’s no mention of Lex Luthor or Kryptonite, and no attempt to situate him in a world of other superheroes (which I always found odd—is he still “Super” if there are others like him?) Lots of bits and pieces are missing.

So when was Superman completed as a character? Who says he ever was? Superman isn’t a character, he’s a process, an endless toppling domino-train of mutations that started in 1933 and continues through to the present day.

There is no dividing line for when Superman “became” himself, just as there’s no point where monkeys became modern humans. Even now, writers and artists are tinkering with him: retconning bad arcs, resculpting his character and world, drawing him in new ways, leaving their metaphorical DNA all over him. Nothing’s safe from revision, not even the logo on Superman’s chest.

Superman exists as a cloudy swirl of ideas, and whenever the sediment starts to settle, we stomp his visage back into mud. He will never stop changing, because he exists as a product positioned in an ever-evolving market. The primitive villainous Superman of 1933 is no more or less the “real” edition than Justice League’s mustache-less Henry Cavill.

The Demon-Eaten Superman

The question of haecceity—the “this-ness” of a thing—exists at the heart of philosophy, and rips that same heart out. Duns Scotus wrote about this issue in Ordinatio: when does a stone stop being a stone? How many pieces can you chip away before it loses its identity as this particular stone? In the early Sino-Indic Buddhist text 大智度論, or Da zhidu lun (of disputed authorship, but traditionally attributed to Nāgārjuna), there is a creepy story:

There once was a man who undertook a distant journey on assignment, and he spent a night alone in an empty hut. In the middle of the night a demon arrived carrying a corpse and placed it down in front [of the hut]. Then another demon arrived chasing after the first, and began shouting angrily at him: “That corpse is mine! How did you end up with it here?” The first demon replied: “It is mine! I brought it here myself!” The second demon said: “In fact I was the one who carried this corpse here.” The two demons each grabbed one hand [of the corpse] and struggled over it. The demon said: “Here is someone we can ask.” The second demon then asked the man: “Who brought this corpse here?” The man thought to himself: “These two demons are pretty strong. If I speak the truth I am certain to end up dead, but if I speak falsely I’ll also end up dead. In either case there is no escape, so what is the point inlying?” So he replied: “The first demon carried the corpse here.” The second demon was furious, grabbed the man’s arm, pulled it off, and threw it to the ground. The first demon took one arm from the corpse and affixed it back on the man. In this fashion, both arms, both legs, the head, flanks, and indeed the [man’s] entire body were changed [into the body of the corpse]. Thereupon the two demons joined in eating the [original] body of the man who had been transformed, wiped their mouths, and departed. The man thought to himself:“ I just saw, with my own two eyes, two demons eat up my body—the body to which my mother gave birth. Do I actually now have a body or not? If I do, then it is wholly the body of another. If I don’t, then what is this body?” Thinking like this made him horribly confused, like a madman.

(Text quoted from The Curious Case of the Conscious Corpse: A Medieval Buddhist Thought Experiment, Robert Sharf 2023, Reasons and Empty Persons: Mind, Metaphysics, and Morality).

It’s a variation on the Ship of Theseus thought experiment. A plank in Theseus’s ship is rotting away, so you pull it out and replace it with a new one. It is still clearly the Ship of Theseus. But one by one, the other planks rot away, and are replaced in turn, along with its sails and fittings. Eventually, none of the original ship remains. Is it still the Ship of Theseus?

Some continental philosophers (Wittgenstein!) try to rescue the notion of haecceic identity by connecting it to a process or a system. Consider the sequence of sets {a, b, c}, {b, c, d}, {c, d, e}, {d, e, f}. The last set {d, e, f} has no elements in common with the first {a, b, c}, but we still recognize it as a variant of the first, because they shared common elements in the middle. Their identity rests on the larger process that they are a part of.

Is the demon-eaten man in Da zhidu lun still himself? Wittgenstein would probably argue “yes”, because he’s part of the same flowing river of consciousness, regardless of what atoms compose his body (it’s important that he remembers “his” old flesh). Likewise, the Ship of Theseus remains the Ship of Theseus, because its planks were replaced one at a time, new caulked against old. If there is a process linking the past to the present, the identity remains.

But then how do you allow room for a thing to become something else? Humans are distantly descended from primitive eukaryotes called protosterol biota, which lived more than 1.2 billion years ago. But humans are not protesterol biota.

And it doesn’t tell you when a thing actually begins. Most philosophers debate the point where the Ship of Theseus stops existing, but it’s equally fraught to say when it starts. When was the moment when the Ship first appeared? When the final nail was hammered in? Surely it could have sailed without that nail. When the lumber was fully milled? But what if a splinter broke off, and was left behind? Did the Ship of Theseus already exist in the trees before they were cut down? Or when those trees were seeds under the soil? When did the Ship of Theseus first amalgamate? Was it always with us, glistening with the dew of creation?

Superman is like that: a creature without start or end. He’s all process and flux, a red metal endlessly beaten into new shapes by the hammers of artists, never allowed to cool.

Biologist Richard Dawkins wrote his “METHINKS IT IS LIKE A WEASEL” algorithm to prove that sophisticated order can emerge out of just two things: mutational pressure and selection. But this algorithm is simplistic: unlike actual evolution, Dawkins’ program has a defined goal to aim toward. Once the program has generated the words “METHINKS IT IS LIKE A WEASEL”, its task is complete.

But real evolution, like art, has no end goal. Living creatures merely seek to survive and propagate more of themselves. So do nonliving creatures. Superman will never be complete, because no creative work is. Some artists say an artwork is “finished”, but they are lying. They merely got tired of working on it. Lock them in a room with their “finished” painting and they’ll last five minutes max before noticing something wrong. Something that offends the eye, and has to be corrected. They’d add a brushstroke, then a second, then one more. Artists are demons, devouring the limbs of their creations and sewing on new ones like hell’s dollmakers.

How Much Did the Fleischer Supermans Cost?

A lot. They cost a lot. You may have heard some variant of the below:

“Superman was a cultural phenomenon, cartoon shorts were booming, and Paramount was ready to pay top dollar for an animated Superman. But Dave Fleischer didn’t want to work on it, so he quoted an absurdly high amount of $100,000 per episode, expecting them to refuse. But they accepted his quote! So then he had to make the shorts, with the biggest budget ever! LOL!”

That is broadly but not entirely correct. Paramount bargained the Fleischers down to a smaller budget of $50,000 for the pilot, and $30,000 apiece for the remaining sixteen cartoons (eight of which were made without involvement from the Fleischer brothers). The total was $530,000, or $3,117 per minute of animated footage.

Compared with the budgets of early animated movies, this was generous but not extravagant. Snow White and the Seven Dwarves (1938) cost $1.5 million, or $18,072 per minute of footage. The Fleischers’ own Gulliver’s Travels (1939) had a $700,000 budget, or $9,210 per minute of footage.

Even by the standards of animated shorts, a $30,000 budget wasn’t unprecedented. Some of Walt’s Silly Symphonies cost around $30,000 to make, for example. (Source: Gabler, Walt Disney: The Triumph of the American Imagination, p214), and that was eight years earlier, during the Depression.

Nevertheless, the Fleischer Supermans are legendary for their high animation quality. That’s the main thing you’ll take from them: how good they look. The machines and the buildings look extremely lifelike: it’s early draft Katsuhiro Otomo, obsessed with gears and rails and steam engines. Things have a wonderful cinematic drama, as well as a Milt Kahlian sense of heft and weight. The Fleischer studio had clearly moved beyond Betty Boop.

The writing is…somewhat less excellent.

Does Superman Suck Balls as a Character?

I don’t “get” Superman. Putting aside the question of whether he’s interesting, how do you write compelling stories around a guy who can do anything? 404 Error: narrative tension not found.

Modern writers have a postmodern answer: debase him. Blacken his eye, and throw mud on his cape. He’s been built up on a pedestal, so pull him off it. Break down everything that’s good and pure in him, and reveal the lie of American exceptionalism. Make him Subman.

But Superman’s early 1930s writers don’t do that. They portray him without a hint of irony. Indeed, once the shorts’ intriguing art deco sci-fi flair wears away, they settle into a familiar trudge. A bad guy will build a visionary techno-gadget, use it to commit some laughably anticlimactic crime (like robbing a bank or stealing jewels), and then Superman will show up, and just…win. It’s never clear how the villain expects to overcome such a mighty adversary. They seem like idiots for even trying.

Also, Superman’s power level wildly fluctuates.

• In “The Mad Scientist”, he lifts a toppling skyscraper.
• In “The Arctic Giant”, he effortlessly overwhelms a Godzilla-sized monster.
• In “Eleventh Hour” he pulls a Japanese battleship to the bottom of the ocean by pulling on its anchor chain.
• In “Terror on the Midway”, he barely wins a fight against an ordinary black panther.
• In “Japoteurs” he cannot get past a locked door.

In “The Mechanical Monsters” he uses X-ray vision, but otherwise the writers seem to forget he has this ability. There’s a maddening sense of nonexistence to Superman. Like he has no traits, he’s just whatever the writers need to get this month’s short out the door. Superman’s only consistent power is “he always wins, perhaps after grunting and straining in a suitably manly fashion”.

What’s the appeal of these stories, and this character? Possible answers:

• Superman is closely entwined with America’s perception of itself, and nobody ever went broke with a theme of “AMERICA! FUCK YEAH!” Particularly not in 1941.
• There are Rooseveltian political overtones to the character that people related to. Often, Superman’s enemy in the early comics will be a slum landlord, or a circus owner who mistreats his animals, or some other populist bugbear. You could call him New Deal Man and nobody would blink an eye.
• It’s provocative to make the greatest American of all a foreigner—a provocation that I’m sure Siegel and Shuster (whose names were once Segalovich and Shusterowich) were both aware of and intended.

…But ultimately, Superman makes sense not as a character, but as the embodiment of an audience fantasy: the little man, empowered. We love Superman, not because he can fly or punch through walls (as we’ve seen, he sometimes can’t do either of those things), but because he’s Clark Kent. Don’t discount the man on the street! He could be a hero!

But that doesn’t work for me. Clark Kent is not the average man. He is a bland boy scout with none of the flawed roughness that makes humanity interest. The average man on the street picks his nose, farts, and masturbates. He laughs at racist jokes, supports questionable political candidates, and leads a double life that’s spectacularly unheroic—the real Clark Kent doesn’t wear a cape, he has a secret second phone hidden at the bottom of his sock drawer, where he hopes Lois Lane won’t find it. I keep hoping to find a flaw in Clark Kent. Instead, he’s perfect. You cannot write a paean to humanity and then fail to depict humanity.

Watch Disney’s Mickey’s Rival, from a half-decade earlier, and you’ll see this done right. Mickey and Minnie Mouse are on a date together, when some smooth-talking jerk pulls up in a fancy car, and makes a pass at Minnie. Unbelievably, Minnie accepts! Mickey (who actually had a character in his early years) fumes and seethes and clenches his fists. Wilfred Jackson’s direction is masterful, and puts us squarely in Mickey’s shoes. He’s weak and he’s pathetic and he’s us. Despite all our rage we are still just a rat on a page. Mickey might be a ghastly quasi-minstrelsy rodent whose malformed visage is injected-moulded onto every hunk of plastic in the Pacific Garbage Patch…yet in “Mickey’s Rival” he’s more human than Clark Kent will ever be.

The truth of Superman is that he’s a children’s character who has attained an unusual place in the pop culture canon due to his longevity and influence. You can analyze him and make him seem deep and profound, but this is like doing X-ray crystallography on a cream soufflé. He’s just a kids’ character.

I find this inspiring. In fact, it’s exactly the tribute to humbleness that Clark Kent isn’t. Even a kiddie comic can lift a god up into heaven.

“Monstrous Tracks of Unknown Origin”

One of the episodes, “The Arctic Giant”, features a dinosaur.

People have always loved dinosaurs, and all they represent. Ancient dirt, vomiting up implications of an ancient and unthinkable past. Dinosaur bones are like shadow puppets gyrating against the wall of the present.

Past generations had their Jurassic Parks. 1933’s King Kong really pushed jazz-age dinomania into the mainstream (King Kong’s influence is also strongly felt in the Superman shorts. In “Terror on the Midway”, the colossal ape makes an uncredited appearance!), but that’s only the most famous example. Dinosaurs feature prominently in Hugh Lofting’s Dr Dolittle books, as well as in Arthur Conan Doyle’s The Lost World. A few decades earlier, the so called “Bone Wars” of early paleontologists had captivated the public. These monsters locked deep in time were awesome, partly because you could imagine them into whatever shape you wanted. There weren’t any of those “dinosaurs were warm blooded!” and “dinosaurs should be depicted with feathers!” types running around back then, damn it.

An important piece of early animation is Windsor McCay’s Gertie the Dinosaur, which is from 1914 and remains highly interesting today. It’s arguably the first 2D animated short to develop a character. Gertie is shy and clumsy. She keeps doing the wrong thing and getting scolded for it. She’s a far more human character than Clark Kent, aside from the moment when she attempts to murder a woolly mammoth by throwing him into the lake—an act that the narrator expresses no disapproval for, oddly enough.

(But seriously, this is starting to worry me. Are there any classic animated icons that haven’t attempted homicide? What about Koko the Clown? Has he killed before? I’m joking. He’s a clown. Of course he’s killed before.)

A collection of 10,000 drawings on rice paper, Gertie not only precedes color and sound, but also the invention of translucent cel mattes. Notice the flicker of the background. There’s a “holy fuck” moment when you realize that they’re drawing the background over and over, thousands of times. I imagine McCay and his assistants breathed a sigh of relief whenever Gertie ate part of the scenery. One less thing they had to draw.

Anyway, Superman also features a dinosaur episode. (I know it’s not a TV show and doesn’t have “episodes”, but my heart insists otherwise). “The Arctic Giant” involves a team of scientists bringing back a frozen “Tyrannosaurus” from the wastes of Siberia. Unfortunately, it unthaws in the middle of New York or Cleveland or wherever we fucking are, and starts rampaging around the city.

The animation on the dinosaur is a bit lackluster (I guess here’s the part where they influenced DiC and Phil Harnage). You might question the scientific believability of a giant lizard living in an arctic climate. Well, obviously it didn’t live. It froze solid. You people need to pay attention.

The Tyrannosaurus lays waste to half the city in grand style (there’s a fantastic shot of the dinosaur stepping on a car and crushing it like a tin can), before Superman gets his butt into gear and overwhelms it. It’s an exciting episode, although the usual plot contrivances (Lois Lane is an idiot who gets into trouble, etc) appear.

The dinosaur is silent! It keeps opening its mouth, as if it’s roaring, but it doesn’t make a sound. I find this very creepy!

Other people apparently find the dinosaur’s silence odd. I found a Youtube fan edit that actually dubs Godzilla roars over the dinosaur’s silence.

Here’s something else creepy. Youtube has a new feature where it shows you the moment people generally spent the longest watching. Wanna guess what that huge bulge at the 1/3 mark represents?

The first time the dinosaur appears? Good guess, but no. It’s the first appearance of Lois Lane. Dinosaurs are cool to look at, but they’ve got nothing on PRETTY GOILS.

Superman as a Propaganda

In late 1941, the United States entered World War II: or was forcibly entered, you might say. The ripples from Pearl Harbor took a while to reach the Fleischer Superman shorts, but reach him they did.

The animation quality remains high in the back half of the Fleischer Supermans, yet the content turns sharply propagandistic. Superman is staged against the backdrop of omnipresent war, with his enemies becoming the Imperial Japanese and Nazis.

“Superman is a ready-made propaganda icon” is so obvious a thought that it’s hardly worth thinking, but it is striking how well he fits in his new role, despite not being created for that purpose.

Most propaganda works the same way: by pairing a triumphant “we are mighty and powerful, the enemy is pathetically weak!” theme with a “we are in an existential fight for survival!” counterpoint. This is a bit contradictory (how powerful are you, to be threatened by such weakness?), but propaganda is not known for its depth of thought.

This zone of cognitive dissonance is exactly where Superman sets up shop. He’s laughably all-powerful, yet strangely useless. He could probably win the war on favorable terms within a few days. But his actions involve capturing a hijacked plane and sinking a battleship. The task of fighting the Japanese is given to the US Army: it never occurs to Superman to volunteer himself as a tactical asset. He still works undercover, maintaining the fiction that he’s a reporter for the Fuckbutt Times or wherever. Like a standard propaganda hero, he’s both omnipotent and impotent.

World War II propaganda, particularly in the United States, tended to work on a “heroes, villains, victims”, model. It establishes the heroes (us!), the villains (Nazis!), and the victims (civilians, Europe) who exist as pitiful moppets who must be defended. It’s basically the “dicks, pussies, and assholes” monologue from Team America: World Police. Superman generally fits this mold as well. He is the hero. The villains are the Nazis. And the victim is Lois Lane—though in a quasi-feminist turn of events, she now has meaningful things to do. Her reporting is portrayed as critical for the war effort.

Hitler makes a brief appearance in the shorts. As I’ve said, the number of classic cartoon characters who haven’t canonically committed murder is quite possibly “zero”, and the same might be true of characters fighting against or collaborating with Hitler.

(According to the ever-helpful DC Fandom Wiki, Hitler’s Powers and Abilities include “Artistry”, “Military Protocol”, and “Political Science”.

Closing Thoughts

I had an idea for a story once. Real life superheroes, whose powers are strange.

1. There’s a man who cannot sign his name. Or he can, but it’s a different signature each time, as though a different person is writing it. Sometimes his signature will look flowery and feminine, other times clipped and jagged. He gets in no end of trouble with the bank and post office, but he simply cannot produce a consistent signature, no matter how hard he tries.
2. There’s a girl who is never noticed at security gates. Ever. She can walk onto a train, and nobody will check her ticket. She can walk into a movie that’s already playing, and the usher won’t ask to see her stub. Many of us have gotten lucky this way. She has gotten lucky hundreds of times.
3. There’s a boy who is ignored by pets. I don’t mean that in the usual “animals don’t give a fuck” sense. I mean pets literally don’t seem to see him. When he pets a sleeping cat, the cat doesn’t purr. When he’s left to dogsit a friend’s mamalute, the dog whines in abject loneliness, pacing the house as if it’s empty, completing ignoring him. You know how animals sometimes seem to see things we don’t? What would it mean if they somehow didn’t see you?
4. There’s a man who doesn’t exist in history. He accomplishes incredible, amazing things…but by next year, they’re gone. He once killed a mugger in a dramatic home invasion. It was all over the newspapers, but when you look up those stories, they are different. No mention of him. He once won a marathon. But on the internet, the winner is listed as a different man. He lives a frustrating existence. He could take over the world, and the world simply would forget it’s ruled by him.

I abandoned this story. For one thing, it was too close to the plot of Mystery Men.

For another, I realized I was revisiting the same theme, over and over. Superpowers that imply you don’t exist.

Perhaps I was subconsciously reacting to Superman. He is the blandest, emptiest nothing I can imagine. In him, I feel the dark sky. The sea’s aphotic midnight swell. The black space that’s devoid of anything, even Kryptonite planets.