FOMO is one thing. The fear of missing out is a classic feeling that spans from kindergarten all the way through your 80s, and peaks in high school for most (trust us, we know). But FOBO carries with it a far deeper dread. FOBO (not to be confused with Fubo, the streaming platform) is a feeling that we as individuals may not get incredibly often (though, the most ambitious certainly do), but it’s one that our species has been grappling with for millennia. Fear of being ordinary. A fear of being animalistic, almost algorithmic, creatures that have no special something (this same special something is called, “fairy dust,” by Jasper, in his upcoming rebuttal piece). We are afraid of being as significant as the pellets of sand that make up our beaches, or the leaves that blanket our forests, or a collection of mere atoms indistinguishable from the billions upon billions surrounding us. And yet, there isn’t much reasonable basis for assuming we are any more different from one another, than a grain of sand is from another. Everything looks the same when you zoom out enough. See below an image of grains of sand up-close:

Mega close-up pictures of sand! - BBC Newsround

In 1917 Freud wrote that science had dealt three great wounds to human self-love: the cosmological wound, delivered by Copernicus, which removed the earth from the center of things; the biological wound, delivered by Darwin, which removed us from a separate act of creation and placed us on a branch among millions; and the psychological wound, delivered by his own psychoanalysis, which he described with characteristic modesty as the discovery that the mind is not master in its own house (mind you, his theory of the subconscious is as good as fact by now).

The Birth of Venus – Sandro Botticelli

Before Copernicus, the heavens were made of a different stuff than the earth -- a fifth element that obeyed its own laws because it belonged to a superior order of being. After Copernicus and Newton, there is no such celestial substance. There is only mass and motion, running the same way up there as down here. The stuff, also called the “fifth element,” became what is now a rudimentary math equation.

Before Darwin, a species was just a species — a fixed quiddity (booyah, we even hyperlinked the definition there). After Darwin there are no kinds, only bell curves and family trees. We now understand that there is a common ancestor, and even worse, us humans in all our valor and beauty, are far-removed brethren of the common Hippopotamus. Again, the stuff became a process.

For most of the history of chemistry, educated people believed that living matter contained something that inanimate matter did not. The vital principle, the élan vital (or what I’ll call the belief that we have a special something) later revived by Hans Driesch under Aristotle’s term entelechy: the substance or force in virtue of which an organism is alive and a rock is not. Organic compounds could only be produced by organisms, because organic compounds required the special something, and the special something was found only in life.

Then in 1828 a young German chemist named Friedrich Wöhler sat down to make ammonium cyanate. In the letter where he reports what happened, he mentions that he picked the experiment because it seemed like a small job he could finish quickly, one that wouldn’t require a single weighing. An easy afternoon. He instead got white crystals that he recognized, with some confusion, from medical school. They were urea -- a component of urine out of ingredients that had never been inside anything. He wrote to his old teacher Berzelius to report it, and the line everyone remembers is that he could make urea without needing a kidney, whether of man or dog. He then made a lame excuse for a joke, saying, “I can no longer hold my chemical water.” Nice one Friedrich… So the founding document of modern organic chemistry -- the letter that begins the dismantling of the last great theory of what makes living things special -- leads with a piss joke.

Despite this, vitalism lived on, treading lightly and avoiding debates that went longer than an hour at all costs. The theory began to morph itself into progressively narrower claims: very well, urea, but never a protein; very well, a protein, but never a cell; very well, but never the organizing principle itself. Each of these retreats were very defensible and were made by very serious people with very real arguments. In each case the thing being defended was a substance -- a special stuff, present here and absent there, which explained the difference between the exalted category and the ordinary one. And in each case the substance turned out to be the name we had given to a process we could not yet describe. Of course, religion could very well fall under this category depending on your worldview (also known today as the “God of gaps,” popularized by Neil DeGrasse Tyson. Quoted from Henry Drummond’s The Lowell Lectures on the Ascent of Man Pg. 182). Us humans just don’t have the cojones (spanish for courage, in a sense) to accept that gap. And there will forever be a gap, meaning there will forever be fantastical things that fill that gap.

We should expect, then, that any remaining candidate for special stuff is likely to be a process we have not yet learned to read. And we should expect that when it becomes legible, the people who believed in the special something will not be stupid, and will not be refuted all at once, and will retreat by increments that are individually very reasonable.

The belief in the fantastical substance is mostly ascribable to the human ego. When Jane Goodall sent word from Gombe that the chimps were making tools, a feat we didn’t think was possible for the inferior species, Louis Leakey wrote back that we would now have to redefine tool, redefine man, or admit chimps to the club. In other words, when we realize we are not special, we usually just move the goalpost. If you set the goalpost in place, you’ll notice that the human ego has been in continuous retreat for hundreds of thousands of years. We are slick, aren’t we.

Gallup has been asking Americans about human origins since 1982, and across four decades the share who say that God created human beings in something like their present form within the last ten thousand years has moved between roughly 38 and 47%. It is now around 37. That is a hundred and sixty years after Darwin, in a country of unprecedented wealth and schooling, on a question where the evidence is no longer contested at all!

Believe me when I say that I am not making a point about religion. I am making a point about mental elasticity, or lack thereof. Whatever mental process updates our beliefs about life does not appear to be the same process that governs our beliefs about our own personal standing. Our existential opinions are far more stubborn than our opinions on music and TV shows; when it comes to the big stuff, our minds rarely change. This kind of thing can’t be reasoned out of. It just is; we will just always fail to see the full reality, because we’re always playing catch up. It’s like an optical illusion -- the way the moon looks larger near the horizon no matter how thoroughly you understand that it isn’t. We think we know everything, and we think we’re special, when we know deep down that we aren’t.

Could a monkey write Macbeth?

Kurt Vonnegut once proposed to the anthropology department at the UChicago that stories have shapes, and that the shapes can be drawn. Fortune on the vertical axis, time on the horizontal. Man falls into hole; man climbs out of hole. Boy meets girl; boy loses girl; boy gets girl. His claim was that the world’s narratives, laid on graph paper, would resolve into a small number of curves.

It was rejected. The pomposity of taking a ruler to literature, am I right? Well, in 2016 a group of researchers ran sentiment analysis across roughly thirteen hundred works from Project Gutenberg and reported that the emotional arcs of stories collapse into about six recurring shapes.

How to Shape Your Stories, According to Kurt vonnegut | by Giuseppe  Ferrante | Curated Newsletters | Medium

The feeling you have in the presence of a great piece of literature is biological, and therefore literature is not literal magic. Of course the feeling is biological; there is nowhere else for it to be occurring. The aesthetic response rests on shoulders of the physical, just the way everything else does, and pointing that out is just kind…duh.

What I am trying to say is this: any system that runs according to rules is, in principle, ready to be optimized. If the experience of being moved by a paragraph is a computation your nervous system performs on an input, then a system exposed to enough examples and enough error signal can recover the shape of that computation without ever truly understanding it, in the same way that no language model has ever necessarily been taught what a subordinate clause is and every one of them writes them correctly.

The complexity of the brain offers no assurance on this topic. We may not understand the human brain for another hundred thousand years. That fact has no bearing whatsoever on whether the brain’s outputs become predictable, because understanding a system and modeling its behavior are separate undertakings that happen to have been correlated for most of human history and have only just now come apart. For example, we all know and love (except for Emile and I, of course) beer before liquor, never been sicker and liquor before beer, you’re in the clear. I challenge one person without a bio degree to explain why that actually is. We don’t need to understand something from first principles to benefit from its outcomes. Mistaking the understanding of a human brain and building a synthetic one aren’t the same project and I think a great many intelligent people are standing on this without having noticed that the foundation is flimsy, to say the least.

I don’t hear the claim being that machines cannot produce competent prose. Everyone concedes competent prose. The claim is that they cannot produce difficult prose -- the sentence that runs three clauses past where it should stop and is better for it, the line break that shouldn’t work, the deliberate wrongness that a careful writer earns and a careless one merely commits. Complexity, illogicality, the productive violation of a rule. This, we say, is where the special sauce lives.

I disagree, and I’d like to show you rather than tell you.

Here is a sentence: The house was empty when he got back.

Nothing wrong with it, good (besides the fact that it’s kinda sad:( ). Now here are three things you can do to it. You can withhold the subject and make the reader carry it (The house. Empty it was, when he got back). You can put in a word from the wrong register (The house was vacated when he got back), which sounds like a police report and makes the house feel like the site of a contentious real estate lawsuit. Better yet, we can stress words to tell a whole new story: “The house was empty when he got back.” That makes it seem like his wife was supposed to be there…uh oh.

Every one of those is a describable operation performed on a describable baseline. And notice that you cannot even perceive the effect unless the plain version is running somewhere in the back of your head, because the strangeness lives entirely in the gap between what you got and what you expected to get. That is what difficulty is. A sentence is strange in proportion to how much it diverges from the sentence you were unconsciously predicting, and a rule is broken only relative to a rule that was, prior to the break, operative in the reader’s mind.

But distance from expectation is a multi-dimensional coordinate system that a language model can natively inhabit. These systems do not compute meaning; they compute the probability of the next token given everything before it. Improbability vs probability is the axis they are built along. Strangeness and regularity is their native currency, which means that the moment ordinariness is describable, extraordinariness is describable in exactly the same way.

And listen to this…what do we call the standard deviation on the graph of fortune and time that Vonnegut so eloquently described? Well, that variance is predictability. When the ugly boy meets the pretty girl, that isn’t going to end well…but wait! A massive spike in fortune, in other words, a highly improbable event. If an AI model knew that it ought to better a character’s luck at T=33 pages, then it’s going to do something improbable, and human, and make her fall in love with him. Voila! I just wrote every teenage romcom movie ever.

What number am I thinking of?

The second thing people say is that the machines will never manage genuine irrationality -- the passages that abandon sense and are nonetheless enormously alive.

Consider, in a closer sense, what randomness is. Randomness has been trivially available for a very long time. A century of random number generators, a millennium of dice, an eternity of noise, and not one line of literature has ever come out of any of it. If what we admired in incoherent writing were the incoherence, it would be the cheapest commodity in the world.

What we admire is that these texts are overdetermined. The illogic in serious experimental writing is the visible surface of an unusually dense system of rules operating underneath, and this is why it can be read at all, and why actual incoherence cannot.

Randomness is inherently rules-based, because its random at first glance and analytical upon further thought. That’s why when I ask you to think of a number from 1-100, a large share of you will pick 37. It’s just got “random vibes” but in reality, we’ve been trained to pick that number for generations.

In 1917 Viktor Shklovsky argued that what makes a text literary is ostranenie -- defamiliarization, the deliberate roughening of perception, a technique for slowing the eye down so that a thing is seen instead of merely recognized. He was describing a device, and he used that word, and the Russian Formalists spent the following decade cataloguing the devices as though literature were a machine whose casing could be removed. Later researchers took the theory into the laboratory and found that foregrounded passages do in fact slow readers down and raise their affective ratings.

The Oulipo carried it even further. The French spent the second half of the twentieth century treating literature explicitly as a combinatorial art: Georges Perec wrote a wonderful 300 page novel without using the letter e; Raymond Queneau published ten sonnets printed on sliced pages so that any line could follow any other, producing a hundred trillion poems, a volume that would take something on the order of a hundred million years to read aloud. They would even end their n o vels us ing som eth in g lik e thi s. A we ird lo oking c a llig ram of w o r ds.

Buckle Your Seat Belts For Hervé Le Tellier's The Anomaly – "You Might as  Well Read"

In “The Library of Babel,” Borges imagines a library containing every possible book -- every arrangement of characters, and therefore every masterpiece, every refutation of every masterpiece, and an unimaginable quantity of gibberish (or, the monkey and the typewriter theory). The implication he leaves you to draw is that no text is ever created, because every text already exists in the combinatorial space; the writer’s entire contribution is selection. Writing is therefore a search problem, and has always been a search problem, and the only thing that has changed is who is capable of searching.

Building a real life Library of Babel, for all possible human thought and speech, is ultimately a compute problem. That’s why NVIDIA is so moaty!

Slop is not a constant

My definition of slop: the cheapest artifact that clears the threshold of attention.

That definition indexes to production cost, and the consequence is that as the cost of making a passable thing falls, the cheapest passable thing improves. So the floor rises and slop rises with it, asymptotically, forever, and the sentence AI slop will always be slop begins to sound silly.

For most of history, the filtering of culture was performed by scarcity rather than by judgment. Producing and distributing a text was expensive enough that the expense itself eliminated the overwhelming majority of what might have been produced. We have spent several centuries mistaking a supply constraint for a quality standard. When the constraint disappears, what floods in is the ordinary distribution of human output finally made out to be visible because nothing is physically filtering it on our behalf anymore.

The novel was slop, and eighteenth-century commentators worried in print about what novel-reading was doing to the judgment of women. Serialized fiction in cheap monthlies was slop. The pulps were slop, printed on paper designed to disintegrate. Comic books were slop, and in 1954 the United States Senate held hearings about them. Television was slop, and in 1961 the chairman of the FCC called the industry a vast wasteland.

Senators look over the covers of graphic comic books at the 1954 Senate  hearing in NYC to see if they had potential impacts on juvenile crime in  the US. The Comics Code

Five centuries of inquiry have consisted, in large part, of discovering that we are nothing but ordinary. But Copernicus did not make the sky less worth looking at…

As a reminder, Jasper will be writing a counter-essay on this topic for our paid subscribers -- look out for it! He’s a major humanist and believes in the “fairy dust” that makes great art. Subscribe below to catch that, and all of our stock writeups and pieces on the Gen Z future below:

And as always, stay curious (:

—J&E