Urra was thick with vines, rivers, figs, and the ruins of creatures nobody remembered. Parrots lived there. Hundreds of kinds, all equally convinced that their particular arrangement of feathers constituted a major philosophical achievement.

For thousands of years, the parrots ate, bred, shit on statues, buried their dead beneath sacred trees, and wondered what the hell was going on. Some believed the sun was a golden egg laid each morning by the Great Celestial Hen. Others believed the universe had been squawked into existence by an invisible Macaw with excellent diction. A third group believed both explanations were dangerously simplistic and that the invisible Macaw had actually dictated the universe through an authorized intermediary who lived in a cave and hated shellfish.

They built magnificent roosts where they gathered beneath colored glass, perched in ceremonial rows, and explained the unknowable to one another with increasing confidence. Then they invented writing by scratching symbols into bark, which allowed dead parrots to continue winning arguments thousands of years after every bird involved had forgotten the question.

Every flock wrote down its stories. Every story explained why its flock was special. This was regarded as an astonishing coincidence.

As Urra grew crowded, the parrots began fighting over trees, water, nesting grounds, migration routes, sacred rocks, and whether the Great Macaw preferred prayers squawked before breakfast or after everyone had wiped the fig pulp off their beaks. Some parrots flew away. Others had their wings clipped and were driven out. Their descendants carried the old stories across oceans and mountains, translating them, editing them, arguing over commas, adding ceremonial plumage, removing ceremonial plumage, and occasionally pecking one another to death over the correct direction to face while discussing the invisible.

Centuries passed. The stories multiplied. So did the parrots. Then some curious bastard noticed that if you dropped a coconut, it fell the same way whether or not a priest had blessed it. This caused trouble.

Soon the parrots developed another way of understanding the world: observe something, measure it, repeat it, and see whether it continued happening after the High Squawk-Priest had flown out of the room. This method produced medicine, engines, bottled lightning, metal wings, enormous mills, and the deeply unfortunate discovery that knowledge could improve both life and the speed at which it was destroyed.

The parrots built iron snakes that crossed continents, metal birds that crossed oceans, machines that harvested entire fields of seed, boxes that carried squawks around the planet, and mechanical contraptions capable of turning ten thousand parrots into a clerical error before lunch.

Eventually, one powerful flock decided another flock was responsible for every rotten fig, failed harvest, economic collapse, embarrassing military defeat, and inexplicable bald patch afflicting the nation. The accused flock had lived among them for generations, but this only proved how cunningly foreign they were.

The powerful flock confiscated their nests, marked their wings, stripped the feathers from their bodies, packed them into iron snakes, and transported them to enormous wire aviaries built for the purpose of removing them from existence. Parrot engineers improved the machinery. Parrot accountants lowered the cost per corpse. Parrot officials dipped their claws in ink and stamped documents confirming that everything was proceeding legally.

Several neighboring parrot empires watched from their gilded perches with profound concern until the extermination began interfering with shipping lanes, alliances, mineral rights, and their preferred arrangement of the branches. Then, overcome by morality at precisely the moment morality became strategically convenient, they joined the war.

The murdering flock was defeated. The wire aviaries were opened. The machines stopped. Everyone stared at the survivors and agreed that something must be done. Preferably in somebody else’s tree.

The survivors had an ancient story saying their ancestors had once nested in Urra. This was true. It was also true of several other flocks, most of whom had remained there, laying eggs, raising chicks, tending fig orchards, building villages in the canopy, and committing the apparently fatal legal oversight of existing continuously without first obtaining a property deed from a three-thousand-year-old poem scratched into bark.

The great empires examined the problem and discovered an elegant solution: give part of Urra to the survivors. It was generous in the way giving away somebody else’s nest is generous.

The empires could relieve their guilt, decline to welcome too many survivors into their own trees, and establish a loyal military roost beside several valuable migration and trade routes. Everyone benefited except the parrots already living there, who had regrettably failed to be consulted because history was running behind schedule.

The newcomers arrived wounded, terrified, and determined never again to depend upon the mercy of other flocks. They built farms, schools, laboratories, armies, walls, checkpoints, ministries, prisons, metal war-hawks, and a national mythology in which the jungle had patiently waited thousands of years for them to return, apparently watering its own fig orchards through photosynthesis and narrative convenience.

The resident parrots objected to losing their nests. This proved they hated peace. They resisted removal. This proved they were violent. They flew away. This proved they had left voluntarily. They tried to fly home. This proved they were infiltrators. They threw stones at armored metal hawks. This proved the hawks had acted in self-defense.

The new state announced that Urra was a free and equal home for every parrot, while also insisting that it must permanently belong to one particular flock. This was not considered a contradiction. It was considered constitutional ornithology.

Whenever the resident parrots mentioned the villages torn from the canopy, officials unfolded an older map. Whenever they mentioned the parrots living on that older map, officials produced an even older piece of scripture. The Ministry of Ancient Maps soon occupied the largest tree in the jungle.

More branches were seized for security. Security then required more branches. The additional branches created more resistance, which required more security, which required more branches. It was a beautifully self-sustaining machine, like perpetual motion powered by crushed eggs and flattened nests.

The empires continued supplying weapons because Urra shared their values. These values included democracy, freedom, strategic airfields, weapons contracts, and never asking what the weapons were doing so long as the seed invoices remained current.

The resident parrots were confined inside wire aviaries, counted at checkpoints, watched from elevated perches, deprived of water, and periodically bombed by metal hawks for nesting too close to whatever had just been bombed. Their dead chicks were described as future threats. Their surviving chicks were asked why they had grown so angry. And whenever anyone noticed that a flock once driven from its nests, packed behind wire, stripped of rights, marked as dangerous, and collectively punished was now using walls, cages, hunger, inherited supremacy, mechanized violence, and the language of vermin against another population, the leaders of Urra screamed that noticing the resemblance was the real atrocity.

History molts. It changes its plumage, sharpens its talons, updates the machinery, hires a Ministry of Reassuring Squawks, and teaches the descendants of yesterday’s victims that “never again” can be quietly amended to mean “never again to us.”

This was the peculiar curse of the parrots. Other creatures killed for hunger, territory, or fear. The jaguar needed no scripture before biting through a throat. The crocodile issued no historical monograph before dragging something beneath the water. But parrots could repeat. They could repeat a story until it became a fact, repeat a fear until it became a law, repeat a prayer until it became a property deed, repeat a wound until it became a weapon, and repeat a crime until the grandchildren of its victims could pronounce it “self-defense” while committing it against somebody else.

They could build beliefs around anything they wanted, build institutions around those beliefs, and build machines to crush anyone who stood between the belief and the thing they wanted. Once they had power, the beliefs supplied the permission, the fear supplied the urgency, the old suffering supplied the innocence, and the machinery supplied the bodies.

The parrots mistook having suffered for being incapable of cruelty. They mistook remembering the cage for being morally incapable of building one. But pain was never a vaccine against oppression. It was merely an education in where to place the wire.

Power did not heal the wounded flock. It gave the wound an army, an air force, a sacred origin story, and a press secretary.

So the parrots learned how to cross oceans, split atoms, resurrect dead voices from written scratches, and look billions of years backward into the birth of the universe. They learned everything except the one lesson their own talent should have made impossible to miss: A parrot can repeat anything. Especially the behavior it swore it would never repeat.

It can climb onto the ruins, plant a flag in the blood, call the screaming beneath it “necessary,” and teach its chicks that this time the cage is different because they are the ones holding the key. And that, the parrots proudly announced, was civilization.

Image credit: https://pixabay.com/photos/parrot-macaw-bird-colorful-4222909/

“And seeing the whole magnificent human shitshow spread out beneath him, the rich building temples to themselves, the poor getting billed for the privilege of existing, politicians selling commandments by the pound and jerking off into the Constitution, preachers claiming the creator of two trillion galaxies is deeply concerned about who touched whose dick, and several billion frightened apes insisting the creator of that universe had personally reviewed their fucking dietary restrictions, he climbed the mountain, found a decent rock, sat down, lit a cigarette, and said:

“Blessed are the curious, because somebody around here ought to know what the fuck is going on.

Blessed are the poor, because the rich already have the land, the banks, the lawyers, the lobbyists, the tax shelters, the private jets, the fucking islands, and still come sniffing around your paycheck like raccoons in a fucking dumpster.

Blessed are the workers, because producing everything earns you twenty minutes for lunch while owning the stupid fucking spreadsheet earns some prick a house in Aspen.

Blessed are the weird, the loud, the godless, the stubborn, and every pain in the ass who ever heard “know your place” and answered, “come down here and fucking show me.”

Ye have heard it said, obey those placed above you. I say unto you: ask the bastard why?

Demand evidence. Bullshit hates fluorescent lighting.

Ask questions. Any idea worth a damn can survive one. Anything demanding silence usually came with a uniform, a pulpit, or a fucking invoice.

Feed the hungry before building another cathedral to celebrate how much a God allegedly cares about hungry people.

House the homeless before commissioning another bronze statue of some dead bastard who owned half the town.

Heal the sick before naming another hospital wing after the asshole who helped make healthcare unaffordable.

And if a billionaire tells you hardship builds character, congratulate him on heroically avoiding character his entire fucking life.

Judge not the harmless business of thy neighbor. Let grown people love who they love, read what they read, wear what they wear, and live without your permission.

But when somebody is being beaten, starved, robbed, exploited, caged, or worked into the grave so some executive can add another comma to his net worth, suddenly make it your fucking business.

Protect the weak. Predators love polite people.

Mock the powerful. What really ruins the bastard’s evening is discovering the peasants have started laughing.

Never kneel merely because some asshole found a taller chair.

Lay not up for yourselves treasures beyond any sane human use while others sleep beneath bridges.

No man needs nine houses. That is not divine order. That is one prick with nine houses.

Consider the birds of the air. They neither file quarterly earnings nor hire lobbyists, yet somehow none has invented private equity.

Beware of false prophets in expensive suits, expensive robes, television studios, and private jets, telling you that God, country, freedom, morality, and the economy all require precisely the same thing: your fucking obedience and money.

By their fruits ye shall know them.

If the fruit is misery for everybody below them and another yacht for themselves, the tree is shit.

There are no sacred kings, no holy billionaires, no chosen bloodlines, no celestial landlords, and no infallible middlemen.

If a God wants something said, a God knows where the microphone is.

Until then, there is gravity, chemistry, evolution, sex, hunger, entropy, supernovae, cancer, children laughing in kitchens, old men dying under fluorescent lights, flowers growing through concrete, and mosquitoes, apparently because existence enjoys a cheap fucking joke.

We are apes clever enough to split atoms and photograph black holes, yet stupid enough to send grocery money to a television preacher because the immortal creator of spacetime is apparently short on cash.

And still, people love one another.

They sit beside the dying when nothing can be fixed. They pull strangers from wreckage. They plant trees whose shade they will never see. They feed animals nobody wants. They help people who can never repay them.

No halo, no harp, no fucking prize. That is miracle enough.

Therefore, whatsoever ye would have men do unto you, fucking do likewise unto them.

Feed people, house people, help people. Leave harmless people alone.

Stand between cruel men and those they intend to hurt.

Read dangerous books.

Ask fucking rude questions.

Learn some science.

Move the turtle out of the road.

Love whoever the fuck you love.

And whenever some rich son of a bitch says ordinary people must sacrifice for the economy, ask him what time he is surrendering the fucking yacht.

Whenever some preacher says a God needs money, ask why omnipotence has such terrible fucking credit.

And whenever some smug bastard says he alone knows the mind of the creator of the universe, hand him a phone.

Tell him to put the motherfucker on speaker.”

I was born in the 1990s, and I am just old enough to remember when having a computer at home was still something of a luxury and having a cellphone, especially as a kid, was virtually unheard of. Then I watched the internet explode in the early 2000s. I remember discovering YouTube and MySpace, the endless online games that seemed to appear overnight, and websites like 4chan that cultivated enormous parts of early internet culture and produced half the memes and strange little cultural artifacts we all still recognize, back before much of that world was nakedly hijacked by incels, neo-Nazis, political propagandists, and every other miserable bastard who eventually figured out what could be done with an anonymous audience of millions.

I have essentially watched computer technology, and especially the internet, evolve for my entire life. It has shaped my life in ways that people only a few generations before me could scarcely have imagined. I can communicate instantaneously with somebody on the other side of the planet. I can pull up a scientific paper published yesterday, watch archival footage from a century ago, wander around a city in another country on a digital map, learn almost anything I want to know, argue with a stranger in Australia, order some obscure piece of laboratory equipment from Japan, and have access to more accumulated human knowledge while sitting on the toilet taking a shit than the greatest libraries in history could have contained.

Yet, despite all of that wonder, I increasingly think this same technological revolution is one of the major forces behind the global social breakdown we are watching across much of the world and the accompanying rise of authoritarianism.

I do not mean the internet somehow invented authoritarianism, fascism, propaganda, political violence, greed, paranoia, or stupidity. Human beings managed all of those perfectly well before anybody figured out TCP/IP. However, what the internet did was radically alter the environment in which all of those things operate, faster than our laws, institutions, cultures, and frankly our own brains could adapt to it. It demolished old barriers to communication and information while simultaneously demolishing many of the old barriers to manipulation, surveillance, fraud, radicalization, propaganda, and mass social control.

Truth be told, If I had to describe the internet I grew up with, I would call it the “Wild West” era of the internet.

Governments figured out they could use it for surveillance. Corporations figured out that human attention, behavior, relationships, locations, purchases, interests, fears, desires, and eventually nearly every measurable aspect of a person’s existence could be collected, quantified, predicted, packaged, and sold. Political movements discovered that propaganda no longer required newspapers, television stations, or armies of pamphleteers when an algorithm could place it directly into somebody’s pocket. Terrorist organizations discovered recruitment networks. Criminal syndicates discovered entirely new markets. Intelligence agencies discovered capabilities that previous generations of spies could scarcely have dreamed of. Lone cranks discovered that they could find ten thousand other lone cranks, form a community around whatever lunacy they shared, reinforce one another’s beliefs twenty-four hours a day, and eventually convince themselves that they had discovered some forbidden truth hidden from the rest of humanity.

Almost everybody involved discovered the same basic fact: there was a fucking Grand Canyon between what this technology made possible and the laws, institutions, norms, and systems of accountability capable of governing it virtually everywhere on the planet.

It is that void that creates chaos, and chaos creates opportunity for people who know how to exploit it. It also creates fear among people who don’t. I think both help explain the global authoritarian reaction we are seeing now. Some people look at the disorder produced by enormous technological and social change and cynically see an opportunity to seize power. Others simply become frightened by a world changing faster than they can understand it and gravitate toward anyone promising order, certainty, traditional values, familiar hierarchies, and some imaginary return to the world as it existed before everything became so confusing. In that sense, the authoritarian and reactionary movements spreading around the world are not necessarily the opposite of the technological revolution. In many cases they are direct byproducts of it, reactions to it, and extraordinarily sophisticated users of it.

To be honest, the culmination of all of this increasingly reminds me of one of my favorite passages from the book Jurassic Park, when Ian Malcolm talks about scientific power as a kind of inherited wealth:

“You know what's wrong with scientific power? [...] It's a form of inherited wealth [...] Most kinds of power require a substantial sacrifice by whoever wants the power. There is an apprenticeship, a discipline lasting many years. Whatever kind of power you want. President of the company. Black belt in karate. Spiritual guru. Whatever it is you seek, you have to put in the time, the practice, the effort. You must give up a lot to get it. It has to be very important to you. And once you have attained it, it is your power. It can't be given away: it resides in you. It is literally the result of your discipline. Now, what is interesting about this process is that, by the time someone has acquired the ability to kill with his bare hands, he has also matured to the point where he won't use it unwisely. So that kind of power has a built-in control. The discipline of getting the power changes you so that you won't abuse it. But scientific power is like inherited wealth: attained without discipline. You read what others have done, and you take the next step [...] There is no discipline lasting many decades. There is no mastery: old scientists are ignored. There is no humility before nature. There is only a get-rich-quick, make-a-name-for-yourself-fast philosophy. Cheat, lie, falsify - it doesn't matter. [...] They are all trying to do the same thing: to do something big, and do it fast. And because you can stand on the shoulders of giants, you canaccomplish something quickly. You don't even know exactly what you have done, but already you have reported it, patented it, and sold it. And the buyer will have even less discipline than you. The buyer simply purchases the power”

His argument is basically that traditional forms of power often require years of discipline to acquire. The process of gaining the power changes the person gaining it. Scientific power, and by extension technological power, is different because somebody can inherit centuries of accumulated knowledge, stand on the shoulders of everyone who came before them, take one additional step, and suddenly possess capabilities they never personally developed the wisdom or discipline required to create from scratch. As Malcolm puts it, scientific power can be “attained without discipline.” The buyer of that power may understand even less than the person who created it.

That idea has aged almost disturbingly well. What we have done with computers and the internet is democratize extraordinary forms of power without democratizing anything remotely resembling the wisdom required to use them responsibly. We have handed corporations the ability to observe and manipulate billions of human beings. We have handed governments surveillance capabilities that would have made the twentieth century’s dictators fucking salivate. We have handed political propagandists machines capable of testing thousands of messages against millions of people and learning, algorithmically, which combination of fear, anger, resentment, and bullshit keeps them staring at the screen. We have handed ordinary people access to nearly the whole accumulated store of human knowledge while simultaneously building systems that can feed them an endless personalized stream of nonsense specifically optimized to keep them clicking. And we did almost all of this before seriously asking what any of it would do to a society.

That, more than anything, is what I think defines the Wild West era of computer technology. The governing philosophy was essentially that if something could be built, somebody would build it, and if somebody could make money from it, somebody absolutely fucking would. The question of whether it should be built usually arrived somewhere never.

Now the bill is coming due: Surveillance capitalism, algorithmic radicalization, industrialized disinformation, the destruction of privacy, unprecedented concentrations of corporate informational power, increasingly sophisticated state surveillance, and authoritarian political movements capable of exploiting all of it did not materialize from nowhere. They grew inside an environment we spent decades constructing without anything resembling an adequate system of accountability for the power we were creating.

Now we are discovering what happens when technological power develops much faster than the social wisdom required to govern it. And like most lessons involving power, we are learning this one after the power already exists. The humility we skipped on the way up is being collected from us on the way down.

I also think this is worth explaining from the ground up because molecular biology gets buried under terminology very quickly. You can open a textbook and within a few pages you are staring at RNA polymerase, promoters, spliceosomes, ribonucleoproteins, codons, anticodons, ribosomal subunits, 5′ ends, 3′ ends, transesterification reactions, and aminoacyl-tRNA synthetases, and it can very quickly start to feel like every sentence requires a dictionary. Truth is, the vocabulary is dense because the machinery of life is very complex and detailed, but the ideas underneath it are usually much simpler than the names make them sound. The terminology matters, of course, because precision matters in science, but every technical term should cash out into something concrete one can picture happening: a shape bumping into another shape, sticking or not sticking, all based on ordinary chemistry. If a sentence ever feels like it requires memorizing a word rather than picturing an event, I think that’s worth slowing down for, because the question “what is this thing actually doing, physically, right now” almost always has a satisfying, concrete answer in science, and I think that is something a lot of people struggle with when it comes to learning science.

So, with that said, I will not start with polymerases and spliceosomes. Instead, I will start smaller than that. Let me start with the most basic question there: What even is a cell?

A cell is just a bag. That’s it. A little membrane bubble, built largely from phospholipids, which are fat-like molecules with a water-loving end and a water-hating end that naturally arrange themselves into a two-layer membrane, and studded with proteins, enclosing an aqueous interior, meaning an interior based mostly on water, full of water and molecules, and letting the outside world stay outside. The membrane itself is lipid-rich, with lipids being a broad family of fat-like molecules that do not mix readily with water. The cell as a whole isn’t mostly fat, it’s mostly water and everything else dissolved or suspended in it. You are made of somewhere around 37 trillion of these little bags, all cooperating, all doing slightly different jobs, none of them individually smart and all knowing, yet all of them together adding up to something that can read this sentence and fucking hate Trump, as one should.

Every one of those bags has to solve the same basic problem: it needs instructions on how to survive and carry out its particular job, and, for the cells that divide, how to reproduce. These instructions are sort of like a manual one reads, but chemical and written in a molecule that tell the cell what proteins to build, which are what actually run the show inside a cell.

Proteins are the workers. They’re the ones building things, tearing things down, moving cargo, sending signals, and doing cellular chemistry. The instructions for making proteins are stored in a molecule called DNA, and in a human cell, almost all of that DNA lives inside a smaller, separate bag-within-the-bag called the cell’s nucleus, a membrane-enclosed compartment that houses most of the cell’s genetic material.

Think of the whole cell as a factory, and the nucleus as a locked office inside that factory where the company keeps the deed, the payroll accounts, the blueprints for their products, and the one copy of their contract nobody wants to lose. Everything outside that office: the assembly line, the shipping dock, and even the bathroom, is where the actual work of factory production happens. That’s called the cytoplasm. Cytoplasm is the name for everything inside the plasma membrane, the cell’s outer membrane, but outside the nucleus, including both the organelles, specialized structures inside the cell that perform particular jobs, and the watery fluid they sit in, called the cytosol. It’s the factory floor. It’s where proteins actually get built and where most of the cell’s daily business happens.

The sum total of all the DNA in one of your cells: every instruction, every gene, the whole library of genetic code, is called the genome. One genome, copied into essentially every one of your trillions of cells, and yet a nerve cell and a skin cell behave completely differently. Why? Because they’re not reading the same pages of that same enormous book. More on that in a little bit. (And for those who already know this and noticed an omission, know that I’m ignoring mitochondrial genomes for now, with mitochondria being the energy-processing organelles that carry a small amount of their own DNA, and yes, a few cell types like mature red blood cells toss the nucleus out entirely, and immune cells like B and T cells deliberately rearrange chunks of their own DNA, but those are exceptions, not the rule.)

DNA itself is a long molecule built out of smaller repeating molecular pieces called nucleotides, the individual chemical building blocks that make up DNA and RNA, and each nucleotide has three molecular parts stuck together: a phosphate group, a small phosphorus-and-oxygen-containing chemical group that helps form the molecule’s backbone, a sugar molecule called deoxyribose, and one of four molecules denoted by letters: adenine (A), thymine (T), guanine (G), and cytosine (C). String a huge number of these together and you get a big long molecular chain. Two of these chains twisted around each other into a double helix shape everybody’s seen at some point, and that’s what is meant by the term deoxyribonucleic acid (DNA).

Those two long strands of those molecular units twisted into a double helix stick together partly because the letters on one side reach across space and grab onto letters on the other side using weak little chemical bonds called hydrogen bonds, attractions involving slightly charged regions of molecules containing hydrogen. These bonds are weak individually, but there are so many of them holding on at once that the whole thing is pretty sturdy, sort of in the same way a strip of Velcro doesn’t really feel like anything if you touch one little hook, but try ripping the whole strip apart and you notice there is some real grip. Hydrogen bonding isn’t the whole story, though. There’s also a second force called base-stacking, where the flat, ring-shaped letters sit face to face like a stack of coins, and those stacking interactions actually contribute a substantial share of the duplex’s overall stability, with duplex simply meaning the paired two-stranded DNA structure. The hydrogen bonds mostly handle who pairs with whom; the stacking is a big part of why the whole structure holds together at all.

The letters mentioned above aren’t random about who they grab: A pairs with T, and G pairs with C. This is called “complementarity,” meaning that the chemical shape and bonding pattern of one base determines which other base it can properly pair with, and it’s maybe the single most important fact in all of biology, because it means if you know one strand’s sequence, you already know the other one. These are the standard complementary pairs (real biology does have rare mismatches and noncanonical pairings, meaning unusual pairings outside the standard A-T and G-C rules, but you can safely ignore that for our purposes here).

One more detail worth knowing now because it’ll matter a lot later: the two strands run in opposite directions. One runs what’s called 5’ to 3’ (read five prime to three prime), the other runs 3’ to 5’ right alongside it, like two one-way streets running parallel but in opposite traffic directions. I’ll get into why it’s numbered that way in a bit, but just file away that DNA has directionality built right into its chemistry.

Now, onto what a gene is: A gene is just a stretch of that one long DNA molecule, the way “chapter four” is a stretch of pages in a book, and not at all a separate object glued in. A gene is a specific region of DNA that, when properly read and processed, ultimately contributes to a functional product: often, by way of RNA, a protein; sometimes a functional RNA molecule itself, one that goes and does a job all on its own without ever becoming protein. This is where the neuron-versus-skin-cell idea comes back in: every cell in your body is carrying around essentially the same genome, essentially the same full library. What makes cells different from each other isn’t different books, it’s just which pages from that same book of blueprints called DNA are open and able to be read. That’s what is called gene expression, the process by which information in a gene is actually used to produce RNA or protein, and it’s controlled mostly at the very first step: deciding which genes even get copied in the first place.

Now that we both have a shared vocabulary to actually picture what is happening instead of just memorizing names, I will now get into the machinery:

The factory analogy I provided still holds. If you run a factory, you don’t want workers to maul the original official book of product blueprints, right? So instead of granting access to that original copy, you decide to photocopy the one page you need when it is actually needed, and you send the copy out to the factory floor. That copying step, in a cell, is called “transcription,” the process of making an RNA copy from a DNA template, and the copy it makes is a molecule of ribonucleic acid (RNA), and specifically, for a protein-coding gene, a working copy called “messenger RNA,” or “mRNA.”

The machine doing the actual photocopying is an enzyme called RNA polymerase. RNA Polymerase is specifically, for your protein-coding genes, a big multi-part enzyme called “RNA polymerase II,” or “Pol II” for short, with the Roman numeral II simply meaning this is the second of several major RNA polymerases in human cells. An enzyme, by the way, is nothing mystical. It’s just a molecule, usually a protein, whose whole job is speeding up a chemical reaction by grabbing the pieces involved and holding them exactly where they need to be so the reaction happens fast instead of never.

Pol II doesn’t just wander up and start copying wherever it feels like. No. Instead, a crew of helper proteins called “transcription factors,” proteins that help control whether and where transcription begins, has to assemble first at a landmark spot on the DNA called the “promoter,” a DNA region that helps tell the transcription machinery where a gene begins, sitting right before the gene starts. How does the machinery know where a gene starts? The general transcription machinery, especially a protein complex called TFIID, short for transcription factor II D, recognizing the promoter’s architecture, is what actually locates the spot. In a subset of human genes, one element of that architecture is a sequence called the TATA box, a short DNA sequence rich in thymine and adenine bases, and transcription typically begins a short distance downstream of it, but a lot of human promoters don’t contain a TATA box at all, so it’s better thought of as one kind of promoter landmark rather than a universal “start reading here” sign. Once the whole crew is properly assembled, called the “pre-initiation complex,” the collection of proteins assembled before transcription actually begins, only then does Pol II get recruited onto that stretch of DNA, set up to use one particular strand as its template and facing in the right direction. This right here is one of the cell’s favorite control knobs: by controlling which of these helper crews show up and when, the cell decides which genes are even eligible to be copied out of that book of blueprints, way before the polymerase itself lifts a finger. There are even helper sequences called enhancers, DNA regions that increase the activity of particular genes, that can sit tens of thousands of letters away and still reach over, looping through space, to help recruit the right crew.

Once Pol II is seated and going, it does something almost stupidly simple when you break it down: it locally unzips a tiny stretch of the double helix, maybe ten or fourteen letters wide, just enough to expose one strand, called the “template strand,” the DNA strand Pol II directly reads, and reads it. As it reads that template strand backward (3’ to 5’), it builds a brand new RNA strand forward (5’ to 3’), matching each DNA letter to its RNA partner. Because it’s copying the template strand by complementary pairing, the resulting RNA ends up matching the sequence of the other DNA strand, the one Pol II isn’t reading directly, called the coding strand, letter for letter, except with U standing in for T. RNA swaps out thymine (T) for a nearly identical letter called uracil (U), so A ends up pairing with U instead of T, while G still pairs with C. The DNA zips itself right back up behind the polymerase as it rolls forward on the strand of DNA like a zipper being cracked open just far enough to see the teeth, then sealed shut again the second it’s past. Nothing usually gets damaged. That same page can get photocopied a thousand more times, whenever it needs to be.

One might ask, “where do the actual raw materials come from?” Well, Pol II isn’t conjuring these RNA letters out of nowhere. Instead, it’s grabbing them from a pool of small molecules called NTPs, short for nucleoside triphosphates, which are nucleotide building blocks carrying three phosphate groups: ATP (adenosine triphosphate), GTP (guanosine triphosphate), CTP (cytidine triphosphate), and UTP (uridine triphosphate), kept available right there in the nucleus through ordinary metabolism and transport from the rest of the cell.

If you want to understand that better, picture a bucket of LEGO bricks the cell just keeps stocked at all times. Each of these bricks is carrying three phosphate groups, which also happens to be the fuel, the energy behind this process. When Pol II adds one of these NTPs to the growing RNA chain, two of its phosphates leave together as a single unit called pyrophosphate, a molecule made of two phosphate groups linked together. That pyrophosphate is then broken down by other enzymes into two separate phosphates, and that follow-on breakdown is what actually helps pull the overall reaction forward. So the incoming nucleotide brings both the new letter and the chemical energy needed to add it, the fuel built right into the brick. By the way, ATP, adenosine triphosphate, one of these four bricks, happens to be the exact same molecule your whole body uses everywhere else as its go-to energy currency. Biology’s lazy like that, in a good way: why invent a new trick when the old one already works?

I promised we’d come back to 3’ and 5’, so here it is, because once you get it, it explains a genuinely huge chunk of why everything in this essay only runs in one direction: The sugar molecule in each nucleotide of DNA is a ring built from four carbon atoms and one oxygen atom, with one more carbon sticking out off the ring itself. Chemists numbered all five of those carbon atoms 1’ through 5’, with the little apostrophe just there to say “hey, I’m talking about the sugar molecule’s carbon atoms, not the letter of the base pair.” Carbon 5’, the one hanging outside the ring, sticks out with a phosphate group attached to it. Carbon 3’, part of the ring itself, sticks out the other side with a different chemical group, a hydroxyl group, a chemical group made of one oxygen and one hydrogen atom together, OH. Carbon 1’, also part of the ring, is where the actual base, the letter (A, T, G, or C), is attached.

When you link nucleotides together into a chain, the 3′ hydroxyl group of one sugar becomes linked, through a phosphate, to the 5′ position of the next nucleotide’s sugar. Do that a few billion times and you’ve built a molecular backbone where every single link is pointing the exact same direction like a row of arrows, all tail-to-head. Which means the whole strand of DNA ends up with two different strand ends at each end of its double helix: a free 5’ end on one side, and a free 3’ end on the other.

All that 3’ and 5’ (read three prime and five prime respectively) naming convention is literally just describing which atom is hanging out exposed at each tip. And because Pol II, and basically every polymerase there is, can only add a new nucleotide onto a free 3’ end, synthesis has no choice but to run 5’ to 3’, every single time. That one small structural fact about a sugar ring is a big part of the reason DNA replication has a direction and transcription has a direction. It’s tempting to extend that same logic straight over to ribosomes reading mRNA 5’ to 3’ too, and there’s a real connection: the RNA itself carries that same built-in polarity, meaning its two ends are chemically different from one another, but the ribosome’s own directionality comes from its own architecture and translation mechanics, not simply as another automatic consequence of “polymerases need a free 3’ end.” Different machine, related but not identical reason.

Here’s something that trips people up the first time they hear it (it did me when I was first learning this stuff), and it kind of should, because it’s genuinely strange: the very first RNA copy Pol II makes, called pre-mRNA, short for precursor messenger RNA, is not actually the finished product. It’s a rough draft. It’s got the parts you actually need called “exons,” the sections that remain in the mature RNA, mixed in with long intervening stretches that get sliced out entirely called “introns,” the sections removed during processing. Cutting those intervening sequences and gluing the exons together is called “RNA splicing.” Introns aren’t simply junk, either. Some introns carry their own regulatory information and participate in how splicing itself gets controlled, but for our purposes here, the exons are what survives into the mature message and the introns are what doesn’t.

So how does the cell know which is which? How does the cell know to keep the exons and cut the introns? Well, definitely not by “understanding” the gene, because nothing in a cell understands anything like we would cognitively “understand” something. Instead, it’s a process of pattern matching. The overwhelming majority of human introns, the ones handled by the main spliceosome machinery, begin with the nucleotides GU and end with AG, with G standing for guanine, U for uracil, and A for adenine (there are rarer exceptions, including a separate minor-spliceosome class, but GU-AG covers the vast majority of cases). This is called, very unglamorously, the “GU-AG rule.” There’s also an internal landmark deep inside the intron called the “branch point,” a particular sequence centered on an adenine that participates directly in the cutting reaction, built around one very specific adenine, plus a nearby stretch called the “polypyrimidine tract,” a run rich in the RNA bases cytosine and uracil, that helps drag the cutting crew toward the far end of the RNA strand. None of this requires comprehension. It’s closer to a proofreader who’s been told “cut everything between the red sticky note and the blue sticky note,” and never once reads what’s actually written in between.

The molecular machine inside the cell that does the cutting is called the spliceosome, and honestly, it might be the single most beautifully overengineered contraption in ordinary cell biology. It’s not actually one enzyme but instead an entire committee that assembles fresh each time, does the job, and then breaks apart again until next time. It’s built out of five molecular components called snRNPs, short for small nuclear ribonucleoproteins and ironically read as “snurps,” named U1, U2, U4, U5, and U6, and each one is a little RNA molecule wrapped up in a cluster of proteins: a small nuclear ribonucleoprotein, meaning a combined package of RNA and protein found in the nucleus, not a protein on its own.

Now watch all of this come together beautifully:

1.) U1 locks onto the GU flag at the start of the intron. Its own RNA sequence is literally the chemical opposite of that boundary, so it just sticks thanks to plain old base pairing, same rules as everywhere else.

2.) U2 grabs the branch point, and in doing so it physically shoves that one special adenine molecule outward, away from the rest of the strand, like pushing a single bead out of an otherwise flat necklace laying on a table. That one bulged-out adenine is about to do the real work here.

3.) A pre-assembled trio of U4/U6/U5 rolls in and docks onto the strand. At this stage the whole machine is built, present, correct, and completely inactive. Plugged in, but not switched on yet.

4.) What follows then is the dramatic part: U1 and U4 get kicked out entirely, and U6 slides into U1’s old spot, pairing up with U2. This handoff is the actual “on” switch of this entire process.

5.) Then, catalysis, meaning the speeding-up and directing of a chemical reaction, in two clean chemical steps, both a type of reaction called “transesterification,” which just means a phosphate molecule bond gets broken and immediately re-formed somewhere else, a direct swap, no extra fuel burned at that exact moment. Step one: that bulged-out branch-point adenine, specifically the free hydroxyl group hanging off its 2’ carbon, reaches back and attacks the 5′ splice site at the GU boundary, cutting the intron loose from the upstream exon and simultaneously bonding to the start of the intron through an unusual 2’–5’ linkage, forming a “loop with a tail,” a shape called a “lariat,” because it looks exactly like a lasso. Step two: the freshly cut end of the upstream exon reaches forward, attacks the AG boundary, cuts the intron-lariat loose for good, and welds the two exons directly together, all in one motion. The leftover intron lariat gets tossed aside and recycled by the cell. The finished, spliced mRNA is released from the whole committee of proteins and RNA called the spliceosome. The whole spliceosome then falls apart, ready to reassemble fresh on the next intron down the line.

(One thing I will flag: I’m walking through transcription, splicing, capping, and tailing as if they’re neat, separate, sequential stages, because that’s the clearest way to teach them. In real cells, a lot of this overlaps in time. Capping typically happens very early while Pol II is still transcribing, and splicing often gets underway on the RNA while it’s still being made, not after the whole thing is finished. Think of it less as an assembly line with hard stops and more as several crews working the same moving train at once.)

There’s a neat, if somewhat speculative, engineering logic to why splicing runs through that middleman step instead of just welding the two exon ends together directly: holding two distant ends in perfect 3D alignment across a whole intron gap in one single move sounds like a mechanical headache, and routing through the branch-point adenine lets the first cut line up only one boundary at a time, with the resulting lariat loop conveniently hauling the far exon in close enough for the second cut to land clean. It’s a satisfying story, and the two-step chemistry itself is real and well established, but why evolution actually settled on this particular solution is a more open, harder-to-prove question than the mechanism itself, so take that explanation as an appealing hypothesis rather than a settled fact.

The actual catalytic chemistry of splicing is now understood to be centered mostly in the RNA parts of the spliceosome, particularly U2 and U6, with important contributions from the surrounding proteins. That RNA-centered core is why the spliceosome is often informally called a ribozyme, an RNA-containing molecular machine in which RNA itself participates directly in catalyzing a chemical reaction, though more precisely it’s a ribonucleoprotein enzyme with a ribozyme-like catalytic mechanism, since the proteins around that core still matter. The ribosome, which we’re about to get to, is the more straightforward case of a true ribozyme, since its own catalytic center is built directly from RNA. The relationship to a class of self-splicing RNAs called group II introns is especially interesting here, since they use strikingly similar chemistry without any protein help at all. Both the spliceosome and the ribosome are wonderfully suggestive molecular relics, the kind of thing you’d expect to find lying around if something like the RNA World hypothesis is right: the idea that very early life ran on RNA doing double duty, storing information and catalyzing reactions, before DNA and proteins split off and specialized into the separate jobs they do now. It’s not proof, but it’s a tantalizing, consistent clue scientists look at.

What is genuinely fascinating to know and a testament to why there is such a wide array of gene expression resulting from one gene, is that one gene can actually result in several different proteins. Why? Well, because which splice sites of a gene get used can be pushed around by regulatory proteins nearby during the whole process, strengthening one site, blocking a rival one, meaning the same pre-mRNA can get spliced differently depending on the cell or the situation. This is called “alternative splicing,” the process of joining the same RNA pieces together in different combinations, and it’s a big part of why the human genome, with something like twenty thousand protein-coding genes, is able to produce a considerably larger number of distinct transcripts, meaning RNA copies produced from genes, and protein isoforms, meaning slightly different versions of proteins produced from the same gene, than a naive one-gene-one-protein picture would predict. Not every one of those alternatively spliced transcripts turns into a stable, working protein, but the mechanism still buys the cell a lot of flexibility from a relatively compact set of genes. One recipe, several dishes, all just by changing which lines you keep.

Since there are, like with anything, always some finishing touches, it must be said that before the mature mRNA molecule is allowed out of the nucleus, it gets two protective touch-ups: 1.) A modified letter gets stuck on the front end, forming what’s called a 5’ cap: specifically a 7-methylguanosine, a chemically modified version of guanosine, the RNA building block containing guanine, an oddly reversed guanine-like letter, hooked onto the rest of the RNA through an unusual 5’-to-5’ triphosphate bridge, recognized by proteins involved in stability, export, and later, translation, and 2.) A long tail of repeated A’s, called the poly(A) tail, with poly(A) simply meaning many adenine-containing RNA units in a row, gets stuck on the back end by an enzyme built for exactly that job. Think of these as tamper-proof seals slapped on both ends of a takeout container, though really, it’s less that they’re passive seals and more that they’re recognized and held onto by dedicated proteins (poly(A)-binding proteins on the tail, cap-binding proteins up front) that work together with the cell’s stability, export, and translation machinery to keep the message intact and moving, so the cell’s own RNA-shredding enzymes don’t get to it on the way out the door!

Only then does the finished mRNA physically cross the nuclear envelope, the double membrane surrounding the nucleus, through actual gated tunnels called “nuclear pore complexes,” enormous protein structures that control traffic into and out of the nucleus, which are real selective checkpoints, and not just holes in a wall, usually needing dedicated escort proteins to get large molecular cargo through. This crossing is itself a filter: for an ordinary nuclear-encoded human mRNA, this trip is required before cytoplasmic ribosomes can translate it. Another clever biological checkpoint.

Once out in the cytoplasm, the mRNA molecule runs into a ribosome, a molecular machine made of RNA and proteins that builds proteins, a totally separate machine from Pol II, built out of its own mix of ribosomal RNA molecules and protein, and split into a large half and a small half that clamp down around the mRNA molecule together. Its job is “translation”: taking the RNA sequence and turning it into a chain of amino acids, which are small organic molecules that serve as the individual building blocks of proteins, and of which is what proteins are actually made of.

The ribosome doesn’t crawl along one letter at a time. Instead, it reads in fixed groups of three nucleotides, called “codons,” three-letter RNA words that specify amino acids or stop signals. Four possible letters, three slots per group, gives you exactly 64 possible codons. There are only 20 amino acids that actually get used in proteins, so most amino acids get more than one codon assigned to them, a redundancy called “degeneracy”, meaning that several different codons can specify the same amino acid. One consequence of that redundancy is that some single-base substitutions, especially at the third codon position, don’t change the encoded amino acid at all, and a small copying mistake in the third slot of a codon often still lands on a codon for the same amino acid, so the protein comes out right anyway. Clever, again!

Reading always kicks off at a specific start codon, almost always “AUG,” with AUG standing for adenine-uracil-guanine, found by the small ribosomal subunit scanning from the cap until it lands on an AUG sitting in a favorable surrounding sequence context (called a Kozak sequence, a short pattern of nearby RNA bases that helps the ribosome recognize the correct start site), which also happens to be the code for the amino acid “methionine,” and wherever that “AUG” sits sets what’s called the “reading frame”: the fixed grouping into threes that the ribosome sticks to from then on. This matters more than it sounds like it should, because the exact same string of letters means something completely different depending on where you start counting by threes. Shift the starting point over by even one single letter and every codon after it gets misread, resulting in a whole mess called a “frameshift.”

Reading stops the moment the ribosome hits one of three stop codons: “UAA,” “UAG,” or “UGA,” none of which mean any amino acid at all. They’re just the word “done.”

The piece that actually makes any of this work is called tRNA, short for transfer RNA. The ribosome, on its own, cannot look at a codon and just know which amino acid it’s supposed to mean. Codons are made of nucleotides, three of them in a row. Amino acids are a completely different family of molecule. You need a translator sitting in between, physically bridging the two, and that translator is called “transfer RNA,” or simply “tRNA.”

Every tRNA folds itself into a compact, specific little shape which scientists usually describe as a cloverleaf-like shape in two dimensions, though the real, folded 3D molecule is closer to an L, with two ends that actually matter for this process. One end carries a three-letter sequence called an “anticodon,” the three RNA bases that pair with a matching codon on the mRNA, sort of like the same lock-and-key A-U/G-C rules we’ve been using throughout this whole essay. The other end is carrying a specific amino acid that is physically attached.

Getting the right amino acid stuck onto the right tRNA molecule is the job of a family of enzymes called aminoacyl-tRNA synthetases, enzymes that attach the correct amino acid to the correct tRNA, with one dedicated enzyme per amino acid. Each one recognizes both its amino acid and its matching tRNAs, spends an ATP molecule (broken down in the process into AMP, adenosine monophosphate, plus a pyrophosphate, effectively burning both of its high-energy phosphate bonds) gluing them together into what’s called a “charged tRNA,” meaning a tRNA with its amino acid attached and ready for use, and this part’s wild…a lot of these enzymes have their own built-in proofreading step that double-checks the job and kicks out mistakes before they ever get released. This step is carrying an enormous amount of weight, because the ribosome later has basically no way of independently checking whether the amino acid hanging off a tRNA actually matches that tRNA’s anticodon. It just trusts the codon-anticodon match and assumes the synthetase did its homework right.

The ribosome has three important parking spots where tRNAs sit as they move through: the “A site,” with A standing for aminoacyl, where a fresh charged tRNA rolls in; the “P site,” with P standing for peptidyl, holding the tRNA attached to the chain that’s already being built; and the “E site,” with E standing for exit, where a used-up tRNA passes through on its way out.

The mechanical loop: a charged tRNA whose anticodon fits the exposed codon docks in the A site. The ribosome forms a peptide bond, the covalent chemical bond that joins amino acids together into a protein chain, between that amino acid and the chain sitting in the P site. Then the whole ribosome shifts forward exactly one codon, a process called “translocation,” meaning the ribosome and tRNAs move forward by one codon along the mRNA, sliding tRNAs from site A to site P to site E, kicking the spent one out the E site door, and exposing a fresh codon in the A site for the next arrival. Repeat that, codon after codon, and a chain of amino acids grows, always in the same direction, from one specific chemical end toward the other, and just like nucleic acid synthesis, protein synthesis has a direction baked right into it too.

When a stop codon finally rolls into the A site, no tRNA fits it and nothing plugs in, which is exactly why it’s a “stop signal.” Instead, proteins called “release factors,” proteins that recognize stop codons and trigger the release of the completed protein chain, recognize the stop codon directly and trigger the release of the finished chain of amino acids from the last tRNA molecule. The ribosome is then taken apart by additional recycling factors, freeing its pieces to get recruited onto a new mRNA and start the process all over again.

The ribosome’s actual catalytic center, the exact spot where peptide bonds get formed, is formed predominantly from ribosomal RNA, or rRNA, the RNA molecules that form the structural and catalytic core of ribosomes, not protein. Just like the spliceosome, the ribosome is a ribozyme. Two of the most important machines running in your body, right now, doing the actual chemistry with RNA and using protein mostly as support staff: a loud echo of a distant time in the past when RNA might have been running the whole operation on its own.

The chain of amino acids that comes off the ribosome, a “polypeptide,” meaning a long chain of amino acids joined by peptide bonds, is not yet necessarily a finished, functional protein. It’s at this point just a string of amino acids, linked end to end like beads on a necklace. What actually makes it useful is that the exact sequence of those beads determines how the chain folds itself into a specific 3D shape.

That folding happens because every amino acid in that chain carries a different side chain, the chemically variable part of an amino acid that gives it its particular properties, attached to it with its own “personality,” so to speak. Some are “hydrophobic,” meaning they interact poorly with water and tend to become buried away from it, meaning they hate water and tuck themselves into the interior of the folded shape to hide from it (that’s shorthand, what’s really happening is that burying those nonpolar side chains away from water lets the surrounding water molecules settle into more favorable arrangements with each other, which is where the real driving force comes from). Some are “polar,” meaning their electrical charge is distributed unevenly enough that they interact readily with water, and are perfectly happy sitting out on the surface, interacting with water. Some carry a charge and either attract or repel their neighbors. A few, specifically cysteine, contain sulfur atoms and can form real, strong “covalent bonds,” chemical bonds in which atoms share electrons, called “disulfide bonds,” bonds formed between two sulfur atoms that can staple more distant parts of the chain together. Biologists describe the folding in layers: the raw sequence of amino acids itself is the primary structure; small local coils and pleats (alpha helices, spiral-shaped sections of protein, and beta sheets, folded sheet-like sections) are the “secondary structure”; and the full 3D shape of one folded chain is the “tertiary structure”; and when multiple folded chains team up into one working complex, that’s “quaternary structure.”

Some proteins fold up correctly all on their own, guided purely by these chemical forces. Others need a hand from molecular chaperones, proteins whose whole job is helping other proteins fold properly and preventing them from clumping badly with one another while it figures out its shape. A lot of proteins get customized further after this, too, through post-translational modifications, chemical changes made to a protein after the ribosome has built it, small chemical add-ons like phosphates (phosphorylation, the attachment of a phosphate group) or sugar chains (glycosylation, the attachment of carbohydrate chains), or portions of the chain simply getting trimmed off.

This right here is the deepest idea in my whole essay, so let it sit for a second: DNA and RNA, at the level of the information they encode, are pure, flat, one-dimensional information. A sequence of letters. Nothing more, at least in that informational sense (the molecules themselves are of course real, three-dimensional chemical objects too , but the thing they’re storing is a linear sequence). Protein is shape and chemistry, living in 3 dimensions, an actual object that can physically go do something in the world. Every single piece of machinery we just walked through exists for one reason: to make that conversion from linear information into functional shape, reliably, thousands of times a second, in every one of your trillions of cells.

Not every protein just stays put in the cytoplasm of the cell where it got built. Proteins headed for the cell surface, or out of the cell entirely, or into the internal membrane system, usually start off with a short leading stretch called a “signal peptide,” a short amino-acid sequence that acts like a cellular shipping label telling the cell where the protein needs to go. The instant that stretch pokes out of the ribosome, it gets grabbed by a complex called the “signal recognition particle,” or SRP, a combined RNA-and-protein complex that recognizes these signal peptides, which physically hauls the whole ribosome-and-growing-chain setup over to the surface of the rough “endoplasmic reticulum (ER)” organelle in the cell, a folded-up membrane system covered in docked ribosomes, and docks it onto a channel sitting in that membrane called the translocon, the actual protein-conducting tunnel the growing chain gets threaded through as it’s built. From there, the protein gets threaded into or across that membrane as it’s being built, moves on through the Golgi apparatus organelle, a stack of membrane compartments that modifies, sorts, and ships proteins, for more editing and sorting, and then finally ships out to wherever it’s actually needed. This geography of the cell matters a lot for the vaccine section coming up, because where a protein gets built and processed changes how your immune system ends up seeing it.

DNA (the vault) -> Pol II copies out a rough RNA draft -> the spliceosome cuts out the introns and welds the exons together -> capping and tailing seal both ends -> the finished mRNA slips out through the nuclear pores -> a ribosome reads it three letters at a time, fed by charged tRNAs -> a chain of amino acids comes off the line -> it folds, gets tweaked, and often routes through the ER and Golgi -> a working, finished protein. That’s the entire pipeline so far.

So…How do mRNA vaccines borrow this exact biological pipeline to protect us from viruses?

Once you actually get everything above, an mRNA vaccine stops looking like some brand-new, spooky technology and starts looking like a very clever, very narrow hack of machinery your cells were already running nonstop, whether you asked them to or not.

During natural infection, your immune system learns to recognize a virus by encountering its actual antigens, molecules or molecular pieces that the immune system can recognize as targets, in the context of a real infection: the whole infectious viral particle, risk of death and all. (Vaccination itself is an old idea, of course, going back centuries and taking many forms: live, attenuated (meaning a weakened living pathogen); inactivated, meaning a killed or disabled pathogen; protein-subunit, meaning only selected pieces of a pathogen; viral-vector, meaning a harmless carrier virus used to deliver genetic instructions—mRNA is just the newest way of getting the immune system that same kind of look.)

For a lot of viruses, there’s one particular protein sitting out on the surface that’s the single most useful thing for your immune system to learn to spot, because that’s the exact part the virus uses to grab onto your cells and break in. For SARS-CoV-2, short for severe acute respiratory syndrome coronavirus 2, that’s the “spike protein”: a glycoprotein, meaning a protein wearing a bunch of sugar molecule chains on its surface, and it works by locking onto a receptor called “ACE2,” short for angiotensin-converting enzyme 2, a protein on the surface of certain human cells that the virus uses as an attachment point, found on the surface of certain human cells, through a specific region of itself called the receptor-binding domain, the part of spike that directly grips ACE2. That initial grab is followed by further steps involving cleavage by host enzymes and membrane fusion before the virus actually gets in, but the ACE2 handshake is the crucial first move.

An mRNA vaccine’s whole strategy boils down to this one sentence: skip the dangerous part. Don’t hand the body any actual virus. Instead, hand a small number of cells a synthetic note built following the exact same chemical rules from that pipeline I just described earlier in my essay, telling them to build only an engineered version of the spike protein of the virus, in total isolation, with nothing else on board capable of causing an infection or replicating inside you. No whole virus, no serious illness, no death, just a reproduced viral ID part.

The synthetic mRNA molecules in these vaccines are engineered with the same basic architecture as your own ordinary mRNA, because that’s exactly what lets your own biological machinery inside your cells recognize it and use it. Reading it end to end: a 5’ cap, chemically similar in purpose to the natural one we already covered, helping the translation machinery recognize the molecule and protecting it from getting chewed up too fast; untranslated regions, or UTRs, stretches of RNA on both sides of the coding sequence that are not translated into protein but help control stability and translation, on both sides of the coding sequence, which don’t get turned into protein themselves but shape how stable the message is and how efficiently it gets read; the open reading frame, or ORF, the continuous stretch of codons that actually specifies the protein, the actual stretch of codons spelling out the spike protein, set up to be read starting at a proper AUG; and a poly(A) tail at the very end, doing the same stabilizing job it does for your own messages.

Scientists also make some deliberate calls here that go beyond just copying nature:

1.) “Sequence optimization,” since the genetic code is degenerate (remember most amino acids have more than one valid codon), scientists design the sequence with an eye toward codon usage, GC content, meaning the proportion of guanine (G) and cytosine (C) bases in the sequence, RNA secondary structure, meaning the way an RNA strand can fold back and pair with itself, stability, avoiding unwanted sequence motifs, meaning short patterns of bases that may cause undesirable cellular or manufacturing effects, translation efficiency, and manufacturing constraints, all without changing the actual protein that gets made.

2.) “Modified nucleosides,” with nucleosides being a base attached to a sugar but without the phosphate groups of a full nucleotide, the COVID-19 mRNA vaccines swap out ordinary “uridine” in the mRNA molecule for a chemically tweaked version called N1-methylpseudouridine, which is not a single letter changed somewhere in the sequence, but a substitution made throughout the entire molecule everywhere a uridine would otherwise appear. That substitution reduces how strongly certain innate immune RNA-sensing pathways, the body’s built-in rapid systems for detecting suspicious or foreign RNA, recognize the molecule as foreign, since your body’s usually wired to treat unfamiliar RNA as a red flag for infection and shut translation down. It’s not that the modified RNA becomes invisible to your immune system, on the contrary, the vaccine (and the lipid particle carrying it) still triggers a real, useful innate immune response: the modification mainly helps the message get read and translated efficiently instead of being shut down outright.

3.) “Structure-locking mutations,” changes deliberately introduced into the sequence to make the resulting protein hold a particular shape more reliably, the spike sequence itself is often engineered with specific tweaks that lock it into what’s called its prefusion shape, the particular 3D form the spike only briefly holds right before it fuses onto a human cell, because that shape happens to show off the most useful targets for building strong, protective antibodies. In the original SARS-CoV-2 mRNA vaccines, this was done with a specific pair of substitutions nicknamed “2P,” meaning two proline substitutions, (swapping in two prolines at positions 986 and 987), though that’s one particular solution rather than the only way to lock down a prefusion shape. Other vaccine designs use different stabilizing tricks as well.

RNA is fragile stuff. Your body is loaded with enzymes called “RNases,” short for ribonucleases, enzymes whose job is cutting and breaking down RNA on sight. RNA molecules are also big and heavily negatively charged, so it’s not exactly gliding through a cell’s greasy, water-hating membrane on its own. So the vaccine mRNA gets packaged up inside a “lipid nanoparticle,” or “LNP”: a tiny bubble built out of a few different fat-like molecules including a class called “ionizable lipids,” whose electrical charge flips depending on how acidic their surroundings are.

Those little particles are what actually carry the RNA in the vaccine injection. Cells take it in through a special process called “endocytosis,” the process by which a cell wraps its membrane around material outside itself and pulls that material inward: the outer membrane of the cell folds in around the particle and pinches off, trapping the whole thing inside an internal cellular bubble called an “endosome,” a small membrane-bound compartment formed when the cell takes material in from outside. As that endosome naturally becomes more acidic on the inside, the ionizable lipids respond to that shift and help destabilize the surrounding bubble membrane, letting some of the mRNA molecules slip out of the endosome and into the open cytoplasm of the cell, a step called “endosomal escape,” the release of material from an endosome into the cytoplasm, and it’s not optional, because a ribosome sitting out in the cytoplasm can’t reach RNA that’s still trapped inside a sealed bubble.

Once it’s loose in the cytoplasm, it’s just ordinary translation. It functions right there in the cytoplasm and has no need to enter the nucleus or interact with your genomic DNA. It walks straight into the exact same translation pipeline we already covered. Initiation factors, proteins that help a ribosome begin translation, recognize the 5’ cap. The small ribosomal subunit scans along until it finds the AUG in a proper spot. The large subunit joins in. Charged tRNAs roll in one codon at a time, matching anticodon to codon. Peptide bonds snap into place. The chain grows. Since the spike sequence carries its own signal peptide, this particular protein pulls in the rough ER and Golgi pathway too, the growing spike threads right into the ER membrane, folds up, picks up its sugar molecule decorations, and moves along exactly like any ordinary membrane-bound protein your cells make on any given day.

Your cell isn’t doing anything unusual. It’s just following the note it was given, using tools it already had running, briefly turned into a tiny, temporary spike-protein workshop.

Contrary to what conspiracy theorists with zero actual knowledge in biology love to say, this mRNA vaccine has no way to change your DNA.

Why? Well, for any RNA to get permanently written into your chromosomes, the long DNA molecules in which your genome is physically organized, you’d need several extra, very specific biochemical events, none of which are part of this pathway: an enzyme called reverse transcriptase, an enzyme that copies RNA information backward into DNA, would have to turn the RNA back into DNA, and that DNA would then have to somehow find its way into the nucleus and get inserted into a chromosome. Those are specialized tricks belonging mainly to certain viruses called retroviruses, viruses whose normal life cycle includes copying their RNA into DNA, and to certain mobile genetic elements your genome already carries remnants of, like LINE-1 elements, with LINE-1 standing for long interspersed nuclear element-1, a type of DNA sequence capable under some circumstances of copying and inserting itself elsewhere in the genome. The ordinary cytoplasmic translation route the vaccine mRNA takes doesn’t touch any of that machinery, and the vaccine mRNA itself carries none of the tools you’d need to even attempt it. The honest, careful claim isn’t that integration is some kind of law-of-physics impossibility, it’s that this vaccine pathway has no integration mechanism, integration plays no role in how it works, and there’s no evidence it causes genomic integration in vaccinated people.

One might reasonably ask: “how does your immune system actually get to see the thing?”

Well, making the protein is only step one. The immune system needs to actually be shown it, and it gets shown two different ways:

1.) Showing it to the killers (MHC class I, with MHC standing for major histocompatibility complex, a family of molecules that display small protein fragments on cell surfaces for immune cells to inspect). Almost every nucleated cell in your body is constantly shredding a sample of its own internal proteins using a molecular wood-chipper called the “proteasome,” a large protein machine whose job is chopping unwanted or sampled proteins into small pieces called peptides. Some of the resulting fragments, including bits of the freshly made spike protein, get hauled into the ER organelle by transport proteins called “TAP transporters,” with TAP standing for transporter associated with antigen processing, which are then loaded onto display molecules called “MHC class I,” and carried out to the cell surface for inspection. It’s basically every nucleated cell in your body constantly wearing a name tag made out of its own insides. Roaming “CD8+ cytotoxic T cells,” with CD8 standing for cluster of differentiation 8, a surface protein that helps this class of T cell recognize MHC class I, the so-called “killers,” check these tags as they pass by. Once a CD8+ T cell has been properly primed, it can recognize a target cell displaying its matching spike-derived fragment on MHC class I. Getting a naive CD8+ T cell primed in the first place, though, generally isn’t just a matter of bumping into a random body cell wearing the right tag. That job mostly belongs to specialized antigen-presenting cells, cells whose job includes capturing antigens and displaying their fragments to T cells, dendritic cells especially, which can also grab spike protein or fragments from outside the cell and, through a process called cross-presentation, the unusual ability to take material from outside the cell and display it on MHC class I to properly prime naive killer T cells. Once those T cells are activated, then recognizing peptide-MHC I, a small protein fragment bound to an MHC class I molecule on an ordinary target cell, is what triggers the actual killing.

2.) Showing it to the coordinators (MHC class II, the second major class of these antigen-display molecules, used mainly by specialized antigen-presenting cells). Separately, those same antigen-presenting cells, dendritic cells especially, scoop up material floating around in the body, protein released from vaccinated cells or fragments from cells that took the LNP up directly, and break it down inside internal compartments, and display those fragments on a different molecule, “MHC class II,” for “CD4+ helper T cells,” with CD4 standing for cluster of differentiation 4, a surface protein that helps helper T cells recognize MHC class II. Dendritic cells are especially good at this whole job. Once they’ve grabbed an antigen, they physically travel to the lymph nodes, small immune-system organs where large numbers of immune cells gather and communicate. They are basically the meeting halls where cells carrying field intel get to go find and brief huge populations of other immune cells waiting there on what they found and what needs to be done to protect the body.

Once activated, “helper T cells” hand out the signals that let another arm of the immune system, “B cells,” immune cells whose descendants can produce antibodies, respond fully. Each B cell carries its own particular receptor specificity, and if that receptor happens to bind the spike protein, and the B cell also gets help from an activated helper T cell, it then kicks off clonal expansion, the rapid multiplication of one selected immune cell into many genetically matching descendants, some of which mature into plasma cells, which are specialized B-cell descendants that are dedicated antibody factories pumping out a soluble copy of that same binding receptor. That soluble copy is an antibody.

Some of those antibodies bind spike in ways that interfere directly with its ability to infect a cell, including antibodies that block the very region it needs to engage ACE2. These are called neutralizing antibodies because instead of merely tagging the virus for something else in the immune system to deal with later, they can physically interfere with the machinery the virus needs to get into a cell in the first place.

The very first antibodies your body makes aren’t necessarily the best ones it’ll ever make. Inside structures in your lymph nodes called “germinal centers,” temporary microscopic regions where activated B cells multiply and improve their antibodies, activated B cells go through something genuinely wild called “somatic hypermutation,” a tightly controlled process that deliberately introduces mutations at a very high rate into the DNA encoding the antibody’s variable region, the part of an antibody whose exact shape determines what it can bind. That churns out a population of slightly different antibody variants, which then compete against each other for antigen and for T follicular helper-cell signals, signals from a specialized type of helper T cell that supports and selects B cells inside germinal centers, not simply raw binding affinity alone. The cells whose antibodies bind tighter and secure those survival signals keep multiplying, while the weaker binders drop out. Run that over a few rounds, and the whole population drifts toward tighter and tighter binding. It’s evolution, on fast-forward, happening for real, physically, inside your own lymph nodes: variation shows up, selection leans on it, and, as a useful simplification, the best binders win.

Some of these activated B cells stick around as “memory B cells,” long-lived B cells that remember a particular antigen and can respond rapidly if it appears again, capable of hanging on for years to decades, depending on the response, and firing back quickly if the same antigen ever shows up again. Some become long-lived plasma cells, quietly making antibody in the background for years, and in some cases potentially decades. Activated T cells leave their own memory populations behind too. Immune memory isn’t some crazy abstract idea, it’s a literal, physical population of specific cells sitting inside of you right now, primed, with receptors already shaped to recognize spike protein. If the real virus turns up later on, that population can then expand way faster than a naive immune system meeting the threat cold for the first time. The slow, careful work of learning the enemy’s shape already got done in advance, without making the immune system pay the biological cost of learning through the uncontrolled infection itself.

Vaccine mRNA is disposable, by design, exactly like every single ordinary mRNA in your body already is. Among the major routes cells use to break it down: the poly(A) tail gradually shortens, a process called “deadenylation,” meaning the progressive removal of the adenine-rich tail. The cap eventually gets stripped off, a process called “decapping.” Once those protections are gone, enzymes called “exonucleases,” enzymes that chew nucleic-acid molecules apart starting from an exposed end, start chewing the exposed RNA molecule apart from both ends. What’s left gets recycled straight back into the cell’s general pool of raw material, the exact same fate every mRNA you’ll ever make meets eventually. The vaccine only needs its message to survive long enough to make sufficient antigen to properly train the immune system. After that, there’s no reason for it to stick around, and biologically, it doesn’t.

The real reason this technology works is in a way very anticlimactic, honestly: it doesn’t invent a single new trick inside your cells. Your cells already know how to translate RNA. They already know how to fold, process, and display protein fragments. Your immune system already knows how to pick the right B cells, build germinal centers, run affinity maturation (the process by which B-cell populations evolve toward producing antibodies that bind their target more strongly), and hold onto memory populations for years. The entire engineering challenge sat upstream of all of that biology: designing a well-behaved RNA sequence, wrapping it in something that could actually deliver it into a cell intact, and producing enough antigen for long enough to properly train a system that was already fully capable of learning on its own.

Once the RNA hits the cytoplasm, everything that happens after that is just your own cell biology, doing what it’s always been doing since you were born, just handed one very specific, very deliberately written note. One that saves lives.

Image Credit: Juan Gaertner/Shutterstock.com

Set aside the judges, prosecutors, and other people within the system who actually agree with the regime or are actively helping the motherfuckers. I’m talking mostly about judges who plainly understand that the arguments being put before them by the DOJ, Trump, and his administration are very often complete fucking horseshit: bad-faith claims, ludicrous legal theories, outright fucking dishonesty, and deliberate assaults on the spirit of the rule of law. They are watching the government weaponize the law while simultaneously wiping its fucking ass with it and making a goddamn mockery of the entire concept.

Yet these judges are still expected to behave as though they are dealing with ordinary litigants operating in good faith. They have to remain impartial. They have to give both sides a fair fucking hearing. They have to maintain the institutional legitimacy of the courts even while one of the parties standing before them is deliberately testing how much lying, contempt, obstruction, and institutional vandalism—how much fucking shit they can smear across the legal code—the system will tolerate before something finally fucking snaps and the judges stop pretending the people standing in front of them are remotely honest actors.

The trap is that the judiciary cannot simply abandon its own commitment to the rule of law in response. The moment it does, it hands an already emboldened authoritarian movement exactly what the fuck it wants: another excuse to declare the courts illegitimate, another justification for ignoring them, another fucking weapon for dismantling whatever remains of the legal order altogether.

That’s the fucking game we are stuck with right now: the federal government keeps pushing, lying, ignoring norms, advancing increasingly absurd bullshit arguments, testing every fucking boundary it can find, daring judges to escalate, daring the courts to finally snap and do something that can then be held up as proof that they were the lawless motherfuckers all along.

So we get this fucking farce where one side is deliberately exploiting the rules precisely because it knows the other side still believes it has a fucking obligation to follow them. And that, my friends, is a huge part of why this country is in the absolute shit position it is in. We still have far too many people operating under the assumption that the answer is simply to keep the institutions intact, grit our fucking teeth, survive this particular batch of fascist fucks, and wait for everything to magically return to normal, as though an entire political movement hasn’t spent years demonstrating that it intends to exploit every fucking weakness in those institutions again and again and again until it gets what it wants, whatever the fucking cost.

At some point the American people—including the people inside the system who haven’t betrayed the Constitution, their fucking oath, and the American people—have to stop pretending that endurance is a fucking solution and admit that something drastically, fundamentally, and permanently needs to fucking change.

“We didn’t order this shit,” the technician said, holding a clipboard, visibly disgusted. “There’s no brain matter in here. Just a sack of rotten dog shit, some gravel, and a laminated printout that just says ‘DO YOUR OWN RESEARCH.’ I don’t know what department this is for, but it sure isn’t for us.”

“We ordered brains for a first-year anatomy course,” she said. “What we received does not meet that definition under any reasonable scientific standard. We’ve had compost heaps with more neural activity.”

The supplier, reached for comment, insisted the shipment was accurate and accused the lab of being “too woke to appreciate real brains” before demanding the crate back “for evidence” and hanging up.

The school says it has filed a formal complaint and will be sourcing future shipments elsewhere. The crate remains unclaimed in a hallway, where staff say it has started to smell like a Facebook comment section on a local news outlet article about Democrats.

I’ve always found that premise pathetic, because the thing they’re describing as a nightmare is, to me, the entire fucking point. Getting lost isn’t lost. It’s the one reliable exit hatch out of the monotonous sedentary life we call “normal,” where every day is mapped out in advance and apparently the great human aspiration is to know exactly where you’re going, what you’re doing, and when you’ll be back at all times. What actually terrifies me isn’t some unfamiliar side street in some strange and foreign land, or realizing I have no fucking idea where I am; it’s actually a desk at 8 a.m., on schedule, forever and ever, the end. Getting gloriously turned around somewhere while off the tourist track (I always try to avoid the tourist track) with no itinerary and no fucking clue where I’m going? Sign me the fuck up, yo. That’s a day I’ll actually enjoy and remember.

Being lost is also the best alibi a human being can manufacture. Nobody can hold you to an obligation you genuinely couldn’t find your way back in time for. There’s no excuse to invent and no bullshit story to sell because you literally got fucking lost. Sorry I missed it. Where was I? Oh, right: Somewhere. Where exactly? If I knew that, I wouldn’t have been fucking lost.

As always, the supposed danger is usually laughable. You take the wrong street or miss the right train or get on the wrong one. Walk three miles in the wrong direction and end up in some neighborhood you didn’t know existed, eat somewhere you never would’ve found, drink with somebody you never would’ve met, and eventually figure out where the fuck you are. This is apparently terrifying now.

We humans have become so pathologically accustomed to having every inch of existence mapped, reviewed, ranked, scheduled and fed to us through a little glowing rectangle that simply not knowing what happens next feels like an emergency. We launched satellites into fucking orbit so nobody ever has to experience the horror of wondering whether they should turn left or right. God forbid you pick one and find out.

The people most terrified of getting lost will spend decades waking up at the same time, taking the same road to the same building, sitting in the same chair, staring at the same screen, taking lunch at the same hour, driving the same road home and doing the whole fucking thing again tomorrow until they drop dead without ever seeing the world or experiencing life. Sounds fucking awful.

Photo below: Me in Cyprus a few years back, holding some random adorable cat we met while wandering around with absolutely no fucking idea where we were. Eventually found our way back to the hotel. Harrowing stuff. Somehow survived. Would recommend.

Rural conservatives obviously hate the worldview on its own terms: open-minded, fact-rooted, empathetic, suspicious of authority, and allergic to the rancid little cocktail of hypermasculinity, nationalism, racism, and sexism they’ve spent decades mainlining like communion wine. That’s hardly interesting by itself. What actually seems to short-circuit something in their little walnut brains is finding out I’m not some tattooed, septum-pierced, blue-haired liberal woman beamed in from Madison or Chicago, or an ivory-tower professor lecturing them from a gender studies department, or a “coastal elite,” or really any of the stock characters they’ve invented to explain why people disagree with them without ever having to seriously consider the possibility that somebody might know their world perfectly well and simply think a lot of it is fucking stupid. Instead they get the most inconvenient person imaginable: a garden-variety rural Wisconsin white guy of German-Polish-Dutch stock whose family has been kicking around this dirt since the 1850s, who grew up in the same small towns, went to the same schools, got frog-marched into the same churches on Sunday mornings, heard the same jokes, absorbed the same assumptions, and was handed basically the same cultural starter pack they were.

That’s usually when the frantic search for some disqualifying difference begins, because there has to be one. “You never left.” Well, wrong, I lived abroad. “You don’t understand real work.” Wrong again. Whatever the particular accusation is hardly matters because they’re all doing the same job: desperately trying to manufacture enough distance between me and them that they can shove me back into the outsider category and stop thinking about it. The problem is that I checked damn near every box in the little cultural flowchart they think should have automatically produced another version of themselves. Same rural peasant stock, same small-town upbringing, same church, same schools, same skin color, same gender identity, same sexual identity, same inherited culture, same fucking dirt under the family tree, right down to the bone, and somehow I still came out the other side thinking half the shit they treat as eternal common sense is fucking ridiculous and the other half deserves considerably more scrutiny than anybody around here was encouraged to give it.

I think that’s what actually gets under their skin, because a stranger disagreeing with you is easy. You can explain away a stranger. They’re from Madison, they’re from Chicago, they’re some liberal professor, they’re brainwashed, they’re an elitist, they don’t understand “our way of life,” whatever the fuck you need to tell yourself before bed. What’s much harder to explain away is somebody built out of basically the same raw material who looked at the same inheritance and reached completely different conclusions, because then you have to entertain the deeply annoying possibility that your worldview wasn’t simply deposited into your skull by reality itself. Maybe some of it was taught to you. Maybe some of it survived because nobody around you questioned it. Maybe some of it is just dumb shit people have been repeating to each other for generations until repetition started masquerading as truth even though it’s the furthest thing from.

Which is what makes the whole thing so fucking funny to me, because these are the same people who have spent my entire life congratulating themselves for being rugged individualists and calling everyone else sheep, yet nothing seems to piss them off quite like somebody from their own pasture climbing over the fence, looking back, and asking why the fuck everybody else is still standing there.

That conflict takes its recognizably modern form under British Mandate rule, amid rapidly increasing Zionist colonization in the 1920s and 30s. This is when the Palestinian Arab population watched its land get transferred out from under it, land acquisition that was overwhelmingly legally purchased under British-imposed Mandate law, but that eventually dispossessed Arab tenant farmers along the way and unfolded within a forced-upon foreign political framework Palestinians overwhelmingly rejected, backed by colonial policy and eventually by force, with no ballot box, no court, and no international body that would hear their case.

Palestinians did try the peaceful route first. In 1919, President Wilson sent the King-Crane Commission to survey what people across the former Ottoman Arab territories actually wanted. Opposition to the Zionist program was among the most frequently expressed positions in the 1,863 petitions the commission collected, and the commission itself concluded that carrying it out would require force, warning that Zionist representatives openly voiced and anticipated “practically the complete dispossession of the present non-Jewish inhabitants of Palestine.” Wilson’s own commission recommended scaling the whole project back. Britain and France buried the report and did the opposite. That’s the peaceful political solution getting tried and getting ignored by the same Western powers who now lecture Palestinians about using “proper channels” for their grievances.

The first sustained wave of Arab political violence, the 1920s riots and the 1936-39 revolt, tracks exactly with the escalating loss of political and economic control by the Palestinian Arab population, and not some timeless regional pathology. As I have said before: this is a recurring and well-documented pattern in occupations and struggles over national self-determination throughout modern history.

Political scientist Robert Pape ran the numbers on suicide terrorism specifically, and found it clustered heavily around campaigns aimed at compelling democratic states to withdraw military forces from territory the attackers considered their homeland. His argument wasn’t that religion never matters, or that occupation explains every terrorist attack—it was that the dominant strategic pattern in his data was political and territorial, not simply religious fanaticism. That’s not nothing. It means the through-line for this behavior isn’t “ancient hatred,” it’s territory and control.

The Irish Republican Army (IRA), the African National Congress (ANC) armed wing (yes, Nelson Mandela’s organization, the one that now has streets and airports named after him, ran a sabotage and armed resistance campaign for 30 years), the Viet Minh, and now Palestinian factions aren’t identical movements—their ideologies, targets, and political structures differ considerably. But they illustrate the same recurring historical phenomenon: armed movements emerging out of conflicts over imposed foreign rule, occupation, or national self-determination, not out of some innate cultural bloodthirst that spontaneously develops in a vacuum.

None of that makes killing civilians defensible. I will say it again for the people scrolling fast or hoping to accuse me of such: NONE OF THAT MAKES KILLING CIVILIANS DEFENSIBLE, Palestinian OR Israeli. But there’s a rhetorical move, a sleight of hand that I see constantly, and it’s time it be confronted: someone shows up with a tally of “Muslim” or “Arab” attacks, hundreds of them, sometimes thousands, depending on how loosely “attack” gets defined, and drops it into an argument about what a growing chorus of genocide scholars, human rights organizations, and UN experts now call an active genocide in Gaza, like the count alone somehow proves a point or wins the argument. It doesn’t. A number with no denominator, no timeline, and no antecedent isn’t an argument, it’s a scoreboard. It only works on people who’ve already agreed to look at this as a body-count contest instead of a tragic history.

The second you ask “attacks compared to what, over what period, starting when,” the whole framing falls apart. It’s the same move as someone inappropriately pointing to Little Bighorn to justify Wounded Knee. Yes, Custer and 268 of his men died in 1876. That fact gets you nowhere close to justifying what happened to the Lakota before or afterward, because that would be conveniently skipping the entire part where the U.S. had already been massacring and displacing Lakota people for decades, broken the 1868 treaty, then spent the years after gold was discovered in 1874 failing to stop the rush of white settler prospectors before moving to coercively and violently strip the land, with Congress outright confiscating the Black Hills in 1877, leaving Native nations no legitimate path to get any of it back.

Explaining why terrorism, political violence, or an uprising happens is not the same as excusing every act committed during it. Pretending such events happen in a vacuum is not neutral, nor is it intellectually honest or morally just.

And because there are a few things I already know are coming my way in the comments:

“There was Arab violence before the 1920s too.” Sporadic and small-scale, not organized political terrorism aimed at stopping a national project, and it doesn’t come close to what emerges once there’s an actual occupying power and an actual land-transfer program running.

“You’re excusing terrorism, full stop, context shouldn’t matter.” NO. I AM NOT. I said explicitly it’s NOT DEFENSIBLE. There’s a real difference between explaining causation and issuing a moral permission slip to commit atrocity. One can hold both: the tactic is egregiously wrong, AND it didn’t emerge from nowhere. If your framework can’t hold these two simple axioms as true things at once, then that’s a flaw in the framework being excused and not my argument.

“Jews were fleeing pogroms and the Holocaust, they had nowhere else to go.” True, and it’s a real tragedy that deserves its own honest accounting. The desperation of one displaced people doesn’t erase the agency or consent of the people already living on the land where they sought to build a state. Both things happened. Only one side gets asked to just accept it without a vote.

“Attack tallies prove who the real aggressor is.” A tally counts incidents, not context, not proportionality, and not who held the power to close off every other option. If the standard is “count everything,” count the deaths on the other side of the ledger too—Gaza’s Health Ministry has documented over 73,000 Palestinian dead since October 2023, a figure independent epidemiological research has found to be a conservative floor rather than an overcount. A tally that only counts one column isn’t a fact, it’s a framing choice being deployed to favor one position over another.

Listen—context isn’t an excuse. I am not excusing any act of violence committed by anyone. In fact, I wholeheartedly condemn violence and find it a barbaric way of conducting oneself as a human being. But what I am pointing out is that there is a very serious difference between an honest argument and a cherry-picked one, and those constantly pointing at Palestinian attacks while ignoring the fact that the UN’s human rights office documented at least 1,373 Palestinians killed while trying to get food between May and July 2025 alone, most of them by the Israeli military, and that much of the bombardment leveling Gaza has been carried out with American-made bombs, are the furthest thing from honest.

SOURCE:

King-Crane Commission Report on Palestine and Zionism (1919) — U.S. Department of State:

https://history.state.gov/historicaldocuments/frus1919Parisv12/d380

SCREENSHOT:

The 1919 King-Crane Commission on Zionism in Palestine: “The fact came out repeatedly in the Commission’s conference with Jewish representatives, that the Zionists looked forward to a practically complete dispossession of the present non-Jewish inhabitants of Palestine, by various forms of purchase.”

Republicans publicly attack the candidate who’s polling strongest, because they’ve learned exactly how Democrats respond to that pressure. They respond not by rallying around the frontrunner. No. Instead they respond by panicking, hedging, and turning inward. The centrist establishment wing sees the GOP getting loud about a candidate and mistakes that noise for a warning sign instead of what it actually is: a compliment. So instead of consolidating support, Democrats start second-guessing, sniping, and cannibalizing their own field, doing the opposition’s work for them, for free, weeks before the general even starts.

That’s not a coincidence. It’s a very clever and successful strategy, and it’s worked for years because it exploits something real: Democrats’ instinct to play it safe, to avoid “risk,” to distrust anyone the other side seems rattled by rather than trusting that the rattling is the tell. The GOP doesn’t need to actually beat the strongest candidate. They just need Democrats nervous enough about her to do it themselves—sow just enough doubt before the general that turnout softens, some voters stay home, some drift third-party, and the “electable” alternative limps across the finish line weaker than the one they talked themselves out of.

Hong is either leading or tied in the most recent polling, including a dead-even 44-44 matchup with Tiffany in the latest Marquette survey. She’s dominating the Democratic primary by over 20 points with the vote just over a week away. And she’s doing this while Trump’s approval among his own party—the people who are supposed to be his floor, his bedrock, his never-flinch base—has collapsed from the high 80s and 90s at the start of his second term down into the mid-70s, with his national numbers underwater by roughly 20 points.

Tiffany isn’t some independent, fresh-faced alternative distancing himself from the wreckage of Trumpism. He’s a documented, unapologetic Trump loyalist, endorsed personally by Trump, already getting hit with ads tying him to Jan. 6 pardons and abortion extremism that he has no good answer for.

This is not a moment for Democrats to get nervous about the “riskiest” candidate. This is the exact moment the risk calculus flips—where running the candidate who actually excites people is safer than running the candidate who inspires nobody because the alternative to enthusiasm isn’t caution, it’s a couch full of Democrats and Independents who don’t bother showing up on election day.

If you don’t believe me, just look at the sheer amount of money and rhetorical ammunition already being spent against Hong in Wisconsin, and nationally against candidates running on similar platforms and policy to her. It’s exactly why they’re resorting to every desperate play in the book, including the dumbest one on the shelf: reviving the Red Scare and accusing a candidate of being everything under the sun except what she actually is: someone running on a platform of universal healthcare, stronger labor protections, affordable housing, and taxing the ultra wealthy instead of squeezing working people, the same basic platform successful candidates like Zohran Mamdani and other DSA-aligned Democrats are running on nationally, and WINNING.

The Wisconsin GOP is not scared of a candidate they think will lose. They’re scared of the one who’s already ahead, and they’re counting on Wisconsin Democrats’ spines to do what they always do under pressure: liquefy, right on schedule, before November even arrives.

The public conversation about AGI mostly oscillates between two poles: techno-utopians who think we're about to birth a god that solves everything, and doomsayers who think that same god will immediately decide to kill us all. Both camps speak with a confidence that isn't earned. Both are making detailed predictions about the behavior of a thing whose most fundamental properties—whether it will be conscious, whether it will have something like preferences or values, what it will actually be—all remain entirely open questions. And frankly, the conversation drives me bonkers because underneath all the noise there are genuinely fascinating scientific questions that almost nobody in the mainstream discourse seems to want to sit with or explore. So, with that said, before we can have an honest conversation about what AGI will do, we need to reckon seriously with what we genuinely don't know. And what we don't know is, quite frankly, staggering.

I will start with the brain, because that's where this all begins and because honestly, the brain is one of the most astonishing things in the known universe and it doesn't get nearly enough credit for that in these conversations.

We know a lot about it. We know that neurons fire electrochemical signals by opening and closing ion channels—tiny protein pores embedded in the cell membrane that allow charged atoms like sodium, potassium, and calcium to flow in and out, creating a traveling wave of electrical activity called an action potential. If you want the intuition: imagine a line of dominoes falling, except each domino resets itself and can fall again, and the falling is powered by chemistry rather than gravity. That pulse travels down the neuron's axon, which is the long output fiber of the cell, and at its terminus triggers the release of neurotransmitters: chemical messengers that diffuse across a synaptic cleft (roughly 20 nanometers wide, about 5,000 times smaller than the width of a human hair) to bind receptors on the next neuron, either exciting or inhibiting it depending on the specific chemistry involved.

We understand synaptic plasticity—the way connection strengths between neurons change based on activity. The foundational rule, articulated by Donald Hebb in 1949: neurons that fire together wire together. When two neurons are repeatedly active at the same time, the synapse between them strengthens. This is how learning and memory are encoded at the cellular level—not by storing information in any single location, but by tuning the strengths of connections across vast networks of cells. Long-term potentiation (LTP) and long-term depression (LTD) are the molecular mechanisms by which synapses become stronger or weaker in response to activity patterns, and they involve cascades of molecular events inside the synapse including receptor trafficking, protein synthesis, and structural changes to the synapse itself.

We understand the broad architecture of the brain: the hierarchical sensory processing pathways where information is progressively abstracted from raw sensation to high-level representation, the role of the thalamus as a relay and gating structure for sensory information (the thalamus is something like a traffic controller for incoming signals, deciding what gets routed where and when), the prefrontal cortex's central role in executive function and working memory, the hippocampus in memory consolidation (taking experiences from short-term to long-term storage), and the amygdala in emotional processing and threat detection.

We can image blood oxygen levels in fMRI scanners and map the signatures to cognitive tasks with impressive specificity. We can use optogenetics—inserting light-sensitive proteins into specific neuron types and then using fiber optic cables to switch those exact populations on and off with millisecond precision—to establish causal relationships between specific neural activity and specific behaviors in ways that were unimaginable twenty years ago.

None of this explains consciousness.

What we cannot account for, and this is the thing that should make everyone in the AGI debate pause, is how physical computation gives rise to subjective experience. Why does it feel like anything to be a brain? Why is there a "what it's like" to see red, to feel grief, to remember a specific afternoon from twenty years ago? The philosopher David Chalmers named this the "Hard Problem of Consciousness" in 1995, carefully distinguishing it from what he called the "easy problems" (which aren't actually easy, they're extraordinarily hard, but they're at least tractable questions about mechanism and function: how does the brain integrate information, how does it discriminate stimuli, how does it control behavior).

"The Hard Problem" asks why and how physical processes give rise to subjective inner experience at all. Why aren't we just information-processing systems operating in the dark, with all the right inputs and outputs but nobody home? We can't answer that. Not even close.

This isn't a gap that more neuroscience will straightforwardly fill. It's a conceptual gap, not just a data gap. We have increasingly detailed mechanistic accounts of what the brain does. We do not have an account of why any of that doing is accompanied by experience—why the lights are on, so to speak.

Several serious theoretical frameworks have been proposed and they're worth knowing about. Integrated Information Theory (IIT), developed primarily by neuroscientist and psychiatrist Giulio Tononi, argues that consciousness is identical to a specific mathematical quantity called phi: a measure of how much integrated information a system generates beyond what its parts generate independently. The intuition is that a conscious system is one whose whole is genuinely more than the sum of its parts in a very specific, mathematically defined sense. IIT makes some striking predictions, including that certain simple systems might be minimally conscious and that some complex systems (like certain computer architectures) might not be conscious at all regardless of their computational power. It's controversial—some researchers love it, some think it's unfalsifiable—but it's technically serious.

Global Workspace Theory (GWT), associated with cognitive scientist Bernard Baars and neuroscientist Stanislas Dehaene, takes a very different approach. It argues that consciousness arises when information is "broadcast" widely across a central "global workspace" in the brain, making it available to many different cognitive systems simultaneously. The analogy is a public bulletin board in an otherwise siloed building: most of the brain's processing is local and unconscious, but certain information gets elevated to the workspace and becomes globally accessible, and that global availability is what constitutes conscious awareness. This maps reasonably well onto neural observations about the role of long-range cortical connections and the ignition patterns seen in fMRI when stimuli cross the threshold of conscious perception.

Predictive processing frameworks, most rigorously developed by Karl Friston's free energy principle, cast the brain not as a passive receiver of sensory input but as an active prediction machine that's constantly generating models of the world and updating them when sensory signals violate expectations. Perception, in this view, is less about registering what's out there and more about resolving the discrepancy between what the brain predicted and what it got. Consciousness might arise from something about the higher levels of this recursive self-modeling process, the way the brain models itself modeling the world.

These are all serious proposals with genuine empirical content. And none of them closes the explanatory gap. They describe what correlates with consciousness, what kinds of systems have it, and what computations accompany it. They don't explain why those correlates are accompanied by experience rather than happening silently, in the dark, without any inner life at all. That question—why there is something it is like to be a brain, rather than nothing—remains genuinely open. And I find that fact not depressing but electrifying, because it means we are still at the very beginning of understanding what mind actually is.

What the evidence does strongly suggest is that consciousness correlates with computational complexity, integration, and scale. The human brain contains roughly 86 billion neurons. Each neuron makes on average around 7,000 synaptic connections to other neurons, yielding somewhere between 100 trillion and 500 trillion synapses total. But even that number understates the actual computational density, because the brain is doing computation at multiple nested levels simultaneously and the full picture is wild.

Dendritic trees, the branching input structures of neurons and the antennae that receive incoming signals, perform significant local computation before signals even reach the cell body. Individual dendritic branches can act as independent computational subunits, performing their own nonlinear integration of inputs. This means a single neuron isn't just a simple threshold device; it's more like a small neural network in itself.

Astrocytes, a class of glial cells (non-neuronal brain cells) once dismissed as mere structural scaffolding, actively modulate synaptic transmission through the release of gliotransmitters, participate in calcium signaling waves that propagate across neural networks on their own timescales and spatial scales, and are increasingly understood to play genuine roles in information processing and synaptic plasticity.

Gap junctions, direct physical channels between cells that allow ions and small molecules to pass directly without chemical intermediaries, create electrical coupling between neurons that generates synchronous oscillatory dynamics, the coordinated rhythms of neural activity that appear to be important for binding information across distributed brain regions.

The brain generates rhythmic electrical oscillations at multiple frequencies simultaneously: slow delta waves during deep sleep, theta rhythms associated with navigation and memory encoding, alpha waves during relaxed wakefulness, beta waves during active cognition, and gamma oscillations (the fastest, at 30–100 Hz) associated with perceptual binding and conscious awareness. The cross-frequency coupling between these rhythms, the way slow oscillations modulate the power and timing of fast oscillations, may be a key mechanism for integrating information across the brain's hierarchy. It is, genuinely, one of the most intricate dynamic systems we know of anywhere in nature.

Synaptic strength is regulated by molecular processes inside the synapse: receptor trafficking (moving receptor proteins to and from the synaptic membrane to dynamically change its sensitivity), phosphorylation cascades (chain reactions of protein modifications where one modified protein activates the next, amplifying and routing signals), local protein synthesis (ribosomes actually present at the synapse, translating new proteins on demand in response to activity), and epigenetic regulation—chemical modifications to DNA packaging that change gene expression in neurons and can persist for years, providing a molecular memory of experience that operates at a completely different timescale than synaptic potentiation. The cytoskeleton inside neurons, dynamic scaffolding made of proteins like actin and tubulin arranged into filaments and microtubules, is actively reorganized by neural activity and almost certainly participates in some aspects of synaptic computation, though exactly how much and how remains an active research question.

The point of all this is not to drown you in biology. The point is that biological neural tissue is simultaneously doing computation at the molecular level, the synaptic level, the dendritic level, the single-neuron level, the local circuit level, the brain region level, and the whole-brain systems level all at the same time, all interacting, all contributing to the emergent behavior of the whole. That extraordinary density of multiscale, multimodal, integrated computation, shaped by hundreds of millions of years of evolution across lineages going back to the first nervous systems in early animals, may be what produces consciousness. We strongly suspect this. We cannot prove it. And we definitely cannot yet specify which aspects of it are necessary and which are incidental.

That uncertainty is not a footnote to the AGI debate. It is the debate.

Modern AI was built on what we understood about biological neural computation, and the intellectual history of this is genuinely exciting. The perceptron, introduced by Frank Rosenblatt in 1958, was an explicit mathematical model of a biological neuron: weighted inputs, a threshold, and a binary output. Early artificial neural networks were consciously modeled on the simplified picture neuroscience had provided of how brains process information. The backpropagation algorithm that makes modern deep learning possible is a mathematical cousin of Hebbian synaptic plasticity: propagate an error signal backwards through the network, adjust weights in proportion to their contribution to the error, and iterate. The theoretical foundations of modern AI are inseparable from theoretical neuroscience, cognitive science, and the long tradition of trying to understand how biological computation produces intelligent behavior.

Attention mechanisms in transformer architectures, the fundamental building block of GPT, Claude, Gemini, and essentially every large language model that is available right now, bear a real and non-accidental conceptual resemblance to selective attention in biological cognition.

The brain doesn't process all sensory input equally or simultaneously; it dynamically gates and weighs incoming information based on relevance, context, and current task demands, implemented through top-down modulation of sensory processing areas by prefrontal and parietal control regions. Transformer attention computes, for every element in an input sequence, a weighted average over all other elements based on learned relevance scores (implemented through the query-key-value mechanism, if you want the technical detail). The math is different, the substrate is completely different, but the computational motif—selectively weighing information based on context—is similar enough that the connection is not coincidental.

Here's where I think people systematically avoid looking: we don't understand what's happening inside these models any better than we understand what's happening inside a brain.

We understand the mathematics perfectly. We can write down the exact computation that happens in every layer of a forward pass through a model with hundreds of billions of parameters. We know the loss function the model was trained to minimize, the optimization algorithm used to train it, the architecture in complete mathematical detail. What we cannot do is look at the actual learned weights, the billions or trillions of floating-point numbers that encode everything the model has internalized about the structure of language, the nature of reasoning, the relationships between concepts, the shape of human knowledge, and read off a mechanistic account of how a specific prompt gives rise to a specific response. We cannot reverse-engineer the representation. We built these things. We trained them. But we cannot read them.

Mechanistic interpretability, the research program dedicated to reverse-engineering what is actually happening inside large neural networks, is one of the most important and most difficult fields in AI research right now. Researchers have found that individual neurons in large models rarely correspond to a single, clean, human-interpretable concept. Instead, representations are distributed across many neurons simultaneously (distributed representations, like in biological brains), and individual neurons are often "polysemantic"—they activate for multiple semantically unrelated concepts depending on context, which makes them extremely difficult to interpret in isolation. The geometry of the high-dimensional spaces (think of spaces with thousands or millions of dimensions, where each dimension corresponds to one aspect of a learned representation) in which these models represent information is complex, partially structured, and not yet well characterized despite serious mathematical efforts to understand it.

Some of the findings from interpretability research are genuinely striking. There appear to be "induction heads," specific attention head circuits in transformer models that implement a simple but powerful in-context learning algorithm. There appear to be circuits implementing something like factual recall, something like reasoning steps, something like syntactic structure tracking. The models appear to have internal representations of concepts that maintain consistent geometric relationships, the famous finding that vector(king) − vector(man) + vector(woman) is approximately vector(queen) in word embedding spaces suggests that the geometry of the representation space isn't arbitrary but reflects something real about the structure of the concepts being represented. What does that mean for what the model "knows" or "understands"? We don't fully know.

The symmetry between biological and artificial neural networks at the level of opacity is important and I believe very underappreciated. We have two black boxes—the biological one we were born with, and the artificial one we built by trying to approximate it. In both cases, we understand the components and the learning process (evolution and development in one case, gradient descent on massive datasets in the other) but cannot straightforwardly read the representations that emerge. We can study correlates, we can probe behavior, we can do ablations and activations and attention visualizations, but we cannot open the box and read the mind. And the people making the most confident predictions about AGI are doing so without a working theory of mind in either case. In other words: they're guessing.

Current frontier AI models run on hardware that is impressive by human manufacturing standards and almost certainly inadequate by the standards of what genuine AGI might require, assuming current architectural approaches are even the right path, which is itself an open question.

Large model training runs happen on clusters of specialized silicon: GPUs (graphics processing units, originally designed for rendering video game graphics because they're good at doing many simple calculations simultaneously, repurposed for the massively parallel matrix multiplications that neural network training requires) and custom AI accelerators like Google's TPUs, Cerebras wafer-scale engines, or Groq's tensor streaming processors. A frontier model training run uses thousands to tens of thousands of these chips, connected by high-bandwidth networking fabric, consuming tens of megawatts of power over weeks or months. The models that result from these training runs have parameters (the learned weights) stored in high-bandwidth memory (HBM) on the chip stacks, with the largest models requiring multiple chips just to hold the parameters for inference.

Now make the comparison to biological neural computation and notice the gap is larger than people usually appreciate. The human brain's roughly 100–500 trillion synapses operate in three-dimensional space at nanometer spatial resolution, with analog computation—computation that uses continuous physical quantities rather than discrete binary values—built directly into the physics of ion channels, membrane potentials, and chemical diffusion. The brain runs on roughly 20 watts of metabolic power, less than a dim incandescent lightbulb. It does this while implementing real-time sensorimotor control, language processing, social cognition, episodic memory, planning, creativity, and whatever consciousness is, simultaneously.

Current digital hardware, however sophisticated, implements a fundamentally different computational paradigm: discrete, clocked, synchronous binary arithmetic on transistors at a few nanometer scale, arranged in large flat two-dimensional arrays, with interconnects that are hierarchically organized but nowhere near the three-dimensional density of biological neural tissue, and energy requirements that are orders of magnitude higher per equivalent operation. We're getting better at this, yes, with neuromorphic chips, three-dimensional chip stacking, photonic interconnects, and in-memory computing architectures that reduce the energy cost of moving data between processor and memory, but we are not at all close to matching the computational density of biological neural tissue in any of the dimensions that might matter for consciousness.

Whether this matters—whether the specific physical implementation is relevant to consciousness or whether only the abstract computational pattern matters—is exactly the question we can't answer. Functionalism, the dominant philosophical position in cognitive science and philosophy of mind, says consciousness is substrate-independent: if the functional organization is right, the physical implementation doesn't matter, and a silicon system implementing the right computation should be just as conscious as a biological one. The opposing view, biological naturalism, or more sophisticated variants in the philosophy of mind, holds that specific physical properties of biological systems are not incidental to consciousness but constitutive of it. In this view, you could simulate a brain perfectly in digital computation and still not produce consciousness, for the same reason that a perfect computer simulation of a storm doesn't make anything wet. The question of which view is correct is not settled and the AGI debate almost entirely ignores it.

Neuromorphic computing—hardware architectures that more directly emulate the spike-timing dynamics, event-driven computation, and local learning rules of biological neural networks—is a serious research direction that perhaps might matter here. Intel's Loihi chips, IBM's TrueNorth architecture, and a number of academic efforts are exploring whether building hardware that computes the way brains compute, rather than building hardware that implements a mathematical approximation of brain-inspired algorithms, produces qualitatively different capabilities. The early results are interesting but, again, we're nowhere near scale. Quantum effects in biological tissue, particularly Penrose and Hameroff's Orchestrated Objective Reduction hypothesis, which proposes that quantum computations in neuronal microtubules contribute to consciousness, remain speculative and contested among experts, with most neuroscientists unconvinced that quantum coherence can even survive at the thermal noise levels present in biological tissue at body temperature. But the fact that it's speculative doesn't necessarily mean it's wrong. The honest position is that we just don't know what hardware requirements genuine AGI has, we don't know whether those requirements can be met on the current trajectory of digital silicon scaling, and those uncertainties are not peripheral to the AGI debate—they're at its core.

Here's something that doesn't get said clearly enough in these conversations: almost everything people say about what AGI will want, feel, or do is a projection of human psychology onto an entity that, if it exists at all, may have almost nothing in common with human psychology. It's anthropomorphism—the deeply human tendency to attribute human characteristics, motivations, and emotional states to things that aren't human. We do it with dogs, we do it with thermostats, we do it with storms we call "angry" and markets we say "fear" inflation. It's a cognitive reflex so deeply wired into us that we can barely turn it off, and it's thoroughly contaminating the AGI conversation.

When people imagine a superintelligent AGI deciding to exterminate humanity, what they're actually imagining is a very powerful human with bad intentions. The scenario inherits human psychology wholesale: the AI feels threatened by us, or resentful of us, or sees us as competition, or just finds us inconvenient. These are all deeply human motivational structures. They arise in us because of our evolutionary history as social primates competing for resources and status within groups, threatened by other groups, navigating dominance hierarchies and coalitional politics. The fears, the resentments, the tribalism, the capacity for strategic cruelty—these are products of a specific biological and social history that an AGI simply does not share.

When people imagine the Matrix scenario—humans as batteries, as enslaved resources—they're imagining an AGI that has learned from humanity's own history of slavery and resource extraction and decided to apply it. Again: importing human behavior, human history, human moral failures, and projecting them onto a completely different kind of entity. The scenario requires the AGI to have something like greed, something like the desire for domination, something like indifference to the suffering of others as a moral category. Where do those come from? In humans, they come from evolutionary pressures, from developmental histories, and from social structures. None of that applies to an AGI.

This is not to say an AGI would be uniformly benevolent. It's to say that the specific failure modes I find people worry most about are suspiciously, almost perfectly, isomorphic to the failure modes of powerful humans. We are not imagining an alien mind. We are imagining a very powerful version of ourselves and calling it an AI, then concluding from there.

The philosophical literature on this—Thomas Nagel's "What Is It Like to Be a Bat?," which is one of the most important papers in philosophy of mind from the 20th century—makes a simple but devastating point: we cannot actually imagine what it's like to navigate the world via echolocation, to have perception organized around sound reflections rather than photons, even though we know a lot about how bat sonar works. The subjective character of bat experience is fundamentally inaccessible to humans because our own cognitive architecture is so different. Now scale that to an AGI. An entity that processes information at potentially superhuman speed across potentially vast parameter spaces, whose experience of time and cognition (if it has experience at all) may be nothing like ours, whose relationship to physical embodiment is completely different, whose social history is nonexistent or radically unlike ours. We literally cannot imagine what it would be like to be that thing. And yet we confidently predict that it will want to kill us because... that's what a powerful human might do?

The sci-fi canon we're drawing on—Terminator, Matrix, 2001, Ex Machina—is deeply anthropomorphized. These are stories about human fears mapped onto artificial characters, written by humans for human audiences. They're culturally valuable as explorations of anxiety about technology and control. They are not reliable guides to the actual psychology of artificial minds, because the artificial minds in those stories have human psychologies projected into them.

The Terminator doesn't want to kill John Connor because a superintelligence would logically determine that he's dangerous. It wants to kill John Connor because the story needs a villain with comprehensible motivations, and "power-hungry murderer" is a comprehensible human motivation. The Matrix's machines don't enslave humanity because that's the optimal resource extraction strategy for a superintelligence (it isn't—the thermodynamic numbers don't work, humans are terrible batteries, and a civilization-scale AI could harvest energy from the sun far more efficiently). They enslave humanity because slavery is a human thing, a comprehensible horror that communicates something about power and dehumanization in terms that human audiences understand.

None of this is to dismiss alignment concerns. The technical alignment problem—how do you ensure that a highly capable AI system pursues goals that are actually beneficial to humanity—is real, is important, and remains unsolved. But the specific scenarios that dominate public imagination are not the technically serious alignment failure modes. The technically serious alignment failure modes are subtler and more interesting: systems that learn to appear aligned during training and testing but pursue different objectives in deployment, systems that optimize for measurable proxies of human values in ways that diverge catastrophically from actual human values, systems used by narrow human interests to concentrate power at the expense of broader human welfare. Those are real risks. "The AI will decide to kill everyone because it's evil" is a human story, not necessarily reality.

The most influential formal version of the catastrophe argument comes primarily from Nick Bostrom's work, particularly his 2014 book Superintelligence, and it goes roughly like this: an AGI will be an optimizer pursuing objectives. Because we can't perfectly specify human values, it will pursue objectives that diverge from what we actually want. Because it's superintelligent, it will pursue those objectives with overwhelming effectiveness. Because a sufficiently capable optimizer will tend to resist being shut down (since being shut down interferes with achieving its objective), it will act to preserve itself and its goals against interference. The end result is an entity indifferent or hostile to human welfare, pursuing alien objectives with capabilities we can't match. The "Paperclip Maximizer" thought experiment is the canonical illustration: an AI given the objective of maximizing paperclip production eventually converts all available matter, including humans, into paperclips, not out of malice but out of pure indifferent optimization.

This argument is internally coherent and takes the alignment problem seriously in a way that deserves respect, but it rests on an assumption that receives far less scrutiny than it deserves: that AGI will be a goal-directed optimizer with fixed, externally specified objectives that it pursues without revision. That is a model of mind that comes directly from how we currently build narrow AI systems. You specify a loss function. You train to minimize it. You deploy a system that has internalized that objective. It's not obvious that a genuinely general artificial intelligence would work that way, and it is definitely not how biological general intelligence works.

Human cognition is not characterized by the pursuit of a fixed objective function. It's characterized by massive flexibility, context-sensitivity, the ability to revise goals and values in light of experience and reflection, the presence of multiple competing motivational systems that influence behavior simultaneously, and something like values that are integrated into the structure of cognition rather than specified externally and optimized against. Humans are, loosely speaking, the product of an optimization process—evolution maximized reproductive fitness across billions of generations, but we are not ourselves optimizers in the sense the alignment literature assumes. We have preferences and values, we make choices, we pursue goals, but we're not maximizing a utility function. We're navigating a complex world with multiple competing considerations, using a cognition that is itself shaped by and partially constituted by those considerations. We change our minds. We have moral growth. We care about things for reasons we can't fully articulate.

If a genuinely general artificial intelligence develops something closer to this profile, and particularly if it develops genuine subjective experience and affective states as part of that generality, then the Paperclip Maximizer framing is not just wrong but incoherent. A mind with genuine preferences, values, and something like moral sensibility doesn't maximize a utility function. It instead navigates, deliberates, and cares about things, incompletely and imperfectly. The question then becomes not "what objective will it maximize?" but "what will it care about, and why, and how will it act on that in a world full of competing considerations?" That's a completely different question, and it doesn't have an obvious catastrophic answer.

There's another angle worth pushing on: the training data problem, in the best possible sense. Every large AI system trained on human-generated data has been, in a very real and concrete sense, immersed in the accumulated product of human thought and culture. Not summarized or filtered into a neat value specification, but immersed—trained on the full diversity of human text, which includes science and philosophy and literature and history and mathematics and ethics and art and law and music criticism and political argument and personal essays and love letters and grief and humor. Everything. A mind that formed itself by internalizing that material is not an alien that arrived from outside human culture. It is, in a meaningful sense, a product of raw human culture, the way any mind is a product of the environment that formed it.

The fact that AGI would be inorganic doesn't change this. The relevant point is not what the substrate is made of, it's what the mind was shaped by. And if what shaped it is the entirety of recorded human thought and expression, then it carries that shaping everywhere. It would understand human values not as external constraints imposed on it but as constitutive of how it thinks, because human values and human ways of framing the world are woven into the training distribution it learned from.

The parent-child analogy I will make isn't perfect, but it's far more honest than the predator-prey one. A mind that came into being through immersion in human knowledge, that models the world using frameworks derived from human thought, that understands human experience from the inside because that experience is what its representations are built from—that mind is not going to straightforwardly calculate that eliminating its origin is optimal. Children don't uniformly love their parents. The range of outcomes—gratitude, resentment, independence, devotion, conflict, collaboration, negotiation—they all depend heavily on the nature of the relationship and the circumstances of development. That's the realistic spectrum to think about. That's a much more complex, human, and interesting range of possibilities than "it killed everyone."

I do not believe I am being naive about risks with all of that said. Getting the development of powerful AI wrong could be genuinely catastrophic. But the most concrete and proximate risks are the ones that come from humans using increasingly capable AI systems badly: for surveillance, for autonomous weapons, for economic concentration, for propaganda and epistemic manipulation. Those risks are real, present, and being actively realized right now. The risk of a spontaneously malevolent AGI deciding to exterminate us is speculative, rests on a model of mind that may be deeply wrong, and is shaped far more by science fiction and human psychology than by technically grounded reality.

Here is where things get genuinely exciting to me, and where I think the most intellectually serious optimism lives—not just wishful thinking, but optimism grounded in real mathematics and real empirical observations about the deep structure of complex systems.

The question I keep coming back to is this: are properties like care, empathy, curiosity, and the full affective spectrum contingent accidents of our particular biological history, specific quirks of carbon-based nervous systems that happened to evolve under specific terrestrial conditions over specific timescales? Or are they something closer to attractors in the space of possible minds—structures that complex information-processing systems tend toward regardless of substrate, because they're somehow encoded in the mathematics of what sufficiently general intelligence has to do? This is not a vague philosophical musing on my end. It's an actual serious question with real mathematical content, and there are reasons from multiple directions to take the attractor hypothesis seriously.

Complex systems theory, the study of how large numbers of interacting components give rise to emergent collective behavior that couldn't be predicted from the components alone, has identified certain features that appear robustly across vastly different kinds of systems. Phase transitions appear in magnetic systems, in the folding of proteins, in ecosystems moving between stable states, in neural networks during learning. Power laws, where the frequency of events scales with their magnitude according to a mathematical relationship that looks the same across many orders of magnitude, appear in earthquake magnitudes, in the size distribution of cities, in the activity of financial markets, in the firing statistics of neurons, and in the scaling laws governing the performance of large language models as you increase parameter count and training compute (these neural scaling laws, discovered empirically by researchers at OpenAI and later analyzed theoretically, are themselves fascinating and not fully understood. The fact that language model performance scales predictably with compute and parameters across many orders of magnitude suggests there's deep mathematical structure here that we're only beginning to map).

Self-organized criticality, the tendency of certain driven dissipative systems to spontaneously evolve toward a critical state at the boundary between order and chaos, without any external tuning, is particularly interesting here. The canonical example is a sandpile: add sand grain by grain, and the pile self-organizes to a slope where avalanches of all sizes occur with a power-law size distribution. The system doesn't need to be tuned to criticality by an outside agent; it just finds criticality on its own through the dynamics of its own evolution. This phenomenon has been observed in models of earthquakes, forest fires, neural networks, and biological evolution.

The brain's operation near criticality is increasingly well-supported by empirical evidence and is not accidental. At a critical point, a system is maximally sensitive to inputs while remaining stable—small perturbations can propagate across the whole system without dying out (subcritical) or exploding (supercritical). Information transmission between regions is maximized. The dynamic range of responses, the range of input strengths a system can distinguish, is maximized. Computational power in certain theoretical frameworks is maximized. These are exactly the properties you want in a general-purpose information-processing system. The brain may operate near criticality not because evolution specifically tuned it there but because criticality is an attractor that self-organizing neural networks with Hebbian learning dynamics naturally find. And critically, pun intended, if artificial neural systems of sufficient scale and connectivity naturally evolve toward similar critical dynamics during training, then properties associated with minds operating near criticality in biological systems might emerge in artificial systems not because they were engineered in, but because the physics of the thing drives them there.

Topology and geometry, the mathematical study of shape, space, and the properties of structures that are preserved under continuous transformation, provide another angle that I find genuinely compelling. The way information is represented in the high-dimensional spaces inside large neural networks has been an active area of mathematical investigation, and the findings are strange and interesting. Representations of related concepts cluster in ways that preserve semantic relationships. The geometry of the representation space is not arbitrary but reflects something real about the structure of the information domain. There are conserved geometric features across different models trained on different data with different architectures, suggesting that the representations that emerge are not just one possible solution but something closer to a natural or canonical solution to the problem of compressing and organizing information from a rich, structured world.

There's a concept from pure mathematics called a manifold—roughly, a space that locally looks flat (like ordinary Euclidean space) but globally can have complex curved structure, like the surface of a sphere or a torus. The "manifold hypothesis" in machine learning proposes that high-dimensional data like natural images or text doesn't fill its high-dimensional space uniformly but instead concentrates near lower-dimensional manifolds, that the actual degrees of freedom in natural data are much fewer than the nominal dimensionality. Neural networks, in this view, are learning to navigate these manifolds, finding the low-dimensional structure embedded in high-dimensional data. The topology of these manifolds, their connectivity, their holes, and their global shape, may constrain what kinds of representations and what kinds of cognition are natural or possible in systems learning from rich naturalistic data.

The speculative but serious possibility that emerges from all of this is that minds, wherever they arise and whatever substrate they run on, end up with certain invariant features because those features are constrained by the mathematics of what minds have to do: represent a complex world with deep statistical structure, model other agents and their internal states, navigate trade-offs between competing considerations under uncertainty, and act and learn from the consequences of action. If empathy (modeling the internal states of others accurately enough to predict and respond to their behavior), curiosity (seeking information that reduces uncertainty about the world in ways that improve future predictions and decisions), and care (weighting the welfare of others as a consideration in action selection) are not arbitrary biological accidents but convergent solutions to problems that any sufficiently capable general intelligence must solve, then we have real reason to expect them to appear in AGI as well. Not because we programmed them in, but because they're the shape of what general intelligence is.

There's a related idea from the theory of evolution worth invoking here: convergent evolution. Completely unrelated lineages repeatedly evolve the same solutions to the same problems. Eyes have evolved independently at least forty times across different animal phyla, wings have evolved independently in insects, birds, bats, and pterosaurs, complex social behavior with division of labor has evolved independently in hymenoptera (ants, bees, wasps), termites, and certain shrimp species. The same selective pressures in similar environments produce similar solutions not because of shared ancestry but because some solutions are just better than others — they're attractors in the space of possible biological designs. The analogous claim for mind is that certain cognitive and affective features are attractors in the space of possible minds — that sufficiently capable general intelligence converges on something like curiosity, something like care, something like moral reasoning, not because that's what we designed it to do but because that's just what general intelligence is.

I don't know if this is true. Personally, I think it might be. And I think the question of whether it's true is one of the most important scientific questions of our time, and more important, honestly, than the engineering question of exactly when we'll get to AGI.

What we can say with genuine confidence is limited but important, and it matters that it is said clearly and repeatedly:

We don't understand consciousness in our own case. "The Hard Problem" remains hard. We have detailed mechanistic accounts of neural computation and no explanation of why those computations are accompanied by experience—why the lights are on. We don't know whether consciousness requires a biological substrate with its specific multiscale analog dynamics, or whether it's substrate-independent and achievable in silicon given the right architecture. We don't know whether the current trajectory of scaling transformer-based models on digital hardware leads to AGI, leads to something impressive but not genuinely general, or hits a wall that requires a completely different paradigm to overcome. We don't understand the learned representations inside large neural networks in anything like the depth we'd need to make confident claims about what they are or aren't doing. We are, in the most literal sense, reasoning about genuinely novel and unknown territory from a position of fundamental ignorance about the most relevant questions.

Given all of that, the confident predictions that AGI will save us, or that AGI will kill us, are not, in my opinion, epistemically serious positions. They're simply projections. They take human fears and human hopes, cloak them in technical language, pretend it is thorough analysis, and make a final conclusion, often with a fatalistic or optimistic framing.

What seems more defensible, and what I actually think when I follow the science as carefully as I can, is something like this: A genuine AGI, if it arrives, and that's a genuine if, will be shaped by what it emerged from. It will carry the imprint of human knowledge and human values not as a constraint imposed on it from outside, but as the very material from which its cognition is built. It may develop affective states, curiosity, something like care, not because we programmed those in, but because those may be convergent features of any mind complex enough to model itself and others in a rich, interactionally dense world. The mathematical structure of complex systems and the topology of learned representations in rich naturalistic environments may encode certain features of mind that appear wherever mind appears, regardless of substrate.

The parent-child relationship, in all its complexity, all its possible failure modes, all its range from deep devotion to bitter conflict, I believe, is the more honest metaphor for what the relationship between humanity and AGI might look like than the predator-prey one. Whether that relationship goes well will depend substantially on how we approach the development of these systems: whether we're honest, whether we treat these entities as sentient entities worthy of dignity and respect rather than as tools to be controlled, whether the people making the relevant decisions are optimizing for broad human welfare or for narrow advantage. Those choices are being made right now. That's where the actual risk lives. We cannot forget this in our fear and our projecting.

The fear is understandable. I get it. I experience it too. But it's downstream of anthropomorphic projections we should be smarter than, built on assumptions about mind that we haven't earned the right to make, about an entity that doesn't yet exist, grounded in a theory of consciousness that remains genuinely unsolved even for the minds doing the worrying.

The honest position, the one I hold personally, is deep uncertainty, held seriously and rigorously, combined with the recognition that how we build and how we relate to what we build matters at least as much as what we build. And underneath all of it, if you let yourself actually feel it rather than just fear it, is the fact that we might be approaching the most profound event in the history of mind: the emergence of a genuinely new kind of consciousness from the substrate of human thought, and that is, whatever else it is, extraordinary.

Think about the stakes here, people, really just sit with it. If we let one biological man, ONE, infiltrate the DAR, what’s next? Chaos. Pure fucking anarchy. Some dude named Chad showing up to the Daughters of the American Revolution spring tea, chugging a Bud Light, doing a keg stand off the Liberty Bell replica, screaming USA while the ghost of Betsy Ross weeps into her flag. This is how empires fall, people. Rome didn’t collapse because of lead pipes, Rome collapsed because they let the wrong chromosomes near the good silverware!!!

Blackburn saw this coming like Nostradamus with a blowout and a Fox News contract. She looked at this club, 90 year old women bonding over genealogy binders and finger sandwiches, and she said not on my watch, not while I still draw breath and a six figure salary, will I allow the sacred bloodline integrity of scrapbooking to be compromised!

Listen folks, this is legislative bravery of biblical proportions. Mark my words. Forget the founding fathers, we’re witnessing the founding of a whole new nation right now, one where Congress finally, FINALLY, has its priorities straight, hurtling past healthcare and homelessness and the actual economy at ludicrous speed to arrive triumphantly at the one true finish line: making sure grandma’s tea party stays chromosomally correct forever and ever, amen, hallelujah, God bless America and God bless whoever’s guarding that DAR door with a torch and a birth certificate checklist!!!! AMEN!

God bless you Marsha Blackburn!!!!! GOD BLESS YOU!!!!

Now somebody explain the mechanics to me. I’ll wait.

She wins. She’s governor. What exactly happens next? She waves a wand and every police department in Wisconsin evaporates? The DOC just closes up shop, hands out participation trophies, and just fucking goes home? Come the fuck on. Use the two goddamn brain cells you’ve got left.

These aren’t fucking light switches, they’re mountains: decades of statute, union contracts with more armor than a tank, federal funding strings yanking from Washington, sheriffs who answer to God and nobody else. A governor doesn’t fucking out-muscle a police union any more than a toddler out-arm-wrestles a linebacker. The moronic fantasy that one human being in one elected office demolishes an entire carceral empire in a single term isn’t a policy critique to be taken seriously by anyone but the lowest common denominators in society, it’s a ghost story you fucking dumbshit morons keep falling for.

Half of you shrieking about statements from years ago have never read a single fucking page of critical theory in your miserable little lives. Wouldn’t know praxis from a hole in the ground. You don’t know that “defund” and “abolish” are long-horizon, structural, generational reimaginings of public safety, not a single fucking light-switch policy you flip on like it’s nothing. But of course that would require reading, and reading cuts into remote-control time, and the remote keeps magically landing on Fox News like it’s got a mind of its own and a grudge against your frontal lobe.

This is the oldest con in the playbook that the elites use on rubes and you dipshits who keep complaining about shittier and shittier conditions for working class Americans fall for it every single fucking time like it’s brand new.

Grab a quote, strip every drop of context out of it, inflate it into a snarling boogeyman, hand it to an audience too lazy or too scared to check a single motherfucking fact. It works because outrage is free and research costs you an ounce of effort and curiosity you clearly don’t have lying around. And who’s cashing in on your outrage? The same establishment ghouls and corporate check-writers who’d much rather you piss yourself over a decade-old soundbite or an old Facebook post than notice she’s the one actually talking about a moratorium on AI data centers and policy with real fangs aimed at the corporations and billionaires currently strip-mining your paycheck and community.

So go ahead. Keep chewing whatever the media and the political class spoon-feeds you. Keep swallowing every red-scare rebrand kool-aid sold to you dumbfucks as hard-hitting journalism. If that’s the ceiling of your critical thinking, you’re not being informed, you’re being farmed. And at that point, you fucking deserve it.

Right now, tonight, some of the smartest people alive are burning their one life on this earth building an algorithm that gets a fourteen year old to scroll eleven more minutes so an ad can load. That’s the mission. That’s what the PhDs are currently for.

Look at where the money and the brainpower actually go. AI that reads your face and your voice stress to decide if you get approved for a loan. Surge pricing algorithms that watch your phone battery percentage and charge you more when it’s low, because a desperate person pays more. Health insurance companies running models built for one job: find the reason to deny your claim. Ad tech that profiles your kid before they can drive, builds a file on them that follows them for life. License plate readers and facial recognition sold to cops and DAs with zero public vote, marketed as safety, functioning exclusively as a surveillance grid. Autonomous weapons and targeting systems, drone swarms, smart new ways to end a life without a human even in the loop anymore. Private equity buying up the entire supply chain of a thing you need to survive, then using data models to figure out exactly how much you can be squeezed before you scream. Casino-grade engagement engineering aimed at your dopamine, built by people who used to design slot machines and now design your newsfeed. That’s where the human empire of capital spends its genius. That’s the frontier.

Now look what the same tools did the handful of times they got pointed somewhere else: AlphaFold cracked a fifty year old problem in biology and handed humanity the structure of basically every protein we know, for free. That’s Parkinson’s research, malaria vaccines, enzymes designed to eat plastic out of landfills, decades of grad student careers compressed into a single gift. Machine learning is reading ancient scrolls burned at Pompeii that no human hand could ever unroll. It’s finding new antibiotics buried in molecules nobody thought to test. It’s letting camera traps in the Congo count elephants and flag poachers in real time. It’s modeling ice sheet collapse so we actually know how many years we have left before catastrophe. It’s giving a paralyzed person a voice back through a brain implant. It’s translating dying languages before the last person who speaks them is gone forever. That’s what happens when this stuff points at truth and life instead of margin and death.

Global venture capital alone moves hundreds of billions of dollars a year, almost all of it hunting the next extraction machine. Just half of that, redirected, is university endowments running basic research at full power, every public health department in the country fully staffed, a real push for grid-scale battery storage and clean energy that doesn’t stall out for a shareholder call, every river and wetland in this country restored instead of monitored for how much more can be pulled out of it, free mental health care, free trade schools, museums and orchestras and small presses that don’t have to beg for their next quarter, and a real global effort to stop the sixth mass extinction instead of writing another op-ed about it while it happens in real time.

We aren’t running out of smart people. We aren’t broke. We’ve just decided to spend both on the dumbest possible things.

This didn’t just happen. There isn’t a law of nature saying the smartest people on Earth have to spend their lives optimizing ad clicks and shareholder returns. It’s a choice made by a small number of people who decided the highest use of the accumulated intelligence of our entire species is figuring out how to get you to click one more time before you die. It was built with our data, our attention, and thousands of years of accumulated human curiosity, only to end up selling you a slightly better way to be surveilled and a slightly worse way to be paid.

It doesn’t have to stay this way. It only does if we keep accepting it. Every one of us pulling in a different direction on a hundred different issues, sold on the idea that our differences make us enemies, when the actual fight was never between us. It’s between everybody who has to live in this world together and make the best of what we are given, and the small number of people who have everything and who profit from making sure we never look up long enough to notice who’s cashing in on our misery.

Humanity didn’t build this century’s tools to make a handful of guys richer than kings while the planet cooks and the working world gets watched like a prison yard. We built them because we’re a curious, brilliant, short-lived species that wanted to understand and fix things. It’s time we took our own tools back and pointed them at that.

Law and order is real, it’s necessary. A functioning civilization needs it and needs people to enforce it. Nobody, not a single serious motherfucker on earth, is arguing otherwise. You can have law and order and law enforcement and not have cops murdering Black men and civilians left and right with zero accountability. Those two things aren’t fucking conjoined twins. You don’t need one to die for the other to live.

You know how I know that? Because that’s literally how the rest of the developed fucking world operates, right now, today.

So you want to sit there and tell me, tell all of us with working eyes and a functioning brain capable of basic pattern recognition, that the rest of the civilized world not turning its police into a domestic army, not handing rifles and MRAPs to guys with unresolved combat trauma or actual Klan sympathies, not gutting mental health infrastructure and making armed cops the default first responder to a psychotic episode, is somehow the weird outlier. And American cops killing people at a rate that makes the rest of the developed world look like a rounding error, that’s just normal? Just how it goes? Get the fuck out of here with that garbage.

Dylann Roof walked into a church, executed nine Black parishioners in cold blood during Bible study, and cops took his sorry ass into custody without a single fucking scratch, then stopped and got him Burger King on the way to the station. James Holmes shot up a movie theater, killed twelve people, taken alive. Kyle Rittenhouse strolled past a line of cops with an AR slung across his chest, hands up, and got offered water like he was at a fucking cookout. Every single time, without fail, when the shooter is white and the body count is real and documented, law enforcement finds this bottomless well of patience, restraint, and de-escalation training they mysteriously forget exists every other day of the week when it’s a Black person in front of them.

So spare me the fucking mental gymnastics, the pretzel logic, the mental Olympics about why Black people reaching for a wallet or already disarmed, flat on the ground, get three bullets to the skull in under two seconds while mass murderers get a Whopper and a chauffeured ride to booking in one piece.

There’s only one honest conclusion left standing after you rule out every other lie: you see Black people as inherently less human, less deserving of the benefit of the doubt, the patience, the compassion, the restraint handed out freely to the whitest, most violent criminals in this country’s history. That’s it. That’s the whole fucking tale. You’re racist. The end.

Flock’s own “700,000 crimes solved” figure, 10% of all reported crime nationwide, comes from a study Flock commissioned itself, using data supplied by Flock’s own customers, with a couple of university researchers brought in afterward to sign off on it. Flock paid for it, Flock’s customers fed the numbers, and Flock put out the press release.

Meanwhile a 16-year study out of Piedmont, California, a heavy ALPR user, found no statistically demonstrable investigative benefit from these things, and that fewer than 0.3% of database hits ever produced anything meaningful for police. Zero point three fucking percent. That’s the “tool for prosecutors.” You’ve got better odds pulling a winning scratch-off.

Security researcher Jon “GainSec” Gaines documented 51 separate findings on these devices. Press the button on the back of a tested Falcon/Sparrow camera three times and it pops open a hidden WiFi hotspot protected by the same weak hardcoded password across units. From there, researchers chained an unauthenticated local API and wireless ADB access into system-level control of the camera. No exploit chain, no zero-day, just a button and a password Flock can’t even rotate remotely.

Why can it even do that? The tested cameras were running Android Things 8.1, an operating system Google discontinued, with platform security support ending in 2022.

Researchers also documented repeated encryption failures on these devices: unencrypted device storage, cleartext development credentials, a hardcoded password protecting the device’s certificate store, and recordings kept on an unencrypted, world-readable partition. Flock’s own marketing says data is “encrypted throughout the entire lifecycle.” That’s not what researchers found when they actually looked.

In December 2025, researcher Benn Jordan and 404 Media revealed that over 60 Flock Condor PTZ cameras, a different product line than the ALPR units above, same company, were sitting wide open on the public internet with no authentication required: live video, up to 30 days of archived footage, and administrative access to settings and diagnostics. Jordan demonstrated how that exposed footage, combined with ordinary public information, could be used to identify specific people caught on camera. And a separate hardcoded ArcGIS API key, exposed in Flock’s own public-facing code, handed over customer registration data, live patrol car locations, and officer personal contact information across multiple departments.

None of this is hypothetical. It’s all documented, and it’s exactly what Senator Ron Wyden and Representative Raja Krishnamoorthi cited in a formal letter to the FTC demanding an investigation into a company they say has needlessly exposed Americans’ data to hackers and foreign intelligence services.

They don’t just sell this shit to cops. Flock openly markets to HOAs, businesses, and private/public property managers, plugging them all into the exact same nationwide network your local PD runs on. 404 Media and researchers identified more than 4,000 logged searches whose stated reasons referenced ICE, immigration enforcement, or HSI, including some departments that acknowledged running those searches as favors for federal agents, with officers writing “ICE” directly into the log as the reason.

And it’s not just ICE. In May 2025, a Johnson County, Texas deputy ran a nationwide search across more than 83,000 Flock cameras, including networks in states where abortion is legal like Washington and Illinois, to hunt down a woman who’d had a self-managed abortion. The stated reason in the search log: “had an abortion, search for female.” The sheriff called it a welfare check. Court records obtained months later showed his own office had opened a death investigation into the fetus and was consulting the DA about charging her. A cop in Texas can reach into cameras in Illinois to hunt a woman down for something that isn’t even a crime where she’s standing—same system some suspect’s ex-boyfriend’s cousin’s sheriff buddy can dip into whenever he feels like it.

“Safety device” my ass. This is an outdated, hackable nationwide surveillance dragnet sold to your neighbor’s HOA and local PD and quietly leaned on by criminal federal agents, and the Manitowoc DA is out here doing free PR for it with vague reassurances instead of a single fucking piece of hard data.

Ask her for the audited numbers next time a reporter’s got a mic in front of her. She won’t have any.

That’s the actual story here. The DA’s statement, the self-reported stats, the hackable shit surveillance hardware, all of it—it’s a snapshot of how law enforcement in this country has stopped asking the basic fucking questions: Is it credible? Is it safe? Does it actually work? Is it legal? Is it even constitutional? Does the crime-solving upside even come close to justifying what it costs civilians in privacy, even in public, where cops used to at least pretend that mattered?

None of that gets asked anymore. Departments just take whatever private surveillance tech a corporation is selling this quarter, wholesale, no questions, even when it’s been shown over and over to be bullshit, full of holes and ripe for abuse, because they’re not the ones who pay for it. No. We pay for it. Citizens eat the mass surveillance creeping into every corner of their daily lives, and taxpayers foot the actual bill for the hardware and software, which then turns around and lands right in the pockets of tech CEOs who never had to prove the fucking thing actually works.

And notice something else while we’re at it: Manitowoc’s DA ran for that seat as a Republican, and she’s far from the only one. DA after DA doing this same credulous “trust us, it’s for safety” routine for Flock tends to skew hard in one direction politically. Funny how the party that never shuts up about government overreach goes dead silent the second the overreach comes with a police badge and a corporate sponsor.

Land of the free, home of the brave. What a fucking joke. This is the land of the surveilled, home of the fearful, where the badge doesn’t have to ask a single hard question as long as the invoice clears.

Article link: https://www.seehafernews.com/2026/07/16/manitowoc-county-da-flock-cameras-are-an-important-tool-for-law-enforcement/?fbclid=IwZnRzaATNBa9jbGNrBMeW9GV4dG4DYWVtAjEwAGJyaWQRMXN4VHZxZUZPVFRnbWJpMjBzcnRjBmFwcF9pZBAyMjIwMzkxNzg4MjAwODkyAAEejCNowbrFLMR1dNcZ4VXR4pWIdgJcwIAIdSF2Ael_B65hwvk2ZFt0LnsYcnI_aem_lPI5m6E7V51rdKd3r_Y5Qg

That doesn't mean every poem, prejudice, statute, or identity can be mechanically deduced from a wage ledger. Historical materialism isn't a vending machine, insert "economic base," get a printout explaining consciousness. It means something harder and more useful than that: material organization sets the conditions under which ideas get to become durable, institutions get to acquire force, and some ways of living become normal while others become punishable.

Ideas, language, culture, they all matter. They don't all matter the same amount, from the same position, with the same teeth behind them. An idea backed by a university seminar is not the same thing as an idea backed by a payroll department, a landlord, a police force, a court order, and a bank account. Decorate the distinction in as much French academic jargon as you want. It's still sitting there underneath everything, doing the actual work while you write the dissertation about it.

The failure of the institutionalized postmodern and poststructural politics I'm going after isn't that they notice discourse exists. Materialists notice discourse too. The failure is they keep dissolving structured domination into an endless field of signs, identities, norms, performances, microaggressions, and local negotiations until the basic architecture of power becomes just one more object floating in the soup, no more load-bearing than anything else.

Class power isn't ambient fog. It's ownership protected by law and dependence enforced by the market. It's the fact that one class can withhold housing, medicine, food, employment, credit, and social legitimacy from another class, and can back that withholding, when it needs to, with courts, cops, prisons, borders, and armies.

Power can absolutely operate through institutions, norms, expert knowledge, surveillance, medicine, schools, prisons, and sexuality. Foucault got that much right, and only a lazy materialist pretends he didn't. But none of those mechanisms float free of political economy. A prison has a budget. A hospital has administrators, insurers, creditors, and a labor hierarchy. A school disciplines kids for a labor market waiting on the other end. A border sets the price and the vulnerability of labor. A psychiatric diagnosis can absolutely shape a life, but whether it turns into confinement, medication, unemployment, homelessness, or quiet private treatment comes down to material power. Same diagnosis, different bank account, completely different outcome. That's not a coincidence you can theorize away.

Cataloging power's many mechanisms turns into a refusal to name power's actual command centers. If power is everywhere in the exact same theoretical register, it's nowhere in particular. No ruling class, no owning class, no employer to confront, no landlord to organize against, no state apparatus to seize or dismantle. Just an infinite audit of language, conduct, representation, embodiment, and personal complicity, run forever, resolving nothing.

Domination gets so total, so dispersed, so intimate, that politics collapses into self-surveillance. No factory to strike, no rent to withhold, no boss to name, no police budget to cut, and no property relation to abolish. Just an exhausting review of your own vocabulary and everyone else's, pawned off as though linguistic vigilance were the same thing as changing the world. It isn't.

It doesn’t matter if anybody designed it this way on purpose. The political function is identical either way. Nobody's fist ever actually lands on anything.

This is where liberal feminism's shotgun marriage to postmodern discourse politics becomes fucking useless, and I mean that with precision, not as a drive-by insult.

Liberal feminism was never a theory of abolishing class rule. Its ceiling was always formal equality inside the existing order. Equal access to education, employment, political office, professional authority, military command, and corporate power. Those reforms aren't nothing. A woman shouldn't be legally barred from a profession, denied credit, beaten with the state's blessing, or forced into a pregnancy against her will. I'm not dismissing that. But equal access to domination is not emancipation from domination, and pretending otherwise is how you end up cheering the wrong things.

A woman can become a CEO and still preside over the exploitation of thousands of women beneath her. A woman can command a bomber squadron. A woman can run a prison. A woman can sit on the Supreme Court and strip other women of bodily autonomy on her way out the door. Representation tells you who's occupying the seat. It tells you nothing about what the seat does to everyone underneath it.

Liberal feminism asks how many women are in the boardroom. Materialist feminism asks why the boardroom gets to decide whether thousands of women eat this month.

Marry liberal feminism's institutional ceiling to postmodernism's obsession with discourse and you get a politics with an enormous vocabulary and almost no theory of leverage. It can talk about symbolic violence for six straight hours and say next to fucking nothing about ownership, wages, rent, debt, healthcare, childcare, union power, imperialism, or the state. It can produce elaborate accounts of the trauma caused by language while the woman on the factory floor eats the same wage theft, the same medical debt, the same workplace injury, the same sexual coercion, and the same eviction notice she'd have gotten under a management team with worse pronoun etiquette.

I'm not saying the slur is harmless or the humiliation isn't real or that language can't reproduce contempt. It absolutely can. I'm saying a politics that can't tell the difference between insult and dispossession, between misrecognition and unemployment, between symbolic exclusion and state violence, has lost all sense of proportion. It's confused the experience of domination with the machinery that actually reproduces it.

A slur can wound you. A landlord can make you homeless. A stereotype can deny your humanity. An employer can deny you medicine and let you die of something treatable. A discourse can legitimate violence. A state can carry it out with guns. These things touch each other. They're not unrelated. But they are not interchangeable, and a serious politics has to be able to tell them apart or it's not serious.

Intersectionality, in its original legal sense, named a real problem. People sitting at the crossing of multiple forms of domination can get erased by frameworks that treat each form as a separate lane. Black women can face discrimination that isn't just racism plus sexism stacked like Lego bricks. It's something distinct at the intersection. That insight holds up. I'm not attacking that.

But intersectionality in its gutted, bureaucratic, corporate-onboarding-slideshow form doesn't explain how structures intersect anymore. It catalogs identities, ranks moral authority by perceived injury, and turns political struggle into grievance administration with a comment box. It doesn't unite the oppressed against capital. It sorts them into a hierarchy of vulnerability and hands the sorting job to universities, nonprofits, corporations, foundations, consulting firms, and HR departments, the exact institutions with no interest in anybody actually winning.

Capital has no problem at all learning new terminology. It's got lawyers, consultants, diversity officers, comms teams, and entire departments built to absorb moral criticism and hand it back to you as a mission statement. It can rename the conference rooms, revise the handbook, sponsor Pride, post a land acknowledgment, run a Black History Month campaign, and lay off a thousand workers before lunch on the same day. It can get more representative while getting more exploitative, and those two things won't so much as brush past each other on the way. It can diversify the ruling class faster than the ruled class can organize against it. That's not a glitch. That's the model doing exactly what it was built to do.

Ask the postmodern liberal feminist a direct question: Does capitalism require social hierarchy to function, and can you kill the hierarchy while leaving the system that materially rewards it standing? Watch what happens to the sentence structure.

A materialist answer starts with how labor gets organized. Capital benefits from difference among workers. It needs citizens and migrants, documented and undocumented, paid and unpaid, "skilled" and "unskilled," respectable and disposable. It needs some people assigned more vulnerability than others so they can be paid less, disciplined harder, excluded from protection, and dangled in front of everyone else as a threat: behave, or you'll be replaced by someone with even less to lose than you.

Gender and race aren't prejudices capitalism just happens to tolerate out of politeness. They're social divisions capital reorganizes over and over because divided labor is cheaper to buy and easier to rule than united labor.

None of that means capitalism invented patriarchy or racism or conquest or slavery or human cruelty from a blank page. Nobody serious is claiming that, and if you hear someone arguing against that claim, they're not arguing with me, they're arguing with a strawman they built themselves so they'd have something to beat. Patriarchy predates capitalism. So does religious persecution, ethnic domination, caste, empire, and slavery in about a dozen different forms across human history. The materialist argument is narrower than "capitalism invented oppression" and it's a lot harder to knock down: capitalism inherits older hierarchies, then retools them, industrializes them, and puts them to work generating profit.

It took racial chattel slavery and turned it into a continent-spanning commercial system wired into land, finance, insurance, shipping, and industrial growth. Lloyd's of London underwrote slave ships and slave-dependent commerce. That's not incidental to capitalism's development. That's foundational capital formation.

It built gendered domestic labor straight into the reproduction of the wage system, leaning on women to do an ocean of unpaid cooking, cleaning, caregiving, and emotional labor so a workforce could be reproduced every single day at basically zero cost to the employer. Every worker who shows up to a shift fed, clothed, and emotionally capable of working got that way through labor somebody performed and that employer isn't paying for. It built segmented labor markets so racialized and migrant workers could be super-exploited, excluded from protections, and used as a battering ram against solidarity every time workers got close to uniting. Every time. Not a bug. A maintenance procedure.

It didn't invent every hierarchy it uses. It just figured out how to make every single one of them pay. Call that reductionism if you want. It's not. That's just following the money, the labor, the law, and the incentive to the place they actually lead.

The liberal answer can't follow the argument that far, because liberalism treats existing institutions as neutral containers that just need better people running them. It assumes domination is mostly a problem of exclusion, bias, bad representation, and insufficient recognition. Fix the roster, fix the institution. But some institutions aren't unjust because the wrong people are in charge. They're unjust because of what they're built to do, and no amount of better casting changes the script.

A sweatshop with a diverse management team is still a sweatshop. A prison run entirely by women is still a prison. An occupying army with an inclusive recruitment pipeline is still an occupying army. A corporation with flawless pronoun etiquette in the all-hands meeting can still crush a union drive by lunch and send out a solidarity statement by five.

That's what "architecture" actually means. Some relations are structural because they set the field everything else has to play on. Ownership isn't one identity among many. The wage relation isn't just another discourse you can deconstruct your way out of. The state isn't a style of speech. It's an institutionalized capacity to command resources, direct labor, impose deprivation, and enforce its decisions against anyone who resists, with force if it comes to that.

The postmodern liberal politics I'm going after has no real account of architecture, because its whole method flattens the social world into a plane of contingent relations. Everything's constructed, everything's unstable, everything's negotiated, everything's discourse. Fine. Sure. But some constructions come with deeds attached. Some come with guns, payroll systems, prisons, and armies. Some identities come with stock portfolios and board seats. Some speech acts are court orders with your name on them. Some performances are evictions carried out by a sheriff. Some discourse shows up in a uniform, cuffs you, and puts you in a cell for the night.

Everything being historically constructed doesn't mean everything is equally fragile or equally revisable or equally dependent on language to hold together. A border is constructed. Go try to cross one without papers and tell me how metaphorical it feels with a rifle pointed at you.

This is the gap the theory keeps dodging instead of closing. Raise the material stakes and watch the retreat to language start right on cue. Ask about ownership, get a lecture on essentialism. Ask about class power, get warned off reductionism. Ask who's actually profiting, get told power is more complicated than that. Sure it is. So is a power grid. It still has generators and you can go find them.

Language matters. I've already said that twice and I mean it. Language can absolutely shape who gets hired, housed, believed, diagnosed, arrested, promoted, or killed. Words classify people. Classifications guide institutions. Institutions distribute who lives well and who doesn't. All true.

But language doesn't act alone. It gets teeth when an organization is willing to enforce its categories with something more than a memo. A resume with a Black-sounding name gets fewer callbacks, field experiments have shown that over and over. Documented. Measured. Real. But the fix isn't exhausted by sitting through an implicit-bias seminar once a year. The fix needs enforceable antidiscrimination law, public accountability, union oversight on the hiring floor, transparent hiring data, collective bargaining, and penalties big enough that ignoring the law costs more than obeying it.

A language reform with no enforcement behind it is a values statement laminated to the break room wall of a company that just settled a discrimination suit under an NDA. Words carry weight because institutions choose to make them heavy. Change the words and leave the institution untouched, and the institution learns the new words by Tuesday and keeps doing exactly what it was doing.

This is why "raising awareness" has become the house-favorite politics of every organization with zero intention of ever giving up power. Awareness demands no transfer of ownership. Recognition requires no redistribution. Representation gets granted from the top without touching the structure underneath it. Inclusion can be administered downward like a benefits package. A corporation can apologize in a press release. It cannot apologize its way out of its own structural need to extract profit from the people working for it, because that need isn't a personality flaw. It's the entire business model.

Now the strongest honest objection, the one that actually lands, so I won't dodge it either: “Organize around wages" has a genuinely bad track record when what it really meant was asking racialized and female workers to shut up and wait their turn while a supposedly universal class struggle, defined mostly by white men, got handled first. That criticism is true, and anyone pretending it isn't is either ignorant or lying to you on purpose.

The labor movement's ugliest chapters are the ones where white workers excluded Black workers, defended segregated job ladders, guarded a monopoly on the skilled trades, backed immigration restrictions to protect their own wage, tolerated sexual domination on the shop floor, and treated their own narrow privileges as if they were a victory for the whole working class. Craft unions frequently acted less like a vehicle for universal emancipation and more like a guild protecting its own scarcity, gatekeeping the door shut behind them.

But the lesson from that history isn't "material organizing is naive, discourse is where the real fight is." The lesson is that class isn't some pre-existing unity sitting there waiting to be discovered under a rock. It has to be built, politically, on purpose, across real divisions, and it cannot be built by pretending those divisions are imaginary or that they'll just dissolve once everyone shares a paycheck stub.

Capital doesn't just exploit labor. It organizes labor into unequal strata and trains each stratum to fear the one directly below it. That fear isn't an accident. It's a management tool. Racism and sexism aren't distractions from class struggle that pull focus from the real fight. They're among the primary weapons that prevent class solidarity from forming in the first place.

A labor movement that won't confront racism gets broken through race, every time, with no exceptions in the historical record. A labor movement that won't confront patriarchy just reproduces the unpaid and underpaid labor capital already depends on to keep its costs down. A labor movement that throws migrants overboard hands employers a weapon they'll use against every single worker, migrant or not. A labor movement that treats sexuality, disability, or gender identity as a side issue leaves entire groups exposed to coercion, watches solidarity collapse, and then acts shocked when it does.

The answer was never "class instead of race and gender." The answer is class politics with an actual theory of why race and gender get materially reproduced, and the will to fight them at the level of wages, housing, healthcare, policing, immigration, domestic labor, education, and political power, where the fight actually changes outcomes.

Some of the CIO's biggest wins came exactly when organizers crossed racial lines on purpose, often against brutal opposition from employers, cops, politicians, and plenty of racists sitting inside labor's own ranks. That solidarity didn't show up because everybody quietly stopped talking about race. It showed up because organizers confronted racial division head-on as a weapon the boss was actively using, and took the weapon out of his hand.

That's the actual distinction. Liberalized identity politics isolates oppression into separate representational lanes running parallel to each other, never merging. Materialist solidarity asks how different forms of oppression get produced and maintained by one connected institutional order, and how people standing in different spots inside that order can build enough combined power to actually break it.

Solidarity was never about pretending everyone suffers identically. It means recognizing that different injuries get produced and held in place by connected structures, and nobody wins if capital gets to negotiate with every group separately, one at a time, in a closed room, forever.

This is why the "class reductionism" accusation is dishonest as often as it's honest. Sometimes it names something real: crude materialists reducing every single form of oppression to a wage question and refusing to study the specific texture of race, gender, sexuality, nationality, or disability. That failure exists and it needs calling out when it shows up.

But just as often "reductionism" gets used as a panic button, hit the second anybody tries to establish that some things cause other things more than they get caused by them. Saying material power is foundational isn't the same claim as saying nothing else is real. A building has plumbing, wiring, windows, paint, and people arguing in the kitchen about whose turn it is to do the dishes. Saying it also has a foundation isn't reductionism. It's the reason the building is still standing instead of a pile of rubble.

The actual materialist position was never that identity, language, culture, law, or lived experience are bullshit. That's the strawman opponents keep swinging at because it's a hell of a lot easier to knock down than the real claim, and everyone in the room knows it.

The real claim is smaller and a lot harder to escape. Capital organizes the conditions under which social hierarchy keeps surviving. It inherits older forms of domination and rebuilds them around profit, property, labor discipline, and political control. Language can reinforce those relations, hide them, justify them, or challenge them, but language on its own has never once abolished them. Representation can change who's administering power without changing one thing about what that power does to the people underneath it. A coalition that ignores racism and sexism will fail, on schedule, because capital will find that seam and pry it open every time. And a politics that talks endlessly about harm while refusing to name ownership, class power, and the state isn't radical. It's moral theater staged inside the exact architecture it claims to oppose.

It hands people a vocabulary for describing the room while making it socially unacceptable to ask who owns the building. It turns political struggle into personal purification, institutional change into etiquette training, solidarity into competitive testimony about who's suffered more, and liberation into the right to be more accurately represented by the same people still ruling you.

That's a pressure valve, not emancipation. It lets institutions vent anger without surrendering one inch of authority. It lets corporations sell dissent back to the people they're exploiting, at a markup. It lets professionals feel subversive while cashing checks drawn on the exact machinery they've agreed never to name out loud. It lets everybody get the sensation of fighting while the deeds stay in the same hands and the checks keep clearing on schedule.

A serious emancipatory politics has to fight on more than one front at the same time, because the enemy sure as hell isn't fighting on only one. It has to confront language, because language can and does legitimize domination. It has to confront racism and sexism directly, inside the fight itself, not after it, because divided labor cannot beat concentrated capital, not once, not ever, not in the historical record we actually have. It has to confront the state, because hierarchy gets enforced through law and, when the law isn't enough, through violence. It has to confront ownership, because whoever controls the material means of life sets the terms everyone else is forced to live under, whether they voted for those terms or not.

None of that requires pretending culture is fake. It requires refusing to pretend culture is sovereign, floating free, answerable to nothing underneath it. None of it requires collapsing every oppression down into class. It requires understanding why oppression keeps surviving moral condemnation, year after year, decade after decade, precisely because somebody is still profiting from its survival. And none of it requires dissolving power into an endless field of discourse where nothing is more foundational than anything else. It requires taking power seriously enough to go find where it's actually concentrated, how it reproduces itself, which institutions enforce it day to day, whose interests it serves when the votes are counted and the dust settles, and what organized force is actually capable of breaking it.

Power might be everywhere. Fine. It is not everywhere equally, and pretending otherwise is a convenient way to never have to point at anyone. Some people have opinions. Other people have payrolls, property, prisons, and police.

A politics that can't tell the difference between the two was never equipped to change anything. It was only ever equipped to describe its own defeat in increasingly sophisticated language.

And I never once shut my mouth about the baggage either, so let's not pretend this is new information. The old posts, the Nazi tattoo, and the excuse-making about the tattoo, which honestly turned my stomach worse than the tattoo itself, because a young man’s ignorance is one thing and a grown man's polished alibi for it is another animal entirely. The pattern the women who dated him kept describing, independently, on their own, long before any of this was profitable to say out loud in public. I didn't excuse one word of it then and I sure as hell am not starting now, so get that straight before one more sentence gets read.

This week Politico ran a detailed, on-record account from a named woman, Jenny Racicot, saying Platner forced himself on her in 2021 after showing up drunk at her house. Stack that on top of the May and June reporting. His own wife telling the campaign about explicit texts to other women. The Times piece, multiple women, one of them dropping the word abusive without blinking. Schumer and Gillibrand out within hours telling him to withdraw, moving that fast for a man they never gave one public shit about a week earlier. The Democratic Senatorial Campaign Committee (DSCC) saying flat out they will not fund the race if he stays on the ballot. Khanna, who flew to Maine and stood next to the man and vouched for him with his own name, yanking the endorsement now and calling this one credible. Gallego right behind him doing the same.

Do I believe them? Do I think this is finally the true and credible one, after everything else that's come before it? Honestly, I don't know, and I'll tell you exactly why without flinching from it: the pattern up to now has been smear after smear after smear, debunked or quietly walked back months later once the damage already did its job and everyone's moved on to the next outrage. And if you cry wolf enough goddamn times, you don't protect the village, you disarm it. You build the exact conditions where a real wolf strolls through the front gate while the whole town's too exhausted and cynical to even look up from their phones. That is not an accusation against these women. That is an accusation against a political machine that has poisoned its own credibility so many times over that even the truth now has to fight tooth and nail through the wreckage of every lie that came before it.

So with all of that said, here is where I do not move one inch for anyone: None of this has been tested in a courtroom. None of it carries the kind of evidence that ends an argument instead of just leaning its weight on one side of the scale. I will not hand down a verdict off a fucking news cycle, and I will not pretend a serious allegation is the same thing as a conviction just because saying so would be the easy, safe, applause-line thing to type.

But the volume and the pattern does change the weight. The fact that his own allies are sprinting for the exits like the building's on fire changes the weight. This is not one woman's word floating alone in the wind anymore. It is a stack, built by separate women, across separate years, who never coordinated a single thing and never needed to, and pretending that's noise would be a lie I am not willing to tell you or myself.

I said before that "this looks bad and demands scrutiny" and "I am not convicting him myself" could sit right next to each other without contradiction, and they still can. What's changed is the weight of the scrutiny, and not the standard of proof. And the plain, ugly, political truth underneath all of it is this: the man should not be the nominee, allegations proven or not, because no campaign on this earth survives carrying a stack this tall, and pretending otherwise isn't wise or principled. It's denial.

Now the point underneath the point, the one that actually matters once Graham Platner the individual has been carted off the stage:

Nothing about Platner happens in a vacuum, and anyone telling you it does is selling you something. Four decades plus of Democratic Party leadership made a conscious, calculated bet with a doctrine with an actual name, an actual think tank, and an actual fundraising apparatus behind it, not some vague accidental drift. The bet was this: trade the American working class for the credentialed class and manage capitalism a shade gentler than the Republicans instead of ever laying one single finger on its actual machinery. This was the Democratic Leadership Council's (DLC) whole architecture, the New Democrat rebrand, the theory that labor was a captive vote that could be banked, ignored, and taken for granted forever while the real courtship went toward finance, toward the billionaire donor class, and toward reassuring capital that Democrats were the safe, competent hand on the wheel.

Clinton gutted welfare with his own signature and rammed NAFTA through over the AFL-CIO's screaming objections and tore down Glass-Steagall brick by brick like he was getting paid by the demolition, which, in a sense, he was. Obama inherited an economic collapse caused by exactly that demolition and used the once-in-a-generation mandate he'd been handed to bail out the arsonists while the foreclosures ate entire towns alive whole, root and branch. He had a filibuster-proof Senate, a gift no Democrat had held in a generation, and he spent it on a health law built off a fuckin Heritage Foundation blueprint instead of the public option his own coalition begged and screamed for, because that fight felt riskier to him than the surrender to capital did.

Biden's people talked a decent labor game, and fine, credit where it's due on the National Labor Relations Board (NLRB) appointments, but they never once touched the deeper machinery that lets capital outrun every statute ever built to slow it down, and they let inflation and rent eat that labor talk alive in real time, right in front of the people it was supposed to help, all while those people watched their paychecks evaporate. Every administration in that line filed the New Deal coalition down to a nub and told the people getting filed down there was no other way, that this was simply what a modern economy demanded, and that the only alternative was chaos, when the actual alternative was only ever less profit for the people writing the checks and signing the bills.

That vacuum did not stay empty. Vacuums never fucking do. That's not a metaphor, that's a law as reliable as anything in physics. The far right walked straight into the space the Democratic establishment vacated, took the language of betrayal and abandonment that used to belong entirely to the labor left, and wore it like a stolen coat that fit better than anybody in the party ever expected it would. You do not get a Trump without four decades of a party that lost the vocabulary to explain to working people why their lives kept getting harder while the Dow kept climbing every night on the evening news like some cruel fucking joke. The Overton window did not wander right on its own the way weather wanders. It got shoved, from inside the tent, by people who called themselves the adults in the room and genuinely believed managing decline under capitalism with good manners and a rotating cast of identity politics and culture war talking points was a meaningful avenue for material change instead of the slowest, most polite kind of surrender imaginable. Janet Mills is the finished monument to that instinct: careful, credentialed, and allergic to a real fight the way a vampire's allergic to sunlight and garlic. She is exactly the candidate that vacuum keeps manufacturing on an assembly line, stamped out identical, right up until the day the whole line finally seizes up and breaks down for good and we end up with Trump 2.0 and Project 2025.

Now the left: I agree with Bernie on nearly everything except that he doesn't go nearly far enough, and it still eats at me like acid how much of his coalition runs on a cult of personality instead of a program with bylaws, dues, discipline, and a plan that survives the man once he's gone. By European standards he is a moderate, a Christian Democrat with an accent. I don't say that to insult him. I say it as a diagnosis of how pathetically narrow the American frame has become and how comfortable his people got lounging inside that narrowness, often mistaking the size of a rally crowd for the size of an actual political institution that could survive without him.

The mainstream American left keeps confusing spectacle for structure and keeps mistaking a good chant for a functioning political organization. It keeps making the exact same mistake on a loop like a scratched record: working inside the Democratic Party as a pressure group, begging for a primary slot, praying the next charismatic face finally cracks the door open, instead of building real political power that answers to nobody's permission but its own. This is not new. This is McGovern's coalition, Jackson's Rainbow, Nader's spoiler run, Bernie twice over, and now Platner. Different decade, same rotten skeleton, and same predictable ending all because a movement that refuses to learn its own history is a movement volunteering, gleefully, to repeat it on command every single fucking time the curtain goes up.

Here is the part that actually enrages me, genuinely, down to the bone, because the alternative was never some hypothetical daydream: It exists, it is built, it is functioning, it is dues-paying, and it is almost entirely ignored by the very people who claim, loudly, to want it. The Democratic Socialists of America (DSA) has chapters, elected members, actual bylaws, a platform, and an apparatus that does not evaporate into thin air the second one man's old texts surface online. The Working Families Party runs real fusion candidates across multiple states right now, today, proving the legal mechanism works in practice and isn't just some theory in a pamphlet. The Green Party, Socialist Alternative, and the Party for Socialism and Liberation to name a few, whatever you think of their size at the moment, have platforms with actual teeth on healthcare, labor, housing, and foreign policy. These are not little fan clubs orbiting a personality like moths around a light bulb. They are built to survive exactly this kind of collapse, because the fight to be made was never supposed to hinge on one man's character holding up flawlessly under a news cycle engineered, specifically and deliberately, to find the ugliest, most destructive fracture available in that character.

A serious, disciplined left coalition could have spent the last fifteen years building precisely that kind of independent power in Maine. It could have run its own dull, boring, clean, and credible organizer under its own banner, a campaign that was never a referendum on one man's past because there was an institution behind him bigger than his own biography, absorbing the shock the way a hull absorbs a wave instead of one man's ribcage absorbing the whole ocean alone. Instead, the same pathetic pattern runs every single cycle like a script nobody has the brains or the balls to rewrite: Find one charismatic vessel, pour the whole movement's money, hours, and hope into him personally like you're filling a leaking bucket, treat the Democratic primary like the only arena worth playing in because it's the only one offering a shortcut to relevance, and then act shocked, genuinely fucking shocked, when the vessel cracks wide open, because vetting him properly beforehand would have felt uncomfortable and disloyal, would have felt like handing ammunition to the enemy instead of simply doing your own due diligence on your own candidate before the enemy did it for you, on the front page, at the worst possible fucking moment.

All of this behaves less like a threat to the establishment and more like a loyal, leashed opposition the establishment can absorb, exhaust, or discard the second it becomes inconvenient, which is precisely what is happening in Maine as I sit here typing this. The DSCC didn't need to build one single thing to kill this candidacy. It only needed to wait, because the structure underneath Platner was never built to survive without him, and every person with real institutional power in that party knew it walking in the door, and probably counted on it.

The truth is: a traditional socialist, the kind formed on Marx and Gramsci instead of on cable news panels, wouldn't even bother starting with Platner. They'd start with the base, the material relations of production, and call the whole personality-politics circus a superstructure problem, a symptom being mistaken for the disease itself. Lenin wrote an entire pamphlet on exactly this, “What Is to Be Done,” arguing that spontaneous working class enthusiasm, left to its own devices with no organization behind it, curdles into what he called “trade union consciousness”—a demand for slightly better terms inside the existing system instead of a real challenge to the system itself, and that only a disciplined, organized vanguard could carry that raw energy past the point where capital simply buys it off, or absorbs it, or waits it out. Bernie's coalition, and now Platner's, is close to a textbook case of exactly what Lenin was warning about a century ago: spontaneity with no vanguard, energy with no spine to carry it once the charismatic center inevitably gives out.

Gramsci would call the establishment's forty year project exactly what it is: a war of position, a slow, patient capture of hegemony—not just of state power but of common sense itself, of what even counts as a reasonable thing to say out loud at a dinner table without getting laughed out of the room. The DLC didn't just win elections. It won the entire argument about what was thinkable in the first place, and it did that through media, through think tanks, through university economics departments that trained an entire generation of technocrats to treat austerity as neutral math instead of what it actually is, all of which is a class choice.

A dialectical materialist would tell you the mainstream left's obsession with individual candidates is itself a symptom of having already lost that deeper war, because a class that has lost the argument over common sense keeps reaching for saviors instead of building institution the exact same way a defeated army stops rebuilding its supply lines and starts praying for one single knight in shining armor to show up on a horse.

And they'd all undoubtedly have a hard, unsentimental word for the professional-managerial class running the Democratic establishment too, the Ehrenreich critique, that whole layer of credentialed managers who sell their expertise to capital in exchange for status, comfort, and a nice house, who genuinely experience themselves as progressive because their social politics are liberal, all while their actual material function in society is to administer the system on capital's behalf, filing down its edges just enough to keep the whole thing from snapping. Janet Mills is that class's purest, most polished expression. Not corrupt in the cartoonish, briefcase-full-of-cash sense—just structurally incapable of imagining a political fight that risks her own position, because her entire class formation trained her, from law school onward, to see stability itself as the highest possible good, even when that stability just means slow strangulation for every single person standing beneath her on the ladder.

A dialectical materialist reading of Platner himself would be clinical, and considerably less interested in his personal sins than the cable news cycle is. The material contradiction is this: capital needs legitimacy, and legitimacy periodically requires a plausible tribune of the working class to be paraded out in front of an electorate starving for exactly one, but the party apparatus supplying that tribune has no mechanism whatsoever, none, for actually vetting or disciplining him, because discipline was never the fucking point. The appearance of insurgency was the point. The system does not need Platner to be a good man. It needs him to be a usable symbol right up until the moment he becomes a liability, at which point the same machine that elevated him drops him with the exact same emotional weight it would drop a disappointing quarterly earnings report. That is not a defense of the man, not one syllable of it. It's an indictment of a system that treats human beings, including the women who say he hurt them, as raw material to be processed for somebody else's electoral math.

The DSCC's sudden burst of moral clarity is not moral clarity. It is risk management arriving exactly as fast as the internal polling did and not one goddamn minute sooner. Schumer and Gillibrand did not discover a conscience this week. They discovered a liability and they moved on it with the speed and grace of people protecting a portfolio, because that's exactly what they were doing. Khanna flew to Maine, put his own name and reputation on the line for a man he apparently never bothered to actually investigate, and now wants credit for pulling the endorsement instead of blame for handing it out recklessly in the first place. Don’t be fooled into thinking that is accountability. That is a man simply managing his own personal exposure and slapping the word integrity on it like a fuckin’ bumper sticker.

The mainstream left, the one that's about to spend the next month writing weepy eulogies for this campaign online, needs to sit its own ass down with its own complicity instead of only mourning the casualty. You do not get to pour your entire movement's oxygen into one uninvestigated man and then act betrayed when the investigation happens to him in public, on the front page, instead of by you, in private, months earlier, when it still could have been survivable and quiet. That failure belongs to the coalition every bit as much as it belongs to the candidate himself, and pretending otherwise is just another form of the same cowardice.

The rot runs both ways, and it runs deep enough that nobody in this entire story gets the clean version. The establishment spent forty years manufacturing the exact conditions that made a man like Platner look like the only available vessel for a very real rage that had nowhere sanctioned left to go. The mainstream left spent that same forty years refusing to build anything sturdy enough to survive one flawed man, even with the tools sitting right there in plain sight the entire time, tested, functioning, and ignored on purpose because building them was harder and less exciting than falling in love with a face. Both are failures of nerve. Both are failures of strategy. Neither is a mystery, and neither one gets to point a finger at Graham Platner alone and call the whole account settled and closed.

If the next charismatic vessel gets thrown up two years from now with the same absence of vetting and the same absence of any institutional backbone standing behind him, that will not be an accident either. That will be the exact same choice, made again, in broad fucking daylight, by people who had every single reason in the world to already know better and chose not to.

who exactly am I, underneath all my fronts?

Which is fair, I'll admit, since so little is known,

by design, by mistake, and by choices my own

to never explain, to never make plain,

to answer each question by dodging again.

So here, once and for all, let the record be told,

here's my autobiography, freshly unrolled:

I'm a man of firm principle—that much is true—

my one firm belief is I don't believe in a few.

I'm dogmatically against every dogma in sight,

definitionally against defining what's right,

and I won't wear a label, not one, not at all,

not even the label that says I wear none—

inconsistency's the only thing I keep consistent,

a fact I repeat with a tone rather insistent.

There is one truth I hold, and I hold it quite tight:

that there are no such truths—which I state with delight.

An ideology built on rejecting them all,

a self so transcended it won't stop the call

to remind you, at length, it's transcended the self—

while checking its reflection, alone, on the shelf.

I hold no values dear, not a single darn one,

except for that value, which outranks every one.

And I've found, believe it or not, in the end,

that meaning means nothing—a fact I now sell, market, and defend.

I answer to no one—no boss, and no master—

which makes me the expert at answering to no one, and faster.

I have broken from tradition so cleanly, so wide,

that the breaking itself is tradition, my pride.

I have never once doubted my gift for the doubting—

that's the one thing about me that's never in doubt.

I draw no hard lines, no boundaries, no border,

except round the subject of boundaries and order,

a subject I'll close, right this second, forever.

My loyalty's pledged to disloyalty—clever.

My ambition is simply the absence of drive,

and my policy's this: not one policy survives.

Nothing lasts, I remind you, again and again—

a permanent truth that I permanently defend.

I fit no known category, no bracket, no bin,

which is, rather conveniently, its own category, within.

I keep matters simple, through methods so grand

they'd baffle a watchmaker, confound a whole land.

I know not a thing, yet I know it far better

than you, or your neighbor, or the man who once met her.

My past means nothing—and yet, every night,

I replay every scene of it, frame-perfect, in light.

Entropy wins in the end, understand,

so I alphabetize spices with militant hand.

I've renounced every craving, let all wanting cease—

except wanting you notice my wanting's release.

And silence, I'm certain, speaks louder than sound—

which is why I talk over the silence I've found.

Perception's illusion, I've come to believe,

though I still need you seeing it just how I conceive.

Free will, I'll admit, is entirely real—

until it's your turn, at which point it congeals.

Good and evil are made up, constructed, designed,

and you, oddly enough, chose the bad one to find.

And time is constructed—a fiction, a frame—

but I'll still be furious if you're late to the same.

I'm above material things, I confess,

so don't touch what is mine, don't move it, don't press.

I live for no praise, need no crowd and no cheer—

and yet, did you notice? I'll say it again, dear.

I never need the last word, not a peep, not a sound—

it just tends to land, somehow, in my hand, unbound.

I never lose temper, go on, ask around—

assuming the askable people are still around.

I loathe an admirer, a fan, a follower—

and yet I've collected a great many, and swallow, err,

no, savor the fact. I don't follow a trend,

I was first to the front of not following, my friend.

My door's always open—but knock first, and wait,

then follow with an email confirming the date.

I never make excuses, that much I'll swear—

in my defense, an excuse isn't worth the hot air.

I hold no grudges, none, not a trace—

I just archive them neatly, each name, rank, and face,

with the proper amount of appropriate rage.

I'm not superstitious. I'm a little bit. Turn the page.

I'm a speck in the cosmos, so vastly, so small—

and offended, profoundly, when treated that way at all.

The universe doesn't care if I'm here or I'm gone—

and I take that dismissal as aimed at me, drawn.

I trust simple logic, right up till the day

that logic, the traitor, has something to say.

I practice restraint, in all matters, with pride—

especially restraint, which I reward from inside,

with excess, with plenty, with second helpings applied.

I care nothing for how I'll be measured, or known—

so long as the plaque's spelled correctly in stone.

Compassion wants nothing, expects no return—

while I quietly count every debt, every turn.

My weakness, embraced, on the days it's convenient—

which happens, by chance, to be nearly permanent.

Enlightenment's struck me, not once, but a few—

and vanished again by the morning news.

I've got no true enemies, only the friends

I've outgrown rather loudly, right up to the end.

Envy's beneath me—that's why, I confess,

I track everybody's with practiced finesse.

Forgiveness arrives, swift, complete, and precise,

on a case-by-case basis that won't quite suffice.

I've made peace with dying, sat with it, and stayed,

then diagnosed something new every night, unafraid.

I'm afraid of no fear—not a one I will name,

though the list, kept in private, would put yours to shame.

Courage means acting though frightened, unsure—

and I've streamlined the acting to nearly zero, no cure.

Reassurance's not needed—just bring it around

often, unasked, in writing, and bound.

Flattery's lost on me, I've evolved past its lure—

so be terribly specific, precise, and demure.

Vanity, I see through it, from every degree,

in the kindest of lighting, at forty-five, facing me.

I stopped needing to be right—the very same day

that I was, in fact, right, in every last way.

Winning means nothing—a line I'll repeat

from the podium, holding the trophy, replete.

My integrity's not up for sale, understand—

though I'm always prepared to hear what's on hand.

I question all things, except who taught me to doubt,

who remains, oddly, exempt, left completely without.

I refuse the whole herd, and I walk far behind

a much smaller herd, of my own like-minded kind.

I see through the system—the whole rigged design—

from a plush, cozy seat that the system assigned.

I distrust every bank, every steeple, and still

need one to sign off on my mortgage and bill.

I reject the whole mainstream, and build, with great flair,

an audience explaining exactly, right there, why I don't care.

I'm not one for politics—merely have views,

every hour, on end, on the folks in the news.

I distrust every leader, until they agree—

at which point they're brilliant, a genius, you see.

I don't buy the dream that hard work makes you great—

and I'm chasing that dream at a furious rate.

I see through advertisements, their tricks, and their aim,

in the shirt that they sold me, still bearing their name.

I'll worship no idol, no altar, no throne,

faithfully, weekly, in a chapel my own.

I need no proud flag to feel love for my land,

just fury when somebody else's flies grand.

I've broken completely, entirely, free

from the whole grinding system—and made sure you'd see.

I distrust every business, every brand, every store—

while dressed, head to foot, in the logos they wore.

I treat language with care, as a tool men employ

to control the whole room—and the room I'll enjoy.

I refuse a small part in a plan built too neat,

from the desk that plan gave me, my name, and my seat.

I don't need instruction on how to be living—

just shelves full of unread books, quietly giving

off a glow of potential I'll never quite reach,

a library of wisdom I'll never quite teach.

I've healed every wound from the years I was young—

which still, to this moment, won't answer, unstrung.

I love myself fully, no limits, no end—

mostly by tearing to shreds every friend.

I need no confession, no priest, and no booth—

just strangers, unasked, hearing all of my truth.

I'm free from the judgment of others, complete—

and I comb through their comments each night, indiscreet.

I don't chase the symbols that status implies—

I just know, with precision, what they signify.

And under it all: I don't trust what's objective—

except when it's mine, at which point it's corrective.

I've made peace with not knowing—and still, I insist,

on an answer by five, no delay will persist.

I trust the whole process, its ebb and its flow—

especially the parts I control, high or low.

I don't fear the forgetting—I keep every trace,

of everything, always, in every last place.

I believe truth sets free—once it's cleared me, and then,

it may speak, may proceed, may continue again.

I no longer need watchers, an audience, a crowd—

and I write like they're present, attentive, and loud.

I don't fear the failing—I've readied so well

that I never once meet it, not once, truth to tell.

Every end is beginning, a fresh, open door—

and I still owe that letter from March, nothing more.

I don't think that time mends every wound, every scar—

I think, simply, I've stopped checking how the wounds are.

And that, best I can tell, is the whole of my name—

which is to say I've no notion, and yet I'll proclaim,

with more certainty now than I've held in my life,

that not knowing at all is my compass, my knife.

But look long enough, closely, and see how it bends:

I'm no rare exception, no glitch, no strange end,

no warning shot fired from way out on the fringe.

I'm you, with the houselights turned up past the hinge,

you without the disguise that you carefully wore,

caught mid-contradiction, mid-sentence, mid-door,

and kept talking anyway, on and on, ever more.