The Monastery at the End of the World

A little place for Bhante Sujato to stay, practice, and work on SuttaCentral

the union of the two towers—how to build a government out of ai

Bhante Sujato 25 November 2024

monks and politics

Monks shouldn’t be political, they say.

Like as not, monks who say that have a framed photo of the king and queen elevated near their shrine. Not quite worshiped, exactly. But when they speak of royalty they say things like, “those great beings from the national pantheon”.

The Thai king’s wealth is counted in tens of billions of dollars. Poor people know exactly what they have, down to the last cent, and they have to account for it all. When it comes to the rich, is it thirty billion? Or fifty billion? Or much, much more? Who can say?

Do you really think that kind of money can ever be clean?

It turns out, monks doing politics is fine, if it’s the right sort of politics—elitist, royalist, misogynist, patriarchal. The kind of politics that reinforces monks’ prestige and wealth.

Or another sort of politics is tolerated, barely: the pleasant-left. The kind of politics that thinks in terms of civil-minded people having reasonable conversations about effective solutions. The kind of politics that might have made sense a decade or two ago. The politics of hope and change, of a future that bends towards justice. The politics that built the UN and the EU, that created the international order of human rights and democracy. The kind of politics we were taught at school, that we were meant to aspire to.

You sit at the table with your people on one side, and rapists, thieves, and murderers on the other, and look forward to a productive conversation. And everyone agrees that there are no bad guys, and the important thing is to stay civil and see both sides. Until a wise woman reminds you that every abuser needs an enabler. And you wonder who the enablers are. And you realize with a cold chill: it’s the people who sit at the table.

No, the two sides are not the same. They have never been the same. But table-sitting has left us with the kind of politics where one side systematically murders tens of thousands of children in Gaza, and the other side is really bad.

I cannot help but wonder, “What sort of people are we that have come to this?”

Our understandings no longer make sense, our moral truisms are betrayed as lies. But even in such a tumbling scree of change, something in the recent US election really shocked me.

The union of the two towers.

why we are the way we are

I’m just a monk. What would I know? Not much, really. I’ve no illusion about that.

I know about early Buddhist texts, and I’m pretty confident when I speak on them. They tell us that we should learn to recognize evil; and when we recognize it, to stop doing it. This, then, is an essay about evil, about how to recognize it, and how to stop participating in it.

I know a little something something about my own mind, too. Well, I picked up a thing or two.

I’m a child of the eighties. Most kids my age were doing the thing, getting a career and a job, listening to Wham! I was too cool for that. I listened to The Clash, or The Triffids, or Do Re Mi, or later, Nina Simone and Charles Mingus. And I helped pass out leaflets warning that Rupert Murdoch was a threat to society.

We had moral conviction, but we passed the time chatting in pubs, our minds too blurred by beer and pot, our lives too comfortable, to ever really do anything that would make a difference.

And here we are. I gave up hope long ago. I’m not expecting to change anything, least of all anyone’s mind. So I don’t really think of myself as political. It’s just who I am. Call it a failure of imagination. I just can’t imagine being the kind of person who would live at a time when the whole world was burning and never say or feel anything.

I like a good scalpel, one that cuts so deep it lets out not blood, but light. I write not to persuade but to make sense; for myself, and maybe for someone else.

As a Buddhist monk, I’m surrounded by nice people who like to imagine that other people are nice too. It’s a comforting thought. I like being around people like that.

But I don’t believe them. I don’t trust comforting thoughts. I see evil all around, rising like the tide, rising like the warming oceans. I fear it. It harrows my soul. I don’t have answers or solutions. All I can do is bear witness with my words.

I’ve come across the writings of Sarah Kendzior. She’s the Cassandra of Calamity, a writer whose every sentence slices and dices to get at the truth. At first I thought she was, well, a bit much. Too dark. Then over time I began to realize that pretty much everything she had predicted was coming true.

She’s not an ideological writer. She looks at how people have behaved and assumes they’ll probably keep behaving that way. And she brings the receipts. Bulging binders of them. Dark, blood-stained receipts of how America became a mafia state.

Her politics is one of character. The character of criminals, for the most part, does not change. If we want to see clearly, we must hold on to our own character, our moral clarity, our love and compassion.

This essay leans heavily on her work. I won’t document the details; better just go to the source. She doesn’t paywall her writing.

I want to understand this moment in time, because it will define the place we will all live in.

I want to understand the union of the two towers.

of dark lords past and present

Saruman was the more interesting of the evil lords. Sauron was just a primal ferocity of evil. Saruman was talented and respected, yet fell prey to weaknesses that were almost adorably childish. Power and dominion, and most of all, pride.

Saruman fell into Sauron’s grasp, but it did not happen easily. He was, by Gandalf’s account, the greatest of the wizards, a pure being from the land of Valinor sent on a mission to redeem the fallen world from evil. He descended into inhumanity by being all-too-human, by continuing to long for the things that men like him are not supposed to long for. All people, saints aside, are liable to such weaknesses, but powerful men more than most.

But there was more to it than that. He didn’t just abstractly covet Sauron’s power. He was connected to Sauron through his choice to use a dangerous item of magical technology, the palantir.

The palantir were created by the Numenoreans of old to manage their vast empire. Large crystal seeing stones, they were distributed at key points throughout the realm for the use of trusted high officials. But, like any object of power, they had a fatal flaw. They didn’t just show you an objective view of truth, they enabled a melding of minds. This was no problem when used as intended. But in the hands of an evil being of great power, the palantir enabled a stronger mind to dominate a lesser, and to make them see what they wanted to see. Even a mind as keen as Saruman’s was no match for the will of Sauron.

So far did he fall in his hubris that he even believed Sauron would share power. Saruman was fool enough to believe he could ally with Sauron in the “union of the two towers”. His overreach was to prove his doom.

The 2024 U.S. election also witnessed a union of the two towers: the conmen, rapists, and mafioso of Trump and his circle with the technofascists Elon Musk and Peter Thiel. (I want you to know that when I call these people “rapists” or “fascists”, these terms are not meant as insults. They are simple factual descriptors. We should use the word for the thing.)

Up till now, the world of technology has retained a skepticism about government, leaning towards a libertarian ethos of innovation and progress, in the fervent belief that the future will be built by the unfettered genius of the engineers. Now they are not merely supporting one side, they have joined it decisively, sponsoring the electoral campaign, mentoring the vice-presidential candidate, and even setting up their own government department.

It is no coincidence that the AI surveillance company owned by Peter Thiel is called Palantir. It is a digital analogue of Tolkien’s seeing-stones. As usual, Thiel seems to miss the point, unaware that he is the bad guy. He controls a system that lets governments see what he wants them to see, in the way he wants them to see it. His systems are ostensibly about information and intelligence, but in fact they create a new layer of unreality, so that a nation’s leaders do not have to respond to the actual people who believe themselves to be citizens. All that matters is their digital representation.

Palantir does not operate in isolation. They partner with an AI firm named Anduril, after the elvish sword wielded by Aragorn. One of their signature products is named Anduril Menace, described as a “family of deployable compute and communications systems to instrument the tactical edge for the government’s secure, large-scale data retention and distribution.” And they are linked to a home security company that is actually named Sauron after the all-seeing Eye, the dark lord himself, faithless and accursed. I guess I respect their honesty. They’re not playing pretend any more.

ai weirding the romanian election

For a long time the technocrats have wanted to replace government with AI. That this will happen is an article of faith, a resigned (or hopeful) “when” rather than “if”. We are now far enough down the road to see how they will do it—or try to.

Already governments are primed to accept their machine overlords. They have tasted the sweet, sweet honey of tech dollars. More than that, governments are mostly run by old men whose chief understanding of technology is to realize that they don’t understand it.

Like most everyone else, they are liable to imagine that AI is smarter than them, so they turn to AI to answer their problems. AI is a perfect government adviser, as it immediately offers plausible-sounding words whose factuality is of little concern. AI makes fools sound smart. Worse, it makes fools think they are smart.

Take Romania for example. 18 months ago, they announced a new government advisory AI named Ion, which apparently scanned social media to tell the government about the sentiment of the Romanian people. Of course, what it would actually gather from this is the sentiment of those Romanian people who use social media. That is, insofar as it was reading the words of people at all, as most social media is made by AI.

At the time, Romanian PM Nicolae Ciuca said:

I have the conviction that the use of AI should not be an option but an obligation to make better informed decisions.

A little over a year later and suddenly Romanian elections are dominated by an extreme right figure who popped out of nowhere, unknown to almost anyone until the current election cycle.

Early speculation that he was a Russian plant was soon confirmed. His rise was also boosted by an unholy combination of crypto, “underworld clans”, paid influencers, even “natural medicine and saints”. But the Guardian is surely correct to attribute his rise to the feckless leadership of mainstream parties. Extremism may well be promoted by foreign powers, but it only finds a foothold in weak nations.

His popularity was entirely due to his massive TikTok presence. TikTok was the first social media based on AI, its uncanny recommendations powered by a machine ultimately trained on Chinese social media data. Romanian TikTok is full of bots, and there is no doubt that a massive AI disinformation campaign boosted his rankings.

Both sides used AI, but in very different ways. The incumbent used AI for what it is marketed as, a reliable source of accurate information. The challenger used AI for what it is actually good for, gaming attention and amplifying delusion.

Far from improving governance, the use of AI merely highlighted the incapacity of current leadership, unable to take a step without consulting the oracle. Did the use of AI pave the way for an extreme right takeover? Well, clearly it did not help. It was posited as the solution to problems, but those problems just got worse.

It’s a pattern. The capacity of humans to lead humanity is eroded, and the solution, somehow, is always, “less humanity, more machines”.

democracy is sick and its doctor is a machine

Democracy has been circling the drain for twenty years now. A 2020 study found that:

Between 1974 and 2005, a majority of states became democratic for the first time in history. However, a global democratic recession began in 2006 and has persisted – and deepened – over the past 14 years. Not only have average levels of freedom (or democratic quality) been declining globally and in most parts of the world, but the pace of democratic breakdown accelerated and the number of democratic transitions declined, particularly in the past five years.

The process is ongoing; the Global State of Democracy 2024 study says, “Declines in the quality of democracy continue to outnumber advances. One in four countries is moving forward, while four in nine are worse off”.

“Democracy” is the “rule by the people”. When the people lose faith in humanity, they look to inhumanity for leadership: a brutal dictator, or, even better, a machine more soulless than any psychopath.

Most governments do not have an official AI adviser. But how many of the politicians, their assistants, their assistants’ assistants, the lobbyists, and the policymakers, are using AI to draft their reams of text and policy, to workshop their ideas, to evaluate hypotheses? I’m guessing the answer is somewhere between “a lot” and “pretty much all”.

In doing so, they not only reveal their collapsing faith in humanity, they advance it. No longer do they trust their own judgment or the judgment of the people around them. They would rather listen to the unthoughts of a machine.

“I could never write this well,” they muse, as the machine spits out yet another stream of statistically-determined, plagiarized unthought. They are content to be the editor and checker of machine dribble, serving its fancy, and every day finding that their role becomes less and less necessary.

We have seen a similar process unfold with smartphones. When they were introduced, people started developing new services and capabilities, offering new ways of paying for things, or identifying yourself, or verifying a service. As time went on, more people got phones, and providers started assuming that almost everyone has one. Other methods were neglected, and ultimately only the phone was left. We never really asked to live in a world where you need a phone to read a restaurant menu; it just kind of … happened. And now we throw away 15 million phones every day.

AI slides imperceptibly from “advising” to “deciding”. A student gets a machine to spit out an essay, then, theoretically, uses their critical thinking skills to fact check, evaluate, and rewrite. Except if the machine output is good already, or if they left it to the last minute, or if they never developed critical thinking skills because machines were always there.

AI targets families and children for the Israeli army to murder, then leaves the decision up to the operator. In theory. But there are thousands of targets and no time so they just press “go” every time. Technocrats promise us a glorious AI future, but those innocents, the fathers and brothers, the infants, the mothers and daughters, they have no future. AI decided that for them. It decided death.

AI advice becomes AI decisions, and suddenly the machine decides who gets health care, who gets welfare, who goes to school.

This won’t affect those who make AI. It is not for them. AI is a gun they hold, aimed at the poor. The rich will still have human nannies, human teachers, human doctors and human nurses. The poor will be fobbed off with crappy machines because of “cost-cutting”, because we as a society have decided that they are not worth the time and attention and care of another human being.

A report by Techtonic Justice called “Inescapable AI: The Ways AI Decides How Low-Income People Work, Live, Learn, and Survive”, found that:

The use of artificial intelligence, or AI, by governments, landlords, employers, and other powerful private interests restricts the opportunities of low-income people in every basic aspect of life: at home, at work, in school, at government offices, and within families. AI technologies derive from a lineage of automation and algorithms that have been in use for decades with established patterns of harm to low-income communities. … essentially all 92 million low-income people in the U.S. states—everyone whose income is less than 200 percent of the federal poverty line—have some basic aspect of their lives decided by AI.

there is only one machine

Governments have always had advisers. What has changed? One difference is that the advisers are machines, sure. But it’s more than that. It’s that the adviser is the same machine. There is, in effect, only one machine, and pretty much everything else is just a skin on top.

Every day that passes, we step further into a world where every government is being run by the same machine.

There are two senses in which this is true. Let me expand this a little.

Coding a modern AI model itself is straightforward. The limiting factors are:

  1. Data
  2. Energy
  3. Processors

A large LLM requires the energy of a city. It feeds on enormous fields of servers, housed in giant warehouses, full of expensive processors running at full speed for months at a time. This is not something that can be optimized away; we have been optimizing chips and energy production for a long time now, and we know how they will improve in future.

This costs billions, so it will forever remain in the hands of a few massive corporations. Nor should such capabilities be available to just anyone. The environmental cost of AI is staggering: the energy, the land displacement, the minerals mined for the chips and other infrastructure, and much more.

So long as we continue to allow them to exist, LLMs will be monopolized in the hands of a few corporations. History shows us that technology tends towards monopolization. (Have you tried buying a phone recently? Did you enjoy trying the Firefox OS, or Ubuntu Touch, or Nokia’s Symbian, or Samsung’s Tizen, or Palm’s webOS, or Blackberry OS, or Microsoft’s Windows Mobile? No, you chose between iOS and Android, between a 3.5 trillion dollar company and a 2 trillion dollar company. All those operating systems lost in time, like tears in rain.) This tendency is true even in fields where scale is not needed—how much more so in a field where scale is the defining feature?

Currently there are a few large players in the field, but over the next few years this will be winnowed down as the biggest players strive to outperform their competition the only way they know how, by throwing even more money at the problem.

There is a deeper sense in which even this modest differentiation is an illusion. Far and away the most important factor in determining the capabilities of an LLM is the data. The differences in the way the model is constructed and run are relatively trivial; they affect the manner of operation, not the underlying capabilities. The data used by these LLMs is not public, since AI companies arrogate to themselves the right to take everyone else’s data but not share their own data. But so far as we know, they are basically using all the data that they can get their hands on. There are obviously massive overlaps in the data between the different AIs, which is why the basic capabilities of the major AI models are similar. It cannot be otherwise.

Thus all major LLMs use much of the same underlying data. They are minor variations on a theme. The same pseudointelligence offering the same pseudoadvice to every government in the world. Soon, to every pseudogovernment.

This is how AI is a fundamentally totalitarian technology. It only works because of scale, and only a vanishingly few people can afford the scale to build a modern LLM. Which means the power of advising the world’s governments will increasingly fall in the hands of a few machines, and those machines are in the hands of a few men. The character of those men—whose ranks include narcissists, fascists, misogynists, and abusers—shapes and informs their work at every level.

This is the world as it is. Except for massive and rapid intervention by national governments, what is to stop us going further down this road? And what when a new generation enters the government, today’s students who have been brought up by AI? To them, turning to AI for governance will be as natural as breathing. It’s all they will have ever known, and all they can know. The more AI is allowed in schools, the more children will fail to learn basic skills of critical thinking. This is not a by-product of AI: it is its goal.

what do the ai overlords want?

With the latest US election, the AI overlords Peter Thiel and Elon Musk have gone even further than I would have guessed a few months ago. They will wield a decisive influence on the next administration through their sponsorship of the winning party, and through their proxy, the incoming VP, J.D. Vance.

Musk wants to start his Department of Government Efficiency. Who knows if it will happen, but the outlines of what they want it to do are becoming apparent.

It’s got nothing to do with efficiency, obviously. The idea, it seems, is to put a DOGE official in every government department to make assessments. (Doge is Musk’s favorite crypto. How I loathe the fact that I can’t even talk about this stupidity without repeating one of Musk’s infantile, corrupt jokes.) Then they’ll use an AI to winnow out the “essential” workers and get rid of the rest.

Musk sacked the safety and ethics people at Twitter. With their departure the place was overrun by Nazis and bots, which, it now seems clear, was the plan all along. It’s what they meant by “free speech”. It seems equally clear that the same plan will be applied to the US government. This should be of utmost concern to every person on the planet. After all, Twitter sends messages, but the US sends bombs.

They’ll claim to use AI to replace the “unnecessary” workers. This will hugely impact basic government services, which will, like most things run by AI, fail repeatedly in new and exciting ways.

Meanwhile, with the government “savings” they will grant even more tax cuts to the rich in order to “balance the budget”. The budget will get even further out of balance as money flows into the coffers of the AI overlords. It is widely underappreciated how much of Elon Musk’s wealth is due to his ability to convince governments to give him money.

The benefactors will include other criminals and oligarchs as well. The crypto world invested big in the Trump administration, so they’ll be paid nicely.

Even worse will be the growing role of AI in the military. It’s the same companies, the same players, the same fascists and psychopaths, who are supplying the tech that pilots drones or targets missiles.

As the military becomes more dependent on AI, it will become more dependent on the men running AI. And as multiple governments deploy competing AI systems, they will engage, machine to machine, in an escalating gameplay of one-upmanship, in which genocide will be a mere side effect. The machines won’t care. And their overlords will reassure us that, if only we grant them more power, their machines will make sure the right side wins.

Thus the downward spiral accelerates: government gets worse, so its capacity to improve is undermined, so more AI is used, which makes it worse still. The enshittification of government.

Will there come a time when people get that they have been hoodwinked, and realize that humans must remain forever the agents of human destiny? For some, surely. But I fear it will be too late, and perhaps not even then. Rather, the sentiment will shift to, “Who even needs a government anyway? What has the government ever done for us?” We will replace the very idea of government with a fantasy of machine utopia.

Let us march crisply towards our glorious future, “all watched over by machines of loving grace”. Of course, it won’t be “our” future, exactly. We will be just another inefficiency, soon to be optimized away, or retained only at the pleasure of the machine.

Perhaps the saddest part of all this is that, as far as Musk’s plans go, it’s all a side issue. It’s not that he wants to create a fascist authoritarian state on earth, particularly. What he wants is to colonize Mars. And to that end, as he has said many times, he must get rid of EPA restrictions on his Starship program. That’s why he is doing what he is doing: don’t be fooled.

a functioning government is so ohio

I realize it’s uncool to like normal things, but it is a fact that for centuries people fought and died so that we could enjoy democracy. They bled oceans as humanity blundered on in our violent, messy way, towards a concept of a state that actually works. I’m no great fan of the nation-state idea as such, but since that is what we have, we should recognize that it is quite possible to run a nation in a reasonably okay way. We were doing it long before AI.

Sure there are problems. But the biggest threat to a well-functioning state is always the same: the corrupting power of the rich. States function well when they rein in that power, moderate inequality, and ensure that the benefits of work go to the people. They tend towards autocracy, or fall apart completely, when the power of the rich is unfettered.

The core neoliberal strategy has always been to undermine faith in government, for a well-functioning government acts in the public good, not in the interests of the rich. AI is the perfect instrument to achieve this aim, indeed, to take it to new heights, as it works by undermining faith in humanity itself. No nation became better off by transferring wealth to the rich—but that is exactly what AI is accelerating. We can now envisage a world where there is not just small government, but no government at all. Only the machines matter.

the two towers

The two towers correspond to the two main kinds of authoritarian government in the world today.

Russia is a mafia state, government as organized crime. This is what Trump and his cronies, all of whom are crooks, want to do to the US. We have heard many times of the threat posed by Russia to American democracy. While Russian interference is real and serious, the bigger problem is that the career criminals running the US government want to do the same thing to the US that Putin did to Russia: loot the state assets and hock them off to oligarchs, retaining an external veneer of government only as a shell to guarantee their impunity.

The technical term for this is “kleptocracy”, “rule by thieves”. Personally, I think “rapocracy” is more precise. Trump and his cronies—rapists and enablers of rape, one and all—aim to do to the US what they have done to so many women and girls: throw them down, strip them of dignity, violate their bodies and their self-worth. “They’re nothing, these girls. Just trash,” said Trump’s sex-trafficking partner Ghislaine Maxwell.

This theme was announced by the little nazi Nick Fuentes when he celebrated Trump’s election with a much-reposted message: “your body, my choice.” A few weeks later, he was arrested for beating a woman.

The other kind of authoritarianism is exemplified by China, a single-party surveillance state variously described as techno-authoritarian or techno-nationalist. There, all citizens are monitored and controlled by a single overbearing entity that uses sophisticated digital tools to control every aspect of life.

These two visions appear quite incompatible, yet they are both part of the incoming Trump administration.

Trump himself is a simple man with simple desires. All he wants is what he has always wanted: to lie on a bed of gold, to destroy his enemies, to bask in the adoration of the whole world, and to hurt women. He wants to destroy government, not because of any ideological belief—he is a subliterate moron incapable of abstract thought—but so he can continue to do crime with impunity. Like he’s always done. First investigated by the Department of Justice in 1973, since then he’s made sure that everyone who matters is bought or silenced, a longterm investment that paid off handsomely when the justice system let his multiple prosecutions run out of time so he could become President and put himself permanently beyond the law.

Trump is too small, too stupid and lazy to really destroy the world. Musk has much bigger ambitions and is therefore much more dangerous.

What does Musk want? He wants to colonize Mars as the first step to colonizing the galaxy.

How does he plan to do this? By building a fleet of a thousand Starships.

What does he need in order to achieve this?

More than any of these, he needs a vast and highly skilled industrial manufacturing capacity. Most US manufacturing has, of course, been moved to China, but Tesla is built in the US. He can’t outsource to China, because then he would be under their thumb. Anyway, he’s a white supremacist.

This is where Russia fails, as their manufacturing and engineering is shoddy. So while Trump and his cronies will be looting the US like Putin and his mates, Musk will be emulating China. Not the communist part, just the ruthless industrial powerhouse part. To do this, he needs to drive down wages, remove labor and environmental regulations, ramp up engineering, mechanical, and manufacturing capacity, and build a top-down, unaccountable system of power, where economic activity serves the will of the great men of vision, namely him.

Musk is no genius. But for all his many flaws he has shown a rather extraordinary flair for business. His business strategy is rational and consistent, and his apparently bizarre choices ultimately serve the same end.

He creates businesses whose success on Earth enable him to colonize Mars.

In taking over the government, he is doing the same thing. He doesn’t care about government efficiency, any more than he cares about public transport (hyperloop, Boring Company), or trucks (cybertruck), or renewable energy (SolarCity), or electric cars (Tesla), or payment systems (Paypal). He does these things because they will help him colonize Mars. And AI is the magic key, the palantir, that lets him connect Trump’s criminal cabal with technofascist control.

The two towers will set AI to the task of looting the state, hollowing out the government until only the shell remains, while at the same time, increasing powers of police surveillance and citizen control (that part is Peter Thiel’s job). They will retain only the essential functions of a state: lavishing money on the rich and pain on the poor and undeserving. AI will get rid of those who stand in the way, namely, government employees.

This is where Musk the technofascist and Trump the sadist come together: they both love firing people. Trump’s fame came from his public display of cruelty—his catchword was “you’re fired!” And Musk sacked 80% of the staff at Twitter, driving it into the ground in the name of efficiency.

Behold the union of the two towers: the crooks and the technofascists, the Russian model and the Chinese, wedded by the machine and united in their rapacious contempt for humanity.

The irony is, if Musk succeeds, he will end up destroying both forms of authoritarianism. Climate change will burn the gangsters and the technofascists along with the rest of us. We are NPCs, he is “Imperator of Mars”. He’ll never reach the stars, but he may just kill us all trying.

They tell us the future belongs to the machines.

What they don’t tell us is that the machines answer not to the common good, to reason, or to humanity. They answer only to the wills of their masters. Governance by AI means governance by tech oligarchs. We are sleep-walking into a future of global machine totalitarianism, oohing and aahing over their toys as their dreams play out on the lids of our own serenely closed eyes.

Can we stop it? For a little while, perhaps. So long as governments remain in the hands of humans, we can pass our own laws. The thing about machines, you can turn them off. Every day that passes, I look at the trees and breathe in the air, and a full-on Butlerian Jihad grows more appealing.

Let us not make machines in the likeness of a human mind.

Related Posts