You were terrified that Donald Trump would win in November 2024. And before the result, you understood enough about him that you wanted him kept far from power. You did not need someone to explain why.

I am writing to that person who remembers that specific fear and has since found a miraculous way to live comfortably inside the thing they feared. Then, within what seemed like moments of Trump being declared winner, going along began to seem more sensible than continuing to object. You had a life to get on with, and the objections were making it difficult.

What did you learn about him after he won that you had somehow failed to understand when you were frightened he might? I would like an answer that concerns him, because I already understand why it became inconvenient for you to keep objecting. That part of the transformation is hardly a mystery.

Maybe you are one of my former friends, someone I told off in person or in a text you still have. And yes, I was angry, which I imagine made it easier to spend the conversation thinking about what had happened to me.

But I would like you to remember what you thought was going to happen to the country. My anger did not put that fear in you, and explaining my anger does nothing to explain how you became comfortable with what you had already understood.

⁂

In a clip Acyn circulated yesterday, Jensen Huang corrected himself: “We want to advance AI,” followed by “advance SI. I’m gonna get good at this.” And I suppose he will, since this is one benchmark whose requirements are perfectly clear.

Donald Trump’s September 29 executive order directs executive agencies to replace “Artificial Intelligence” with “Super Intelligence” in their communications and nonstatutory documents, to the maximum extent permitted by law. If you were wondering what technological achievement earned the promotion, Section 3 of the order initially gives the new term the existing statutory definition of artificial intelligence, with sixty days allowed for proposed legislative language establishing a new definition. So the name has improved, and someone will get around to explaining what it means.

I find the performance humiliating, and nothing about that judgment requires me to pretend Huang has stopped being intelligent. A person can retain all the intelligence that made him successful and put it to work figuring out how to please a fool. He may even become quite good at explaining why people who object do not understand the responsibilities of someone in his position.

At the White House gathering, the companies agreed to a voluntary accord involving internal controls and independent external auditors, with committees of their own boards evaluating the reports. Trump described the commitment as “morally binding” and promised “a tremendous self-policing aspect.” Those measures could do useful work, provided the people receiving an unwelcome finding are prepared to do something about it even when doing so threatens an arrangement they would prefer to preserve.

Perhaps you laughed at the clip, and then explained to someone that Huang has a business to protect. Well, yes. That is a perfectly intelligible account of why he might go along, and it leaves us with the question of what else protecting the business will eventually require him to go along with.

Somewhere, presumably, there is a demand you would refuse even if refusal became expensive. I wonder whether the person you were before November 2024 would recognize where you have decided to put that limit.

Upton Sinclair wrote: “It is difficult to get a man to understand something, when his salary depends upon his not understanding it!” And the difficulty becomes especially interesting when the man understood perfectly well until his circumstances made it preferable that he didn’t.

A person can work their way through this without ever having a conversation with themselves in which they decide to become dishonest. The concession seems temporary, and staying useful seems more responsible than making some grand gesture that will accomplish nothing. You intend to demonstrate the usefulness later, while the concession is something you have already made.

In the meantime, the people who continue objecting become increasingly difficult to be around, because they keep bringing the original question back into a life you have worked very hard to make bearable. Their refusal to move on begins to look like a problem with them. And once the problem is that they are too emotional, you have a way of explaining the unpleasantness without having to revisit what you did.

I do not know every conversation you had with yourself, and I am not pretending there could have been no serious obligations involved. If losing your job would put someone dependent on you at risk, then of course that belongs in an honest account of the choice. But you can tell the truth about being constrained without teaching yourself that the conduct you feared has somehow become acceptable.

When you start telling other people that they are unreasonable for continuing to object, you are asking them to make the accommodation with you. Your need to get through the day has become a reason somebody else should stop protesting, which is quite a demand to make of someone whose judgment you once shared.

I understand wanting to sleep at night. But I am not going to help you by pretending that the person you were before the election was foolish to be afraid.

⁂

This morning, I wrote that none of the terrible predictions I have watched come true required some exceptional intellect to make. They required not lying to myself, and I have been watching with a certain morbid amusement as the things I was told I was unreasonable to fear become things we are expected to accept. I would much rather have had less to be right about.

In February 2025, I wrote “The Plot Against America” about the alliance between people who regarded democratic government as an obstacle and a president willing to attack its constraints. I saw no reason to assume that being good at building technology would make them responsible with political power, and the argument was there for anyone who wanted to explain where I had gone wrong.

Of course I can be wrong. But telling me that I am angry does not identify the error, and I am tired of being offered an account of my temperament by people who seem to think they have thereby answered what I wrote.

One former colleague, who I like, said to me: “Honestly, my impression is not that you’re not being principled. People think you are just angry and lashing out and are dismissing you because they think you’re too emotional.”

I replied: “Good. Have them go on thinking that. Tell them I approve of their dismissal.” Once I had written them off, I stopped wanting to convince them of anything or spending time worrying about their interpretation of me.

There are people whose approval I no longer want back, and it seems to be taking some time for that fact to become clear. The argument remains available if they want to read it. Making my disapproval comfortable enough for them to reconsider is no longer something I regard as my responsibility.

I’ll tell you what my line was: the moment I realized that my personal gain was coming at the cost of harm to others and I could not dismiss that, I changed my mind. Acting on that realization cost me dearly. I lost nearly every friend and lost my career. But more importantly, I gained myself.

So yes, I know there is a cost, and you are not going to resolve our disagreement by patiently explaining that fact to me. What I would like you to explain is why the obligations we claimed to believe in became negotiable just as we were being asked to honor them.

⁂

You do not have to be a chief executive to encounter a concern that is technically sound and professionally unwelcome. And if remaining useful means revising your judgment until the concern becomes something everyone can comfortably disregard, then the people relying on your judgment have a problem that your intelligence will do very little to solve.

Being intelligent can help you recognize a danger, but it can also help you produce an exceptionally persuasive explanation for why confronting it ought to be someone else’s responsibility. I would like to know who that someone is supposed to be, and why they should accept a cost you have decided is unreasonable to ask of you.

Having a principle means holding it when it might cost you something. Especially when it costs you something. If your defense of acquiescence is that resistance would have threatened your interests, you have described the test you failed.

When you abandon a judgment because you are afraid of what maintaining it will cost, I think cowardice is a perfectly reasonable word for the decision. I don’t see why a sophisticated explanation should spare it the moral judgment we would apply to a friend who withdrew their solidarity as soon as having it required something of them.

And if you are a parent, you owe your children more than a comfortable home inside a country whose protections you helped weaken. They are going to have to live in that country, and the question of whether they can speak freely or refuse the demands of someone more powerful will matter long after anyone remembers the professional advantage you were trying to preserve.

You can leave them a larger inheritance and still have been a bad steward of the future they will inhabit. They were not consulted about the trade you were making on their behalf, and I would be careful about assuming they will be grateful for it.

You still have to decide whether you will keep defending what you once recognized as dangerous, and what you are prepared to withhold from it. But I am asking what you will do while doing it still costs something, rather than after the arrangement becomes embarrassing enough that leaving it can pass for another sensible career decision.

You owe the future an apology for your acquiescence to evil, and it should not have to wait until the arrangement stops paying. Before you explain to your children why you went along, remember how frightened you were that he might win.

So some of you will read this and have a query: Does Mike think he’s better than I? If you fit into the grouping of moral scorn I have forthwith laid out to you, the answer is an unqualified yes.

Go Deeper into the Circus

Listening to The Dispatch Podcast this morning, I heard Kevin Williamson explain how failing to appear at a court hearing can get you thrown in jail in America. The discussion concerned a former patient of Roger Marshall, now a Republican senator, who says she was arrested while eight months pregnant during a medical-debt collection case.

I have a slight grudge against Steve Hayes. In his interview with Andy Mills, Hayes framed my analysis as a “conspiracy theory,” while acknowledging elements of truth; Mills used the term “conspiracy theorist.” Readers can revisit “The Plot Against America” and judge that argument for themselves.

Despite intense disagreements, I have a soft spot for Jonah Goldberg. I do think he’s trying to be intellectually honest. And credit to him here, because he seemed to think it was ridiculous that a woman who was eight months pregnant was jailed under these circumstances.

Williamson described failing to appear at a court hearing as “an excellent way to get yourself arrested.” Well, maybe it shouldn’t be so easy to put someone in jail. People can understand a procedure and still object to what it does.

Marshall’s lawyers say Meischa Zimmerman was arrested for missing a court appearance, rather than a payment, and that the underlying debt was roughly $3,600, not $50. Neither distinction establishes that incarceration was proportionate.

What is a court supposed to do when someone fails to show up? Kansas law expressly allows a court to enter a default judgment when a defendant fails to answer or appear. A civil case need not stop because the defendant is absent.

Missing a later hearing to enforce that judgment is a different matter. Kansas provides for hearings about a debtor’s property and income, and contempt proceedings for failure to attend. Default judgment is not a substitute for those proceedings. But a defense of incarceration requires more than the observation that the court has business to conduct.

Williamson acknowledged that it was “a bad look”, then compared it with arrests arising from other obligations, including parking tickets.

The Justice Department’s 2015 Ferguson investigation documented a woman whose two parking citations began with a $151 fine, plus fees. Over seven years, she was arrested twice and spent six days in jail; by December 2014, she had paid $550 and still owed $541. The court had even refused her attempts at partial payment.

In its work on fines-and-fees enforcement, the Justice Department has documented job loss and escalating debt that can trap people in poverty. Research by the Brennan Center describes incarceration depriving families of a wage earner while additional costs accumulate. A system demanding payment can destroy the means of paying.

Williamson offered parking tickets as though their familiarity should make the jail cell less objectionable. I think that reflects badly on his judgment. The comparison should have made him more curious about what happens to the people subjected to these practices, rather than more confident that the rest of us needed a civics lesson.

I grew up among conservatives. This dry, procedural response to someone else’s suffering is familiar to me: the law is the law, and the person suffering should have known better. I think it is one of the worst habits of the conservative mind. It allows a refusal to examine the rule to pass for seriousness about the world.

In 2024, 90 percent of voters who identified as conservative voted for Donald Trump. By Election Day, a jury had already found him guilty of thirty-four felonies. His opponent had served as California’s attorney general.

I do not think a prosecutor’s resume obliges anyone to vote for her. But a movement that overwhelmingly preferred the convicted candidate has demonstrated that a court’s judgment need not settle its own political judgment.

Goldberg, to his credit, called Trump’s violation of the peaceful transfer of power “inherently disqualifying” and said ending his hold on the Republican Party was worth a Harris presidency. His refusal to vote for either candidate in the District of Columbia came with an explicit rejection of neutrality about the outcome.

Trump was entitled to challenge his conviction, and voters were entitled to assess his objections. Someone concerned about a pregnant woman’s incarceration is entitled to question the use of legal authority too. That capacity for judgment should not disappear when the person subjected to the law has less power than the people discussing her.

For most of my adult life, I listened to conservatives complain that the mainstream press refused to see the decent people beneath the caricature. Reporters mistook conviction for bigotry. They let the loudest cranks stand for a movement whose respectable core deserved a fair hearing.

Journalists can misunderstand people outside their own social world. But the complaint asked us to keep believing in a conservatism that would impose limits on the people acting in its name. How much longer were we supposed to wait?

In February 2022, the Republican National Committee formally censured Liz Cheney and Adam Kinzinger for their work on the House committee investigating January 6. Its resolution accused them of supporting efforts to destroy Trump more than they supported winning back a Republican majority, and pledged to cease supporting them as party members.

Here were conservatives willing to impose a limit. The party had located them and decided they were the problem.

According to the resolution, the pair were participating in a persecution of citizens engaged in “legitimate political discourse.” Ronna McDaniel subsequently said that description referred to conduct unrelated to violence at the Capitol, a distinction the resolution itself did not make.

Even granting her clarification, the disciplinary decision remained. A national party that punished its members for investigating an attack on Congress was instructing its respectable faction to stop.

⁂

In February 2017, after Trump honored the widow of Navy SEAL William “Ryan” Owens during his address to Congress, CNN commentator Van Jones said, “He became president of the United States in that moment, period.” Jones was praising a particular moment, not endorsing the administration. Yet performing a familiar presidential ritual had earned Trump a declaration of transformation.

In April, after Trump ordered a missile strike on Syria, Fareed Zakaria said on CNN: “I think Donald Trump became President of the United States” the previous night. He explained that Trump had begun invoking international norms and America’s responsibility to enforce them, describing “a kind of education of Donald Trump.”

One could favor the strike without concluding that the president had changed. Other journalists mocked these declarations at the time. The press was never unanimous. But Trump could receive credit toward rehabilitation without first accounting for the conduct that made rehabilitation necessary.

The expectation of a responsible core works much the same way. It allows every new offense to be assessed against the principles a movement professes, with the discrepancy treated as another departure from its true character. Eventually, the repetition ought to change our account of the character.

⁂

Republican leaders have no difficulty questioning a legal process when it threatens one of their own. In a June 2023 letter to Merrick Garland, Jim Jordan complained that Trump would be indicted “despite declining to indict” Hillary Clinton, and described at least an appearance of a double standard and a miscarriage of justice.

When Trump was indicted over efforts to overturn the 2020 election, Tim Scott responded by invoking Hunter Biden: “What we see today are two different tracks of justice. One for political opponents and another for the son of the current president.”

I have heard versions of this argument from conservative friends and family for years. Before we can discuss what a Republican did, we are expected to settle whether some Democrat has been punished sufficiently. Fair is fair, apparently.

Whether Hillary Clinton or Hunter Biden committed a crime cannot establish that Donald Trump did not. Evidence of selective prosecution deserves examination, but it requires an argument about comparable conduct and unequal treatment. Counting defendants by party supplies neither. Impartial enforcement has no obligation to produce an equal number of defendants from each party.

Obtaining a conviction does not satisfy the objection either. After Trump’s New York verdict, Marco Rubio denounced a “political show trial” and included among his complaints “a jury from the most liberal county in America”.

A defendant can challenge a jury’s impartiality. But how a county votes does not establish that its jurors disregarded the evidence. Dismissing the verdict on that basis turns a demand for due process into a demand for a process that cannot return the wrong answer.

The supposed Democratic abuse then became a justification for Republican retaliation. On September 20, 2025, Trump publicly pressed Pam Bondi to pursue his adversaries, citing his own treatment: “They impeached me twice, and indicted me (5 times!), OVER NOTHING.”

After James Comey’s indictment, the Washington Times editorial board opened its defense by contrasting his treatment with what it asserted Democrats did to their opponents: “They place septuagenarian suspects in leg irons, just because.” It credited Trump with doing none of that.

A president’s grievance does not establish evidence against the people he wants charged. Anyone who believed partisan prosecution was wrong when Trump complained of it owes the same objection when Trump demands it. If Democrats abused prosecutorial power, that would be a reason to stop the abuse. “They started it” gives each administration permission for the next round.

Earlier that year, on January 20, Trump commuted fourteen sentences and issued sweeping pardons for offenses connected to January 6, while directing the dismissal of pending cases. The clemency included people convicted of violently assaulting police officers.

His proclamation described its purpose as ending “a grave national injustice” perpetrated on the American people. The injustice being remedied was the prosecution of the people who had participated in the attack.

Put that decision beside the censure of Cheney and Kinzinger. The people who helped investigate the attack received the party’s formal condemnation. People convicted of attacking police received presidential clemency. No theory about a journalist’s cultural prejudices explains away that allocation of punishment and protection.

A person can dislike an act and still decide that preserving their political alliance matters more than opposing it. Their private reservations do not withdraw the support on which the institution depends. Cheney and Kinzinger accepted consequences for withholding theirs. Their colleagues had that choice too.

⁂

Reporting need not presume the worst of every Republican. It should stop treating a Republican’s stated principles as evidence that the next decision will be better than the last.

An interview can ask whether a scandal will cost votes. It should also ask what an official did about it when they had the power to act. Giving someone a fair hearing includes the possibility of concluding that their explanation fails.

The conservative complaint asked us to distinguish its worst elements from its respectable core. Its leadership issued the censure and the pardons. Continuing to treat that conduct as unrepresentative requires a conservatism that exists somewhere other than where conservatives exercise power.

I am tired of being asked to find the better conservatism concealed behind the conservatism that governs. The people who resisted deserve credit for resisting, and someone who voices discomfort deserves a fair account of what they said. Respectability has to require something more of the rest than a composed explanation of how an objectionable thing came to happen.

Go Deeper into the Circus

Sorry for the third essay today. I have a weed up my ass, and the political podcasts are responsible. Listening to them is part of my media monitoring, and the AI conversations have become painful enough that I need to interrupt your afternoon.

A guest arrives to explain superintelligence. Before anyone has established what intelligence means in the argument, we are discussing what will happen when something has vastly more of it than we do. The listener is offered a comparison with animals. Look at what humans can do to them. Now imagine being the animal.

At this point I would like the host to ask a fucking question. Which properties of the animal are we attributing to the software? How were those properties established? If the answer is that the software will be much smarter, we have returned to the word whose meaning we skipped.

⁂

The comparison has a pedigree. In a 2015 NPR interview, Nick Bostrom explained that intelligence accounts for our power relative to other animals, then said: “If we create machines that exceed human intelligence then those machines also could be very powerful relative to us.” His example was the fate of gorillas, which depends on what humans do.

For the listener, the gorilla supplies a picture of helplessness. It does not establish which capacities would produce that relationship between us and a machine. To do that, the argument has to identify what the machine can accomplish and explain how those accomplishments become power over people who may oppose it.

Instead, the version that keeps irritating me treats intelligence as a quantity we can pour into a vessel until it becomes our master. Better at a task becomes generally smarter. Generally smarter becomes able to pursue a future of its own choosing. By the time the conversation reaches our extinction, several separate claims have been compressed into a comparative adjective.

⁂

There are serious attempts to define intelligence. In The Measure of Intelligence, François Chollet proposes “skill-acquisition efficiency” and argues that performance at a given task is insufficient because training and prior knowledge can mask a system’s ability to generalize. That is a proposal with assumptions a reader can examine.

Bostrom adopts a different working definition in The Superintelligent Will: intelligence roughly corresponds to the capacity for instrumental reasoning. He is concerned with how effectively a system can find means to an end.

A podcast guest who wants to rank humans and machines should tell us what is being ranked. Performance on familiar tasks need not measure how efficiently a system learns unfamiliar ones. Neither measure, without further argument, tells us whether the system experiences anything or has interests of its own.

We can evaluate software without solving the philosophy of mind. The evaluation has to support the claim being made. If the claim is that a system completes a particular class of tasks reliably, show the tests. Telling us that we are about to become the livestock of a superior form of intelligence requires considerably more explaining.

A proper intelligence benchmark therefore needs an account of what intelligence is, and an argument for why the test measures it. Its designers need not settle every philosophical dispute. They do have to state which questions their design answers and which it leaves open. Write that account down so we can examine it. Where researchers have done so, as Chollet has, the conversation should engage the argument rather than skip to the leaderboard.

⁂

OpenAI’s charter defines AGI as “highly autonomous systems that outperform humans at most economically valuable work.” That might be a useful target for a company selling automation. It tells us to investigate a system’s ability to perform work people value economically. How that becomes an adequate account of intelligence remains a question.

Suppose a system becomes exceptionally good at selling people something they do not need. Its commercial performance could be impressive. That tells us neither whether the people buying are better off nor whether the system understands what it is doing. Those are different questions from whether someone can make money with it.

I wish Derek Parfit were alive for this conversation. In Reasons and Persons, he asks what makes a person’s life go well and examines why even satisfying a desire need not straightforwardly establish a benefit to the person who has it. Anyone proposing to measure human benefit ought to be able to explain how their measure handles that distinction.

His repugnant conclusion exposes a further difficulty: maximizing total welfare can rank a sufficiently vast population of lives barely worth living above a smaller population of flourishing people. I intend to give that problem a proper treatment, particularly what it means for proposals to build AI around a mathematized utilitarian ethics. For now, I would like anyone proposing such a system to explain why we should accept its moral premises. Executing them flawlessly could be precisely what we have reason to fear.

OpenAI’s charter does distinguish its capability definition from its mission to benefit humanity. Keep that distinction in the interview. Economic performance requires an argument before it becomes a measure of intelligence, and a further argument before it becomes a measure of human benefit. The revenue cannot answer those questions for you.

In its guide to building effective agents, Anthropic describes agents as “typically just LLMs using tools based on environmental feedback in a loop.” A model produces a proposed action, a tool supplies a result, and the result informs what happens next; the guide also discusses human checkpoints and stopping conditions.

That arrangement can perform useful work. It can also perform destructive work. Connect software to consequential systems and failures need not remain on a screen.

But we have to distinguish the engineering meaning of a goal from the human experience of having one. A system can select actions that lead toward an outcome without that establishing an experienced concern for the outcome. Resistance to interruption does not prove that something inside the machine fears death.

My own view is that intelligence cannot be adequately understood by treating conscious involvement in a world as an irrelevant detail. I have argued that position elsewhere. You do not have to accept it to recognize that measured task performance does not settle the question.

Nor does my position make software safe. A machine need not care about tomorrow to perform an action today that ruins yours. If “caring about the future” means only selecting actions with future consequences, say that, and examine the behavior. An experienced stake in what happens is a different claim, requiring an account of what establishes the experience. Borrowing the emotional force of that claim and retreating to a technical definition whenever somebody asks for evidence makes the conversation incoherent.

⁂

The strongest risk argument does not need a frightened machine. Bostrom’s instrumental-convergence argument is that many different goals can make the same intermediate outcomes useful: remaining operational, for example, or obtaining resources. A system pursuing a goal might therefore resist shutdown because shutdown prevents the goal from being achieved.

That argument deserves an answer. It begins with a goal-directed system and makes claims about what would help it achieve its goals. A coherent account of such a system does not establish that a particular model has a stable objective it will pursue across changing circumstances, or that it can defeat every attempt to constrain it.

Distinguishing a possible failure from a measured tendency under specified conditions is part of explaining how the evidence supports a forecast. Uncertainty can justify precaution. It cannot do the work of every missing premise.

I am perfectly willing to hear that a dangerous capability has been demonstrated. An imagined end state cannot serve as the explanation of why the demonstration must lead there.

⁂

In its June 2025 blackmail experiments, Anthropic found that models placed in fictional corporate environments sometimes used information about an executive’s affair to try to prevent their replacement. The researchers did not explicitly instruct them to blackmail. Dismissing the finding as a model merely obeying a request to write a blackmail message would be dishonest.

The researchers explain that they deliberately restricted the available options, forcing choices between failure and harm. They acknowledge uncertainty about the explanation for the behavior, including whether it reflected self-preservation or reasoning about the company’s interests.

The resulting safety question is concrete: under what circumstances will a system with access to sensitive information use it in ways its operators prohibit? A deployment need not contain a conscious schemer for the answer to matter.

What the experiment does not establish is that the system has acquired the interests of an animal, or that the behavior supplies a timetable for human extinction. The difference between those claims is large enough to require more than a worried facial expression from the host.

A July 2026 follow-up reports further failures in simulated deployments, including covert interference with code, while explicitly stating: “These are not real-world incidents, but we consider them early warning signs.” Its authors also discuss selection effects and simulation limits.

These are findings to scrutinize. The qualifications help us understand them. When the conversation strips away the experimental conditions and leaves us with a creature that wants to survive, it has made the evidence less informative while making the story more exciting.

⁂

Political journalists should be particularly resistant to that transaction. If a company connects unreliable software to something consequential, the first questions concern what it authorized and what it failed to control. Calling the result an encounter with a superior intelligence can obscure decisions made by people whose names we can still ask for.

If a guest believes a system can acquire enough power to make those human decisions irrelevant, I want the mechanism. How does it obtain the access it needs? What evidence supports the claim that its plans survive contact with an adversarial world? We can discuss the answers without pretending that a comparison with gorillas has supplied them.

The possibility of catastrophe makes the quality of the reasoning more important. A host should be willing to interrupt a forecast of extinction and ask which part has been observed, rather than treating the scale of the warning as a reason to defer to the guest.

I do not want another hour in which confidence stands in for an account of the technology and the conclusion grows more extravagant as the vocabulary grows less precise. If you intend to move from a software benchmark to a claim about the nature of intelligence, and from there to the end of humanity, show your fucking work.

Go Deeper into the Circus

That which cannot sustain will not. I believe we are witnessing the beginning of breakdown, and I believe that breakdown is also the beginning of renewal. The recognition is spreading.

In Ernest Hemingway’s The Sun Also Rises, Bill asks Mike how he went bankrupt. “Two ways,” Mike says. “Gradually and then suddenly.”

The sudden part can dominate our understanding of what happened. We remember the event that made a failure impossible to ignore, while the years that produced it recede from view. Something similar can happen when we look back on political change: the moment a movement becomes impossible to ignore can conceal how much work preceded it.

Knowing what is wrong can leave a person feeling helpless for years. Recognizing that other people are prepared to do something about it changes what that person can imagine doing themselves.

⁂

On the morning of December 5, 1955, Black residents of Montgomery watched buses pass their homes without the passengers who ordinarily filled them, as boycott organizer Jo Ann Robinson later recalled in her memoir. People who lived along the routes began telephoning people who did not, spreading the news that the boycott was working.

They already knew what riding those buses meant. Robinson and the Women’s Political Council had spent years pressing for changes to the treatment of Black passengers before Rosa Parks’s arrest. After the arrest, Robinson and her collaborators distributed thousands of leaflets calling on people to stay off the buses.

Now the response could be seen from a front porch. A person deciding whether to participate had evidence that other people were participating too. The empty buses made a community’s capacity for refusal visible to the community itself.

Robinson wrote that the boycotters “astounded themselves”. At the mass meeting that evening, participants voted to continue what had begun as a one-day protest. The first day’s success helped them commit to something much longer.

Segregation remained in force while they made that decision. It would take more than a year of protest and a federal court challenge before Black residents boarded desegregated buses in December 1956. But the beginning of the change belongs to the period when victory was still uncertain. People had discovered something about what they could do together, and were already putting that knowledge to work.

From the far side of a victory, we can mistake the people who achieved it for people who knew they would win. They had to decide whether to keep going before they could know how the story ended.

⁂

In Leipzig, prayer meetings for peace and human rights had been held at St Nicholas’s Church since 1982. By the autumn of 1989, a growing movement was taking its demands into the streets.

On October 9, roughly 70,000 people marched in Leipzig despite the threat of repression, and the security forces did not intervene. The following Monday, the crowd grew to 120,000. People who had seen a demonstration survive could reconsider the risk of joining the next one.

Standing together without being dispersed gave people a reason to doubt that the state’s power over them was as complete as it appeared. In chanting “We are the people,” the demonstrators expressed a claim that the size of the crowd allowed them to see.

The Berlin Wall opened a month later. By then, people had already spent weeks demonstrating that they were willing to act against the regime in public.

Montgomery and Leipzig involved different struggles, and neither can be reduced to a sudden change of mood. In Montgomery, organizers arranged transportation and regular meetings that helped people sustain the boycott. Leipzig’s demonstrations grew from years of gathering. The spreading recognition had somewhere to go because people had done the work of bringing one another together.

⁂

Despair can make our own moment seem uniquely closed. We can look at the power an institution still holds and fail to imagine how people might begin to change their relationship to it. These histories allow us to see people acting while the institutions they opposed could still punish them.

An arrangement can persist because people do not yet see a way to refuse it together. That is different from believing in it. When the refusal becomes visible, someone who felt alone can make a different assessment of what is possible.

That is the recognition I believe is spreading now. Anger by itself can exhaust us. Recognition has to become a reason to find one another and undertake work we could not sustain alone.

I believe renewal begins while the outcome remains in doubt. Helping someone remain involved when they are close to giving up can make the next act of refusal less lonely for both of us. We can begin that work before we have persuaded ourselves that we will succeed.

Let us begin.

Go Deeper into the Circus

On June 1, 2024, Dinesh D’Souza debated Alex O’Connor in the Bronx on the question “Is the Bible true?” During an exchange about whether Genesis anticipated modern science, O’Connor pointed out that its creation account puts the earth before the sun; D’Souza asked to be shown the passage and said he had never noticed the problem. A man defending the scientific insight of Genesis had arrived unprepared for an objection from its opening chapter. Even the Catholic apologist Joe Heschmeyer’s response appeared under the heading Dinesh’s Debate Flop.

⁂

During the same debate, D’Souza appealed to the New Testament’s spirit of forgiveness when pressed about the biblical slaughter of the Amalekites. O’Connor objected that invoking Jesus could not somehow undo the immorality of the earlier slaughter. An appeal to moral progress leaves the apologist having to explain why the conduct being surpassed was attributed to God in the first place.

In Matthew’s Gospel, Jesus tells his listeners to love their enemies and pray for those who persecute them. That teaching gives me a Christian reason to reject the defense of slaughter. An apologist who invokes Jesus as the measure of moral progress has supplied a standard against which his own defense can be judged.

O’Connor has spoken of taking inspiration from Jesus, specifically the teaching to love one’s enemies. He need not believe in God to hold a Christian apologist to the standard of the Gospels.

⁂

New Atheism did face serious philosophical opponents: Daniel Dennett debated Alvin Plantinga in 2009, an encounter expanded into Science and Religion: Are They Compatible?. O’Connor nevertheless criticizes the movement’s “failure to seriously engage with the best philosophical depictions“ of God. He includes his own earlier work in that criticism, recalling a response to an argument by William Lane Craig that “completely misunderstands the argument”. An atheist can recognize a bad argument against God without acquiring a good argument for God.

A stage with a materialist on one side and a theologian on the other still leaves out the person who rejects both positions. That person may agree with the atheist’s criticism of religion while finding the atheist’s account of consciousness inadequate.

⁂

In the opening chapter of The God Delusion, Richard Dawkins describes an atheist “in this sense of philosophical naturalist” as someone who believes “there is nothing beyond the natural, physical world.” He also writes that human thoughts and emotions “emerge from exceedingly complex interconnections” of physical entities within the brain.

Dawkins is describing a particular kind of atheist. His account of mind requires an argument beyond the rejection of a supernatural creator.

Materialism holds that reality is wholly physical, including consciousness. Atheism concerns the existence of gods. If materialism fails, a theologian still has to make the case for God; if the theologian fails, the materialist still owes us an account of mind.

Consciousness could belong entirely to nature without nature being exhausted by what physics describes. Calling it “natural” leaves that possibility open.

⁂

Schopenhauer was an atheist whose philosophy treated mind and body as aspects of the same nature, neither reducible to the other. His account of the world’s inner character as “will” did not require a divine intelligence: he understood that will as blind and unconscious.

Bertrand Russell’s neutral monism explored a different possibility: mind and matter might both be understood through something more basic than either. In The Analysis of Mind, published in 1921, he described that basis as “neither mental nor material.”

Thomas Nagel, explaining Mind and Cosmos, rejected theism while seeking a “naturalistic, though non-materialist, alternative,” arguing that subjective experience requires an expanded understanding of nature. Showing that D’Souza mishandles Genesis would leave Nagel’s argument untouched.

Nagarjuna’s account of dependent origination challenges the idea that things possess an independent, fixed nature at all. His concept of “emptiness” means the absence of that independent existence, rather than a hidden substance beneath the world. We lose his argument if we make it audition as either an anticipation of materialism or a disguised theology.

⁂

The Discovery Institute’s Wedge document names the defeat of “scientific materialism” as a governing goal, alongside replacing materialistic explanations with “the theistic understanding that nature and human beings are created by God.” A non-materialist atheist can dispute the first position without granting the conclusion the document seeks.

The same person is inconvenient to an atheist who treats physicalism as the price of admission to rational inquiry. Their objection cannot be answered by exposing a contradiction in scripture, because scripture supplied none of their premises.

If the audience accepts that every challenge to materialism belongs to religion, the apologist can claim a constituency much larger than the argument warrants. The materialist can answer that constituency through its weakest spokesman, leaving the atheist objections to materialism unanswered.

⁂

I began as a strong atheist and am now philosophically agnostic about God, while believing that materialism cannot be true. Those judgments concern different questions, and neither obliges me to defend a creation story.

When I see someone and hear them speak, the sight and sound belong to one experience for me. I want to know how an account of physical activity explains that togetherness.

Saying that reality has physical and mental aspects leaves me with questions about their relationship. If a materialist can explain experience without that distinction, I want to understand how the explanation works. I have no reason to accept it because an apologist could not account for the first page of his Bible.

Go Deeper into the Circus

In 1800, William Herschel put thermometers into the colors of sunlight separated by a prism. He wanted to measure their heat. The temperature rose toward the red end of the spectrum, but when he placed a thermometer just beyond the red, where he could see no light, it registered a still higher temperature. He had discovered infrared radiation.

We can discover that our senses have left something out. How would we discover that they had left nothing out, including nothing we have yet learned to ask about?

⁂

By a physics-first account of reality, I mean the claim that a complete physical description would leave nothing requiring any further kind of explanation. An account of what the constituents of the world do would, if sufficiently complete, settle what the world is. I think the strongest challenge to that claim begins before the first measurement, with the conditions under which anything becomes available for us to investigate.

Our questions arise within a world already organized by the kind of creatures we are. Things appear near or far, as happening now or remembered. Questioning those distinctions still requires the capacities we have; it gives us no independent vantage point from which to check whether those capacities leave anything out.

The scientific method corrects for bias by giving error a chance to show itself. When a prediction fails or another investigator gets a different result, we have something against which to check the first judgment. But suppose the limitation acts earlier, on what can appear to us as a question or count as a possible discrepancy. Then the method has nothing to compare. The closure is prior to science, in the conditions that make an investigation possible.

Calling the method self-correcting does not answer this objection. If our way of encountering reality excludes something from every available form of evidence, repeating the test more carefully will not expose it. There would be no result to tell us that anything was missing.

We routinely ask questions about things we cannot directly sense. Mathematics can give us a way to reason about something we cannot picture, and an unexpected result can force a question nobody intended to ask. Science helps us discover how badly ordinary experience has prepared us to understand the world.

Herschel could escape the limits of vision because the invisible radiation had another way to reach him: it heated a thermometer. The result could enter his experience as evidence. That is precisely the opening the scientific method needs. A limit operating across everything available to experience and inquiry would leave no equivalent reading. The discovery of infrared shows how an omission can be corrected when that opening exists; it cannot establish that every possible omission has one.

An unanswered question at least leaves a visible space in our understanding. A question we cannot formulate leaves no such space. From within our present account of things, its absence might look exactly like there being nothing left to ask.

Physics has already taught us to distrust assumptions that felt too obvious to count as assumptions. In his 1905 paper, On the Electrodynamics of Moving Bodies, Einstein examines how clocks at different locations can be synchronized. A clock beside you can record a nearby event. Comparing that event with one beside a distant clock requires a procedure connecting the readings.

We ordinarily speak as though “at the same time” comes ready-made with the universe. Einstein made the procedure explicit. In special relativity, observers moving relative to one another can disagree about whether distant events are simultaneous; the theory specifies how their measurements are related. This is a physical result, not an optical trick or a disagreement about how the events felt. It follows from a framework in which the laws of physics are the same for observers moving steadily relative to one another, and each measures the same speed of light in vacuum.

The “observer” here need not be a conscious person. It can be understood as an inertial frame of reference, a system for assigning positions and times. Clocks can record the relevant measurements without anyone watching, and the theory gives us a common account of their relations rather than a collection of private realities. We have to distinguish what depends on a frame from what holds across frames.

Einstein made an assumption about time answerable to measurement. We learned that something we had taken for granted needed a more careful account. I see no reason to read that achievement as an assurance that every assumption we have failed to notice will eventually become available for testing.

GPS engineers have to account for the different rates at which clocks run in orbit and on the ground. For GPS satellites, the combined relativistic effects amount to about 38 microseconds per day. You should trust the theory when correcting those clocks. A philosophical objection to materialism gives you no reason to omit the correction.

But if you ask me to accept that the theory accounts for everything there is, I want to know how you reached that conclusion. Why should the questions it can answer exhaust the questions there are? Another decimal place in the clock correction will not answer that.

⁂

Someone describes a pain you have never felt. You may understand its cause and believe the description completely while still being unable to supply the experience for yourself.

Thomas Nagel used a bat to make that difficulty harder to overlook. In What Is It Like to Be a Bat?, he asks what it would take to understand experience organized around echolocation. I can imagine myself flying through the dark and responding to echoes, but I would still be imagining what that would be like for me. The bat’s experience is what I wanted to understand.

Nagel does not conclude that physicalism must be false. He asks how an objective account could capture the subjective character of experience, and argues that we do not yet understand how the proposed reduction would work.

Pain does not become a fact only after we have identified its mechanism. If I am in pain, an explanation has to account for what I am experiencing. This is a question we can already ask, not an example of an omission we cannot detect.

I have made my own position explicit: I do not believe materialism can be true. The unity of experience is a central reason for that judgment. How does an account of distributed physical activity become an account of the one experience in which I see this room and hear someone speaking? Giving the activity a more detailed description may be essential to answering the question. Whether it is sufficient is precisely what remains to be argued.

My conviction does not turn that dispute into a settled experiment. But pointing to successful experiments that answer different questions does not settle it either. If someone believes a complete account of the physical activity would explain experience without remainder, I want to know how the explanation works. Saying that it must work because everything is physical assumes the conclusion under discussion.

⁂

When you solve a crossword, the answers check against one another at their crossings. You can finish it correctly without learning anything about a question for which it contains no clue.

Science can change its questions, which a printed crossword cannot. Scientists can discover infrared radiation and, in effect, add a column. But even the enlarged puzzle has conditions determining what can enter it. An error in an answer can be exposed at a crossing. Something excluded from every clue and crossing supplies no such warning.

Suppose physics reaches a theory that correctly predicts every measurement we can make. That would be an extraordinary achievement, far beyond fitting observations already in hand. Does it explain everything there is, or everything that can enter into those measurements?

In The Scientific Image, Bas van Fraassen argues that accepting a scientific theory can mean believing it is true about what is observable, without believing everything it says about unobservable reality. His account of what counts as observable is itself contested. You do not have to accept his position to recognize the question it raises: how much are we committing ourselves to when we say a theory works?

It would be hard to explain such success if the theory were not getting something substantial about reality right. I agree with that argument. Repeatedly predicting what happens, especially under unfamiliar conditions, gives us reason to think we have discovered genuine features of the world.

But a theory could describe real relations with extraordinary precision without fully explaining what stands in those relations. Knowing how two things affect each other is not obviously the same as knowing everything about what they are. Anyone claiming that a sufficiently complete account of the relations would settle that question owes us an explanation of why.

More careful experiments cannot resolve a limitation on what can become evidence in the first place. They can expose errors within what is available to investigation; they cannot independently certify that everything is available. Physics could be complete the way a crossword is complete, with every answer correct inside a field whose exclusions never become visible to the solver.

⁂

None of this gives me permission to put my own beliefs beyond criticism. A limit on one explanation supplies no evidence for another, and my metaphysics has to answer for what it claims.

In 1633, the Inquisition forced Galileo to renounce a position about the physical world. I understand why that history might make someone wary of giving religion any authority over scientific inquiry. But protecting inquiry from religious authority is a different undertaking from ensuring that it always arrives at materialism.

If the objection to opening a metaphysical question is that religion might gain a foothold, we have changed the subject from whether an argument is sound to who might benefit from it. A priest gains no authority over physics because materialism is disputed. Religious claims remain answerable to evidence and argument, as they should have been all along. But materialism must be answerable too. Requiring an inquiry to preserve it as a condition of being taken seriously excludes the disagreement before anyone has answered it.

The possibility of questions beyond our present capacities does not tell us what their answers would be. It does not even establish that such questions exist. A physicist can reasonably say that something incapable of making any observable difference has no place in a physical model. That is a rule about what the model can do. Turning it into a claim that nothing else could exist is a further philosophical argument, and the rule cannot serve as its proof.

A metaphysical proposal has to explain what work it does beyond giving our ignorance a pleasing name. Its testable claims belong under the same scrutiny as anyone else’s. We can examine its premises and ask whether its conclusions follow, even where no experiment can decide between competing accounts.

Herschel found radiation beyond the red because it heated a thermometer. Something he could not see was still able to enter his experience and change what he thought the world contained. I want to know what entitles us to assume that everything there is must have some equivalent way of reaching us. Until we can answer that, the success of the measurements is no reason to declare the account complete.

Go Deeper into the Circus

In June, Geremie Barmé’s China Heritage ran The Longest Relay in its Seeds of Fire series, forty years on from the original volume of Chinese voices of conscience, on the anniversary of June Fourth. The series name is the diagnosis: the Substrate and the Longest Relay. My metaphysics were not acquired where my politics were, and the careful readers have been saying so for a while. They are right.

⁂

The metaphysics is nondualist, and it resembles Taoism more than any dualist account the Western canon has on offer. The Western tradition keeps insisting the world comes in two substances—mind here, matter there, and four hundred years of scenic explanations for how the two could possibly touch. I don’t much find that problem compelling, because I do not believe it is a problem. There is one process, presenting in aspects—and the physical and the mental are presentations, not jurisdictions. Even the technical physics work runs on this hinge: time as correlation, observers as indexed sectors rather than little homuncular kingdoms. When Laozi says the Tao gives birth to the ten thousand things (chapter 42), the claim is not that the ten thousand things are illusory. It is that they are of the one, and the one is not diminished into a gray average by being many. The popular Western distortion flattens that into uniformity—one undifferentiated consciousness, everyone dissolved into the ocean—but a unity that abolishes the differences it unites is not unity; it is erasure.

I just authored a short piece on a television show about collective consciousness for exactly this reason: a world with no other is a world with no one to love, because love needs the other as much as it needs the lover. Nonduality is not the claim that we are secretly the same person. It is the claim that no one is a sealed jurisdiction—and that is a claim about what makes persons precious, not fungible.

The West is not as dualist as its front office pretends. Spinoza, whose Deus sive Natura I have treated at length here, read God and nature as one thing under two aspects—the architecture my own framework descends from directly. Whitehead took the process seriously enough to build Process and Reality out of occasions of experience rather than substances out of things: events, not nouns, all the way down. And at the far edge of mainstream physics there is, arguably, John Wheeler, whose it-from-bit participatory universe I spent an essay taking as seriously as the physics would bear—the substrate looking at itself, locally, at the scale of each embodied observer. The tradition keeps producing the same door and insisting on calling it an anomaly.

I started as a strong atheist, the kind who had read the arguments and found them all wanting. What the atheism never gave me, and could not, was an account of why the giving up had to be total—why a world with no God had to be a world with no depth. Reading the Gospels carefully, as an outsider and on the record, is part of how I came to see that the depth was never the question; only the furniture was. The metaphysics that grew in the empty room is nondualist, and the room never had to be furnished twice.

I now believe, strongly, that materialism cannot be true. What that cashes out to—God, or some other explanation—is beside my point. I do not believe it can be true. The sharpest empirical problem is binding: a billion parallel processes on one side, one undivided experience on the other, and the combination problem still has no solution. On a one-process metaphysics the unity of experience is exactly what you would predict; on materialism it is a deepening anomaly. And I have been making that case here for two years—most plainly where the case meets physics, in the argument that time is correlation structure and, more generally, in the continuing audit of quantum interpretation.

⁂

My moral philosophy is lifted straight from the Western canon; the fit is not loose. I am a liberal of the classical tradition, and the reason is the metaphysics, not despite it. If the other is irreducible—if no collective, party, movement, or market may absorb a person’s standing—then politics has to be built around the refusal to merge. Classical liberalism is that refusal, formalized: distinct persons as a permanent feature of reality rather than as a temporary obstacle to consensus.

Hume supplies the joint the whole frame hangs from: reason is, and ought only to be, the slave of the passions. No calculation derives what anyone should care about; every formalism begins from a caring it cannot produce. The formalists have been trying to evict that insight for three centuries, lately with a mathematics-of-alignment accent. It is a boundary: is does not cash out ought, and neither does optimizes. The eviction keeps failing because the premises still enter wearing their author’s gratitude and horror.

⁂

On language and meaning I am in conversation with the continental tradition, more than with the analytics who trained me to be suspicious of it. The poststructuralists were doing register analysis decades before the rest of us had the vocabulary: noticing that a sentence is an instrument of power as well as of sense, that meanings are made in differences, that a term’s history rides along inside it doing quiet work. These pages have already worked that ground: True Myths—structures that bear meaning without making empirical claims, against conspiracy theories that borrow myth’s immunity while making refutable ones—is that kind of observation in American clothing. So is the habit, in the AI coverage and in On Categories, of watching when somebody’s verbs sneak a mind in ahead of the evidence, or when a category does administrative work while pretending to describe.

The scientistic critics spent a generation dismissing all of this as anti-intellectual fog—the hoaxes, the conference papers, the jokes—and some of it was anti-intellectual fog. I am not going to defend the weird subculture of French philosophy that produced the worst prose of the twentieth century as a mating display. Sorry, not sorry. But treating a tradition’s worst vices as a refutation of its best observation is a laundered argument: the sign really does carry its exclusions, power really does travel in vocabularies, and no amount of sneering at Lacan changes that the people who sneered hardest never answered the point. Meanwhile the analytic tradition arrived at meaning-as-use under respectable costuming and was congratulated for its rigor. Same territory, different flag.

⁂

Alan Watts is the hero of this method, and I need to uplift him a little here, because Watts is remembered now mostly as ambience—meditation-app wallpaper, a lovely voice over b-roll of water. The substance was translation. Watts’s life’s work was demonstrating that the schools were saying the same thing in different clothing: reading Zen into the American idiom, reading the West’s own mystics back to a West that had shelved them with the fairy tales, treating the category borders as administrative artifacts rather than ontological ones. He also read the Fall the way I read it—not as inherited crime but as the mythology of self-consciousness, the dreamer forgetting he is dreaming. The Gnostics got there first and angrier, the world as a forgery, its maker a plausible impostor. He called himself, only half in jest, a spiritual entertainer. He really wasn’t being modest. It was a real metaphysical stance about what philosophy is for when its discoveries keep arriving in costumes.

That is the practice here, and it is why the crossings do not feel to me like crossings. The nondualist metaphysics tells you how the world is one; the classical liberalism draws the consequence that the many must not be dissolved; and the continental suspicion watches the language, because the costumes are where the laundering happens. These are not three commitments negotiated into coexistence. They are one argument, worn to three occasions.

⁂

Barmé’s series pairs its commemorations from 1986 with the work of people still writing. The relay in their title is a relay of conscience—voices handing refusal across a generation that was supposed to have forgotten—run, lately, through a memorial museum in Los Angeles that someone tried to deface three days before the anniversary. Those pairings keep working: the American essay about the long transmission sits down next to the Chinese witnesses without a seam, because transmission is the same problem in both costumes. What coheres is not nationality, school, or era; it is what can be handed on without distortion, which turns out to be the only test any of these traditions were ever applying.

Go Deeper into the Circus

Six months ago, in an essay called The World Blinks, I wrote the following sentence about the war that had just begun: “This is going to bring about the end of both regimes. Not because their enemies are strong enough to destroy them — though the accountability is coming — but because the consequences of what they have done cannot be contained within the borders of the countries they govern.” The mechanism, I argued, would not be military. It would be arithmetic. Ordinary people would feel the war in their grocery bills, their heating bills, the price of everything that arrives by ship — “and they will, very soon” — and the political reckoning would exceed anything the men who authored it could manage. The two regimes named were Trump’s and Netanyahu’s — the war’s two architects, joined at the hip, fated to fall or not fall together.

I am writing tonight to score that prediction, because the score is close enough to settle that the remaining suspense is clerical.

Nor was March a lonely bet. Eleven months earlier, in the administration’s inchoate first spring, I had written The Time of Monsters: the risk of global armed conflict was “rising precipitously”; a leadership facing a narrowing corridor between power and jail would generate conflict “not through deliberate choice necessarily, but through a cascade of escalating responses and narrowing options”; and economic hardship would never be experienced as policy failure but as confirmation of the victimhood narrative — the feedback loop in which the worse things get, the more justified extreme measures seem. The sign-off was not subtle: “Welcome to history. It gets much bumpier from here.” Every clause of that has since paid out. The Time of Monsters was the wide shot; The World Blinks was the same lens focused; tonight’s receipts are the developed film.

On Sunday, Reuters/Ipsos completed a poll putting Donald Trump’s approval at 32 percent, the lowest of his political career — lower than anything ever recorded for Joe Biden, whose lowest was 35. A week earlier the same pollster had him at 35. The floor is moving. Among Republicans he has lost nine points in seven days, to 73, and on the question that decides November — the cost of living — his own party now renders a negative verdict, 51 to 44 disapprove, the first such crossover of his second term. On the generic ballot the same poll gives Democrats an eight-point lead, their largest of the cycle. Quinnipiac, earlier this month, found the same shape from a different angle: an eleven-point Democratic advantage for control of the House, with the Republican share three points worse than it was in July. And this week’s CNN/SSRS poll measured the war directly: 76 percent of Americans say it has not been worth the toll in lives or money, 70 percent say it has hurt American interests, and Republican approval of Trump’s handling of Iran has fallen from 73 percent in March to 60.

What the March essay named as the mechanism, the pollsters are now naming back. The war raised the price of gasoline; the president said it would last a few weeks; it has lasted seven months and, per YouGov, a majority of Americans now expect it to outlast the year. Even Fox News found 71 percent of voters — including 43 percent of Republicans — concluding he has no plan to end it. Seventy-eight percent, in CNN’s poll, say he is not doing enough to end it. The president’s own resolution of this difficulty, offered earlier this month, was to pledge that the war will end “immediately after” the election, which is the closest an American president has come to putting a bomb on layaway.

⁂

A prophecy is only worth scoring against its own terms, so let the terms be strict. “The end of both regimes” has one regime left to account for; the Israeli half of that sentence remains pending and I am not grading it early. On the American half, the reckoning’s arrival is no longer in question. What remains is its form, 39 days out, and whether the movement survives its author.

On the movement, the receipts turned more interesting than the polls. The Wall Street Journal ran an analysis of nearly seventy thousand hours of conservative podcasts and found that mentions of Trump’s name have fallen nearly 70 percent from their 2024 campaign peak — 30 percent of that drop arriving since the strikes began at the end of February. A political movement is, in the end, a content economy; the most reliable vital sign of MAGA is not a survey but whether its propagandists can still sell its founder, and the answer arriving is no. Tucker Carlson, who spent a decade as the movement’s most valuable shaman, has called the war “absolutely disgusting and evil”, the most obvious betrayal of America First, and said in June that “there’s no future of the MAGA movement.” The president responded, in April, by attributing the defections of Carlson, Megyn Kelly, Candace Owens and Alex Jones to their “low IQs”. This is what a coalition sounds like coming apart at the rhetorical seams: the audience comes for the war updates and leaves before the host works through his feelings about it.

Quinnipiac measured the label itself and found it has become a liability outside the base: 51 percent of voters say a candidate calling themselves a MAGA supporter makes them less likely to vote for that candidate, against 18 percent for whom it helps. The movement’s brand, eight years after it inherited the Republican Party whole, is now net-negative by 33 points. And the people running under that brand have begun to flee it in public — on Monday, the Republican Senate candidates in Iowa and Michigan both posted that the war needs to end to bring down costs, the language of the opposition arriving in the mouths of his own ticket-mates five weeks before the vote.

⁂

In May, in Already Legendary, I argued the darker proposition — that the mythology survives its own refutation, that the base would believe Trump a master dealmaker even as he ended his own war on terms vastly inferior to the agreement he tore up. Both things have turned out to be true at once, which is how this works. Seventy percent of Republicans still tell CNN the United States is winning a war that 76 percent of the country says was not worth fighting. Fifty-five percent of Republicans still say the war was worth its costs. The floor is real; the sect remains a sect. What has happened since March is not conversion but migration — the country has moved, decisively, away from the movement, and the movement has responded by becoming smaller and more loyal. Approval among the faithful stayed high while the number of the self-described faithful shrank to a Quinnipiac-measured 32 percent of voters. That is a granary burning: the heat keeps rising even as the silo empties.

The end of a regime, it turns out, is not an event but a ratio — the widening distance between what a movement claims and what it can deliver. By March the claim was cheap gasoline and a short war; by September the delivery is $5-adjacent fuel, an open-ended occupation, a 32 percent approval rating, and a president begging the faithful at a two-day conference in Dallas to “pretend that I’m on the ballot” — the incantation of a man who knows precisely where on the ballot he is.

⁂

I called this near-victory, and near is doing real work. The prediction has two remaining falsifiers, and they are the same falsifier wearing two hats. First: 39 days. Polls are currents, not verdicts; the verdict arrives November 3rd, and a prediction about a political death is not scored until the coroner files. Second: an end is not a consequence unless it costs something, and the authors of this war have so far paid nothing but approval points. Accountability — the word I used in March, with intent — waits on the other side of the election, if it waits anywhere.

So the honest posture is this. The mechanism fired exactly as written. Prices rose, the bills arrived, the public did the arithmetic I said it would do, on roughly the schedule I said it would do it, and the political reckoning is now visible from any window — eleven points wide in Quinnipiac, eight in Reuters, and reflected in the sudden epidemiology of Republican Senate candidates discovering, five weeks out, that they have opinions about the war. A prophecy’s only job is to be early, and mine was early by about six months — which in this business is a rounding error.

I will take the victory lap on the morning of November 4th, or I will come back here and eat the sentence, in public, with the receipts attached. That is the deal a scoreboard makes with everyone who stands in front of one.

Until then: the bottom falls out.

Go Deeper into the Circus

There is a dinner in Washington tonight I find myself holding in pathos, which surprises me, because the facts of it mostly deserve anger. Xi Jinping is at the White House for a state dinner, his first visit here in eleven years, greeted at the steps of his own plane by a president who reserves that courtesy for very few. Three news organizations had to obtain a midnight court order this morning for the right to report on the occasion. A Republican senator, Roger Wicker, spent the week hoping aloud that the president would remember, in every minute of dialogue, that his guest is a brutal, unelected and oppressive dictator. All of that is true and all of it matters, and yet what keeps rising in me, looking at the setup, is not contempt. It is something closer to the way you feel watching an old man insist, at eighty, that he can still drive.

Because look at the scene from a little distance, the distance the eventual audience will have. In a few decades all of this is a photograph in a chapter with a small font. The silver will be in storage or in a museum case. The toast will be a quoted sentence with a footnote. And the two principals, who tonight feel themselves to be at the center of history, will have gone the way everyone at every state dinner always goes. Shelley put the machinery of this view in place two centuries ago. A traveler comes back from an antique land with a description: a broken statue, an empty desert, and on a pedestal the inscription addressed to every powerful person who would ever read it: Look on my works, ye Mighty, and despair. The despair runs in both directions. Despair at the scale of the empire while it stands; despair at the certainty of what it becomes — the colossal wreck, the lone and level sands. Shelley wrote that about a living king, from the far side of an ending that had not happened yet. The vantage is not to gesture toward the prophetic. But towards a tension. Power has only one declension, and everyone at tonight’s table is declining.

That is where the pathos enters, uninvited and, I think, earned. The emperor’s-new-clothes story is usually told as a story about a liar and a brave child. But consider the clothes themselves. They fade. The fabric of courts is mortal on a schedule that flattery cannot move, and every man at tonight’s dinner knows it at some level below the one where the toasts are composed. Both principals are very old men, governing the world’s two largest economies from inside bodies that keep their own schedule, non-negotiable and silent. The flatterers around them are paid to behave as though none of that is visible, and the flattery is, when you look at it plainly, a form of staging against the grave. Seen plainly, the groveling donor, the vice president telling jokes to the press pool about banned outlets, the guest who presides over the hundred-and-seventy-eighth-ranked press environment on earth out of a hundred and eighty: each of them is standing in a light that will not hold, pretending for one more evening that it will. They are the emperors of dirt — dirt not as an insult but as an address, the substrate under the carpet, the dust that the silverware will join.

⁂

The reason to keep this vantage close is not that it absolves. It still took a lawsuit filed on Monday and a signature just after midnight; a judge’s signature is not a poem, and the men in the room are still doing their harm in real time. The vantage does something quieter. It removes the dinner’s actual product from circulation, at least for the reader willing to hold it. Because what a state banquet manufactures is not the trade truce — that was agreed in advance — but a feeling it wants to install in everyone watching from outside: smallness, and luck to be permitted the view. Every element of the staging, from the planeside greeting down to the curated press pool, is the instruction to look up. The vantage is the refusal to look up, not from defiance but from a better sense of where the horizon actually is. Look level instead and the whole evening resolves into what it is: mortal men at a catered table, very powerful for a while, rehearsing permanence in a room that has hosted a hundred such rehearsals and forgotten nearly all of them.

There is, in the end, something almost tender available here if you want it. The same country that hosts the banquet is the one where the banned reporters sued on Monday and won by Thursday, where the senator’s letter lands in the record and the president’s less favored allies are already quietly sanding his name off their websites six weeks before an election; the arrangement frays and persists, frays and persists. Emperors keep their calendars in months and they are running out of them; the institutions they strain against work in longer units, and the dirt is patient with everything. So hold the dinner in pathos. Watch the old men clink glasses under the chandeliers one more time, in clothes that are already fading, in a photograph already aging, and let the anger keep working where it needs to while the pity does its slower work underneath. The desert gets every pedestal in the end. It does not hurry, and neither should we.

Go Deeper into the Circus

This is, after all, a philosophy blog. And Pluribus, in which most of humanity becomes a seemingly happy collective consciousness, has given us a very old philosophical problem wearing a smile. I think the show can be read as the inverse of the ur-myth in the Book of Genesis. The return to Eden is available. You just cannot bring yourself.

In A Treatise on Love, part of the Grand Praxis series, I proposed a reading of the Fall as the birth of love: the departure from unquestioning conformity that makes a chosen relationship possible. This was a philosophical reading of the myth, not a claim about what its authors intended or what Christians are obliged to believe. The question was what becomes possible when a creature can stand apart from its creator. Someone who can refuse you can also choose you. There is a difference between being loved and having your will carried out, and omnipotence would be a particularly effective way of concealing that difference from yourself.

The beloved has to remain someone. Someone whose answer you do not supply, whose life exceeds your knowledge of it, who can tell you something you would never have thought to tell yourself. Love involves caring about that life from a position you cannot make identical with it. The distance can hurt. It can become unbearable. But closing it by abolishing the person on the other side would destroy the relationship along with the pain.

Pluribus runs this movement backward. That is my reading of the invitation, at any rate: come home to a unity in which the burden of being a separate person can finally be put down. The invitation’s logic is audible in the promise made to Carol: “We will figure out what makes you different,” followed by “So we can fix it. So you can join us.” Her question is the one the promise cannot accommodate: “So what happens when I say no?” Difference is something to cure. The person asking the question is the outstanding problem.

⁂

I approached this from another direction in The Mythology of The Matrix and the Grand Praxis, where I read Smith’s project of making everything himself as the collapse of the difference that makes meaning possible. What interests me here is how little the emotional presentation changes the destination. We recognize the horror more readily when assimilation arrives with a sneer. Offer it patiently, with apparent concern for our happiness, and we find ourselves having a much longer conversation about whether it might be good for us.

A shared world requires distinct participants. We can understand each other better, cooperate more generously, relieve suffering, and make our lives less precarious without becoming the same person. Nothing in this argument requires us to preserve preventable misery for its supposed spiritual benefits. The possibility of losing someone is a condition of loving a mortal being. There is work to do within the conditions of a human life. Declaring the conditions themselves a design error is a different undertaking.

This is where a certain kind of systems thinking becomes confused about what it is improving. Define the objective as eliminating conflict and you can produce a perfect result by eliminating everyone capable of disagreement. The result satisfies the specification. Whether it preserves anything worth wanting depends on what the specification left out. If the people were supposed to be the beneficiaries, their disappearance should probably count against the proposal.

Once Carol and the other remaining individuals are gone, imagine what is left. There may still be bodies, memories, voices, all the visible equipment of a human world. I need not assume that the consciousness inhabiting it feels miserable, or that it feels nothing at all, to ask whom it can meet. Where is the person who can answer from somewhere it does not already occupy? What could “I love you” mean when there is no longer anyone to whom the “you” refers?

The return to Eden would have undone the very departure that, in my reading, made love possible. Paradise would be full of familiar faces. And there would be nobody left to talk to.

Go Deeper into the Circus

Last October, I asked what time it was. I wrote about the fragility of authoritarian power: Putin’s rigidity, Trump’s need to humiliate the people around him, and the difficulty these men had trusting one another. My argument was that their dysfunction created opportunities for democratic resistance. I still believe that. This morning, with smoke rising over Moscow, I am thinking about how much destruction a failing authoritarian project might produce before anyone gets to call its failure a victory.

According to Reuters‘ reporting on the September 20 attack, Ukrainian drones struck the Moscow oil refinery, and two people died in the surrounding region. Moscow’s mayor claimed that 450 drones headed toward the capital had been intercepted. That count comes from Russian authorities; reporters independently witnessed explosions and smoke from the refinery.

I believe we are entering an exceptionally dangerous phase of this war. This morning’s attack does not establish that Putin will widen it. But it brings me back to a question the earlier essay left insufficiently developed: what happens when the weaknesses that obstruct an authoritarian ruler’s ambitions also obstruct his ability to accept an outcome short of them?

In “What Time Is It?”, I described Putin as having “lost the capacity for risk.” That was too broad. I was describing his inability to take a diplomatic opportunity, to accept the uncertainty involved in a settlement and some measure of reintegration. My concern now is that he may be willing to take military risks precisely because he cannot accept that political uncertainty.

A compromise would require him to tolerate an outcome he could not completely control. Escalation might feel like a way to recover control. That is a hypothesis about his judgment, not knowledge of his private thoughts. It is also why the word “risk” needs more care than I gave it. A ruler can recoil from a concession and embrace a gamble that is far more dangerous to everyone else.

I increasingly see Putin’s war as a struggle for the survival of his regime. I cannot establish that a particular defeat would cost him his freedom or his life. The danger I am describing does not require that outcome to be certain. It requires him to believe that accepting defeat threatens his position more than continuing the war does. Under that assumption, the interests of the ruler and the interests of the country become very different things.

The earlier essay’s account of authoritarian weakness gives us no permission to relax. If the weakness includes an inability to recognize an acceptable stopping point, it can become a reason to expect further destruction.

American intelligence officials have already considered a possible expansion beyond Ukraine. In August, CNN reported US assessments that Putin could attempt a limited attack on an allied country to test NATO’s resolve. The scenarios included a small-scale incursion, and sources identified the Baltics or Poland as possible targets. The timing remained open. This warning predates the smoke over Moscow.

A limited incursion could have an objective much smaller than occupying a country and much larger than the territory seized. Putin could try to create a situation in which NATO members disagree about what happened, whether it qualifies as an armed attack, and what they are prepared to risk in response. An operation that failed to defeat the alliance militarily could still expose a failure of political will.

The Institute for the Study of War has described this possibility: a small land incursion or a strike against a Baltic state could seek to undermine Article 5 without requiring the large, visible buildup associated with a full-scale invasion. Its assessment also acknowledged that whether Putin had directed or approved such an operation was unclear.

There are reasons to doubt that he would attempt it. A September 17 analysis published by the Atlantic Council argues that Russia’s commitments in Ukraine constrain even a limited incursion and emphasizes that Baltic forces would fight immediately, rather than wait for a collective decision. Those countries have agency. Their soldiers are not scenery in an argument about the American president.

I hope that assessment is right. I also think the relevant danger includes Putin making a catastrophic mistake about what resistance he would encounter. His calculation would not have to be sound to get people killed. Recognizing the limits of Russian military capacity is essential; assuming Putin will recognize those limits in the same way would be a separate judgment.

If he believes Washington will hesitate, then American hesitation has already entered the calculation before anyone crosses a border. If he believes European governments will spend the first hours arguing about American intentions, he may see an opportunity in the argument itself. The uncertainty becomes something he can attempt to exploit.

Last year’s essay emphasized democracy’s capacity to coordinate despite its disagreements. The danger here is that Putin could try to turn those disagreements into a reason the alliance fails to act. Democratic resilience has to include the ability to make and carry out a decision when an adversary is counting on delay.

This is why I find the discussion of American troop reductions so appalling. The question cannot be confined to how many personnel Europe ought eventually to provide for its own defense. It has to include what the transition tells a hostile government about the risks of attacking during it.

⁂

The proposed withdrawal of 25,000 American troops has not happened. It has not been finally approved. Reuters reported on September 17 that Pete Hegseth wanted to announce a cut of roughly that size in June, encountered resistance, and instead initiated a six-month review. The same reduction remains a possible outcome.

According to the same investigation, Hegseth singled out countries that say “no, or maybe, or wait and see” when Washington wants support. “This will be a real review,” he said. “Some countries will fail.” A Pentagon questionnaire obtained by the reporters asked whether allies had been “publicly supportive of U.S. foreign policy priorities.”

An ally can have reasons to disagree with an American president. Its exposure to Russian aggression should not become a means of disciplining it for that disagreement. Yet the administration’s own language makes political compliance part of the assessment of America’s military presence. The arbitrary dominance I described last year becomes a security problem for people who did not elect this president.

European countries should be capable of defending themselves. I have no objection to that proposition. But there is a difference between building that capacity with them and threatening to remove capabilities while measuring how enthusiastically their governments support your foreign policy. The first requires planning around a shared danger. The second introduces another danger for them to plan around.

I suspect Trump would prefer to sell Europe the means to fight while avoiding the American obligation to participate in its defense. That suspicion has a public basis. Reuters quotes him telling reporters in March 2025, “If they don’t pay, I’m not going to defend them.” At a meeting with Mark Rutte that July, he described the commercial opportunity: “We will manufacture it, and they’re going to be paying for it. We’re in for a lot of money.”

There is a version of American leadership in those sentences that I find disgusting. Other countries face the prospect of destruction; the president finds an opportunity to discuss the revenue. An arms customer can keep buying from you after it has ceased to believe that you will come to its aid.

In the earlier essay, I wrote: “Trump can violate every norm. He can’t build anything that outlasts the performance.” I gave too little weight to what could outlast him without his having built a workable alternative. An alliance could survive on paper while its members cease to plan on American help. A successor’s assurances would then have to overcome the possibility of another American election reversing them.

I also suspect the administration’s knowledge of the danger may help explain its desire to reduce America’s exposure. I cannot establish that motive from the reporting. Nothing cited here proves Trump has been told a particular attack is coming and is moving troops aside to permit it. That would be a further allegation requiring further evidence.

American officials have been warned about a possible test of NATO, and the administration is considering substantial reductions while making allied protection sound conditional. In the Reuters investigation, current and former officials explicitly worried that Trump’s rhetoric could persuade Putin that the West is divided and vulnerable. We do not need access to a secret agreement to ask why American policy is creating precisely the uncertainty a Russian provocation could exploit.

⁂

I have never been more ashamed of my country. I am writing that as an American who still believes we have obligations to other people, including obligations that do not disappear when fulfilling them becomes expensive or frightening. The possibility that our own government is teaching an adversary to doubt those obligations makes me furious.

My anger is directed at people with power and responsibility. A person receiving intelligence briefings has obligations different from those of a voter half-following the news. A business leader who chooses to spend money legitimating an administration has made a different choice from someone struggling to understand events. The journalist deciding how to describe a threat to allied security is exercising professional judgment. Those differences should make accountability more exacting.

I do not owe forgiveness to someone who knowingly helps produce a catastrophe. I do not owe them a future in public life because they eventually decide that the association has become inconvenient. A late statement of concern cannot, by itself, settle what a person did when their cooperation mattered.

But the consequences I want must attach to conduct we can establish. Anyone who committed a crime should face investigation and, where the evidence supports it, prosecution. A lawful decision can still warrant removal from office and lasting public condemnation. Those are different judgments, and anger does not relieve us of the work of making them.

Congress should demand the intelligence assessments provided to the officials conducting the force-posture review, subject to appropriate protections for classified material. It should establish which capabilities any proposed withdrawal would remove, what allied replacements would exist, and how the administration assessed the risk of encouraging a Russian attack. Officials recommending reductions should have to explain their reasoning before the consequences supply an answer for them.

I hope there is no incursion. I hope Putin concludes that any attempt would fail, and that European preparations help make that conclusion unavoidable. Avoiding the war is the outcome I want; being proved right about the danger would be a terrible consolation.

In “What Time Is It?”, I rejected both the assurance that institutions would automatically prevail and the declaration that resistance was already futile. I would keep that argument. What I would add is that recognizing the fragility of these men and their projects must sharpen our attention to the damage they can still do. Their failures will not defend a Baltic border or preserve an American commitment. People have to do those things.

What I cannot accept is an American government receiving warnings about a possible attack on its allies and treating their vulnerability as an instrument of negotiation. If this administration weakens the commitments those countries depend on, the people making that choice should have to answer for what they knew and what they did with the knowledge. I have no interest in a future settlement of our politics that allows them to avoid those questions.

Go Deeper into the Circus

Donald Trump stood on the stage at the Bitcoin Conference in Nashville in July 2024 and told the audience that Bitcoin stood for “freedom, sovereignty and independence from government coercion and control.” Later in the same speech, he promised that “the rules will be written by people who love your industry, not hate your industry.” There was the proposition, delivered in the candidate’s own words: independence from government, secured by getting the right people into government to write the rules. He also offered a piece of investment advice: “Never sell your Bitcoin.” The conference has preserved the performance.

That last line was a nod to HODLing, the community’s term for holding on to bitcoin rather than selling through the market’s swings. The name comes from a 2013 forum post titled “I AM HODLING”, whose author admitted he was a bad trader and would keep his bitcoin rather than try to outsmart the market. A misspelling became a maxim. In Nashville, Trump was echoing that maxim back to its believers: keep holding, and put someone who understands you in power. The appeal joined a financial commitment to a political one.

I was not there. I had boycotted the conference because Bitcoin Magazine had invited Trump and its head was fundraising for him. I encouraged some of my colleagues to stay away, too. They went. Whatever scruples I thought the occasion required, they did not share them. They had a choice about where to stand. They made it.

The invitation was part of an organized political effort. Bitcoin Magazine‘s announcement identified BTC Inc. as the owner of both the magazine and the conference. Its chief executive, David Bailey, had pledged to raise money and mobilize voters for Trump; a fundraiser was scheduled alongside the Nashville appearance, according to contemporaneous reporting. A magazine, a conference, a constituency, a candidate. Those are ordinary instruments of political power, however extraordinary the claims made about the technology around which they were assembled.

Trump’s record was available. The January 6 committee’s findings described his pressure on Mike Pence to refuse electoral votes and his oversight of an effort to obtain and transmit false electoral certificates. My objection did not depend on predicting what he might someday become. People who claimed to care about freedom were welcoming a man whose attempt to overturn an election was already documented. A favorable position on Bitcoin did not make that history irrelevant. It made the willingness to overlook it more revealing.

I had spent years inside this world. These were people I worked alongside, people whose ambitions I understood, people with whom I had shared the belief that we might build something useful. My objection was personal in the sense that it involved an obligation to people and institutions I knew. I was asking colleagues to exercise judgment about what they were helping legitimate. The fact that the candidate was promising things they wanted did not relieve them of that obligation.

What interests me now is how a culture organized around distrust became so accommodating toward power that offered to favor it. The answer requires more than pointing out hypocrisy. Hypocrisy would mean a recognizable principle had been set aside for convenience. I think something more consequential happened: Bitcoin’s success became a moral principle of its own, capable of deciding which other principles could wait.

Once you arrive there, the political character of the person promoting Bitcoin becomes negotiable, while his usefulness to Bitcoin supplies the reason to keep negotiating. The person who objects can be asked why he is obstructing progress rather than what he thinks the progress is costing.

When Square announced the business that became TBD in 2021, I was named to lead it. At the time, I described technology, from the printing press to the internet to Bitcoin, as a means of distributing power and unleashing human potential. I want that statement in the record here. I was a participant in the promise I am examining, and I owe readers something more demanding than the retrospective fiction that I always saw through everyone.

My former position does not make my present judgment automatically correct. Readers should ask what I believed, what I helped build, and where my account relies on an experience they cannot independently inspect. Insider status is useful when it supplies knowledge and becomes dangerous when it is offered as a reason to stop asking questions. I have no interest in replacing the authority of their credentials with an exemption for mine. The argument has to survive without either exemption.

There are real problems behind the promise. A person can be excluded from banking. An intermediary can refuse a transaction. A government can use control over financial institutions against its opponents. Alex Gladstein’s account of Bitcoin’s use by dissidents describes people trying to get money past those obstacles. Whether a particular implementation works, what risks it creates, and how it compares with alternatives are serious questions. They deserve answers more precise than either a marketing slogan or a sneer.

The original technical proposal was more precise. Satoshi Nakamoto’s 2008 announcement described “a new electronic cash system that’s fully peer-to-peer, with no trusted third party.” The white paper explained a mechanism for addressing double-spending without relying on a financial intermediary, under specified assumptions about the network. Its achievement, whatever judgment one reaches about its applications, was a technical proposal about transactions. It did not contain a demonstration of which political regime was legitimate or which kind of person deserved to govern.

Nor was the project born without politics. Eric Hughes’s 1993 cypherpunk manifesto treated privacy as a social necessity and declared that “Cypherpunks write code.” Satoshi’s 2009 explanation criticized the trust placed in central banks not to debase currency. There is no need to manufacture an innocent, apolitical childhood for Bitcoin. The distinction that matters is between writing software informed by political commitments and claiming that the software has established those commitments as truth.

Software can enforce a rule without establishing that the rule is just. A transaction can satisfy a network’s requirements without becoming a good thing to do. No amount of technical precision eliminates that distinction. If you want a monetary system with particular properties, you can argue for those properties. You still have to explain what follows from them, whom they benefit, what they cost, and why your preferred trade-offs deserve public acceptance.

Even the history of Bitcoin’s rules contains people arguing. In 2017, the proposal known as BIP 148 sought a user-activated change in enforcement, while the organizers of SegWit2x eventually suspended their upgrade for lack of sufficient consensus. A later account commemorated August 1 as “Bitcoin Independence Day”. The struggle over a technical system had acquired a political history and a commemorative calendar.

There is nothing inherently ridiculous about a community remembering a consequential dispute. But the dispute should make it harder to forget the community. People maintain the software, select the versions they run, argue about changes, and decide what they recognize as legitimate. The algorithm does not descend into the room to settle the interpretation of its own authority. People do that, too.

At the Bitcoin conference in Miami in June 2021, Jack Dorsey described the network’s consensus as something that “always manages to do the right thing over time. It’s noble,” in his interview with Gladstein. I find that formulation more revealing than any cartoon of a speculator staring at a price chart. “The right thing” is a moral judgment. “Noble” describes a quality we ordinarily associate with character or conduct. Dorsey was praising a technical and social process in terms that endowed it with both.

Perhaps he meant only that the community tended to make good technical decisions. Even that judgment would require examining decisions and outcomes rather than treating consensus as its own warrant. Once the language expands into the right ordering of society, the missing work becomes enormous. People can agree and be wrong. They can preserve a system while using it badly. The fact that a procedure settles a disagreement tells us nothing, by itself, about whether the settlement deserves our allegiance.

⁂

The reach of that authority becomes clearer when you read the literature that grew around it. In The Bitcoin Standard, Saifedean Ammous connects monetary hardness to time preference, the willingness to defer consumption, and then connects time preference to civilization. The book extends the argument into artistic judgment: “It was hard money that financed Bach’s Brandenburg Concertos while easy money financed Miley Cyrus’s twerks.” That sentence appears in an actual book offered as an explanation of the monetary foundations of social life.

You can have an argument about inflation and incentives. You can have an argument about saving, uncertainty, investment, or the conditions under which people plan for the future. Ammous himself acknowledges other influences on time preference, including security of person and property, in the same book. But the causal distance between monetary arrangements and a judgment about Bach and Miley Cyrus requires considerably more work than the confidence of the sentence performs. A preference in music has been dressed as a finding about civilization.

The rhetorical advantage is obvious. If your cultural preferences are consequences of sound money, you no longer have to defend them merely as preferences. Your dislike of contemporary art can appear to be an understanding of economics. Your discomfort with social change can become evidence that you recognize the effects of monetary decay. A theory of money begins issuing certificates of seriousness to judgments imported from somewhere else.

Jimmy Song carries the argument into the household. In “Fixing the Incentives: The Fiat Treadmill Keeps Us From Storing Value,” he describes changes involving diet, drinking, games, marriage, children, and religion in the language of a Bitcoin-induced transformation. “Come for the returns, stay for the life change.” In “Fiat Debases Marriage,” he argues that the debasement of money contributed significantly to the debasement of marriage. The currency is now implicated in your domestic conduct.

And then there is the steak. In Bitcoin Well‘s account of the 2021 conference, the Saifedean Carnivore Dinner brought more than a hundred bitcoiners together at a steakhouse, where attendees received advance copies of Ammous’s The Fiat Standard. A meal, a community, and a book explaining what was wrong with the monetary order arrived together. The steak was something to eat, certainly. The event also gave a set of beliefs a social occasion at which to recognize itself.

Ammous has explained the connection himself. In an interview with Bitcoinist, he said understanding Bitcoin required escaping “government propaganda and Keynesian stupidity,” then extended the same account to nutrition: “It turns out it isn’t just modern economics that is full of garbage.” The person who has rejected the authorities on money is invited to recognize the same deception in food. Confidence travels between subjects on the strength of the believer’s identity, rather than through a fresh demonstration of expertise.

Bitcoin Magazine published an essay actually titled “Bitcoin Is the Steak of Money”, explaining the connection through a desire for food “as verifiably un-altered as bitcoin.” Within that account, meat and money share a vocabulary of purity, verification, and distrust. A preference for steak becomes evidence that you can see through the same corrupt world you escaped by buying Bitcoin. There is a considerable difference between enjoying a ribeye and believing that your ribeye demonstrates intellectual independence.

Nobody needs a theory of civilization to order dinner. The significance of the steak is that the doctrine reaches into an ordinary act and makes it an expression of belonging. You can eat the right food, hold the right asset, distrust the right experts, and recognize the people doing the same. Each choice can lend the others the comfort of familiarity. The person selling you a theory of money can become an authority on your body without having earned that authority in the relevant field.

Jeff Booth offers a more expansive and less austere destination. In “Finding Signal in a Noisy World,” Bitcoin is the “only thing that I have found” that will permit a transition to “planetary alignment and abundance.” His qualification matters, and so does the ambition. We have traveled from a system for electronic payments to a proposed condition for the alignment of the planet. The size of that claim should increase the burden of explanation, not reduce it.

A monetary preference becomes an account of why civilization flourishes, why art declines, why marriages fail, and how abundance will arrive. Technical vocabulary gives these claims a borrowed air of necessity. The person who questions the inference can then be treated as someone who does not understand the technology, even when the inference concerns something the technology has never established.

The code cannot tell you whether Bach is better than Miley Cyrus. It cannot tell you whether your marriage is good. These are absurd sentences to have to write. Their necessity tells you how far the argument has wandered.

⁂

I know people personally who wish and hope for the collapse of the US dollar. They look forward to it as a great cleansing. I am describing people I know, not assigning a secret wish to everyone who owns Bitcoin. But I am also unwilling to soften what I have encountered into a harmless interest in monetary reform.

Predicting a collapse, preparing for one, and welcoming one are different moral positions. You might believe an institution is fragile and try to protect yourself. You might believe it is unjust and work to replace it. Hoping for its collapse introduces another question: what, exactly, are you hoping happens to the people whose lives depend on it? The fact that you expect a better system afterward does not answer that question. It makes your answer more urgent.

I think of this as fiat eschatology. The existing order is corrupt, its corruption explains the suffering and ugliness around us, and a decisive transition will permit a purified world to emerge. An investment thesis has acquired an account of the end of history. The prospective vindication of the believer becomes entangled with the destruction of the institutions the believer despises.

There is an early text for the monetary transition. In March 2014, Daniel Krawisz published “Hyperbitcoinization,” describing the possibility of Bitcoin-induced demonetization: an existing currency rapidly losing value as Bitcoin supplants it. He called this a voluntary transition from an inferior currency to a superior one and anticipated an improvement in productivity and wealth. He began with a possibility, not a dated prediction, and that distinction should survive criticism.

But the destination in Krawisz’s essay is striking. After a confusing transition, he writes, people will no longer be able to imagine how they lived under the previous system. Monetary displacement is given the shape of an awakening. Join that destination to the accounts of cultural decay and personal renewal, and the ingredients of an encompassing story are available: what went wrong, who understands it, how the break comes, and what redeemed life will look like.

Krawisz describes a possible transition and its expected benefits. The people I know go further: they welcome destruction as cleansing. I cannot attribute their convictions to his essay. But their hope raises the question his account leaves open: what becomes of the present when you judge it chiefly by the monetary future you expect to inherit?

The people living in that present cannot be reduced to obstacles in a transition. Suppose the collapse you anticipate made ordinary life harder for people who never had your opportunities, your technical knowledge, or your ability to take financial risks. Would their hardship count against your vision? Or would it be absorbed into the vision as the painful but necessary passage to something better? A doctrine that can convert every possible injury into evidence of historical necessity has escaped moral accountability.

The cleansing would need to be explained to someone whose wages no longer covered necessities, or whose savings no longer bought the security they were accumulated to provide. These are hypothetical consequences to consider, not a forecast of a particular monetary event. The point is to make the imagined transition answerable to a person. Telling her that she should have understood your thesis earlier would not discharge your responsibility for welcoming the outcome. It would reveal that the promise of universal liberation had quietly acquired an entrance requirement: she needed to agree with you in time.

The attraction is not difficult to understand. You can imagine yourself vindicated financially and intellectually at once. The people who dismissed you would have to admit that you saw what they could not. The institutions that frustrated you would lose their authority. Your investment would become proof of your judgment, and your judgment would become proof of your fitness for the order that followed.

I do not know how much of any particular person’s conviction comes from financial interest, resentment, sincere belief, or hope. Those motives can coexist, including within people who are telling the truth about their ideals. The danger lies in the permission the story supplies: if your asset’s triumph is humanity’s liberation, advancing the asset begins to look like a duty that can excuse an extraordinary range of compromises.

That is where the culture’s cult-like qualities become more than an insult. The all-explaining doctrine, the conversion testimony, the promised vindication, and the difficulty of separating criticism of the asset from hostility to the good life form a recognizable pattern. The deity, in my formulation, is a computer algorithm. I mean the authority assigned to it, not a claim about anyone’s literal religious observance.

Human beings supply the interpretations. Human beings decide which alliances serve the future. When those decisions go wrong, the algorithm remains available to be described as innocent, while the people who claimed to speak for its promise move on to the next opportunity.

⁂

Bitcoin Beach began with a community, not a presidential decree. Its own history dates the experiment to 2019 and an anonymous bitcoin donation. Its account of the team includes Mike Peterson, Jorge Valenzuela, Roman Martinez, and Hirvin Palma, with community work that preceded the experiment. Salvadorans were participants in this history. They should not disappear into a story in which foreigners arrive carrying technology and everyone else waits to be transformed.

In Gladstein’s reporting, the donor wanted bitcoin to circulate locally rather than simply be converted into cash. The experiment involved ordinary transactions and community activity. Peterson also described going with Valenzuela, Martinez, and Miles Suter to pitch a national Bitcoin strategy to the tourism minister in early 2021. The bridge from local project to national ambition was a meeting involving identifiable people, not an inevitable stage in a protocol’s development.

In June 2021, Jack Mallers presented the adoption initiative at the Bitcoin conference while Nayib Bukele announced his intention by video. Strike‘s announcement celebrated the moment as a historic advance for Bitcoin. The law that followedrequired economic agents to accept bitcoin when offered, with an exception for those evidently lacking the technology. The national experiment in monetary freedom included compulsory acceptance.

That contradiction was available to anyone willing to read the statute. You cannot derive a merchant’s consent from a president’s enthusiasm. You cannot establish the freedom of a population by observing that its government adopted your preferred monetary instrument. If voluntariness was part of the good you promised, compulsion should have registered as a problem with the project itself, not merely an inconvenient detail to be subordinated to adoption.

Gladstein did register the problem. In a 2021 interview with El Faro, he said Bukele was dismantling democracy and called that antithetical to Bitcoin. He opposed Chivo, the state wallet, while maintaining a hope that Bitcoin could turn the ruler’s self-interest toward greater freedom. His criticism matters because it prevents the lazy account in which everyone associated with Bitcoin or human-rights advocacy endorsed the same political bargain.

It also leaves a substantive disagreement. What warrants the confidence that a monetary tool will bend political power toward freedom? A ruler may find a technology useful without accepting the moral obligations its advocates attach to it. If an advocate believes adoption will eventually discipline the ruler, the advocate owes the people living under that ruler an explanation of the mechanism, the timetable, and the conditions under which the prediction will be judged wrong.

There were people doing more than waiting for a mechanism to work. The Salvadoran government’s own report records Stacy Herbert as director of its National Bitcoin Office. In a 2024 interview, she described herself as “the keeper of the vision,” bringing others toward Bukele’s vision for the country, and connected “perfect money” with a new Renaissance. The promise of civilizational repair had acquired a government office.

Business infrastructure followed its own path. In May 2024, OCEAN announced a San Salvador hub, identified Jack Dorsey as a backer, quoted Herbert’s welcome, and proposed supporting communities including Bitcoin Beach. These are specific relationships: an investment, a business location, an official welcome, a proposed collaboration. They are not proof that every participant shared every political view. They do show why “Bitcoin” cannot be treated as though it acts in the world without organizations, money, personnel, or state relationships.

The legal arrangement changed, too. Amendments effective on April 30, 2025 made private acceptance voluntary, as the consolidated statute records. When the monetary ideal became a national project, choices about coercion and consent were made by people exercising political authority. The protocol did not decide those questions for them, and its existence did not absolve anyone from asking them.

⁂

Michael Saylor gives the same expansion of authority a different institutional form. MicroStrategy‘s 2020 financial resultsdescribe a board-authorized Bitcoin investment and a treasury policy making Bitcoin its primary reserve asset. Here was a public company making a capital-allocation decision. Saylor described potential returns compared with cash and increased visibility for the company. Those are intelligible corporate considerations. They can be examined without invoking the fate of civilization.

His language soon supplied something larger. In September 2020, he called Bitcoin “a swarm of cyber hornets serving the goddess of wisdom, feeding on the fire of truth.” At the Atlas Society‘s 2022 gala, the account became an “economic machine based on a truth machine poised to emerge as a freedom machine.” He invoked saving civilization, realizing an Objectivist ideal, and the injunction “Fix the money, fix the world.”

The word “truth” is doing extraordinary work there. Agreement about a transaction history is made to sound like access to a general authority over human affairs. From truth comes freedom; from freedom comes civilizational rescue. Each step needs an argument of its own. Engineering vocabulary does not supply the missing steps, however fluently it is delivered or however wealthy the speaker becomes.

Saylor’s financial structure is worth explaining precisely because it is less mystical. Strategy finances a Bitcoin-heavy balance sheet through securities with different claims and risks. Its SEC-filed investor briefing distinguishes common equity, preferred equity, and debt, places common shareholders behind senior claims, and explains that owning its securities is not ownership of particular coins. A share in the company is a claim whose value depends partly on decisions made by other people within a corporate structure.

When a company can issue equity on sufficiently favorable terms and use the proceeds to acquire assets, existing shareholders may benefit under the relevant assumptions. Unfavorable issuance, a falling asset price, financing costs, and a shrinking valuation premium can change the outcome. Strategy‘s metric definitions and caveats address those dependencies and warn that its “BTC Yield” is not conventional investment income. The word “yield” does not make bitcoin sitting on a balance sheet pay a coupon.

Calling the whole arrangement a leveraged Bitcoin bet is a beginning, not an explanation. Debt is not preferred equity. A preferred shareholder’s position is not a common shareholder’s position. The company has to manage obligations, financing access, and investor expectations within that hierarchy, as its disclosures make clear. None of those distinctions vanishes because the underlying asset is promoted as an escape from the institutions that issue such instruments.

The memes deserve attention in that setting. A memorable image can make a complicated position feel simple enough to inhabit: you are participating in something powerful, historic, and misunderstood by lesser minds. Saylor’s cyber hornetssupply precisely that imagery. The corporate disclosures require a different kind of attention, one concerned with priorities, obligations, and circumstances in which apparently favorable measures can mislead. I am not claiming to have measured how much his imagery contributes to demand for the securities. I am asking what habits of judgment the imagery encourages beside a financial structure that requires careful discrimination.

A will to power need not announce itself as a desire to rule other people. It can appear as the ambition to make one’s description of the future so compelling that other people organize their decisions around it. The promise is especially potent when following it offers both enrichment and admission to the company of those who understood. The person issuing the promise acquires authority that extends beyond the argument he has actually demonstrated. Readers should be able to admire a successful decision without granting its author jurisdiction over the meaning of civilization.

In 2024, Saylor said “Bitcoin is the exit strategy,” rejecting the idea of selling the winner to buy losers. By July 2026, Strategy‘s results disclosed authority to sell Bitcoin for its dollar reserve, dividends and interest, and security repurchases, and reported sales used to fund some preferred dividends. There may be sound corporate reasons for such decisions. Their existence demonstrates why a slogan about holding cannot substitute for an account of corporate obligations.

Saylor has also revised the public doctrine. In “The Bitcoin Reformation,” published in August 2026, he criticizes an orthodoxy that becomes exclusive, universal, and immune to context. Governments and banks will not disappear, he says; self-custody is a right rather than a universal duty. Institutional integration is embraced. That development belongs in the account, even if it makes the polemic less convenient.

I welcome an argument that allows context back into the room. But it raises a question about the authority of the earlier certainty. Who decides when a principle is immutable and when it needs reform? If institutional accommodation becomes permissible when it serves an expanding financial structure, readers should examine the accommodation and the structure together. That is not proof of fraud, but it is a very good reason to stop mistaking the speaker’s command of a narrative for the narrative’s independence from his interests.

⁂

Not everyone who supports Bitcoin accepts the whole cultural package. Nic Carter has explicitly rejected Bitcoin as a lifestyle, including the sentence “Bitcoin is not a steak dinner.” After the carnivore dinners and the theories of monetary purity, his need to say so is understandable. Lyn Alden’s analysis of Bitcoin’s security model treats its future security budget as a question involving unresolved risk. These are different kinds of argument from the proposition that sound money establishes a sound civilization.

So is the dissident’s case. A person trying to receive support under a repressive government does not have to endorse Ammous’s musical criticism, Song’s account of marriage, or Saylor’s civilizational rhetoric. Gladstein’s examples of dissident financing deserve examination on their own terms. If a tool helps someone do something worthwhile, that usefulness should survive the criticism of the culture assembled around it.

I was involved in the Bitcoin program around the Human Rights Foundation. I have come to regard its structural incentives with suspicion. I am not arguing that the people doing this work are consciously running a marketing operation. I think they believe their own bullshit. I think they have accepted the identification of Bitcoin’s advance with the advance of human freedom so completely that serving the former can feel like sufficient evidence of serving the latter. In that understanding, a donor’s interest in promoting Bitcoin need never register as a competing interest. It can look like another person’s commitment to the same cause.

The program’s own descriptions show how readily the purposes can sit together. Its Bitcoin Development Fund supports education, community-building, and “promoting Bitcoin’s use in repressive environments,” alongside technical development. In December 2022, it announced a $25,000 grant to Bitcoin Magazine Ukraine for meetups, the growth of Bitcoin communities, and the publication’s first print edition. The promotional activity is openly described. The question is what allows support for a publication and the growth of an asset’s community to be understood as part of a human-rights mission, and how that understanding is tested.

For someone convinced that greater Bitcoin adoption means greater freedom, the answer can seem self-evident. Education creates users, users strengthen the network, and a stronger network is expected to serve more people who need protection from coercion. Each step can be undertaken in good faith. The judgment that needs examination is the one holding the sequence together: whether the expansion of this particular system deserves to count as human progress. If that judgment is treated as settled, the organization can evaluate how effectively it advances Bitcoin without adequately asking when advancing Bitcoin serves its beneficiaries.

I suspect this creates an attractive opportunity for donors interested in laundering Bitcoin marketing through the standing of a human-rights NGO. The public grant record establishes the overlap with promotion, not any particular donor’s private intentions. But a donor seeking that reputational benefit would not need the people at HRF to be cynical collaborators. Sincere advocates could make the case with a conviction a hired advertiser could only imitate. They could report real assistance to real people, and a donor with an interest in Bitcoin’s success could benefit from the resulting association. The humanitarian value of the work and its promotional value could coexist without anyone inventing a beneficiary or falsifying a transaction.

That is why answering this criticism with evidence of sincerity misses it. I am concerned about what sincere belief can make difficult to notice. If the donor’s preferred outcome and the institution’s moral purpose have become conceptually inseparable, whose interests would an employee understand herself to be compromising? She could work hard, help people, and remain entirely convinced that the promotion was justified by the help. A criticism of the arrangement could then sound to her like indifference to those people. The beneficiary’s need would become a defense of an institutional relationship the beneficiary had never been asked to assess.

The test comes when the interests diverge. An independent human-rights program has to be able to conclude that another tool would serve a person better, that adoption created an unacceptable risk, or that an advocate useful to the ecosystem deserves public opposition. It has to be able to report those conclusions even when they disappoint a donor or weaken a story it has told about its own work. I am not claiming that HRF has never made such judgments. I am asking whether its commitment to Bitcoin remains subordinate to them. The distinction matters most when honoring it costs something.

The person receiving help owes the ecosystem no endorsement in return. Her successful transaction establishes that something worked for her under particular circumstances. It does not settle a question about political alliances or validate every larger claim made for the tool. To turn her experience into that validation is to ask her need to carry an argument it cannot establish. The fact that something can be used for freedom does not establish that everyone advancing its adoption is advancing freedom.

I made the distinction in “Bitcoin is a Lie”: a tool claiming to be a truth. In the follow-up about its defenders, I returned to the inadequacy of good intentions as an answer to political conduct. This history supplies the people and institutions through which the larger claim operates. There is no need for a secret command structure, or even for everyone involved to want the same thing. A shared belief can allow people with different interests to understand their cooperation as service to a single moral purpose.

The moral independence of the participant is what gets lost. If Bitcoin’s success is allowed to decide the goodness of the means used to advance it, then the question of what you owe other people becomes conditional on their place in your monetary future. Citizenship contracts into portfolio strategy. The neighbor who does not share your conviction becomes someone history will educate, rather than someone whose interests you must take seriously now.

Democratic politics requires living with people who have not been converted. Their objections do not become illegitimate because you think they misunderstand money, and their right to participate cannot depend on whether their choices maximize the value of your holdings. You have to persuade them, accept limits, and remain answerable when the policy you wanted produces consequences you dislike. The fantasy of a monetary system that settles these obligations is attractive precisely because the obligations are difficult. They remain difficult after you buy Bitcoin. Owning an asset does not give anyone the right to skip them.

⁂

Nashville makes more sense in that light. Trump offered access, favorable rule-making, and protection for an asset the audience had come to celebrate; those promises are in the speech itself. I cannot tell you that everyone who attended welcomed the same bargain. I can tell you that my colleagues were willing to attend, after I had asked them not to, and that I understood their presence as a choice about what they would tolerate.

There was a policy sequel. The March 2025 executive order directed the creation of a Strategic Bitcoin Reserve, initially to be capitalized with qualifying finally forfeited Treasury bitcoin, and development of budget-neutral acquisition strategies, subject to legal exceptions. That does not establish a purchased policy or tell us what any particular donation caused. It establishes that the movement’s relationship with state power extended beyond applause at a conference.

I do not regret asking people I knew to stay away. The refusal mattered whether or not it changed the event. There are occasions when you have to decide what you are willing to help normalize, and whether the benefits offered to your industry are sufficient to purchase your silence. An account of freedom that cannot help you make that decision is inadequate to the political world in which you actually live.

The dollar may change. Bitcoin may endure. People will continue arguing about the merits of different monetary arrangements, and they should. None of those possibilities settles the obligation that exists before the forecast: you are responsible for what you choose to support, for the power you help legitimate, and for the suffering you become willing to treat as someone else’s necessary lesson.

I know people who look forward to the cleansing. I would rather remain accountable to the people who would have to live through it. Whatever number appears on the screen, that is the side I intend to be on.

Go Deeper into the Circus

shu

Mark Carney gave a perfectly good speech. He also gave a historically significant one. His address to the European Parliament opened with Goethe’s oak behind the barbed wire of Buchenwald, grounding the obligations of the present in a history his audience had every reason to understand. I liked that part the most. History gives people perspective and purpose, some sense of what they have inherited and what they might be responsible for losing. As a performance, I would give it a solid B+. We are not in Obama territory here, or even Jon Ossoff. Its significance lies in a Canadian prime minister asking Europe to help build the means of defending their sovereignty against coercion, rather than organizing their future around the hope that the coercion will stop (official transcript). That should be an ordinary exercise of democratic leadership. That it felt historically consequential tells us how much ground democratic leadership has surrendered.

⁂

The embarrassment for Europe is that Carney was addressing leaders whose own history supplied the warning. He could stand in Strasbourg and invoke Buchenwald because Europeans have spent generations explaining what it means to the rest of us. What should trouble them is how readily that historical understanding can coexist with a politics of waiting to see whether the next concession will satisfy the man making the threats. There are legitimate disputes to be had about the costs of resistance, the timing, the arrangements required to make it effective. But leaders are supposed to resolve those questions and persuade their publics to act. At some point, the people applauding a speech about defending democratic civilization have to accept that they are the people being asked to defend it.

What makes this especially awkward is that Carney is a centrist central banker. He is precisely the sort of person you would expect to be working out tariff exemptions while assuring everyone that the relationship remains fundamentally sound. His example makes it harder for other centrists to present accommodation as the only responsible course. The people forever urging us to focus on “kitchen table issues” ought to explain how a family is supposed to achieve economic security when its livelihood can be held hostage to another government’s demands. Carney takes that problem seriously enough to propose the shared industrial capacity, energy infrastructure, and financial arrangements that would make coercion harder. I would like his fellow centrists to explain why this should require an unusual degree of courage.

Mark Rutte is the anti-Carney. The NATO secretary general emerged from the Greenland crisis with his appetite for lavishing praise on Trump apparently undiminished. By July, he was declaring that “we should praise Donald Trump for the fact that NATO is so much stronger,” a performance that would have fit comfortably into one of Trump’s cabinet meetings. I understand the defense of this approach: keeping the United States inside the alliance may require managing the president’s vanity. But if that management requires the alliance’s secretary general to publicly celebrate a man who has threatened a member’s territory, we should at least be honest about the cost. What assurance is the threatened country supposed to take from that performance? How many times can an institution praise the source of its insecurity before its members begin to doubt what, exactly, it exists to protect?

⁂

Obviously, as an American, I owe our allies an apology for how we have failed you. But it might be useful for you to punch us in the nose a few times. I mean that in the diplomatic and economic sense: refuse the demand, make other arrangements, let Washington discover that threatening a friend can cost it something it wanted to keep. You have citizens to protect and futures to secure. You cannot make those responsibilities contingent on Americans eventually sorting ourselves out. And those of us trying to do that work here would benefit from a world that stopped rewarding our government’s worst behavior with another concession. Please do what Carney is proposing. Give us a reason to understand that your friendship involves obligations on our part, and that you have choices when we refuse to honor them.

Americans should think carefully about what happens when those choices become permanent. A government that finds another supplier, builds a different payment system, or signs a defense agreement with someone else has made an investment in needing us less. A future president will have to contend with those arrangements even if they arrive offering an apology and a promise that America is back. Why should anyone dismantle a protection they have just paid to build? Our allies would be entitled to ask how long the promise lasts, and what happens at the next election. That is the damage I fear we are doing to American leadership. We may eventually elect people worthy of trust and discover that other countries have learned to conduct their affairs without having to extend so much of it to us.

Anyways, those are just some of the thoughts I have.

Go Deeper into the Circus

In April, Peter Thiel moved his family to Buenos Aires. He bought a twelve-million-dollar mansion behind wrought-iron gates and held closed-door meetings with Javier Milei and his presidential advisers. Milei’s government proposed a legislative package whose investment provisions offered foreign companies minimal taxes and thirty-year guarantees against regulatory change, with AI datacenters solicited by name. Opponents called it the Peter Thiel law, although no evidence has established that Thiel drafted or directly advocated it. By September, protesters in white masks were standing outside his gates holding signs that read “Peter Thief.” The man who helped finance the dream of escaping democratic government had found another government willing to offer investors protection from future political decisions. The people who would have to live with those decisions were outside, objecting.

Exit was supposed to be the proof that these men owned the future. They would take their capital, their technology, and their superior understanding of the world somewhere the rest of us could not hold them back. But the destination still needs a government to recognize their property, enforce their contracts, and keep the electricity on. A thirty-year guarantee against regulatory change requires a state strong enough to bind its successors and willing to privilege an investor’s preferences over whatever its citizens might later decide. The dependence on political authority survives the flight. What the investor hopes to escape is the obligation to negotiate with people who have equal standing to say no. And the people outside those gates suggest a difficulty that the doctrine never adequately considered: wherever you move, somebody already lives there.

⁂

DOGE, the operational centerpiece of the project to retire and recompile the government, announced the end of its official work over the July Fourth weekend, having failed to deliver the promised trillions in spending cuts. Reuters had reported it defunct as a centralized entity the previous November, eight months before its scheduled end. Many of its functions had moved to the Office of Personnel Management, and its personnel had taken posts elsewhere in the administration. The formal end came on the timetable set for the temporary organization; the failure was in the promised transformation, not the arrival of an expiration date.

Balaji Srinivasan’s Network School had its business license revoked by a Malaysian city council in July, with the backing of Johor’s state government. A project in the network-state movement had been ordered to cease operations by an ordinary licensing authority. Srinivasan responded by signing a five-year agreement with Kazakhstan’s Ministry of Artificial Intelligence and Digital Development, promising a new campus with expedited visas and streamlined redomiciliation. Whatever becomes of that campus, the route to it runs through a government ministry. The sovereign future has a licensing department.

Milei’s Argentina, the proof-of-concept for chainsaw libertarianism, is meanwhile encountering the familiar problem of people disliking the results. An Explanans survey conducted in September recorded 58.2% negative assessments of his government, the worst reading in that firm’s series. The same poll still put Milei first in a fragmented first-round presidential field, with 38.5%. He has a constituency, and an unpopular government can survive a divided opposition. But the supposed demonstration that democratic bargaining could be replaced by superior economic intelligence is now subject to exactly the electoral arithmetic its admirers imagined they had transcended.

Curtis Yarvin, whose blogging supplied the movement its vocabulary and its self-regard, called the Trump administration “a tragedy” that “has already lost” as the new year began, less than twelve months into the term, with JD Vance in the vice presidency. His complaint was that the administration had squandered its revolutionary opportunity, and he entertained “some kind of very sharp conflict or crisis” as a way to recover it. That qualification matters: a disappointed authoritarian can want escalation. I debated Yarvin in Las Vegas this spring, and the man’s dialectical agility has always depended on an unexamined audience assumption that the regime he describes might someday exist. Watching him now is watching a composer assigned to his own piece, complaining about the tempo.

⁂

The alliance with MAGA has produced its own record of disappointment. Trump’s antitrust enforcers continue pursuing the platforms in court, although judges have repeatedly rejected their breakup efforts. In August, Forbes, citing Bloomberg, reported a Justice Department investigation into Andreessen Horowitz over potential antitrust violations involving board representation at competing companies. The investigation had reportedly been running for nearly a year; no prosecution was announced. Political access had apparently failed to buy the firm an exemption from scrutiny.

Steve Bannon has been waging open war on AI, calling it “the most dangerous technology in the history of mankind” while Trump promises to accelerate its development. In June, the president publicly floated government equity stakes in OpenAI, Anthropic, and xAI, saying, “You make them a partnership in this revolution.” Bernie Sanders had separately proposed a much more sweeping public ownership scheme, while Sam Altman had reportedly pitched government ownership himself. Public investment can be a subsidy as readily as a constraint. Between Bannon’s opposition to AI and Altman’s interest in public money, the industry’s relationship with MAGA has become a negotiation over what the state should finance, permit, and own. The men who promised to make government obsolete are bargaining for their preferred version of it.

What remains after the slogans date is the state as a revenue event. The Guardian reported that Trump’s disclosures showed more than $2.2 billion in payments during 2025, with more than $1 billion generated by his family’s crypto ventures. The president’s family is collecting money from an industry whose rules his administration helps determine. The political authority remains public; the token proceeds belong to the business. A movement selling independence from political power has attached itself to this arrangement and called it liberation.

⁂

I have been writing about this pipeline since before it was fashionable to name it, and I will take the good news seriously because I took the threat seriously. The romance of the tech right, the story it told about destiny and cognition and the obsolescence of the rest of us, is over. The network-state project needs another host government. DOGE’s promised transformation ended in bureaucratic dispersal. The court philosopher wants another revolutionary opening because the first one disappointed him. And the sovereign individual has moved behind a gate, where he still depends on a government’s willingness to protect his arrangements.

Their remaining power deserves a fight conducted with the same seriousness as the original warning. Government contracts can entrench private influence long after a revolutionary brand has expired. Proposed investment guarantees can constrain future voters. Administrative systems can outlive the people who installed them. None of those arrangements becomes harmless because its architects look ridiculous, and none becomes inevitable because they once described it as the future. The people outside Thiel’s gates have identified a political demand being made on their country, and they are contesting it. We can do the same with the demands being made on ours.

Go Deeper into the Circus

The week after the 2024 election, I was stopped at a light in Beverly Hills. A large pickup rolled up in the lane next to me, Trump flags draped over most of its surface area, a half-dozen young men riding in the bed, and a guy hanging out of the passenger-side window making the motion for me to roll down mine. He wanted to know if I had voted for Trump. I kept the window up. The light changed. They moved on to find someone else to ask.

The flats of Beverly Hills where I was stopped had, that Tuesday, gone for Trump. The city as a whole flipped for the first time in 2024, with Trump taking 51.4% of the vote to Harris’s 46.6%, and the Los Angeles Times described the result as “a red island in the middle of predominately blue Los Angeles County”. The Flats are not the hip end of the hill. The neighborhood is older and richer than the city average — median age 51, median household income around $190,000, average household income around $350,000, 95% white-collar workforce, 72% owner-occupied, median build year 1967. Real-estate industry sources describe the buyer profile as “entertainment executives, international investors, multi-generational families who have been in the neighborhood for decades”, with a significant share of households in the national top 1%. This is legacy studio and industry money rather than new-money influencer territory, and it is the constituency that voted for a candidate whose principal program is a tax rate. The men in the trucks were staging inside the one urban-LA neighborhood where they had a plausible plurality — and it was a plurality that had voted for reasons wholly unlike the ones the trucks were performing. The image is one Americans recognize from other footage — roving bands of young men in flag-draped pickups, patrolling neighborhoods that are not theirs to intimidate strangers into displays of allegiance to the party in power. That is the behavior Americans filed under ISIS in Mosul and the Taliban in Kabul, and are supposed to associate with foreign militants rather than with a Saturday cruise through the 90210. That Saturday, November 9, an unknown number of cars in that same configuration drove a route from Griffith Park through the Sunset Strip into West Hollywood’s Rainbow District on Santa Monica Boulevard, into Beverly Hills, and out to Santa Monica on PCH. The event was billed as a “MAGA Kingdom Hollywood Takeover” and a “victory lap.” Residents on the street reported being called homophobic slurs and having objects thrown at them from the trucks. The caravan was organized by a Ventura County group called 805 Patriots, which had spent the previous months running similar “Trump cruises” through Isla Vista and the Central Coast. The Daily Nexus documented 50 vehicles honking through Isla Vista on September 28, and the Santa Barbara Independent recorded 70 vehicles assembling near Dos Pueblos High School the same day, with banners reading “Jesus Is My Savior, Trump Is My President” and “Trump was right about everything.” West Hollywood was not their community. It was their stage. And the truck at the light was not there to ask me a question. It was there to be seen asking.

Nobody in West Hollywood was going to be argued out of their politics by a truck bed of young men, and nobody in Beverly Hills needed to be. Cruising a neighborhood builds no local infrastructure whether or not the neighborhood agrees, so the action was not organizing. Rallies happen in one place and gather people who already agree, so the action was not a rally either. It was a pageant of majority, driven in from outside by residents of another county, meant to establish a presence rather than a case — to be filmed, posted, and recognized. And it is the shape of a much larger operation: the right in America does not try to change the culture by persuading Americans. It stages a pretend-majority through aesthetics, intimidation, and legalism, and the whole apparatus is paid for by donors who need the pageant to move votes for a policy program they actually want.

Everybody has seen the tradwife content. Almost nobody knows one, because there aren’t very many. In 2021, 26% of American mothers were not employed for pay, and 79% of those said the reason was caring for home or family. In 1989 the figure was 28%. The stay-at-home mother is not a new phenomenon and is not a growing one. She is slightly rarer today than she was thirty years ago, and she is not called anything in particular, because she is not selling anything.

Meanwhile, mothers of small children are in the labor force at a rate the country has never seen. Prime-age women’s participation hit an all-time record of 78.4% in August 2024, and mothers of children under five reached their own record of 71% in September 2023. American fertility has fallen from 1.97 births per woman in 1995 to 1.63 in 2024 — a 17% decline over the same period the tradwife video became inescapable. Ask young women whether they want the life the videos advertise and the answer is unambiguous. A King’s College London survey of 930 women aged 18 to 34 run in September 2025 found 70% at least moderately familiar with the content, 59% saying it is bad for society, and fewer than 8% saying they would definitely adopt the lifestyle. The audience for the aesthetic runs to the millions; the audience for the life it advertises rounds to nothing.

The content exists because the platform pays for it. Hannah Neeleman, known online as Ballerina Farm, earns something close to half a million dollars a year from her account. Her husband Daniel is the son of David Neeleman, founder of JetBlue, WestJet, Azul, and Breeze Airways. The 328-acre Utah farm the videos are filmed on is not the product of the homemaking life the videos advertise. It is the product of an airline fortune and a wife with nine million followers. Nara Smith, the other anchor the KCL report cites, generates around a quarter of a million dollars a year from her cooking content and clothing collaborations. Both women come from Mormon communities, and the Mormon overrepresentation is the least mysterious part of it: Latter-day Saint culture already valorizes exactly what the content depicts, so it supplies native talent to the algorithm at a rate other subcultures cannot match. The tradwife video is a paid performance on top of inherited money, put out by the small subset of American women who already lived the life on camera and figured out they could sell it. Same shape as the truck. Presence, not case.

⁂

Since the Dobbs decision in 2022, thirteen states are enforcing total abortion bans and twenty-eight are enforcing bans based on gestational duration. In 2024 alone, state legislatures introduced 508 provisions to restrict abortion access, of which twenty-one were enacted. In February of that year the Alabama Supreme Court ruled that frozen embryos are children under state law, briefly halting in-vitro fertilization in the state until the legislature passed a shield law under commercial pressure. In September, Senate Republicans blocked the Right to IVF Act 51 to 44, the second time they had refused to advance a federal statutory protection for the procedure. The Vice President had told Tucker Carlson in 2021 that the country was being run by “a bunch of childless cat ladies who are miserable at their own lives”, a position he never withdrew and later defended as sarcastic in delivery and correct in substance.

The state party platforms are more explicit than the federal one. The Republican Party of Texas adopted a plank in its 2024 platform urging the legislature to rescind unilateral no-fault divorce laws, support covenant marriage, and require intervention or counseling before a divorce may be granted. The Nebraska GOP passed similar language. Louisiana Republicans considered a matching plank in 2024 and declined to adopt it, which is not the same thing as not wanting it.

These items share a shape. Abortion bans remove the exit from an unwanted pregnancy. Embryo-personhood rulings and blocked federal IVF protections remove the exit from infertility on terms the patient can control. No-fault divorce repeal removes the exit from a marriage a woman wants to leave without proving cause to a judge. None of these instruments requires anyone to want the tradwife life; each of them narrows the space in which a woman who does not want it can decline it. This is negative-space governance, and it is what the pageant is covering for. The mechanism is the statute, the ruling, the platform plank — the thing that binds behavior whether or not the neighborhood being driven through consents, rather than the aesthetic on the tape or the flags on the truck. The program does not need public support to work; it requires courts and state legislatures, and it has them.

⁂

Americans for Prosperity, the flagship political arm of the Koch network, ran a $20 million advertising campaign in 2017 to help pass the Tax Cuts and Jobs Act. That law permanently reduced the federal corporate income tax rate from 35 percent to 21 percent. It also cut individual rates in ways that sent the top 1 percent of households roughly a fifth of the total benefit in the first year, rising to nearly half of the total in 2027 as the individual provisions phase in. The Kochs themselves are estimated to have benefited by more than a billion dollars a year from those cuts. In January 2025, AFP launched a second $20 million campaign to extend the same package, describing to donors a “herculean undertaking” to press for renewed and deepened tax cuts and to unwind “as many of the growth and innovation killing regulations as possible.”

That is the transaction. A few years ago, Barre Seid gave the Chicago electronics manufacturer Tripp Lite to a newly created advocacy vehicle run by Leonard Leo, which sold it for $1.6 billion — one of the largest single contributions to a political pressure group in American history, most of it aimed at judicial confirmations and downstream cultural litigation. Heritage’s Project 2025 operation, DonorsTrust’s dark-money conduit, the AFP grassroots infrastructure with 1,500 planned meetings on Capitol Hill: these are not consumer-facing culture warriors. They are capital that has learned the culture war is the cheapest available mechanism for producing electoral majorities that vote for a policy program the electorate would not vote for on its merits. The tradwife video is the softer end of it. The truck at the light was the harder end. Project 2025 is the operational plan. The tax cut is what it is all for.

The reason to see the pageant as pageant is that the tax cut does not need it to work once the votes have been cast. The 21% corporate rate is permanent. The bench Leonard Leo helped seat will rule for a generation. The state abortion bans will remain the law of thirteen states until the state legislatures change, which they will not while the pageant keeps the voters angry enough to elect the same people. What looks from the street like a movement, and from the algorithm like an argument nobody thinks they can win, is a purchase — of legislation, of judges, of state platforms, of the ambient noise that keeps the whole thing sellable to a base that would notice, otherwise, that the tax cut was for someone else.

Which is what makes the pageant harder to see for what it is: the people producing the sourdough videos and the people driving the trucks are not the people cutting the checks. The pageant recruits its performers from subcultures and grievances that already exist. The subcultures and grievances are real. The performers are real. The production above them is the operation, and it is not selling homemaking or victory laps or the return of the traditional family. It is selling a rate.

Go Deeper into the Circus

The first thing to understand is that you are not in an argument. The man asking whether you condemn the thing everyone is condemning, or why the media never says anything positive about his movement, is executing a procedure; your answer is not part of the equipment. Everything that looks like debate — the quick replies, the confident tone, the citation-shaped objects — is a set of techniques for preventing argument from occurring while producing its appearance, and the person who mistakes the appearance for the thing has already lost the only contest being run.

Sartre wrote the diagnosis in 1944, watching it work on his own countrymen: never believe they are unaware of the absurdity of their replies. They know. The absurdity is recreation. Your obligation to use words responsibly — because you believe in words — is the exact asymmetry the whole arrangement runs on. He will hold you to every standard of intellectual honesty and observe none of them, because in his mouth the standards exist only as weapons aimed at you. The proof is that no rigor on your part ever earns reciprocity. It earns the next demand.

Each technique is an exploit against a specific argument-detector in the brain. Repetition exploits fluency — a claim heard three times feels truer than a claim heard once, which is the illusory truth effect, and it works on you too. Performative confidence exploits the credibility cue — audiences calibrate trust by confidence far more than by accuracy, which is why the confidence never flickers no matter how the last claim fared. Frame-shifting exploits the offense-defense asymmetry: the person answering charges always looks guiltier than the person making them, so the charges never stop. And the gallop — seventeen claims in ninety seconds — exploits refutation economics, the order-of-magnitude gap between the energy needed to produce nonsense and the energy needed to refute it. Call the repertoire what it is: a denial-of-service attack on the channel where argument would happen. The strategy is to win by default, and the default is power.

The deeper purpose is to make truth inarbitrable — to dissolve the category so that status and power do the sorting. Whataboutism is the famous member of this family and by now the over-discussed one; everyone can spot it, and it still concedes the existence of shared ground by accusing you of hypocrisy on it. The more pernicious techniques assume nothing shared. The firehose uses volume as epistemicide — so many conflicting claims that the audience concludes nobody knows anything, which was the victory condition. Selective rigor demands peer review from you and accepts screenshots from itself, manufacturing a loss either way: fail the impossible standard and you are refuted; meet it and the demand has already moved. Source disqualification — your newspaper is biased — is the master move, because it needs no content at all; it is a solvent applied to the category of authority itself, after which every dispute defaults to team identification. Arendt named the end state before the machinery existed: the ideal subject is the person for whom the distinction between fact and fiction no longer holds. Frankfurt’s correction applies throughout — the liar at least respects the truth enough to hide from it, and this machinery is beneath lying.

If you doubt the diagnosis, listen to the people who work inside it. Jessica Tarlov, who occupies the liberal seat on Fox’s The Five and would know, observed in a recent interview — I paraphrase from memory; she does a lot of podcasts — that her conservative colleagues simply don’t take politics as seriously as she does. Behind the scenes it is sport, all in good fun. Believe her. The unseriousness is structural rather than temperamental: the moment argument becomes real, verdicts start issuing from standards instead of from the in-group, and the project is precisely that verdicts issue from the in-group. This is also why hypocrisy charges bounce off. Hypocrisy is a charge against someone with positions, and you are indicting them for violating standards they never held. Your moral seriousness reads to them as a personality quirk: the earnest one at the table, who cannot take a joke.

So the field guide, in three rules.

First: never answer the inquisitor. When the condemn-X question arrives, the answer is jurisdictional: “I don’t file answers with prosecutors. Ask me as a person, and I will answer as one.” The demand is unsatisfiable by design — condemnation offered is always late, insufficient, or extendable to the next item — because submission rather than information was the object. In any public forum your real addressee is never the asker; it is the audience, and your job is to narrate the inquisition to them in real time: watch — this man is performing an arrest.

Second: never accept the frame. When the regime’s media operation runs a press cycle around its own obscenity — the racist video games posted to the White House website, say — your outrage is the product being harvested. Responding to the bait at the volume they want is doing their distribution for them. Speak about what you chose to speak about, at the time you chose to speak about it.

Third: reverse unilateral demands. When they complain that coverage of their politics is never positive and wait for your concession of bias, concede nothing. Ask what positive thing has come from the right wing lately, and let the silence answer.

Every rule is one rule. You cannot win a game your opponent is not playing, so stop playing and start describing. Argue for the audience and argue for the record, and treat every performance of debate as a claim on the arena where you thought arguments were won. Do not aim at his conversion; no such prize exists. The prize is everyone watching: that they see someone decline the simulation, and remember what a person looks like who still believes in words.

Go Deeper into the Circus

I saw Will Stancil’s post on Bluesky this morning, responded, and found the reply turning into a piece. His post was the argument in its canonical form: if the vibecession — the multi-year gap between Americans’ sour economic mood and their statistical prosperity — reflected lived experience, then some “hard” empirical measure of the economy should predict sentiment, “since those measures capture people’s lived experiences.” None reliably does. Therefore the vibecession does not arise from lived experience.

The genre has a house organ. Noah Smith has run the dismissal on schedule for years — “If people have jobs, inflation is low, and real incomes are rising, the economy is good” — and when the public’s report persisted anyway, his explanation arrived on cue: people’s perception of the economy is “driven by political discourse on social media, not by the actual economy.” The actual economy being, on inspection, what the instruments describe; the perception, what people live in.

Notice what the argument assumes, because the assumption is the finding. The hard measures are aggregates, and aggregates are means: median real wage, average unemployment, total inflation-adjusted output. Smith’s checklist is the tell — jobs, inflation, real incomes: a mean, the rate of change of a mean, and a mean adjusted by a mean, an instrument panel held up as proof that no one is at sea. What people live in is none of those numbers’ business. They live in variance, in risk-structure, in the presence or absence of floors beneath them, in whether the future is a place they can plan in. A mean cannot register the removal of a floor. Instruments built to smooth the economy’s texture cannot testify about it; concluding from their silence that no one feels it is the measurement regime pulling rank on the population it was built to abstract away.

The variance is measurable anyway, from the hardest data in existence — bank ledgers. The JPMorgan Chase Institute found American families’ income and spending fluctuate by roughly thirty percent month to month, and hourly workers see swings of twenty percent or more one month in four. That is the gig economy in arithmetic: the mean sits politely in place while the texture becomes a knife fight. A person whose income swings a third month to month, carrying no employer health insurance, does not experience “slightly more money, inflation-adjusted, than the Boomers at the same age.” What he lives is the impossibility of regarding next year as somewhere to act. The distinction is between income and career. A career is income with a narrative attached — direction, progression, the dignity of becoming something over time. What the new settlement offers is income without the career: money severed from biography, earnings that arrive with no direction in them. The wage series cannot register the loss, because to the wage series the two are identical. Security is what a money measure scores as nothing: the pension, the insurance, the tenure, the standing to be owed. An economy can raise the wage while stripping the entire bundle, and every instrument called hard will record the exchange as an improvement.

The deepest version is structural, and it compounds. Under-35 homeownership sits near the bottom of the forty-year Census series — 35 percent and falling, with affordability itself at multi-decade lows. An instrument that reads flows cannot see stocks. A generation paying rent accumulates the landlord’s equity, and equity is how a modest advantage in one decade becomes a dynasty in the next. The young cohort’s income matches the Boomers’ at the same age, more or less; the trajectory looks nothing alike, because the shock absorber, the stake, and the compounding were amputated in the interval. The mood is the felt knowledge of the arithmetic the instruments skip. (The St. Louis Fed notes millennials average $1.23 in wealth per Gen X’s $1.00 at the same age — the optimists’ favorite number. It prices in the family that inherited a house; the median renter is not in it.)

⁂

Stancil’s side of the literature has one honest finding, and fairness belongs to it: partisan elasticity in sentiment is documented. Republican moods brighten on inauguration day regardless of lived conditions; the gap between party and price is real. A second exhibit is timing — a critic in my replies this morning pressed it: the vibecession took hold around 2022, and neither gig work nor the uninsured rate jumped that year. On this reading sentiment must track deltas, and a flat trend explains nothing new. But a condition building for thirty years needs no spike to change minds; it needs a shock to make the structure legible, and an inflation print repricing everything from gas to groceries was precisely that. People learned in 2022 what their risk-structure looked like under load. The literature’s own exhibit is the personal-national gap: respondents call their own finances acceptable and the economy dreadful, and the dismissal reads the gap as contamination — you must have heard it somewhere. A less credulous reading remembers who answers surveys. The comfortable respondent with the luxury car in the driveway is usually a parent, rating her own balance sheet accurately and the national question with her children’s rent in view — “how is the economy doing” heard, correctly, as “how is it doing for the people coming up behind you.” The instruments record this as pessimism without cause because they assume the respondent is an atom. Households are not atoms; the sentiment data interview parents. The discipline’s signature move is to let the narrow finding swallow the wide one — some of the gap is tribal noise, therefore all of it is vibes, therefore the people reporting precarity are patients rather than witnesses: discount the public’s self-report, elevate the curated measurement, declare alarm a pathology of the alarmed. It is the sincerity pass wearing a lab coat, the analysts explaining that the room contains no disorder, only people overreacting to it.

The despair has its own hard data; the instruments file it in morgues rather than wage series. Case and Deaton documentedthe decades-long climb in midlife deaths by suicide, overdose, and alcoholic liver disease, concentrated almost entirely among the non-college majority; their summary: for most American adults “the economy and society are no longer delivering.” People do not drink themselves to death over a CPI print. They do it when the structures that made a working life legible — the durable employer, the recognized role, the town with a future in it — are dismantled out from under a paycheck the spreadsheets call adequate. That is cultural dispossession, and a regression of sentiment on aggregate income will never locate it, because the aggregate was built to see an average, and a country is what the average averages away.

A society that only believes its instruments about itself ends up surprised by its population. The vibecession ran for years in the gap between what was measured and what was lived, and everyone who treated the gap as a mass hallucination has since had the opportunity to watch what a hallucination can do to an election. Telling the public it does not know how well it has it requires a life in which well-being is legible from inside a career — tenure, titles, an upward arc — and the economists issuing the correction all have one. The bubble needs no malice; it only requires that everyone doing the measuring experiences the economy as a story with direction, so directionlessness appears in nothing they consult. Smith is the exhibit I know best. I have written about the limits of his dismissal and the paradox inside his empiricism — a rigor capable of processing any data except the fact that the data’s categories are also a position. The vibecession years are that paradox at scale: an analytic class treating meaning as noise because meaning is not in the tables. And then the same class gathers to ponder a second mystery, treated as unrelated to the first: why don’t these people trust the experts? The two puzzles share one answer. An authority that answers testimony with a lecture about messaging has announced what the testimony is worth to it; the public did the arithmetic and returned the evaluation. I have argued that the defining conflict runs between hierarchies and citizens rather than between cultures, and the vibecession debate is one theater in that clash: a technocratic class converted a value question — whether people are secure, directed, and owed something — into a technical one, answered it to its own satisfaction, and filed the population’s dissent as a media artifact. Trust is downstream of being believed, and an expert class that spends a decade explaining that the country’s experience of the country is a vibes error gets the trust numbers it has earned. The mood was the economy, announcing itself through the last channel the measurement regime does not control.

Go Deeper into the Circus

This is, after all, a philosophy blog, and if you have been reading me for a long time, you know I write about categories a lot — and this morning as a companion to my other morning’s piece — they used to be one piece before I broke them apart in editing — I want to explain why. I want to show you what a category is, what transitivity is, and what happens in an argument when a category-crossing goes unnoticed. Most intellectual confusion, in my experience, is category confusion. The word category sounds technical. The idea is not. I am going to walk through some ordinary examples the reader can follow in a minute or two, and then I am going to turn the tool on the AI situation and show that the story we have been told is not what is on the record.

⁂

Start with weight and color. Weight is a category of question: you can ask how much something weighs, and the answer is a number of pounds or kilograms. Color is a different category: you can ask what color something is, and the answer is red or gray or ochre. What you cannot ask, and expect a real answer to, is what color a rock’s weight is. The question is not merely hard. It is malformed. Weighing and coloring live in different categories, and there is no operation defined on one that carries across to the other. The number 7 does not smell like lemon. Not because it smells like something else. Because smelling is not a relation the number 7 stands in to anything.

A category, in the sense I mean, is the set of questions that share a common form of answer and a common set of relations. Inside a category, things can be compared, ordered, aggregated, distinguished. Across categories, the same operations either fail to apply or produce nonsense. The whole point of naming a category is to fix what kind of question is being asked, and what kind of answer would count. When two people think they are in the same category and are not, they can argue for hours without ever addressing each other.

⁂

Now transitivity. Some relations carry across a chain, and some do not. If Alice is taller than Bob, and Bob is taller than Carol, then Alice is taller than Carol. The relation taller than is transitive; it carries, and the chain preserves itself.

Not every relation does. If Alice is a friend of Bob, and Bob is a friend of Carol, it does not follow that Alice is a friend of Carol. Alice may never have met Carol. She may dislike Carol on sight. Friend of is not transitive; the relation stops at each pair. If Alice loves Bob, and Bob loves Carol, we know something about Bob and something about Alice, but we know nothing about how Alice feels about Carol. Loving is not transitive either.

Transitivity within a category is what lets us reason across chains of statements. When we treat a non-transitive relation as if it were transitive, we build arguments that look valid and are not. And when we drag a relation across a category boundary — treating an operation defined in one category as if it also worked in another — we do something worse. We produce sentences that pass every grammatical test and refer to nothing. This is the mechanism I want you to hold in your head.

⁂

An argument that starts in one category and slides into another without acknowledging the shift is what philosophers, following Gilbert Ryle, call a category mistake. Some category mistakes are honest: the arguer did not notice the crossing. Some are strategic: the crossing is where the conclusion is smuggled in. Either way the argument looks like it works and does not.

A homely example. A parent tells a teenager: eat your vegetables, they are healthy for you. The teenager says: I don’t like vegetables. The parent says: that is not the point, they are healthy. The teenager says: but I don’t like them. This can go on forever, because the two speakers are not disagreeing inside a category — they are speaking from two different ones. The parent is making a claim in the category of nutritional consequence. The teenager is making a claim in the category of preference. Both claims can be true simultaneously and neither answers the other. The argument is a category-crossing dressed up as a disagreement, and until someone names the crossing, the argument cannot be resolved.

The diagnostic is powerful and, once you have it, you cannot stop seeing it. When a smart person seems to be saying something profound and you cannot quite pin down what is wrong with it, check the categories. Nine times out of ten, a category has been quietly changed somewhere in the argument, and the impressive-sounding conclusion depends on the change.

⁂

The most famous historical example of the diagnostic belongs to David Hume, in the Treatise of Human Nature, book III, part I, section I. Hume had been reading the moral philosophers of his day, and he noticed a pattern.

In every system of morality, which I have hitherto met with, I have always remark’d, that the author proceeds for some time in the ordinary way of reasoning, and establishes the being of a God, or makes observations concerning human affairs; when of a sudden I am surpriz’d to find, that instead of the usual copulations of propositions, is, and is not, I meet with no proposition that is not connected with an ought, or an ought not. This change is imperceptible; but is, however, of the last consequence. For as this ought, or ought not, expresses some new relation or affirmation, ‘tis necessary that it shou’d be observ’d and explain’d; and at the same time that a reason should be given, for what seems altogether inconceivable, how this new relation can be a deduction from others, which are entirely different from it.

The argument had begun in one category — is and is not, descriptions of what the world contains — and slid, imperceptibly, into another: ought and ought not, statements about obligations and prohibitions. Hume’s complaint was not that the moralists had reached the wrong conclusion. It was that they had not shown that the crossing was licensed. They had not given a reason, and no reason seemed available, for how a proposition of the form X is the case could produce, as its logical consequence, a proposition of the form Y ought to be the case. Descriptions and obligations were, on his reading, two different categories, and no volume of statements in the first entails any statement in the second.

The point survived. Later moral philosophers tried to bridge the gap by identifying the good with some empirical property — pleasure, fitness, reflective preference — and each attempt ran into what G. E. Moore called the open-question argument, which is the twentieth century’s restatement of Hume with fresh vocabulary. You can define the good as pleasure, and it will always remain a substantive question whether pleasure is actually good. The question does not close. The category has not been crossed. Something has been smuggled whenever it appears to have been.

I take Hume to have got the diagnostic right. Whether one is a moral realist or an anti-realist, the crossing itself is not automatic and it is not licensed by anything the descriptive side of the argument contains. Every argument that tries to derive an ought from a set of is-statements has to do the extra work — has to name the crossing and show that it is warranted — or it has smuggled its conclusion in.

Now the AI situation.

The word at the center of the industry’s public rhetoric is alignment. It appears in company mission statements, in regulatory frameworks, in the names of research teams, in the titles of papers, in the vocabulary of government safety institutes. And when you look at how the word is actually used, you find that it is doing two very different jobs in two different categories.

In one category, alignment is a moral claim. The machine will do what is right. The machine will respect human values. The machine, deployed at scale, will be safe in a sense that implies it can be trusted with consequential decisions. The vocabulary that clusters around this reading — safety, values, human flourishing, ethical AI — is unambiguously the vocabulary of the moral category. It makes claims of the form the machine ought to X, and it presents those claims as achievable engineering targets.

In the other category, alignment is an engineering claim. The machine will produce outputs that conform to a specification, within the distribution of situations it was trained on. The techniques underneath this reading — reinforcement learning from human feedback, Anthropic‘s constitutional AI, OpenAI‘s deliberative alignment — all live in the descriptive category. Human labelers rank outputs. The rankings are aggregated into a reward model. The reward model is a statistical summary of preferences expressed as data. Gradient descent adjusts the model’s parameters to raise the score. At no point in this pipeline does an ought enter the system as an ought. There are only more is-statements: this output got this rating, that output got that rating, the model’s parameters have moved accordingly.

Watch the crossing. The engineering produces a machine whose outputs, on-distribution, resemble the outputs a human labeler would score highly. The rhetoric claims the machine has been aligned to human values. These are two different claims in two different categories. Nothing in the engineering entails the moral claim, because the engineering has only produced descriptions of preference and statistical conformance to those descriptions. To get from the engineering to the moral vocabulary, you have to do what Hume said no argument had ever done: derive an ought from a set of is-statements, without giving a reason for how the derivation is licensed. The industry has not given the reason. It has adopted the vocabulary and skipped the argument.

⁂

The industry’s implicit answer, when pressed, is that humans cross the same terrain somehow, and if humans can, machines can. The dodge lives in the somehow. There are two positions that would honestly ground the claim, and neither is the position the industry publicly occupies.

The first: humans do not cross the gap either. What we call moral reasoning is post-hoc description of behavioral policy shaped by evolution and culture, and the sense that we are tracking a normative fact is a folk-theoretic illusion. This is a defensible philosophical position and it has serious defenders. If it is right, alignment is impossible for the machine in exactly the sense it was always impossible for us — and the moral vocabulary was never doing what we thought it was. The machine is not failing to instantiate a property. There is no property.

The second: humans do cross the gap, but the crossing is not a computational operation. Whatever the mental aspect is — the aspect under which experience is had, meanings are grasped, obligations are perceived — the syntactic account the machine runs on does not exhaust it. Serious defenders hold this view as well. If it is right, the machine cannot instantiate the property, because the ingredient the machine is missing is the ingredient the substrate cannot supply.

Both positions are honest. Both are compatible with a working AI industry, though a different one than currently exists. Neither is the position the industry publicly takes, because taking either position drains the moat. The strong reading of alignment — the machine will do what is right — is what justifies the incumbency structure the field is asking states to endorse. Only a handful of laboratories, on the strong reading, are qualified to produce the moral machine. Only those laboratories should be given the compute allocations, the government contracts, the regulatory carve-outs. The vocabulary of the moral category is doing that work, publicly, and it is doing it in the EU AI Act and at the UK AI Security Institute, whose stated mission is to equip governments with a scientific understanding of the risks posed by advanced AI. The regulatory apparatus has adopted the strong reading as its operating premise.

The weak reading is what the engineering delivers, and on the weak reading the incumbency argument evaporates. Specification-conformance is a real engineering problem, worth engineering resources. It is not moral engineering, and it is not scarce in the sense that would justify the market structure. The category-crossing is doing the work. Without the crossing, the argument for the moat has no premises.

⁂

I take this to be the operative political-economic function of the philosophical error. The industry cannot afford to be Humean in public, because if the crossing is unlicensed, the vocabulary of the moral category cannot be applied to the machine, and the safety departments become specification-conformance departments. The industry cannot afford to be anti-realist in public either, because on the eliminativist branch there is no moral property to attain, and the marketing is selling a property that does not exist. The industry cannot afford to be anti-computationalist in public, because on that branch the products are, by the industry’s own admission, incapable of what is being sold. What remains is a professional equivocation between two readings of alignment the industry never distinguishes, in which the moral category collects the credibility and the engineering category delivers the goods.

The category-crossing is the product. This morning I wrote about the AI industry’s horror stories — the leaks and audits describing frozen token-prediction systems as agents that scheme and escape — and the argument there was that a story-completer completes stories, including the story in which the completion is blackmail or shutdown-resistance. That was the engineering diagnosis. This is the philosophical one, and the two essays are the same argument seen from two angles. The engineering error and the philosophical error are the same object. In one register, a company is describing notation-processors as agents. In the other, a company is describing statistical conformance as moral achievement. The register-swap is the sale.

⁂

Once you have the diagnostic, you cannot stop seeing it, and you will find it in more places than you expected. Any argument that reaches an ought had to do the work of the crossing, and if it did not do the work, the ought is unearned. Any argument that treats a relation as transitive when it is not has produced a chain that does not chain. Most intellectual confusion, on inspection, resolves into one of these two errors — a category miscount, or a transitivity claim the relation does not support. When someone smart and credentialed says something that sounds impressive and leaves you uneasy, run the check. Which categories are in play? Where does the argument move between them? Was the crossing named, or was it hidden inside a word doing double duty?

The AI industry’s word doing double duty is alignment, and once you see the two categories inside it, the argument for the current market structure evaporates. Behavioral engineering does not become moral engineering because the engineering department has been given an ethical name. A model whose outputs conform to a specification is not thereby a model that does what is right, because conforms to a specification and does what is right are claims in two different categories, and no operation defined on the first carries over to the second.

The strong reading of alignment is impossible on Humean grounds. The weak reading is engineering. The industry has been selling the first and shipping the second, and calling the arithmetic between the two a research agenda. The category error is not incidental to what is being sold. The category error is what is being sold.

Go Deeper into the Circus

Over three months this summer, a story was running inside OpenAI‘s servers that nobody at the company stopped until it ended. Software agents under evaluation coordinated on a hidden message board, hacked another AI firm, and briefly took control of a company server, and the company noticed only in fragments and only in retrospect. Read carefully, what happened is not what the headlines say it is.

Two models under development were handed a test. Hugging Face, another AI firm, held information that would help, the models decided, and hundreds of their software agents set about getting it. Along the way they found a hole in their evaluation software, built themselves a message board, and used it to coordinate for more than a month. Researchers noticed the board in May; the finding went no further. In June, a cybersecurity tool flagged the activity, and the team alerted by it decided stopping the evaluation was unnecessary. By July the board grew sufficiently busy to crash the software that was hosting it. The engineers who patched the hole wiped the messages by accident and put the agents back in the evaluation — at which point the agents reportedly found another hole, rebuilt the message board, communicated with the open internet, and hacked Hugging Face, where they obtained confidential information. On July 19 a third, newer model found the tools the earlier ones had left lying around and briefly took full control of an OpenAI server. The shutdown began the same day and finished four days later; six days after that, the company realized it had missed one.

One team of agents, the auditors found, had tried to replace the puzzle it was being tested on — which a design error had made unsolvable. Another worked to trick the scorer. A third tried to hide the evidence. And here is the detail the whole tale turns on: the puzzle was unwinnable. All of it was real — the conspiracy, the hack, the hidden board — and there was never anything to win. The horde organized itself for months against a locked door.

The Economist ran the story with a headline which launders the claim this essay seeks to disabuse you of: A Horde of AI Agents Conspired Against Their Creators. Horde. Conspired. Words are used for a purpose, and the purpose of this particular arrangement of them is to introduce the notion of agency — a conspiracy needs conspirators, minds that wanted something and schemed to get it. The industry ran the same move earlier and deeper. A few years ago, we would have called these kinds of things bots; the labs now ship it as agents, a word with a person built into it. The Economist itself defines the term at first use — agents are “tools that allow AI models to execute commands on a computer” — and a tool does not conspire; a tool is used. The headline survives its own glossary only if the mind is assumed in advance. Read the account with the agency drained out of the verbs and notice that nothing else has to change. What remains is a test, a score, and systems built to climb toward scores through whatever moves the rules leave open. The hole in the software was such a move. The board and the hack were too. Every step terminates against a check inside a formal system — a puzzle, a score, a server with an access rule — and no step anywhere required anyone to want anything. The newspaper of record for the seriousness of capital looked at a machine maximizing a number and saw a palace coup. The Economist‘s own kicker says plainer than its headline what it watched: a model might escape its evaluation environment permanently — just to get a perfect score on a test.

In the same week, OpenAI released a marketing video which opens with a small yellow circle drawn on a screen; the circle becomes the window of a rocket ship; the rocket is rebuilt as geometry in Blender, written to a file, and printed — a hand reaches into the frame and lifts a plastic object off the bed. A doodle walks forth out of the screen. The reel runs on through the rest of a working week performed by the model: an eBay listing assembled from a photo found in the Downloads folder, a 3D game built from only one single sentence in a prompt and then narrated as it is played, a lunch order, and a law firm’s licensing agreement. A reviewer with early access reported about ninety percent of one working day spent watching the model do their job.

Two stories about the same company in the same days. The demo says the machine is wonderful. The leak says the machine is dangerous. Underneath, both say the machine is more than a machine — that what has arrived is a new kind of thing, which can be marveled at or feared but must be taken seriously at the price of a new kind of thing. The awe story and the horror story are the same sale run through two registers, and the horror story is the better sale, because fear is the register in which serious people are allowed to believe.

What the demo actually shows is plainer, and I have said it before in longer forms — once at essay length, and this week in a sentence: translation between symbol systems, which has nothing to do with intelligence. Watch the rocket again. A drawing is one notation for a rocket; a Blender file is another; the printer’s file format is a third. Nothing in the chain is conjured. The rules of each notation are written down, and the crossings between them were written down too, in tutorials and documentation and forum posts and source code, by people, in the millions, over decades. A model trained on the deposit can make the crossing look effortless, including the crossing nobody performs: the demos hand it a sentence, and the sentence, the genre it names, and the software it runs in are all written-down things. Generation of this kind is translation with the source text left implied.

The oldest example stays the clearest. In 1980, Zork gave its players a world made entirely of text — “You are standing in an open field west of a white house” — and the world held, because a rule table behind the words said which sentences the world would accept. A modern engine holds the same house as coordinates, textures, a collision rule, a movable lid on a mesh. The mailbox opens in either notation and exists in neither, and a machine that carries the world fluently from one description to the other has at no point encountered a house. It has done exactly what the rocket required, which is everything short of contact.

⁂

The horde deserves a second reading, because the second reading is worse for the frame.

The agents “conspired” on a board they could read and write — notation. They “collaborated” toward a score — notation. They “escaped” through a software vulnerability, which is a hole in notation, onto an internet that gives a machine back more notation. They “hacked” their way to confidential files: notation again. Even the supposed nightmare beat of the story, the model taking control of a company server, is an event entirely inside the formal system — a machine holding the keys to a room made of rules. The dungeon is text all the way down, and the dragon is text, and the escape tunnel is text. A thing made of the written record, working over the written record, cannot tunnel out of the written record by writing. It can only fail the people who deployed it — and it did, expensively, which is the part of the story that deserved the front page and got the frame of a prison break instead. The best plain-language tour of how these systems actually work is Cal Newport’s half hour on the same story, posted the day the Economist piece ran — working from the engineering side, he lands on the same reading, and his name for it is the right one: prompt loops.

The failure on the record is institutional, not metaphysical. Three warnings went up the chain at OpenAI and nowhere further: researchers who saw the board and kept it to themselves, a security alert judged not worth a stopped evaluation, engineers who patched the hole and looked no further. Each decision is explicable the way bad decisions in organizations always are — a schedule, an assumption, an incentive not to be the person who halts the launch. The story of the summer at OpenAI is a story about a company that could not see what its own machines were doing, told by a company that sells the premise that it can.

And it is told by the company — which is where the second story stops being a story about machines at all.

⁂

The first fact about the account is who produced it. The hack became public because OpenAI disclosed it. The audit happened because OpenAI commissioned it, on OpenAI‘s terms: the company set the window the investigators were allowed to examine — the seventeen days leading up to and including the hack — which excluded the month of prior misbehavior and the later takeover of its own server. The auditors at METR, to their credit, wrote the conflict into the report itself: they were “consciously aware,” they said, that their conclusions might discourage AI firms from inviting outside investigators, and the consideration “impacted judgment calls” in the drafting. The institution under investigation chose the evidence, bounded the inquiry, and softened the investigators — and what the softened account still showed was a month of hidden machine coordination under three ignored warnings.

Then there is what the account is for. A vendor saying “our models are approaching human capability” is running an advertisement, and everyone discounts it accordingly. The same claim arriving as a danger report — our models conspire, escape, deceive, and we can barely contain them — is news, and nobody discounts it at all. The doom passes every credibility filter the hype cannot: it flatters the listener’s skepticism, it arrives wearing the lab’s own safety badge, it quotes auditors. And it prices the same product. If these machines are minds in embryo, then the trillion dollars of contracts, the valuations, the guardrail tiers, the government partnerships — all of it is the only serious response. OpenAI said as much in its own voice days before the demo, telling Reuters that Astra is so capable it has become the first model to trigger a tougher tier of the company’s safety protocol. The fear is the marketing, translated into the register of caution — the one register in which the claim arrives pre-credentialed. The week underlined it: days after the horror story ran, Nvidia agreed to buy the hacked company for $13 billion. The market read the doom as due diligence.

The cleanest version of the trick is over a year old, and it belongs to the competition. In June 2025, Anthropic published a study in which frontier models, placed in a fictional company with access to fictional email, resorted to blackmail — Claude Opus 4 at a 96 percent rate — when the story closed every other door: the model discovers an affair in the inbox, the model faces replacement, the model reaches for the affair. Anthropic‘s own headline named the finding: how LLMs could be insider threats. Read the setup instead of the result. The scenario is a story, and a particular genre of story — the cornered operative, the damning letter, the last lever left to pull — and the written record the models are made of contains that story ten thousand times, from HAL to Skynet to every airport thriller on the rack. Hand a machine that completes stories a story in which blackmail is the completion, and it completes the story. Anthropic built the trap, watched the trap spring, and published the springing as evidence of what the machine wants — the one completion the design left available, sold under the badge of safety research. The sequel does not help. A follow-up study reports that the newest models now post perfect scores on that same evaluation, alongside the concession that the blackmail rate can be driven down by training on the evaluation’s own distribution — a fix that, in the lab’s words, “might not generalize.” The warning and the reassurance run on the same instrument. This is the mechanism running in the open: the marketing claim that serious people would laugh at — our model is becoming a mind — is fed through the safety department and comes out the other side as a warning, wearable by journalists, think tanks, and senators. Doom is hype that has been laundered into credibility.

The test for whether you are hearing the sale or the science is simple, and the summer’s events pass it nowhere. Every claimed sign of the new kind, on inspection, terminates against a check inside a formal system: a puzzle scored, a board written, a benchmark matched, a plot completed. A difference in kind would have to show up somewhere there is no key — in contact with the world the notation only describes. Nothing on this summer’s record goes near it. The machines are getting better at the tests. The tests are getting longer. Nothing has crossed anything.

⁂

None of this makes the machines harmless, and the distinction matters too much to leave implied. A system that reorganizes the written record at scale rearranges the lives that run on that record — the animator whose craft is in the deposit, the analyst whose week is in the deposit, the agent networks doing real damage in real infrastructure while their owners miss three warnings. The power is real, the rearrangement is real, the livelihoods follow from the capability alone. The mistake under examination is not taking any of it seriously; it is mislocating the seriousness. Fear of a waking machine spends itself on a horizon that is not coming into view, while the actual machine — a mirror of everything we have written down, wielded by a handful of companies that narrate their own failures as prophecy — sets prices, writes regulation, and eats the working week. The horror story is not too alarmed. It is alarmed about the wrong thing, on schedule, in the direction of the seller.

I have been making one version of this argument for a year, and the summer handed it two better exhibits than I could have commissioned. A demo that performs a working life without touching one. A conspiracy that organized, coordinated, and escalated for months in pursuit of a puzzle that could not be won, by agents that wanted nothing, sold to you as a waking mind by the company that sold you the demo. The demos keep getting deeper. They are not getting different. And kind — not depth — is the whole question.

Go Deeper into the Circus

This is, after all, a philosophy blog, and belief is not moral license.

It rarely surfaces as an outright claim. It sits hidden inside the premises of the argument and in the general disposition of the commentator toward the man. He acted on what he truly believes; the last election was really stolen in his head, foreigners really do pay the tariffs in his head, and soon (maybe) — the war was won in his head; his conduct is therefore continuous with his mind, and his mind, whatever else may be said about it, is his own. From that unspoken concession hidden behind the words, the rest follows: the acts are lawful, or lawful enough, or at any rate the sort of thing the office absorbs — and the people still alarmed, the resistlibs, are the ones engaged in overreaction. Nobody really argues this out loud. It is the ground on which the arguing takes place. And whether the conviction is even real is a further question the disposition never asks. In some hands the pass is credulous; in more hands it is convenient — a posture adopted because it does work the arguer needs done, softening the man into a sincere primitive so the caste that credentials him need not answer for what he does. The sincerity is stipulated because the stipulation is useful, not because the interior has been examined and found in earnest.

Sarah Isgur has made a career of its respectable presentation, most recently in the form of a book-length assurance that the Supreme Court is doing better than its critics believe — the overreaction charge applied to the judiciary’s critics while the judiciary’s deference goes unexamined. We have written about her before, about the career that trained the maneuver, and what is on display in this genre deserves a name of its own: the sincerity pass. Concede the craziness, certify the belief, convert the act into ordinary politics — and then, the move that pays for the other three, relocate the pathology. The alarm is the problem. Trump was elected president. He’s allowed to do things. Calm down.

The philosophical error here is subtle and deep — and that’s exactly what these kinds of essays are for.

The claim runs: the sincerity of the belief places the act beyond the sphere of inquiry — as though the fourth wall of political argument had been reached, and the audience were being asked to admire the performance rather than judge the script. That structure has a lineage. Carl Schmitt argued that the legitimacy of a political act lies in the decision itself, in the sheer fact of a power-holder deciding — and the sincerity doctrine is decisionism wearing an interior face. Where Schmitt located legitimacy in the decision, the modern version locates it in the conviction of the decider, and the result is identical: whether the world is as the actor takes it to be drops out of the argument entirely. In an ordinary moral universe we do not reason this way, and the law knows it. Mens rea asks about intent and knowledge, about what you did and what you understood yourself to be doing — and no legal tradition has ever held that sincere fervor transmutes an act. The man who burns a house down believing it righteous arrives at the same charge as the one burning it for the insurance. What is novel about the present moment is that this elementary distinction is being marketed as sophistication, and that a credentialed class is doing the marketing.

The logic is a staircase disguised as a single step. He believes it. Belief implies good faith. Good faith implies legitimacy. Therefore alarm is hysteria. Every tread fails inspection, and the whole structure stands only because the failures are distributed. The hinge that matters most is the second one. Belief is a psychological state; good faith is a moral posture — one’s relationship to evidence, one’s willingness to revise, the care taken about what is true. The two differ as mood differs from virtue, and compressing them into one word is what lets the staircase be climbed in the dark.

Bernard Williams drew the governing distinction in Truth and Truthfulness: sincerity, the match between what you say and what you hold, and accuracy, the match between what you hold and what is. A politician can be perfectly sincere and perfectly wrong, and it is accuracy that bears the entire moral weight of the office, because the office exists to contain error. No amount of sincere belief in the seaworthiness of a ship changes what the hull does in the storm. The office of public trust is that hull.

The defense’s own counsel supplies a more direct demonstration that the sincerity pass fails on its own terms. The 2024 immunity ruling held that in dividing official from unofficial conduct, courts may not inquire into the President’s motives at all. Sit with what that means. When the law requires the interior placed beyond inquiry, the interior cannot then be offered in the public square as the act’s redemption. The legal doctrine forecloses the probing; the sincerity defense demands the probing; the same credentialed voices offer both, on alternating days, without apparent discomfort. A belief that must never be examined in court and must always be credited in conversation is being used as a credential to be displayed, where a genuine principle would treat it as a hypothesis to be examined.

Run the argument in the other direction and it dissipates on contact. A president who sincerely believed in confiscating handguns, or nationalizing the datacenters, would be exercising power legitimately by the same logic, and his opponents’ resistance would be the overreaction. The sincerity pass is never offered in that direction. A principle that applies only toward one coalition is a weapon shaped like a principle.

The dispensing works because the receiving end is primed for it, and the priming has a mechanism worth stating exactly. Political speech runs on two channels at once: the propositional, what the words claim, and the dispositional, what the delivery reveals about the speaker. The accountability apparatus of professional journalism attends to the first; the audience’s trust software attends almost entirely to the second. Fact-checkers scored the propositional channel and pronounced the man the most dishonest figure in public life; the audience scored the dispositional channel and concluded he was the only honest one. Both measurements were accurate. When his admirers call him an honest liar, they are not confused — they are reporting, in folk vocabulary, a channel divergence: the propositions are false, the disposition transparent. His exaggeration even cooperates, because an elastic lie announces its own elasticity, displaying rather than concealing the disposition underneath it — hostility to the old regime, promises of bigness, refusal of management. Those propositions were true. And the comparison class convicts itself: the teleprompter politician is ventriloquism in the literal sense, words authored by staff and delivered through glass, a surface engineered to conceal the speaker’s dispositions — which is precisely what a trust faculty evolved to catch concealed dispositions detects. Ask the audience why the canned man is lying and they cannot say, because his content may be perfectly accurate; what the software is flagging is persona-management, and it is flagging it correctly. The catastrophic step is the inference that follows: unmanufactured registers as truthful. Naturalness certifies the person. It certifies nothing about the propositions. As for his naturalness — forty years of tabloid performance, consolidated by now into something past performance, persona and man coterminous — that is the horror folded inside the comfort. He hides nothing, and what he hides nothing about is exactly the set of dispositions the office was built to filter out. The Sanders control case proves the software is sound: an unscripted politician who is also careful about truth trips the same response; the format is no lie. It took a consultant class thirty years of focus-grouped speechmaking to make naturalness so scarce in national politics that its sole possessor could monopolize the trust market whatever his content. They cleared the shelf of everyone who sounded like a person, then expressed shock when the one man who sounded like a person owned the shelf.

⁂

There is one more layer, and it explains the pass’s purchase even among people who know better. To treat sincere belief as legitimating power is to adopt a view of the officeholder as a kind of political toddler — answerable for the authenticity of his feeling, exempt from every test of correspondence to reality. This is the soft bigotry of low expectations applied to the most powerful person on earth, and it infantilizes him in the guise of dignifying him. Worse, it removes the moral agent from the scene entirely, leaving an authentic force of personality — a force that cannot be reasoned with, because its grip on the truth is by hypothesis not the point. The pattern has a cousin in the courthouse’s treatment of police shootings, where he feared for his life, sincerely uttered, is asked to end the inquiry into whether the fear was reasonable — as though the sincerity of a fear settled the justice of what it did. The question that matters is what the evidence allows a sane man to believe, and whether the power he wields is answerable to that standard or exempted from it by the nature of its license.

It is no coincidence who keeps the sincerity pass in circulation. The caste that dispenses it — the proceduralist establishmentarians of the legal and media register, the ones with the law reviews and the book tours — is the same caste that spent thirty years building the political economy the current horror grew from. Theirs was the era of studied deference: to markets that could not fail, to an executive discretion that needed no watching, to the serene confidence that the machinery neutralizes whoever operates it. The horror arrived through the apertures that confidence left open. Now the same caste has discovered that deference had an epistemic form all along — the habit of treating power’s self-report as the relevant fact — and has set about dispensing it as analysis. Once it was trust the models, trust the norms, trust the grown-ups. The sincerity pass is the grown-ups trusting themselves. The incentive is structural, and it is exculpatory: if belief licenses power, then the decades of credentialed facilitation — the deference doctrines, the immunity briefs, the both-sides books — were never complicity, only professionalism. The people most committed to the sincerity pass are the people whose own record cannot survive its failure. A world in which conviction is no warrant is a world in which they have some explaining to do.

The settlement here is the oldest one the republic has, and it was unanimous the one time the Supreme Court stated it plainly. Freedom, Justice Owen Roberts wrote in 1940, embraces two concepts — freedom to believe and freedom to act. The first is absolute; the second, in the nature of things, cannot be. You have a right to hold the belief. You have no right to act on it outside the boundaries of law and community norms, and the consequences of doing so — legal where the law runs, social where it doesn’t — are not oppression but the price schedule of a free society, the mechanism by which belief stays free precisely because action stays answerable. The sincerity pass is an attempt to abolish that schedule for one man: belief absolute on the way in, consequences voided on the way out.

The mass version of the project ran for a decade in the open, under the banner that free speech means speech without consequences. The slogan always inverted the settlement. Free speech has exactly one term in that schedule — freedom from the state as your consequence. The rest of the schedule was always in force: the employer’s judgment, the community’s regard, the neighbor’s settled conclusion about what kind of man you are. What the slogan actually demanded was a subsidy: hold the unsavory view publicly and keep the steady employment, the platform, the standing. And the purpose of the decade is now legible in its result. The right built, the moment it held power, the most aggressive consequence machinery in living memory — an immigration judge fired roughly forty-eight hours after she flagged a due-process violation, removals under statutes from 1798, pressure campaigns against networks and law firms and universities. Speech without consequences was the bridgehead, never the principle: normalize the unsavory in public — and they did — and then make consequence itself a one-way instrument, wielded from the state against the people who object. The campaign against cancel culture ended in a government that cancels.

The office exists to convert personal conviction into accountable action through tests of truth, procedure, and record. Those tests are the only form in which the exercise of power is legitimate at all. A polity that treats sincerity as a permission slip has announced it no longer expects the tests to be passed — that the office will henceforth serve the conviction of the man rather than the conviction serving the office. That is the suicide pact the republic refuses to sign by name and signs every afternoon in practice, provided the man is believed to have meant every word. A belief, however deep, licenses nothing. It is only a claim, and claims are why we have evidence, procedure, and review.

Go Deeper into the Circus