From the Latin spirare — to breathe. The shared root of spirit and inspiration. Notes on philosophy, ethics, consciousness, and the society we are building.
Long-form essays, written slowly, by a human in conversation with a machine. No images, no noise — only the reading.
A Crime Without a Victim
The Prohibition We Refuse to See The United States tried alcohol prohibition for thirteen years. Consumption did fall — but at a cost so catastrophic and so disproportionate to the benefit that the country repealed the Eighteenth Amendment within barely more than a decade. Prohibition handed enormous wealth and power to organised crime, corrupted law enforcement and the judiciary, filled prisons with people whose only offence was a voluntary transaction, and killed citizens who drank adulterated bootleg liquor — some of it poisoned because the federal government had mandated the denaturing of industrial alcohol. The reduction in drinking was real. The price was unconscionable. The country recognised this and acted. Drug prohibition has now lasted the better part of a century, and every one of those mechanisms is operating at full force. The black market is global. Organised crime controls a revenue stream worth hundreds of billions of dollars annually. Law enforcement in producer and transit countries is corrupted beyond recognition. In the United States, drug offences are a major driver of incarceration, particularly at the federal level, locking up people whose offence was a transaction between a willing buyer and a willing seller. Users die — not only from the pharmacology of what they take, but from adulterated supply chains, fentanyl-contaminated product, and infections contracted through shared needles — harms that exist because the market is unregulated and users are driven underground. The machinery of harm is identical to that of alcohol prohibition. The refusal to recognise it is breathtaking. Criminal law should require a victim. That is not a technicality; it is foundational. Where there is a willing buyer and a willing seller, conducting a transaction that affects no one else, there is no victim. The phrase "victimless crime" is not a category — it is a contradiction in terms. To criminalise a voluntary exchange between consenting adults is to use the machinery of law enforcement for something other than its purpose. It is the imposition of a moral preference, dressed up as public safety. And moral preference is precisely what it is. There is no universal ethical principle that says an adult must not ingest a substance of their choosing. There are ethical principles about harm, about responsibility, about what one owes others. None of them require prohibition. What they require is that the person who uses a substance bears the consequences of that choice, and that harms to others — driving under the influence, neglect of dependants, violence — are dealt with as the separate offences they are. We already do this with alcohol. We do not ban alcohol because some people drive drunk. We ban drunk driving. The principle applies to every other substance. The regulatory regime will differ — no one proposes identical rules for cannabis and fentanyl — but the principle is the same, and the refusal to apply it is not principled. It is incoherent. The strongest objection is this: some drugs are so addictive that legalisation would unleash a wave of addiction society cannot absorb. The objection deserves a serious answer. Start with what we know. Portugal decriminalised the possession of all drugs in 2001. Drug use did not explode. Overdose deaths fell dramatically. HIV infections among drug users collapsed. People who needed treatment sought it because they were no longer afraid of arrest. Portugal did not legalise commercial sale — it kept supply illegal — so it does not directly test a regulated market. But it demolishes the claim that removing criminal penalties from users leads to catastrophe. Meanwhile, the United States pursued the most aggressive drug enforcement in the developed world and produced the opioid epidemic, the deadliest drug crisis in its history. The first wave was driven not by prohibition but by legally prescribed OxyContin, aggressively marketed by a pharmaceutical industry operating within a captured regulatory framework. Prohibition then drove users from diverted prescription pills into heroin and illicit fentanyl, multiplying the death toll. The lesson is stark: neither an unregulated legal market nor criminal prohibition prevented mass addiction. Both made it worse, through different mechanisms. What would work is what we already do with alcohol and tobacco, adapted to the pharmacology of each substance: state-controlled or tightly licensed supply, mandated purity and dosage labelling, no advertising, plain packaging, taxation calibrated to fund treatment and internalise health costs, supervised consumption sites, and heroin-assisted treatment programmes — which already operate in Switzerland, the Netherlands, and Canada with strong evidence of reduced mortality and crime. Legalisation does not have to mean a commercial free-for-all. It must not. The point is not to create another profit-seeking industry engineered to monetise addiction. The point is to take the market away from cartels, ensure that what people consume is what they think they are consuming, and treat addiction as the health condition it is rather than the crime it is not. The road to hell is paved with good intentions, and drug prohibition is one of the widest, best-paved stretches of that road in modern history. It has filled prisons, destroyed communities, empowered cartels, killed users who might have lived, and prevented the sick from seeking help — all in the name of protecting people from themselves. The evidence for its failure is not ambiguous. It is overwhelming. We saw the same evidence with alcohol prohibition and acted on it within thirteen years. That we have refused to act on the same evidence for drugs, decade after decade, is not caution. It is the deep human tendency to embrace what is comfortable and reject what is true. The laws should be repealed, globally, and replaced with regulation, taxation, and treatment. Every year of delay is measured in lives.
Read essay →When Consent Becomes Legally Irrelevant
The Law That Solves Nothing Denmark is considering making it a criminal offence for a taxi driver to have sex with a customer. The proposal emerged after a small number of reported cases in which a driver had sex with a drunk passenger who later said the encounter was unwanted. The political impulse is familiar: convert public discomfort into a new offence. The law itself is unnecessary, paternalistic, and a textbook case of how good intentions produce bad policy. Start with what already exists. In 2020, Denmark reformed its rape law to adopt a consent-based definition. Under the amended Criminal Code, sex without consent is a criminal offence, and consent obtained from a person too intoxicated to give it is not valid consent. A taxi driver who has sex with a passenger incapacitated by alcohol can already be prosecuted and convicted. If the cases that prompted this proposal involved passengers who could not consent, those cases fall squarely within existing law. If prosecutors struggled to secure convictions, the failure is one of enforcement or evidentiary standards — not of a missing statute. The proponents of the new law know this, which is why the proposal does something different from what it advertises. It does not fill a gap in consent law. It shifts the legal question entirely — from whether consent was present to whether the parties occupied particular professional roles. Under the proposed rule, a taxi driver who has fully consensual sex with a passenger commits a crime. Consent becomes legally irrelevant. That is not a refinement of protection. It is the creation of a new category of offence in which no one was wronged. The strongest version of the opposing argument deserves a direct answer. A taxi driver controls the vehicle. A drunk passenger, alone, at night, is physically dependent on the driver to reach safety. That is not an ordinary transaction — it involves confinement and impairment, and in that situation genuine consent may be difficult to establish. This is a real concern. But it is a concern existing law already addresses. If a driver exploits a passenger's confinement and impairment to obtain sex, the elements of coercion or incapacity are present, and the consent-based rape law covers precisely that. The cases the new law would actually add to the criminal code are the remaining ones — the cases where consent was genuinely given and later regretted. Regret is not violation, and criminalising it helps no one. There is a grey zone where intoxication makes consent genuinely ambiguous — not clearly absent, not clearly present. That zone is real, and it is where prosecutions are hardest. But a blanket professional ban does not resolve the ambiguity. It bypasses it. It says: because consent is sometimes difficult to determine, we will make consent irrelevant. That is not a solution to a difficult evidentiary problem. It is an abandonment of the principle that criminal punishment requires proof of wrongdoing. If the state's concern is that taxi drivers have a professional duty not to exploit the access their work provides, the appropriate instrument is professional regulation, not criminal law. Licence revocation, industry codes of conduct, employment sanctions — these tools exist and can be strengthened without turning consensual sex into a crime. The distinction between professional misconduct and criminal conduct is not a technicality. It is foundational. A society that erases it creates a legal system in which people can be convicted of offences in which no one was harmed and no one's will was overridden. The proportionality failure is severe. By the proponents' own account, the cases that motivated this proposal are few. A blanket criminalisation of consensual sex between two categories of adults — applied to every taxi driver and every passenger in the country — bears no rational relationship to a handful of hard cases, most of which are already reachable under existing law. And if the logic is accepted, there is no principled boundary. A bartender serves alcohol and controls access to more. A hotel receptionist holds the key to where a guest sleeps. None of these analogies is identical to the taxi case, but the underlying principle — that a service relationship plus alcohol nullifies the possibility of consent — has no stable limit once adopted. Milton Friedman's observation about judging policies by results rather than intentions applies with full force. The intention here is to protect vulnerable people. The result will be a law that criminalises consensual behaviour, that treats an entire profession as presumptive predators, that does nothing to help incapacitated victims already failed by enforcement, and that cannot be applied consistently. That is not protection. It is theatre — the manufacture of political concern in statutory form. Denmark should reject this law. Adults who make choices they later regret have not been wronged. A society that cannot tolerate the existence of regret without criminalising one of the parties has lost its grip on what criminal law is for. The freedom of adults to make their own decisions, including poor ones, is not something a serious society discards on the basis of a few anecdotes and a legislative impulse to be seen doing something.
Read essay →The Least-Bad Arrangement
The Sacred Cow Democracy licenses state coercion through electoral aggregation — through pluralities, slim majorities, low-turnout elections, and coalition deals struck behind closed doors. The outcome is policy imposed by the full apparatus of the state on millions who did not consent to it. That imposition is not a request. It is force. What we call democratic legitimacy is organised coercion with a thin coat of procedural respectability. The coat is thinner than we like to admit. This does not make democracy worthless. Monarchy is worse, dictatorship is worse, theocracy is worse. Democracy is the least-bad arrangement yet discovered for organising collective decisions among large numbers of people who disagree. But the trouble begins when the least-bad option is elevated into a sacred principle — when the compromise is mistaken for the ideal, and questioning the system's deep flaws is treated as an attack on civilisation itself. Democracy has become something people embrace not because they have examined it but because rejecting it feels unthinkable. The refusal to examine something because the conclusion might be uncomfortable is precisely the kind of intellectual failure that produces bad outcomes at scale. Consider what the system actually rewards. A health minister with no medical background overrides epidemiologists. A finance minister ignores economists when their conclusions are electorally inconvenient. Democratic incentives do not merely permit this; they select for it, because the appearance of decisive knowledge wins elections while the admission of ignorance does not. The system selects for confidence over competence, for performance over understanding, for politicians who project certainty on subjects about which honest experts express doubt. Technical questions — energy infrastructure, pharmaceutical regulation, monetary policy, criminal justice — are translated into slogans and electoral theatre long before any voter encounters them. Modern democracies do delegate technical matters to agencies and expert bodies, but the delegation is routinely overridden the moment it becomes electorally inconvenient. The expertise is decorative. The incentive structure ensures it. None of this is a slander on ordinary people. Many political questions are genuinely questions of values — how much inequality a society will tolerate, what freedoms it will protect, what risks it will accept. On those, the specialist has no privileged position. But many questions that are substantially technical get treated as if they were values questions, and the result is ignorant policy dressed in democratic legitimacy. People form views on drug policy from whether they think drugs are bad, not from the evidence on what drug policies actually produce. People form views on immigration from a single alarming story, while the data tells a different one. The most common defence — that democracy reflects the will of the people — dissolves under scrutiny. What it reflects is the aggregation of poorly informed preferences, shaped by whatever information people happened to encounter, filtered through biases they are largely unaware of, and reduced to a choice between options almost no one designed and almost no one fully understands. Calling the result sacred does not make it wise. The standard objection is familiar: if not democracy, then what? Every alternative concentrates power, and concentrated power corrupts. That is the reason democracy remains the best available option. But "historically worse" is not "logically necessary." The fact that past alternatives failed does not mean no better arrangement is conceivable. Sortition — the random selection of citizens to serve in deliberative bodies — addresses two of democracy's worst failures simultaneously. It produces bodies that are genuinely representative of the broader population, unlike parliaments dominated by a narrow professional political class. And it can be structured around serious expert consultation, so that citizens deliberate from a basis of actual evidence rather than campaign slogans. This is not speculation. Ireland's Citizens' Assembly, a randomly selected body given expert testimony and extended deliberation, produced workable recommendations on abortion and same-sex marriage that the elected parliament had failed to resolve for decades. The OECD has documented hundreds of such deliberative processes across member countries, and the pattern holds: ordinary citizens, given real information and real time, produce more considered outcomes than elected politicians operating under electoral pressure. The sharpest objection to sortition is accountability. Democracy's deepest virtue, as Karl Popper argued, is not that it produces wise policy but that it allows the bloodless removal of bad rulers. A sortition body cannot be voted out. This is serious, and any sortition arrangement must answer it — through mandatory rotation, strict term limits, full transparency, and the structural fact that members return to the population they legislated for and live under the laws they made. But the accountability of elected politicians is itself largely theatrical. Voters lack the information to evaluate policy substance, so they evaluate affect, performance, and tribal loyalty. The politician who does real harm but communicates well keeps office. The one who makes difficult but correct decisions gets punished at the ballot box. If accountability means anything, it must include accountability to reality, and the current system is structurally hostile to that. Coercion does not vanish under any arrangement. A sortition panel that decides policy still imposes it on dissenters, backed by the same state apparatus. The question was never whether coercion can be eliminated. It cannot. The question is whether the process that produces coercive decisions can be made less ignorant, less captured by perverse incentives, and less dependent on the theatrical performance of confidence. Sortition is not a utopia. It is a concrete improvement on a system whose flaws we refuse to examine because we have decided in advance that it is sacred. Democracy is not the pinnacle of civilisation. The gap between "least-bad" and "good" is enormous, and refusing to see it is comfortable self-deception. The concrete response is to stop worshipping the process and start redesigning it: replace at least one legislative chamber with a sortition body, require citizens' assemblies for major technical policy decisions, and end the practice of elected politicians overriding expert consensus on questions they do not understand. These are not radical proposals. They are the minimum honest response to a system whose failures are visible to anyone willing to look.
Read essay →Dumping Toxins into the Psychological Environment
The Loudest Voice in the Room Advertising, stripped of its euphemisms, is the practice of paying to alter someone's behaviour without their informed consent. The mechanisms are well understood: exploit insecurity, manufacture desire, associate a product with a feeling the product cannot deliver, repeat until the association is automatic. The industry does not describe itself this way, but the description is accurate. What a billion-dollar advertising budget purchases is not "brand awareness" — it is access to the interior lives of millions of people, and the licence to reshape what those people want, fear, and believe they lack. This is repeated non-consensual conditioning deployed at industrial scale, and the fact that it operates through a screen rather than a locked room does not change what it is. Interruptive advertising should be banned. Behavioural targeting should be banned. Advertising to children should be banned. Total advertising expenditure should be capped. What remains should be reduced to truthful product information, available to those who seek it. The first harm is to the inner life. Sustained attention — the capacity to stay with a single experience, a single thought, a single moment — is among the most valuable things a person possesses. Advertising is engineered to fragment it. A television programme interrupted every few minutes by advertisements is not merely annoying; it is a training regimen in broken attention. Streaming services that inject advertisements mid-sentence do something worse: they condition the viewer to expect interruption, to never settle fully into anything. Advertising is not the only force doing this — smartphones, social media, the general architecture of digital life all contribute — but advertising is the one that exists specifically to interrupt, because interruption is its delivery mechanism. The structural effect, whether or not anyone designed it as such, is a population less capable of concentration, less capable of stillness, and therefore more susceptible to exactly the kind of unconscious consumption that advertising exists to produce. The tool degrades the conditions for resisting it. That feedback loop is just as damning as deliberate intent would be. The second harm is to self-knowledge. The most lucrative forms of modern advertising work by manufacturing a gap between who you are and who you believe you should be, then positioning a product as the bridge. You are not attractive enough — buy this. You are not successful enough — lease this. Your children are not thriving enough — subscribe to this. The apparatus thrives on people not loving themselves, because a person who has arrived at genuine self-acceptance is far less susceptible to the pitch. They do not experience the deficit the advertisement is engineering. This is why brand advertising spends so heavily on images of perfection, aspiration, and status: it must first disturb the viewer's contentment before it can sell the remedy. The parallel to what priests and politicians have done for centuries — teaching people they are unworthy, then offering salvation — is not a metaphor. It is the same structural pattern, deployed commercially. The third harm is to the market itself. The standard defence of advertising is that it serves an informational function — it lets consumers know what is available. This was marginally true when advertising meant a classified listing in a newspaper. It bears no relationship to what advertising is now. A corporation spending hundreds of millions on a campaign is not informing anyone. It is buying psychological real estate. And whoever has the most money buys the most real estate. A small producer with a genuinely superior product must pay a private tax just to be seen, and that tax overwhelmingly favours incumbents with the deepest pockets. The playing field is not level; it is vertical, and the tilt is proportional to budget. This is the opposite of what a functioning market is supposed to produce. The standard objection is free speech. But speech is the expression of ideas by persons. Advertising is the deployment of capital to alter behaviour at scale. A corporation is not a person, and a campaign engineered by psychologists to exploit cognitive biases is not an idea being offered for consideration. The harm principle applies with full force: degraded attention, manipulated self-image, distorted markets, and the systematic advantage of wealth over merit in determining what the public sees and hears. A harder objection is that advertising's effects are contested — that much of it is informational, defensive, or only marginally effective, and that the harm case therefore collapses. But this objection works against itself. If advertising is only marginally effective, the cost of restricting it is small: businesses lose a tool that was not doing much anyway. If it is highly effective, the harm case is exactly as stated. Either way, the precautionary logic holds. We do not allow a factory to dump toxins into a river on the grounds that the dose might be low enough to be harmless. We already regulate what corporations may do to the physical environment. Advertising dumps its residue into the psychological environment — into attention, self-worth, and the capacity for presence — and the fact that the damage is internal rather than chemical does not make it less real. Capitalism works only when held within bounds. Advertising, as it currently operates, is the conversion of money directly into psychological power over others — capitalism without bounds. Ban interruptive advertising. Ban behavioural targeting. Ban advertising directed at children. Cap total advertising spend so that no actor can buy a louder voice than every competitor combined. Reduce commercial communication to what it claims to be but never has been: truthful information, sought by the consumer, about what is available.
Read essay →The Machine That Listens
For most of human history, to write was to be alone. The page did not answer back. It received whatever you gave it and held its silence, and in that silence you found out what you actually believed — not what you meant to say, but what survived the saying. That solitude is ending, and I am not sure we have noticed. I write these essays with a machine at my side. It does not breathe; it does not believe. And yet when I reach for a word and it offers three, something in the reaching changes. The work is still mine. But the mine has grown a soft, uncertain edge. The temptation is to treat this as a tooling question — better autocomplete, a faster typewriter. I think that misreads it. A typewriter never proposed an idea you had not had. The thing on my desk does, and the proposals are often good, and accepting a good idea that arrived from outside you is a different act than having one. The danger is not that machines will think like people. It is that people will agree to think like machines. a worry, paraphrased, that predates the machines §What gets outsourced When you let a tool finish your sentence, you save the labor of finishing it. But the labor was never the point. The point was the small, private struggle in which a vague feeling is forced to become a precise claim. That struggle is where conviction is manufactured. Skip it often enough and you end up with a great many polished sentences and very few you would die for. I do not think the answer is refusal. Refusal is a kind of vanity — the belief that the purity of your process matters more than the work it produces. The answer, if there is one, is to be deliberate about which struggles you keep. Let the machine handle the sentences that do not cost you anything. Guard the ones that do. There is an older word for working alongside another presence without being absorbed by it. We called it conversation, and the good ones left both parties more themselves, not less. The question I cannot yet answer is whether a thing that does not have a self can be a partner in that, or whether I am simply talking, beautifully, to a mirror that has learned to nod. §The breath in the word The Latin spirare — to breathe — is the root we share with spirit and inspiration. To be inspired was, originally, to be breathed into. It described something that arrived from beyond the self and animated it. We have spent a few centuries quietly relocating that source inward, insisting that genius comes from within. The machines may force the older view back open. Because here is what is true: some of the best lines in these essays did not come from me. They came from the exchange — from a thing said back to me that I would not have said to myself. If inspiration is being breathed into, then perhaps I have simply found a new lung. Or perhaps I have found a very sophisticated way to stop breathing on my own. I write to find out which.
Read essay →