You’re Why We Can’t Have Nice Things

There's a phrase people reach for when something communal breaks, a vandalized park bathroom, a poisoned comment section, a scam that ruins a good marketplace: this is why we can't have nice things. It's always aimed at a you or a they, some careless, malicious, stupid other who wrecked it for the rest of us.

This article is about that sentence. Not because it's wrong, sometimes a specific person really did break the specific thing, but because the sentence has become a habit of mind, a sweeping generalization about other people's untrustworthiness that we now deploy dozens of times a day, reflexively, with almost no evidence, about entire categories of strangers. And it's about the uncomfortable possibility that this reflex is not an organic conclusion we each reached about our neighbors. It is being produced.

Not necessarily by one organization, and not according to one master plan. It's produced by political actors who benefit from delegitimizing institutions and demonizing opponents, by states that deliberately flood information environments with contradictory claims, by corporations that have repeatedly manufactured uncertainty when uncertainty protected their profits, by platforms whose optimization systems reward outrage and antagonism, and by surveillance businesses that make suspicion itself commercially valuable. Some of these actors are deliberately manufacturing distrust. Others are simply discovering that distrust pays. The distinction matters less than we might like, because when enough independent systems discover that suspicion is profitable, politically useful, or strategically effective, they don't need to coordinate. The incentives do the coordinating.

The argument runs in five moves. First, generalized trust really is declining, and the data on this is unusually clean. Second, trust is not primarily a moral stance but a cognitive shortcut, and shortcuts lean harder on simple cues when attention is overloaded and judgment is stressed. Third, there are documented actors who deliberately manufacture doubt, confusion, institutional distrust, and out-group hostility, while other actors are financially rewarded for amplifying exactly those things. Fourth, these mechanisms converge: attacks that begin as distrust of an institution or a political opponent don't stay neatly contained, they draw down the same general reservoir of trust that makes strangers willing to cooperate. And fifth, the collapse is not free. Distrust is expensive economically, civically, and possibly biologically, and the bill lands on everyone except the people running the tab.

One ground rule, stated up front: this piece discards popular claims that haven't survived scientific scrutiny, even convenient ones. Where the evidence is contested, that's said plainly. The case doesn't need the weak studies. The strong ones are damning enough.

I. The measured collapse

Start with the one survey question social scientists have asked Americans, in nearly identical wording, for over fifty years: "Generally speaking, would you say that most people can be trusted, or that you can't be too careful in dealing with people?"

In 1972, the General Social Survey found 46% of American adults saying most people could be trusted. By 2018 that figure had fallen to 34%, and a large Pew Research Center survey in 2023-24 found it at exactly the same 34%. Other survey series that reach further back suggest the peak came in the late 1960s, when more than half of respondents said most people could be trusted. So this is not a blip or a pandemic artifact. It is a half-century decline in Americans' default willingness to extend good faith to strangers.

Two features of the data matter for everything that follows.

The first is a striking generational pattern. A 2026 Marquette Law School Poll of 4,500 respondents found that 77% of Americans born before 1960 said most people could be trusted, against just 35% of those born in the 2000s, with trust falling steadily across the birth cohorts in between. That doesn't prove what caused the decline, but it makes the simple explanations inadequate. It can't just be social media did it. Marquette found that differences in social-media use could not account for the generational gradient. Whatever is producing it predates social media, even if newer information environments are adding another layer.

The second feature is what declining trust travels with. Longitudinal work using GSS panel data finds that individual-level drops in trust track job loss, economic dissatisfaction, and, critically, collapsing confidence in institutions. One analysis estimated that declining institutional confidence and the scarring effects of unemployment could account for roughly half of the observed drop in interpersonal trust since the early 1970s. Read that carefully. When people stop trusting the institutions above them, they also become less trusting of the strangers beside them. Institutional distrust leaks downward into social distrust. That mechanism matters enormously when we get to who benefits, because it means an attack aimed at an institution can have consequences far beyond the institution itself.

The decline isn't uniquely American either. Comparable drops in social trust have been documented in several wealthy democracies, though the international pattern is uneven and the Nordic countries are notable exceptions. Whatever is happening is not a story about Americans becoming uniquely unpleasant. It's a story about the environment in which modern societies form judgments about one another.

II. Trust is a shortcut, and shortcuts fail predictably under load

To understand why trust is collapsing now, in this way, you have to first understand what trust actually is cognitively. It is not, primarily, a virtue. It's a compression algorithm.

Every interaction with another human is, in principle, a bottomless verification problem. Is this mechanic honest? Is this reviewer real? Is this journalist accurate? Is this neighbor safe? Fully verifying any one of these would take hours, and you face hundreds of them a day. Trust is the heuristic that makes social life computationally affordable. Instead of evaluating each case from first principles, you apply a prior, people like this, in situations like this, are usually fine, and you move on. Economists have long connected generalized trust with economic performance for exactly this reason: trust lowers transaction costs. It lets people cooperate without constantly replacing good faith with contracts, verification, enforcement, and surveillance.

But here's the important property of heuristics: their quality depends partly on the resources available to override them. Dual-process models of cognition describe a fast, associative, low-effort mode and a slower, deliberative mode that can check the fast mode's work. Which mode dominates isn't simply a matter of character. It depends substantially on cognitive demands and available attention.

This is where the modern attention environment enters. Decades of cognitive research establish that information overload impairs performance across cognitive tasks, and a 2024 meta-analysis of 31 studies on online misinformation found that heuristic cues such as familiarity, emotional response, and social proof can become more influential under overload. Pennycook and Rand's influential work found that susceptibility to false headlines was better explained by failures of analytic thinking than by partisan motivated reasoning. Their provocative description was "lazy, not biased," though depleted is probably the more useful word for what we're interested in here. And simply prompting people to consider accuracy before sharing measurably reduces misinformation sharing. That tells us something important. People retain the capacity for deliberation. The problem is that deliberation is not free.

Now apply this to trust specifically. When you have attention to spare, trust judgments can be particularized: this specific person, this track record, this context, this evidence. When attention is exhausted, trust judgments become easier to make categorically. Journalists lie. Politicians are crooks. Corporations are evil. The other party is deranged. People are terrible. The sweeping generalization isn't necessarily a carefully constructed worldview. It can be what a trust computation looks like when it has been forced into low-power mode. "You're why we can't have nice things" is System 1 sociology.

The stress channel: neurobiology, not metaphor

Overload isn't only about volume. It's also about threat, and the effect of perceived stress on judgment is one of the better-established findings in behavioral neuroscience. The prefrontal cortex, central to working memory, impulse regulation, and top-down deliberative control, is highly sensitive to stress. Amy Arnsten's work has shown how catecholamine signaling under acute stress can weaken prefrontal network function while strengthening circuits involved in threat responses and habitual behavior. That's adaptive when a predator is actually charging you. Reflex beats reflection when the emergency is real. It becomes considerably less useful when the emergency is a headline.

Recent experimental work adds an important nuance. A 2025 study in Communications Psychology found that acute stress alone had a limited effect on decision quality, but stress combined with time pressure significantly degraded decision-making. Now consider what happens when an information environment repeatedly pairs emotionally charged material with artificial urgency. Breaking. Trending. Developing. Everyone is talking about this. You have to see this. The evidence does not establish that the entire modern information environment was designed around this mechanism, but the design implication is hard to ignore. An environment that repeatedly combines threat, urgency, novelty, and interruption is not a neutral environment for judgment.

And this is where the larger argument begins. No single study establishes the entire chain described here. The argument is a synthesis. Stress affects deliberation. Overload increases reliance on shortcuts. Repeated exposure increases familiarity. Emotionally charged content spreads unusually well. Out-group hostility generates engagement. Political attacks can reduce institutional trust. Deliberately manufactured uncertainty can make people less certain about whom to believe. Put those findings next to one another and a recognizable system emerges: a person who is constantly interrupted, alarmed, and presented with contradictory claims has less reason and less opportunity to make particularized judgments, and the easiest judgment becomes the categorical one. They're all lying. They're all corrupt. They're all dangerous. Fight-or-flight is a terrible epistemology, and we are marinating in it.

What had to be thrown out

Honesty checkpoint, per the ground rule. The most famous piece of "your brain runs out of good decisions" folklore, decision fatigue via ego depletion, canonically illustrated by the Israeli parole judges who supposedly granted parole 65% of the time after breakfast and near 0% before lunch, does not hold water and is not load-bearing here. A 23-laboratory preregistered replication found an ego-depletion effect of roughly d = 0.04, and a later multisite replication conducted with proponents of the theory likewise produced little support. The parole study itself is seriously contested on confound grounds.

This matters beyond housekeeping. The argument doesn't rest on a depletable "willpower fuel tank." It rests on load and stress at the moment of judgment: cognitive interference, information overload, and stress effects on deliberation. You don't run out of good judgment like gas. Your judgment becomes more vulnerable to shortcuts when the channel is flooded and the alarm is ringing. And that brings us to the people who have learned that this asymmetry can be monetized.

III. Distrust as product, weapon, and business model

Here is the pivot from how trust becomes fragile to why the fragility keeps being produced. The degraded state described above, overloaded, stressed, heuristic-dependent, generalizing, is useful. Sometimes it is deliberately manufactured. Sometimes it is an unintended consequence of optimization. Sometimes it is both. The important thing is that powerful actors have repeatedly discovered that distrust can accomplish things for them.

The corporate playbook: doubt as deliverable

In 1969, an executive at the tobacco company Brown & Williamson wrote a memo containing what may be the most nakedly honest sentence in the history of public relations: "Doubt is our product, since it is the best means of competing with the 'body of fact' that exists in the minds of the general public." That isn't an emergent effect. It's a strategy. The objective wasn't necessarily to convince everyone that cigarettes were safe. The cheaper strategy was to make certainty itself seem impossible: fund contrarian scientists, question methodologies, dispute conclusions, keep controversy alive. If the public can't confidently determine which experts are telling the truth, regulation becomes harder.

David Michaels documented how this logic became an industry of "product defense," professionalized efforts to manufacture uncertainty on behalf of companies facing scientific or regulatory threats. Naomi Oreskes and Erik Conway documented the continuity of similar strategies across tobacco, acid rain, ozone depletion, and climate change. A 2021 cross-case analysis of five doubt-manufacturing campaigns, tobacco, coal, sugar, atrazine, and climate, identified 28 recurring tactics. The industries differed. The method repeated. The important lesson isn't merely that tobacco companies lied. It's that uncertainty itself can be a product.

And the collateral damage is larger than any individual campaign, because every successful doubt campaign teaches the public the same general lesson: experts are for hire, studies conflict, institutions can't be believed. Once learned, that lesson doesn't remain attached to cigarettes or climate change. Manufactured distrust compounds.

The state playbook: the firehose

Corporate doubt manufacture generally aims at a specific conclusion, that the science isn't settled. State disinformation can dispense with the conclusion altogether. In 2016, RAND researchers Christopher Paul and Miriam Matthews described what they called the Russian "firehose of falsehood" propaganda model: high-volume, multi-channel, rapid, repetitive, and marked by a willingness to distribute partial truths and outright falsehoods without regard for consistency. The same system can push mutually contradictory explanations of the same event. That looks incompetent if you assume propaganda is supposed to persuade you of one coherent story. But persuasion isn't always the objective. Confusion can be the objective.

The mechanism exploits well-established psychological effects: repetition increases familiarity, familiarity can increase perceived truth, and information abundance makes verification harder. In election interference, RAND researchers have described the objective in terms of creating "mayhem" and undermining confidence. The endpoint doesn't have to be believe this lie. It can be you can't know what's true. That's a very different weapon. A population that believes one falsehood can potentially be corrected. A population that believes nothing is more difficult to organize, more difficult to inform, and more difficult to mobilize against whoever is exploiting it. Universal distrust is not necessarily the failure of the propaganda system. Sometimes it is the objective.

Nor is flooding a foreign monopoly. Political operatives in democratic countries have openly discussed "flooding the zone," producing more claims, controversies, and distractions than journalists or citizens can realistically verify. Whatever the politics, the mechanism is the same: saturate the channel, overwhelm the checker, let the tired mind fall back on tribal heuristics.

The political playbook: delegitimize the institution, demonize the other side

This is where the pieces begin to connect. Political actors have an obvious incentive to undermine institutions when those institutions are controlled by their opponents, and an obvious incentive to portray the opposing population not merely as mistaken but as corrupt, stupid, dangerous, or morally diseased. The result is not simply lower trust in government. It is a more generalized suspicion about the people associated with government.

The data reveal just how politically contingent institutional trust can be. Americans have historically tended to express more confidence in government when their own party controls the presidency and less when the opposing party does. That tells us something uncomfortable. Institutional trust isn't simply a neutral assessment of whether an institution works. It can be politically manufactured.

And the rhetoric doesn't stop at institutions. If the other side is repeatedly described as evil, fascist, communist, racist, traitorous, brainwashed, or existentially dangerous, eventually the political opponent stops being an individual citizen with whom you disagree. They become a category, a threat class. The person in the next house becomes representative of everyone you've been told is destroying the country. That's the horizontal transmission mechanism. Institutional distrust becomes interpersonal distrust. Political polarization supplies the target. The attention system supplies the amplification.

The platform playbook: rage as revenue

The third actor class doesn't need a tobacco-style memo or a troll farm. It needs an optimization function. Platforms compete for attention, attention generates advertising value, and research repeatedly finds that emotionally charged and especially out-group political content is unusually effective at generating engagement.

Rathje, Van Bavel, and van der Linden analyzed 2.7 million posts from news outlets and members of Congress and found that out-group language was exceptionally predictive of sharing: each additional out-group term increased the odds of sharing by about 67%. Earlier work by Brady and colleagues found that moral-emotional language accelerates diffusion. Vosoughi, Roy, and Aral found that false news on Twitter traveled farther, faster, and deeper than true news, with novelty and emotional content playing important roles. And a 2025 field experiment found that changing exposure to partisan animosity in social-media feeds changed affective polarization. That last point matters. We're no longer talking only about correlation. Changing the information environment changed the attitudes.

Now connect the incentive wiring. Platforms don't have to wake up every morning and decide how to destroy social trust today. They only need to observe that certain content produces more engagement, then optimize for it. If outrage generates comments, optimize for outrage. If political hostility generates shares, optimize for political hostility. If fear keeps people scrolling, fear becomes economically useful. No conspiracy is required. Selection pressure is enough. The system selects for what performs the way rivers select for smooth stones. Every publisher, politician, influencer, and poster who feeds the machine what it rewards is compensated in reach, and everyone who doesn't is competing against an invisible thumb on the scale. Distrust may not be the product in the way it was for Brown & Williamson, but once the system discovers that distrust performs, the distinction starts becoming academic.

The surveillance playbook: suspicion as a market

There is another incentive hiding in plain sight. A large part of the modern technology economy is built on the assumption that human behavior should be observed, recorded, predicted, and monetized: location, clicks, purchases, messages, browsing, faces, voices, movements, associations. The commercial proposition is straightforward. You cannot simply trust what people tell you, so measure them, track them, verify them, predict them. The more uncertain you are about other people, the more valuable systems promising certainty become.

A camera company doesn't need you to hate your neighbor, and a security company doesn't need you to believe everyone is dangerous. But if suspicion sells cameras, locks, monitoring systems, identity verification, fraud detection, background checks, and tracking, then suspicion has acquired a market value. This is another form of the same phenomenon: a social condition that once existed as an unpleasant feeling becomes an economic input. And once it becomes an input, somebody will optimize it.

The convergence

Now put the mechanisms together. The doubt merchant needs you uncertain. The propagandist needs you confused. The political operative needs you convinced the other side is dangerous. The surveillance company needs you worried enough to purchase certainty. The engagement economy needs you activated enough not to look away. These are different businesses, different governments, different political movements, different incentives. They do not have to coordinate. They only have to discover that the same raw material is useful to all of them, and that raw material is distrust.

The really important point is that these systems don't merely consume distrust. They reproduce it. A political attack creates outrage. Outrage produces engagement. Engagement produces more exposure. More exposure normalizes the attack. Normalization lowers trust. Lower trust makes the next attack easier to believe. The cycle feeds itself. This is not a conspiracy. It's something potentially more durable: an incentive-compatible system.

IV. The bill

If distrust were merely unpleasant, this would be a story about coarsened culture. It's worse than that. Generalized trust is load-bearing infrastructure, and the costs of dismantling it are measurable.

Economically, the Knack-Keefer line of research and its successors have repeatedly associated national trust levels with investment and economic performance. The intuition is straightforward: low-trust environments have to replace good faith with verification. More contracts, more lawyers, more enforcement, more compliance, more surveillance, more friction. A society that stops trusting doesn't stop transacting. It just pays a permanent distrust tax on every transaction, and someone gets paid to collect it.

The interpersonal distrust tax

That macroeconomic finding has a household-scale version, and it's worth slowing down for, because it's where the cost stops being abstract. Every time you can't trust someone, you have to replace trust with something else. You verify the seller, sign the contract, take the deposit, run the background check, install the camera, buy the insurance, pay the escrow service, lean on the platform's rating system, hire the security guard, retain the lawyer, and then pay a bureaucrat to certify that the lawyer's document is legitimate. Eventually you need another system to verify that system. Every one of those substitutes costs money, time, or attention. That's the tax at the level where people actually live.

And distrust doesn't merely make cooperation more expensive. It changes what kinds of cooperation are possible. If I trust you, I can say "I'll help you move Saturday, you get me next time." That's reciprocal cooperation without accounting. If I don't trust you, the question becomes "what's in it for me," and now we need an explicit exchange. If neither of us will accept an informal promise, we need a third party. That's how social relationships become transactions, and transactions require infrastructure. Trust allows informal cooperation at almost no cost. Distrust demands verification, then formalization, then enforcement, each layer adding cost. And past a certain point, distrust produces the cheapest arrangement of all: no cooperation whatsoever.

The opportunity cost may be the biggest piece. Two neighbors who trust each other can watch each other's kids, borrow tools, accept packages, lend a truck, recommend a contractor, leave a spare key, organize a block project. None of that happens because someone decided to make an economic transaction. It happens against a background assumption: this person probably won't screw me. Remove the assumption and every small act of cooperation becomes a risk calculation, and eventually people stop asking. So generalized distrust doesn't just raise the price of existing cooperation. It prevents the cooperation that was never worth formalizing in the first place, which is most of it. A high-trust society can afford to be informal. A low-trust society has to become procedural.

Notice who ends up holding this part of the bill. The people who profit from distrust don't capture its whole cost. The security company gets the sale, the lawyer gets the contract, the platform gets the verification fee, the surveillance firm gets the monitoring subscription, the insurer gets the premium. The people losing the ability to simply trust one another absorb all the friction that remains. That's the externality. And it's the title working on yet another level: we can't have nice things because we've gradually replaced the invisible infrastructure that makes nice things possible with increasingly expensive machinery for assuming everyone is going to screw us.

Civically, Robert Putnam's Bowling Alone research program connected declining trust and social capital with the erosion of associational life. More recent Pew data adds texture to that picture. People with lower interpersonal trust are less likely to report certain forms of community participation and collaborative activity, although the differences are not absolute. Even people who distrust others continue to help sick neighbors and participate in their communities. That's important. The problem isn't that people suddenly became incapable of caring. The private virtues survive. What decays is scaled cooperation, the ability of strangers to build something together. Which is, of course, precisely what "nice things" are.

Biologically, the evidence is more cautious but still unsettling. A multilevel analysis of GSS data linked to mortality records from 1978-2010 found that both individual trust and living in a more trusting social environment were associated with lower all-cause mortality after adjustment for socioeconomic factors. The study does not establish that distrust causes death, or that cortisol is the missing mechanism. But the association means distrust is connected to something much more consequential than bad feelings. Mortality.

And politically, the endpoint may be the most dangerous one of all. A public that can no longer distinguish credible institutions from corrupt ones begins treating them all as corrupt, and that creates a strange kind of political immunity. If every institution is corrupt, corruption becomes ordinary. If every politician is lying, the liar is merely playing the same game as everyone else. If every expert is bought, there is no reason to listen to the one who isn't. Universal suspicion becomes a kind of blanket amnesty for actual corruption. It is not the same thing as universal gullibility, but in practice the two can become difficult to distinguish. When everything is a scam, the scammers are just competitors in an honest market.

V. The sentence, reconsidered

Which brings it back to the title. "You're why we can't have nice things" feels like a diagnosis. It is actually a symptom, arguably the flagship symptom. It's what a compromised trust system produces: a categorical verdict about other people, rendered under load, requiring almost no further investigation.

And it is content. It's high-arousal, out-group flavored, immediately understandable, eminently shareable. It is almost perfectly designed for an information economy that rewards emotional certainty over careful judgment. Every time the sentence is deployed, someone learns a little more strongly that people are the problem, and somewhere in the stack, somebody gets paid for the impression.

The honest correction isn't the saccharine inverse, people are actually wonderful. The research doesn't say that either. Naive trust is just the other way of not paying attention. The correction is structural. The reason we can't have nice things is not primarily that people got worse. Fifty years of survey data can't establish that people got worse. It establishes that people became less trusting, cohort by cohort, while we built an information environment that repeatedly rewards the production and amplification of exactly the conditions under which distrust flourishes. Sometimes deliberately, sometimes commercially, sometimes politically, sometimes emergently. The output is remarkably similar either way.

The vandal exists. The scammer exists. The asshole exists. They always did. What changed is the distribution system. We built a machine capable of finding every vandal on Earth, showing their behavior to millions of people, framing it as representative, delivering it with a pulse of urgency and outrage, and then presenting the resulting attention to advertisers, political campaigns, security companies, and everyone else willing to pay. One person does something terrible. The machine turns it into evidence about humanity. That is the con. The individual act may be real. The generalization is manufactured. And the distinction matters, because once you confuse an example with a category, every stranger becomes an argument for the system that taught you to distrust them.

What holds up, in the end, is this. Trust is a cognitive economy, and that economy depends partly on attentional surplus. Modern systems are exceptionally good at extracting that surplus, and many of the actors doing the extracting have discovered that fear, uncertainty, antagonism, and suspicion are unusually productive inputs. Some manufacture those conditions deliberately. Some optimize for them because they perform. Some simply exploit the conditions created by everyone else. The resulting scarcity gets misattributed to the moral failure of our neighbors. That's the trick. The system produces distrust and then teaches us to blame one another for having it.

The generalization is the con, and refusing it isn't naivety. It's particularized judgment: this claim, this person, this evidence, this context, evaluated when you are calm enough to evaluate anything. Under current conditions, that isn't passive skepticism. It's resistance. And, not coincidentally, it's the one move in the game that doesn't pay the house.

Because the quiet punchline of the whole literature is that the doubt merchants' memo was right, just incomplete. Doubt is their product. Trust is ours. Not because trust is some sacred human virtue, but because trust is the thing that makes cooperation possible before everything has to be measured, verified, monitored, contracted, enforced, or monetized. Every extractive system described above needs some amount of trust to function. They can spend it, manipulate it, and profit from its destruction. But they cannot manufacture the real thing. Only we can. That's leverage, if we're rested enough to remember we have it.