Scale, Baby, Scale
On Proportionality, Curated Outrage, and Why I’m Gobsmacked We Haven’t Had Thermonuclear War
I have aphantasia and severely deficient autobiographical memory. I cannot visualize. I cannot run the mental movie of suffering that makes one dead child more emotionally real than a hundred thousand statistical deaths elsewhere. I process the world in present-tense arithmetic rather than narrative arc. I cannot feel the pull of the story that says this tragedy is the one that matters, this atrocity is the one that defines the era, this body count is the one that demands response.
This is probably a deficit. It might also produce a kind of clarity that I want to try to share, because I think the clarity is useful even if the condition that produces it is not desirable.
Here is what the clarity produces: I am genuinely, completely gobsmacked that we haven’t had thermonuclear war.
Not rhetorically gobsmacked. Not performatively gobsmacked. Actually gobsmacked—surprised in a way that updates my model of the world, that makes me think the world is more stable than the rhetoric around it suggests, that makes me suspicious of the rhetoric precisely because the predicted consequences of the things being described as civilization-ending keep not arriving.
The Existential Crisis That Keeps Not Happening
Every major political development of the last decade has been described, by serious people with institutional credentials, as an existential threat to democracy, to civilization, to the planet, to the future.
Trump’s election in 2016 was going to end American democracy. It didn’t. There was an election in 2020. Biden won. The transfer of power was chaotic and ugly but it happened.
January 6th was going to end American democracy. It didn’t. There was an election in 2024. Trump won. The transfer of power was smooth.
COVID was going to kill tens of millions and collapse healthcare systems globally. It killed millions—real deaths, real suffering, not dismissible—but the mortality rate, once the data was in, was substantially concentrated in people with significant comorbidities and in the elderly, and it did not produce the civilizational rupture that the early projections suggested. The healthcare systems bent but did not break in most places. The economy recovered faster than predicted.
Climate change is going to make the planet uninhabitable. Maybe. The timelines keep shifting. The catastrophic tipping points keep being revised. This does not mean climate change isn’t real or serious—it is real and serious—but the specific predictions of specific catastrophes at specific dates have a poor track record relative to the confidence with which they were made.
The opioid epidemic was going to hollow out an entire generation. It did enormous damage—real damage, real deaths, real communities destroyed—but the generation it was supposed to hollow out is still here, substantially functional, if scarred.
I am not saying any of these things were not bad. I am saying that the gap between the described consequence and the actual consequence has been consistently large, and that this gap is itself data that is not being incorporated into the models.
I am gobsmacked we haven’t had thermonuclear war. I am gobsmacked there hasn’t been a dirty bomb in a major Western city. I am gobsmacked that cafeterias and farmers markets and shopping malls are not being blown up weekly given the existential register of every political conversation I have had in the last decade. Either the crisis is not as total as advertised, or the people performing the crisis have more to gain from the performance than from the explosion.
Probably both.
Scale, Baby, Scale
Seventy-five thousand people have died in Gaza as of this writing.
This is a real number. These are real people. Every one of them had a life, a family, a future that ended. I am not dismissing this.
I am going to apply arithmetic to it, which is apparently controversial but which I think is the only honest way to evaluate claims about what this number means and demands.
Seventy-five thousand is fewer people than die in American traffic accidents in two years. The United States loses approximately forty thousand people per year to car crashes. Nobody is calling for the abolition of the automobile as an ongoing genocide. Nobody is marching in the streets about highway deaths with anywhere near the volume or intensity of the Gaza protests. The arithmetic does not produce the same moral response, because the arithmetic is not what is producing the moral response.
Seventy-five thousand is a rounding error compared to the twenty-seven million Soviets who died in World War II. It is smaller than the death toll of the Bengal famine of 1943, the Congo Free State atrocities, the Rwandan genocide, the Cambodian genocide, the Great Leap Forward. It is a fraction of the deaths produced by the Belgian colonization of the Congo, which killed somewhere between ten and fifteen million people in what is now one of the least-discussed atrocities in Western historical memory.
The current conflict in Ukraine has produced approximately one and a half million casualties—dead and wounded—as of this writing. This number receives roughly one quarter of the media oxygen that Gaza receives, despite being numerically larger by a factor of twenty.
The ongoing conflict in the Democratic Republic of Congo has been running for decades and has killed somewhere between five and six million people depending on how you count the deaths attributable to displacement and disease. It receives approximately one percent of the media attention Gaza receives.
Yemen. Sudan. The Uyghurs in Xinjiang. The Rohingya in Myanmar. Each of these conflicts involves civilian death tolls that are comparable to or larger than Gaza, and each of them exists at a small fraction of the public attention, the protest volume, the institutional response.
The Narrative Is Not The Body Count
The reason Gaza gets the volume it gets is not arithmetic. The arithmetic does not support it as uniquely deserving of the response it receives. The reason is narrative.
Gaza maps cleanly onto existing Western guilt frameworks. The conflict has identifiable villains and victims that correspond to categories that Western progressive culture has spent decades developing moral responses to: colonialism, apartheid, ethnic cleansing, the displacement of indigenous peoples. The visual language of the conflict—the rubble, the children, the bombed hospitals, the desperate families—photographs well and travels well on social media. The story is legible within frameworks that Western audiences have already internalized.
The Congo does not map onto those frameworks as cleanly. The Uyghurs are Muslim victims of a communist state, which scrambles the ideological categories. The Rohingya were largely ignored until they were so catastrophically displaced that ignoring them became difficult. Sudan is complicated. Yemen is complicated. The complexity is not the reason for the reduced attention—plenty of things are complicated—but the absence of clean narrative fit is.
The outrage is curated. The scale is selective. The dead bodies that move Western progressive opinion are the ones that fit the story being told, not the ones dying in the largest numbers.
This does not mean the dead in Gaza don’t matter. It means the people invoking them most loudly have made choices—conscious or not—about which dead matter and which don’t, based on political utility rather than arithmetic. And those choices are not being acknowledged as choices. They are being presented as moral clarity.
The BBC Sound Design Problem
I have seen Wag the Dog enough times to have developed a healthy skepticism about the mediation of conflict.
Wag the Dog is a 1997 film in which a presidential scandal is covered up by manufacturing a fake war using Hollywood production values. It was made in the green screen era, when the technology required to fabricate convincing war footage was already available to people with sufficient budget. What the technology looks like now, in the age of generative AI and deepfakes and real-time video manipulation, is something I do not know in detail and suspect I do not want to.
I have also caught the BBC in enough framing choices—the hushed voice-over, the careful selection of which images to show and which to omit, the “boom boom rat tat tat” sound design that creates an atmosphere of proximity to violence that the correspondent is not actually experiencing—to have developed some epistemic humility about what I am seeing when I see conflict footage.
This does not mean I think all conflict footage is fake. It means I think all mediated information is shaped by the choices of the people doing the mediating, and those choices are not neutral, and claiming to report neutrally while making non-neutral choices is its own form of manipulation.
I do not live in Gaza. I do not live in Eastern Ukraine. I have never seen any of it with my own eyes. I am receiving a curated selection of images and narratives produced by people with their own interests and filtered through platforms with their own algorithms. The selection is not random. The algorithms are not neutral. The interests are not aligned with mine.
I know what I can verify and what I am being asked to take on faith. Survivalism means I only move on what I can actually see. Everything else is someone else’s arc, and I have never been able to see arcs.
The Rate Versus The Volume Problem
Here is a specific arithmetic point that I want to make carefully because it is easy to misread.
People choose rate or they choose volume. If you are choosing volume—if the absolute number of deaths is what matters to you—then Gaza is serious but is not the most serious thing happening. The Congo is more serious. Ukraine is more serious by raw count. The decisions about which number gets the volume of response it gets are not being made by arithmetic.
If you are choosing rate—if the rate of killing relative to the population is what matters to you—then Gaza is more serious, because the population of Gaza is small and the casualty rate relative to population is high. But if you are choosing rate, you also have to apply that metric consistently, which means the Rwandan genocide, which killed approximately ten percent of the Rwandan population in one hundred days, is the worst atrocity in recent human history by this metric. It does not receive the retroactive response that this ranking would suggest.
The point is not which metric is correct. The point is that no metric is being applied consistently. The metric shifts to match the conclusion. The conclusion is reached before the arithmetic.
This is not unusual. This is how human beings process moral information. We have strong feelings about specific situations and we reach for numbers afterward to justify them. What I am arguing is that we should be honest about this—that “I care about these people because their situation is legible to me within frameworks I already have” is a more honest statement than “the arithmetic demands this response.”
The arithmetic demands a lot of responses that are not being produced. The selective application of arithmetic-based moral urgency is itself a form of manipulation, even when it is unconscious.
The SDAM Advantage, If It Is One
I cannot feel the narrative pull that makes one child’s death more real than a hundred thousand statistical deaths elsewhere. This is the aphantasia-SDAM combination at work. I cannot run the mental movie of the child’s face, the child’s future, the specific texture of the specific grief. I process it as data.
This sounds monstrous. I want to be careful that it doesn’t read as monstrous, because that is not the intention.
What I mean is this: I cannot be moved differently by the death of a child whose photograph I have seen versus the death of a child whose death appears only in a statistical table. To me, both deaths have the same weight, because I do not have the neurological capacity to make one more vivid than the other through mental imagery. The photograph does not make the child more real to me than the statistic.
This means I am, in a narrow sense, immune to one of the most powerful mechanisms of curated outrage: the identified victim effect. Behavioral economists have documented this extensively—people donate far more to save a single identified child than to save statistical thousands. The identified victim effect is why charities use photographs of specific children rather than graphs showing aggregate suffering. It works because the photograph makes the suffering vivid and present in a way that the number does not.
I do not experience this differential. The number and the photograph have the same weight.
I am not arguing that this makes me more moral. I am arguing that it might make me more arithmetically consistent, and that arithmetical consistency is underrated in evaluating what actually matters and what demands response.
What I Am Actually Gobsmacked By
Let me be precise about the gobsmacked argument, because I want it to land correctly.
I am not saying the world is fine. I am not saying the crises being described are not real. I am saying the gap between the described stakes and the actual outcomes is persistently large, and that this gap is itself information.
If the world is as described—if we are genuinely living through the most dangerous moment in the history of democracy, the worst humanitarian crises in living memory, the terminal stage of a civilization-ending environmental catastrophe, the rise of fascism in the most powerful country on earth—then some specific predicted consequences should be materializing at a rate consistent with those descriptions.
They are not.
I am gobsmacked by the stability. Not comforted by it—stability is not the same as good, and a world that is stable at a high level of injustice and suffering is not a world to celebrate. But gobsmacked by the gap between the described instability and the actual stability, and suspicious of the people who benefit from describing the instability in maximally alarming terms.
The war profiteers of outrage are real. The attention economy rewards alarm. The fundraising email subject lines are always “this is the worst thing that has ever happened.” The cable news chyrons are always “BREAKING.” The think piece titles are always “Why [X] Could Be The End Of [Y].”
This does not mean X is not happening or Y is not at risk. It means the people describing X and Y have structural incentives to describe them in maximally alarming terms, and those incentives do not go away when the alarm is calibrated accurately. They push constantly toward overcalibration.
I am gobsmacked we haven’t had thermonuclear war not because I think the world is safe. I am gobsmacked because I was told, repeatedly, by credentialed people with institutional authority, that the conditions for it were present and the trajectory was toward it. The conditions may be present. The trajectory has not arrived at the destination. The gap between the description and the reality is the data I am looking at.
Scale, baby, scale. Apply arithmetic consistently. Ask who benefits from the curated selection of which numbers matter. Ask why the Congo doesn’t get this response. Ask why the Uyghurs don’t get this response. Ask why Yemen doesn’t get this response.
And then ask what you actually know, versus what you have been shown, versus what you have been told to feel about what you have been shown.
The answers are not comfortable. They are, I think, more honest than the alternative.
APPENDIX
Context, Notes, and Glossary
The Identified Victim Effect
The identified victim effect is a well-documented phenomenon in behavioral economics and social psychology, first formally described by economist Thomas Schelling in 1968. The core finding: people donate significantly more money and express significantly more concern about a single identified victim (with a name, a photograph, a specific story) than about statistical victims (numbers representing aggregate suffering).
Subsequent research has confirmed and extended this finding across multiple contexts. Charitable giving, disaster relief, policy support, and media attention all show the same pattern: the identified individual generates dramatically more response than the equivalent number of statistical deaths.
This effect is not a flaw in moral reasoning that can be corrected by information. It is a feature of how human emotional systems process loss. We are wired for empathy with individuals and not for consistent proportional response to aggregate suffering. This is probably adaptive in small-group environments where most relevant suffering is local and individual. It produces consistent distortions in a media environment where the selection of which individuals to identify is controlled by parties with their own interests.
The implication for this essay: the moral intensity of public response to any given conflict or atrocity is substantially a function of which individuals are being identified and by whom, not a function of the arithmetic scale of the suffering. This is not a conspiracy—it is the predictable outcome of human emotional wiring combined with media selection effects.
The Congo: What Scale-Consistent Moral Response Would Require
The Democratic Republic of Congo has been experiencing continuous armed conflict since the mid-1990s. The Second Congo War (1998-2003) is estimated to have killed between 3.8 and 5.4 million people, making it the deadliest conflict since World War II by most estimates. Violence has continued at lower intensity since the formal end of the war, with estimates of ongoing deaths from violence, displacement, and disease-related excess mortality running into the millions.
The Congolese conflict involves multiple armed factions, significant involvement by neighboring states, complex ethnic and political dimensions, and the exploitation of mineral resources (including coltan, used in smartphones and electronics) by multinational corporations operating through supply chains that connect to Western consumers.
Despite the scale of suffering—larger than Gaza by a factor of fifty to one hundred—the Congo receives a small fraction of the media attention, protest activity, institutional response, and public moral intensity that Gaza receives.
This is not because the Congo is less important. It is because the Congo does not map onto the narrative frameworks that Western progressive culture has developed for processing moral outrage. There is no clean colonizer-colonized binary. The perpetrators and victims do not correspond neatly to racial categories that Western audiences have already-developed responses to. The supply chain connection to Western consumers exists but is diffuse and requires more cognitive steps to process than a military occupation by a Western-aligned state.
Scale-consistent moral response would require approximately fifty to one hundred times more attention to the Congo than to Gaza. This response does not exist.
COVID Mortality: What The Data Eventually Showed
Initial projections for COVID-19 mortality varied widely but included scenarios projecting millions of deaths in the United States alone in the first year. The Imperial College London model published in March 2020, which was highly influential in shaping policy responses globally, projected up to 2.2 million deaths in the United States without intervention.
Actual US COVID deaths as of the end of 2023 were approximately 1.1 million, with ongoing debate about how to count excess mortality and COVID-adjacent deaths. This is a real and serious number. It is also approximately half the most-cited worst-case projection.
More importantly for this essay: the age and comorbidity distribution of COVID deaths was dramatically concentrated. The CDC’s data shows that the vast majority of COVID deaths occurred in people over 65, and that having multiple underlying health conditions was a significant risk multiplier. This does not mean younger, healthier people faced no risk—they did—but it means the distribution of risk was highly non-uniform in ways that affected the accuracy of early population-level projections.
The economic recovery was faster than most projections. The healthcare system, while severely stressed, did not collapse in most developed countries. The vaccines, developed at historically unprecedented speed, were more effective than most reasonable prior estimates.
None of this means COVID was not serious. It was serious. It means the specific catastrophic predictions were more wrong than the confidence with which they were made suggested they should be. This is relevant to this essay’s argument about the gap between described stakes and actual outcomes.
Ukraine Casualty Estimates
Casualty estimates for the Russia-Ukraine war are highly contested and vary significantly depending on source and methodology. As of early 2026, estimates from Western intelligence sources and independent analysts suggest Russian military casualties (killed and wounded) in the range of 500,000 to 700,000, with Ukrainian military casualties estimated at somewhat lower levels due to defensive positioning advantages. Civilian deaths are separately estimated, with the UN documenting tens of thousands of verified civilian deaths with actual totals likely significantly higher.
Combined, the conflict has almost certainly produced casualties in the range of one to two million people by any reasonable accounting methodology—killed, seriously wounded, and displaced into conditions of significant hardship.
This is roughly twenty times the scale of documented Gaza casualties, occurring simultaneously, with substantially less media oxygen. The disparity is not a function of importance—Ukraine’s strategic implications for European security are enormous—but of the narrative fit within existing Western moral frameworks and the visual accessibility of the conflict to Western media.
The Wag the Dog Problem
Barry Levinson’s 1997 film Wag the Dog, based on Larry Beinhart’s novel American Hero, depicts a political consultant and Hollywood producer manufacturing a fake war to distract from a presidential sex scandal. The film was released three weeks before the Monica Lewinsky story broke and six weeks before the Clinton administration launched missile strikes against Sudan and Afghanistan—a coincidence that generated significant public commentary.
The film’s central argument: that the technology and institutional capacity to manufacture convincing conflict narratives already existed in 1997, and that the people with incentives to do so had both the means and the motive.
The argument has not become less relevant as the technology has improved. Generative AI can now produce convincing video of events that did not happen. Real-time deepfakes are demonstrated and improving. The cost of producing convincing conflict imagery has dropped dramatically while the distribution infrastructure for such imagery has expanded globally.
This essay is not arguing that specific conflict footage is fabricated. It is arguing that the appropriate epistemic response to mediated information about events one cannot directly observe includes some degree of calibrated skepticism about the production choices of the mediators—not paranoid dismissal of all mediated information, but awareness that all mediated information is shaped by choices, and those choices are not neutral.
Glossary
The identified victim effect — The documented tendency of human beings to donate more, care more, and respond more to a single identified individual with a name and photograph than to statistical thousands. A product of human emotional wiring optimized for small-group environments. Exploited systematically by charities, media organizations, and political campaigns. Produces consistent distortions in the allocation of moral attention and response at scale.
Scale consistency — The application of the same metric to all comparable situations when evaluating the moral significance of a body count or atrocity. The consistent application of scale reveals significant distortions in which conflicts receive moral attention: the Congo is fifty to one hundred times larger than Gaza by casualty count and receives approximately one percent of the attention. Scale-consistent moral response does not exist in practice because moral attention is generated by narrative fit, not arithmetic.
Curated outrage — The selective application of moral intensity to situations that fit existing narrative frameworks, while comparable or larger situations that do not fit those frameworks receive less attention. Curated outrage is not necessarily conscious or coordinated—it emerges from the intersection of human emotional wiring (identified victim effect), media selection effects (which stories photograph well and travel on social media), and political incentive structures (which situations can be used to advance existing agendas). The curation is real whether or not the curators are aware of it.
The narrative fit problem — The structural tendency of media and public attention to concentrate on conflicts that map cleanly onto pre-existing moral frameworks, regardless of the arithmetic scale of the suffering. Gaza maps onto colonialism/apartheid frameworks. The Congo does not map as cleanly. The Uyghurs scramble the ideological categories (Muslim victims of a communist state). This produces systematic distortions in which suffering receives response.
Gobsmacked stability — The condition of being genuinely surprised by the gap between the described consequences of ongoing crises and the actual consequences that have materialized. Not complacency about real problems; calibrated skepticism about the gap between description and reality, and suspicion of the structural incentives that push toward overcalibration of alarm.
The outrage economy — The structural incentive system that rewards maximally alarming descriptions of current events. Attention economy metrics (clicks, shares, engagement) reward alarm. Fundraising emails reward alarm. Cable news ratings reward alarm. Think tank relevance rewards alarm. These incentives do not go away when the alarm is accurately calibrated—they push constantly toward overcalibration. The outrage economy is a mechanism that produces systematic upward bias in described stakes relative to actual stakes.
Mediated skepticism — The calibrated epistemic position that all mediated information about events one cannot directly observe is shaped by the choices of the mediators, and those choices are not neutral. Not paranoid dismissal of all mediated information; awareness that selection, framing, sound design, image choice, and narrative structure are all choices made by people with their own interests, and that those choices affect what the information conveys.
Rate versus volume — The two primary metrics for evaluating the scale of an atrocity. Volume: absolute number of deaths. Rate: deaths as a proportion of the affected population. Different metrics produce different rankings of historical atrocities, and the choice of metric is often made post-hoc to support a pre-existing conclusion rather than applied consistently. The Rwandan genocide ranks near the top by rate (approximately ten percent of the population killed in one hundred days). The Second Congo War ranks near the top by volume. Neither receives the moral attention that consistent application of either metric would demand relative to Gaza.
SDAM (Severely Deficient Autobiographical Memory) — See previous essays. Applied here: the condition produces immunity to the identified victim effect, because the mental imagery that makes identified victims more vivid and emotionally present than statistical victims is unavailable. A photograph of a dead child and a number representing one hundred thousand dead children have approximately the same weight. This produces scale-consistent emotional response as a byproduct of neurological condition rather than moral superiority. The result may be more arithmetically honest than typical human moral response.
Aphantasia — See previous essays. Applied here: the inability to visualize means the curated imagery of any particular conflict cannot produce the visceral emotional response that visual imagery is designed to produce. This is a loss—empathy requires imagination and imagination requires some capacity for mental imagery. It is also an accidental insulation against one of the primary mechanisms of curated outrage.
Chris Abraham is the founder of Gerris Corp and writes at chrisabraham.substack.com. He is gobsmacked. He remains gobsmacked. He will probably continue to be gobsmacked.