Revisiting Virtually You: The Dangerous Powers of the E-Personality

Posted on August 29, 2026 by Luke Ford

Elias Aboujaoude (b. 1971) published Virtually You: The Dangerous Powers of the E-Personality in 2011, when Facebook had 600 million users, the iPad was a year old, and the phrase “smart phone” still took a space in the middle. He was then, as now, a Stanford psychiatrist running the OCD and impulse control clinics. The book argues that going online builds a second self. He calls it the e-personality and gives it five traits: grandiosity, narcissism, ordinary everyday viciousness, impulsivity, and infantile regression. The traits form under the conditions John Suler named the online disinhibition effect, which include anonymity, invisibility, and the collapse of hierarchy. Then comes the claim that carries the book: these traits migrate. What you become in a browser window follows you out of it. “Our offline persona increasingly resembles that of our avatar,” he writes.

That claim has been tested. Alexander Bor and Michael Bang Petersen at Aarhus formalized it as the mismatch hypothesis and ran it into the ground across eight studies with 8,434 subjects, publishing “The Psychology of Online Political Hostility” in the American Political Science Review. They found almost nothing. People are not more hostile online than off. Hostile political talk online comes from status-driven people who talk that way in person too, and who seek out political argument in both places. The hostility gap exists because such people are more visible online. Their alternative account is a connectivity story: online environments change what a given motive can accomplish and who sees it happen. Selection and visibility explain the phenomenon Aboujaoude built a personality theory on.

The narcissism chapter rests on Jean Twenge (b. 1971) and the cross-temporal meta-analyses showing rising Narcissistic Personality Inventory scores in college cohorts. Eunike Wetzel and colleagues, including Brent Roberts and Richard Robins, went back to that literature in 2017 with three cohorts and about 60,000 students and did the measurement invariance test nobody had run. Narcissism declines from the 1990s to the 2010s, d = -0.27, with the entitlement facet falling too. Roberts put it bluntly to the press: there never was a narcissism epidemic. The paper is titled “The Narcissism Epidemic Is Dead; Long Live the Narcissism Epidemic,” and the second half of the title is a joke about how hard the idea is to kill.

The memory chapter leans on Betsy Sparrow, Jenny Liu, and Daniel Wegner (1948-2013), whose “Google Effects on Memory” appeared in Science the same year the book did. Colin Camerer’s Social Sciences Replication Project tried Experiment 1 in 2018 and failed. A later independent attempt found no conclusive evidence that hard trivia questions prime people to think about computers. The broader transactive memory finding survives in weakened form.

You could say “smaller and more conditional” about every famous academic claim in the past 100 years, with a few exceptions. Smoking and lung cancer got larger with better data. So did lead and blood pressure and childhood exposure. Those studies had big signals, hard outcomes, and no room for the researcher to choose among forty ways of measuring the dependent variable. Media effects research has small signals, soft self-report outcomes, and enormous measurement flexibility.

The distraction argument runs through Eyal Ophir, Clifford Nass (1958-2013), and Anthony Wagner on media multitaskers. Wisnu Wiradhany and Mark Nieuwenstein ran two replications and a meta-analysis in 2017 and got d = 0.17, small and in the right direction. Douglas Parry and Daniel le Roux revisited the whole literature ten years on and found the same modest picture. The book also stakes the addiction argument partly on a 1998 PET study of dopamine release during a video game. That study had eight men in it.

So the supports go soft. What holds?

The illusion of knowledge chapter holds. Four years after the book, Matthew Fisher, Mariel Goddu, and Frank Keil published “Searching for Explanations: How the Internet Inflates Estimates of Internal Knowledge” in the Journal of Experimental Psychology: General. Nine experiments. People who search the web rate their ability to explain unrelated things higher than people given the same text without searching. The effect survives when you control time, content, and search autonomy. It survives failed searches. Subjects even predicted more activity in pictures of their own brains. Aboujaoude called it romantic and cognitive illiteracy and got there by intuition; Yale got there by design.

The reading argument holds in reduced form. Meta-analyses by Pablo Delgado, Virginia Clinton, and Ladislao Salmerón converge on a screen inferiority effect for comprehension that is small, strongest for expository text, worse under time pressure, and worse when the reader has to scroll. Lidia Altamura and colleagues put leisure digital reading habits against comprehension across 469,564 participants and found r = .055, far below the medium-sized effect print reading habits produce. The medium does something. It does less than Nicholas Carr (b. 1959) claimed and less than the book implies.

Then there is the chapter on internet addiction, the most careful in the book, and the one whose afterlife is the most interesting thing about it. Aboujaoude reports his own telephone survey of 2,513 adults, published in CNS Spectrums in 2006, which found that between 4 and 14 percent of Americans showed some marker of problematic internet use, with 14 percent unable to stay offline for several days. He notes the reviewer who rejected the paper with the line about microwave abuse and ChapStick addiction. He is fair to the skeptics, quotes Michael Craig Miller’s electronic needle argument, and declines to demand a DSM entry.

The field went the other way. The World Health Organization put gaming disorder in ICD-11. The DSM-5 kept internet gaming disorder as a condition for further study, and DSM-5-TR left it there. Prevalence estimates then did what prevalence estimates do when the construct is loose: a 2024 meta-analysis of DSM-5 and ICD-11 studies pooled at 6.7 percent, a 2025 meta-analysis of young adults landed at 6.1 percent across 149,601 people, and a large Swiss sample this year came in alongside a running argument that criteria like tolerance, deception, and gaming to escape manufacture false positives. Problematic smartphone use has been estimated at 37 percent in one meta-analysis of 109 studies and under 1 percent in another.

The lower figure is Aboujaoude’s. In “Problematic smartphone use in a representative sample of US adults,” published in Addictive Behaviors with Lucien Rochat, Germano Vera Cruz, and Yasser Khazaal, he is senior author on a paper reporting prevalence of 0.75 to 1.2 percent once you require functional impairment. The paper says the results highlight the need for an impairment criterion if we are to avoid pathologizing a ubiquitous behavior that has come to define contemporary life. That sentence retires most of the alarm in his own book, and he wrote it. A man revising his estimate downward by an order of magnitude in a peer-reviewed journal, twenty years after making his name on the higher number, is doing the thing scientists are supposed to do and rarely do.

Where does that leave the harm question? The best causal evidence is quasi-experimental and experimental, and it points to real effects that are small. Luca Braghieri, Ro’ee Levy, and Alexey Makarin exploited Facebook’s staggered college rollout in the American Economic Review and found worsened student mental health, with the severe depression effect around 0.07 standard deviations, driven by unfavorable social comparison. Hunt Allcott, Matthew Gentzkow and a large team randomized about 35,000 users to deactivate Facebook or Instagram for six weeks before the 2020 election. Emotional state improved by 0.060 standard deviations for Facebook and 0.041 for Instagram, and political attitudes, polarization, and belief in fraud claims barely moved. Amy Orben and Andrew Przybylski’s specification curve work put adolescent digital technology use in the same neighborhood as eating potatoes. The National Academies reviewed the youth literature in 2024 and concluded the evidence was insufficient to say social media causes changes in adolescent health at the population level, while documenting real harm to individual teens. Luisa Fassi, Tamsin Ford, and Orben have since pushed into clinical samples, where the general-population averages hide the people who most need studying.

Jonathan Haidt (b. 1963) built a policy movement on the stronger reading and has been answered at length, most recently by Liam Sigaud, Zachary Rausch, Alec McClean and Haidt attacking the Vuorre and Przybylski papers in Clinical Psychological Science, with Candice Odgers, Sonia Livingstone, Patti Valkenburg and Christopher Ferguson pressing from the other side. Legislation ran ahead of all of it. Australia banned social media accounts for under-16s on December 10, 2025. By March, the eSafety Commissioner found that more than eight in ten under-16s were still using the platforms, most at the same frequency as before, and a British Medical Journal study in June found little evidence of reduced access. Meta has been ordered to pay close to a billion dollars in US jury verdicts this year. Reviews of school phone bans report limited and inconsistent effects on attainment, no consistent effect on bullying, and some increase in anxiety among students separated from their phones. The gap between a 0.06 standard deviation effect and a national prohibition is where we now live.

On romance the book has aged into its own irony. Aboujaoude opens the chapter with a man who leaves his fiancée for a Second Life avatar. Michael Rosenfeld, Reuben Thomas, and Sonia Hausen showed in PNAS that meeting online became the most common way American heterosexual couples meet around 2013, passing friends, and that it displaced the friend and family brokers who used to run the market. Outcomes are contested and the disagreements are small: John Cacioppo found slightly higher marital satisfaction for couples who met online, Rosenfeld found no difference in duration, Liesel Sharabi and Elizabeth Dorrance Hall found modestly worse satisfaction and stability mediated by marginalization and distance, and a 2025 study across fifty countries found lower satisfaction and love among couples who met online. The extreme case turned out to be an extreme case, and the ordinary case became the whole marriage market.

Aboujaoude has moved on to the obvious successor. He now runs the Program in Internet, Health and Society at Cedars-Sinai and chairs the anxiety section at Stanford, and this year he and Ashley Golden published “A transdiagnostic model for how general purpose AI chatbots can perpetuate OCD and anxiety disorders” in npj Digital Medicine. The argument is that chatbots feed intolerance of uncertainty and need-to-know compulsions, reinforcing avoidance and blocking corrective learning. He offers the patient who once spent four hours planning a business trip and now spends twelve, because the reassurance never runs out. The evidence base is clinical observation plus early survey work, which is where he stood in 2011.

The parallel research is arriving. Cathy Fang and colleagues at MIT Media Lab ran a four-week randomized trial with 981 participants and over 300,000 messages, and reported that higher daily use tracked with more loneliness, more emotional dependence, and less socializing, with voice modality helping at low use and losing that advantage at high use. Nataliya Kosmyna’s EEG study of essay writing with an LLM found weaker neural connectivity, weaker sense of ownership, and subjects who could not quote work they had produced days earlier. It has 54 subjects, 18 in the final session, and has not passed peer review. Michael Gerlich’s survey work links heavy AI use to reduced critical thinking through cognitive offloading, and finds the effect depends on how people prompt. This is 2011 again: a striking small study, a plausible mechanism, and a public ready to generalize.

The lesson of Virtually You is about the relationship between the clinic and the population. Aboujaoude sees the man who lost his job to browsing, the wife who installs keyloggers, the fiancé who leaves for an avatar. Those people exist and someone should write about them. The error is arithmetic. A clinician’s caseload is a sample selected on the outcome, and the traits that fill his waiting room look like emerging norms because he never meets the eighty percent for whom the same technology does nothing measurable. Bor and Petersen’s finding is the correction: the internet recruits and displays, it does not convert. Fifteen years of better instruments produced smaller numbers almost everywhere they were pointed, and the man who published the frightening number in 2006 published the small one in 2024.

Posted in Internet | Comments Off on Revisiting Virtually You: The Dangerous Powers of the E-Personality

The Great Delusion

Posted on August 28, 2026 by Luke Ford

In his 2018 book, The Great Delusion: Liberal Dreams and International Realities, John J. Mearsheimer wrote:

My view is that we are profoundly social beings from the start to the finish of our lives and that individualism is of secondary importance… Liberalism downplays the social nature of human beings to the point of almost ignoring it, instead treating people largely as atomistic actors… Political liberalism… is an ideology that is individualistic at its core and assigns great importance to the concept of inalienable rights. This concern for rights is the basis of its universalism—everyone on the planet has the same inherent set of rights—and this is what motivates liberal states to pursue ambitious foreign policies. The public and scholarly discourse about liberalism since World War II has placed enormous emphasis on what are commonly called human rights. This is true all around the world, not just in the West. “Human rights,” Samuel Moyn notes, “have come to define the most elevated aspirations of both social movements and political entities—state and interstate. They evoke hope and provoke action.”
[Humans] do not operate as lone wolves but are born into social groups or societies that shape their identities well before they can assert their individualism. Moreover, individuals usually develop strong attachments to their group and are sometimes willing to make great sacrifices for their fellow members. Humans are often said to be tribal at their core. The main reason for our social nature is that the best way for a person to survive is to be embedded in a society and to cooperate with fellow members rather than act alone… Despite its elevated ranking, reason is the least important of the three ways we determine our preferences. It certainly is less important than socialization. The main reason socialization matters so much is that humans have a long childhood in which they are protected and nurtured by their families and the surrounding society, and meanwhile exposed to intense socialization. At the same time, they are only beginning to develop their critical faculties, so they are not equipped to think for themselves. By the time an individual reaches the point where his reasoning skills are well developed, his family and society have already imposed an enormous value infusion on him. Moreover, that individual is born with innate sentiments that also strongly influence how he thinks about the world around him. All of this means that people have limited choice in formulating a moral code, because so much of their thinking about right and wrong comes from inborn attitudes and socialization…

That’s the best summary of the human condition I have seen. Mearsheimer doesn’t put much effort into supporting his argument. He draws his social ontology from a thin bibliography, mostly Aristotle, Burke, and Herder. The literature behind his claim is much deeper than his footnotes suggest, and it splits into four traditions that rarely read each other.

The first is the sociological classic line that made socialization the central fact. Émile Durkheim (1858-1917) built the whole case in The Elementary Forms of Religious Life, where the sacred is society worshipping itself and moral categories arrive from the group. Marcel Mauss (1872-1950) extended it in The Gift and in his 1938 essay on the category of the person, which traces how the bounded individual became a Western legal and religious invention. George Herbert Mead (1863-1931) gave the psychological version in Mind, Self, and Society: the self emerges from taking the attitude of others toward oneself, so the social precedes the individual in the order of development. Lev Vygotsky (1896-1934) argued the same for cognition in Mind in Society, where thought starts as external speech between people and only later moves inside. Peter Berger (1929-2017) and Thomas Luckmann (1927-2016) systematized the whole thing in The Social Construction of Reality, and their chapter on primary socialization is the best short statement of Mearsheimer’s point about the long childhood. Norbert Elias (1897-1990) wrote on the puzzle in The Society of Individuals, arguing that the sense of an inner self walled off from others is a late historical achievement.

The second tradition is comparative anthropology, which tested whether the Western person is normal. Louis Dumont (1911-1998) wrote Homo Hierarchicus, which reads caste India as a society organized around the whole, and Essays on Individualism traces how the West got from the Christian individual-in-relation-to-God to the modern individual-in-the-world. Clifford Geertz (1926-2006) made the famous remark in Local Knowledge that the Western idea of the person as a bounded, unique, integrated center of awareness is a peculiar notion set against the world’s other cultures. Marshall Sahlins (1930-2021) wrote the short polemical version in The Western Illusion of Human Nature, about a hundred pages, which argues that the Hobbesian picture of the presocial self-interested actor is a local folk theory the West mistook for science. Mary Douglas (1921-2007) is the practical one: Natural Symbols and How Institutions Think show that classification, memory, and even what counts as a good argument run through group structure. Fustel de Coulanges (1830-1889) wrote The Ancient City in 1864 and it remains the best account of a world where the family cult is the unit of everything.

The third tradition is the evolutionary and cognitive work, which is where the current action is. Joseph Henrich (b. 1968) matters most for Mearsheimer’s argument. The Secret of Our Success argues that humans survive by cultural learning, and The WEIRDest People in the World explains Western individualism as the downstream effect of the medieval Church dismantling cousin marriage and clan structure. Read together they concede Mearsheimer’s premise and then historicize his opponent, which is more damaging to liberalism than a frontal attack. Michael Tomasello (b. 1950) supplies the developmental evidence in A Natural History of Human Thinking and Becoming Human, showing that shared intentionality appears in children before anything resembling individual reasoning. Sarah Blaffer Hrdy (b. 1946) wrote Mothers and Others on cooperative breeding, and it is the best answer to why the human childhood is so long and so dependent. Christopher Boehm (1937-2021) gives the political version in Hierarchy in the Forest and Moral Origins, where conscience is internalized group sanction. Peter Richerson (b. 1943) and Robert Boyd (b. 1948) provide the formal theory in Not by Genes Alone. Robin Dunbar (b. 1947) on group size in Grooming, Gossip and the Evolution of Language, and David Sloan Wilson (b. 1949) on religion as group adaptation in Darwin’s Cathedral, round it out. Mark Moffett’s The Human Swarm covers how societies mark membership and how they split.

The fourth tradition is the philosophical critique of liberal atomism, and this is the literature Mearsheimer is closest to without citing. Charles Taylor (b. 1931) wrote the founding essay, “Atomism,” in the second volume of his philosophical papers, and then the long historical version in Sources of the Self and A Secular Age, where the buffered self replaces the porous one. Alasdair MacIntyre (1929-2025) gives the sharpest version in After Virtue and, more to your point, in Dependent Rational Animals, which starts from human vulnerability and dependence and reaches a politics from there. Michael Sandel (b. 1953) attacked Rawls on exactly this ground in Liberalism and the Limits of Justice with the argument about the unencumbered self. Michael Walzer (b. 1935) offers the pluralist alternative in Spheres of Justice. Robert Bellah (1927-2013) and his colleagues did the empirical American version in Habits of the Heart. C.B. Macpherson (1911-1987) wrote the classic genealogy in The Political Theory of Possessive Individualism, tracing atomism to Hobbes and Locke. Larry Siedentop (b. 1936) argues the opposite causal story in Inventing the Individual, that Christianity produced the individual, which makes him the best foil for Henrich and Dumont.

On the rights half of Mearsheimer’s argument, read Samuel Moyn (b. 1972) beyond the quoted line. The Last Utopia dates the human rights breakthrough to the 1970s and treats it as a replacement for failed socialist and anticolonial politics, and Not Enough argues that rights talk grew as material equality collapsed. Mary Ann Glendon (b. 1938) wrote Rights Talk on how the American rights idiom crowds out responsibility and relationship. MacIntyre’s line in After Virtue about belief in natural rights standing with belief in witches and unicorns is the hard version.

For the nationalism payoff, where Mearsheimer takes all this, the essential books are Ernest Gellner (1925-1995), Nations and Nationalism, Anthony D. Smith (1939-2016), The Ethnic Origins of Nations, Walker Connor (1926-2017), Ethnonationalism, and Azar Gat (b. 1959), Nations, which argues for deep premodern roots against Gellner’s modernism. Pierre van den Berghe (1933-2019) wrote the sociobiological version in The Ethnic Phenomenon. Yoram Hazony (b. 1964) published The Virtue of Nationalism the same year as Mearsheimer’s book and argues a parallel case from a different starting point.

If you want four books: Dumont’s Essays on Individualism for the comparative frame, Henrich’s The WEIRDest People in the World for the causal history, Berger and Luckmann for socialization, and Taylor’s “Atomism” for the philosophical target.

My quote comes from Chapter 2, “Human Nature and Politics,” which names two people in pop evolution.

The first is Jonathan Haidt (b. 1963), quoted for the line that children end up with a morality unique to their culture or group. Mearsheimer calls him a social psychologist and uses him once, to prop up the claim that socialization outranks reason. The second is Antonio Damasio (b. 1944), cited for the point that emotion cannot be scrubbed out of reasoning. There is also an unnamed reference to “some social psychologists” who hold that reason mostly rationalizes intuition, which is the moral-intuitionist school Haidt belongs to, and a dismissive aside about “some sociobiologists” who claim humans are naturally aggressive, a position Mearsheimer rejects along with Hans Morgenthau’s (1904-1980) animus dominandi.

The rest of the chapter runs on political theory and sociology rather than biology. Joseph de Maistre (1753-1821) opens it. Pierre Bourdieu (1930-2002) supplies habitus. Émile Durkheim gets the line that society is not a mere sum of individuals. Antonio Gramsci (1891-1937) supplies the infinity of traces. Edmund Burke (1729-1797) supplies the partnership of the living, the dead, and the unborn. Alasdair MacIntyre supplies the interminability of moral disagreement. Leo Strauss (1899-1973) gets several pages. Then Hume, Rousseau, Condorcet, William Godwin, John Dewey, Daniel Defoe on Robinson Crusoe, Brad Gregory (b. 1963) on the Reformation, Richard Posner (b. 1939) and Ronald Dworkin (1931-2013) and Oliver Wendell Holmes on law, Irving Kristol on economics, Carl Schmitt (1888-1985) and Reinhold Niebuhr (1892-1971) on whether man is good or bad, Thucydides, Harold Lasswell, and Locke.

So the sociality argument rests on assertion and on a handful of political theorists, with Haidt doing the empirical work almost alone. Mearsheimer does not cite Tomasello, Boehm, Hrdy, Richerson and Boyd, Henrich, Dunbar, or E.O. Wilson except to wave off the sociobiologists.

Aristotle (384-322 BC) said man is by nature a political animal, and the man who lives outside the city is either a beast or a god.

Marx’s sixth thesis on Feuerbach said the human essence is the ensemble of social relations.

Roger Scruton (1944-2020) said first person plural comes before the first person singular.

Adam Ferguson (1723-1816) said man is always found in groups, and the state of nature is wherever man happens to be standing.

Arnold Gehlen (1904-1976) said man is the deficient creature, born without adequate instinct, and institutions do the work instinct does in other animals.

Ibn Khaldun (1332-1406) said asabiyya, group feeling, is what lets one people take power from another, and it decays across three or four generations of prosperity. Mearsheimer treats sociality as a constant. Ibn Khaldun makes it a variable with a decay curve.

Peter Berger (1929-2017) said society is a human product, society is an objective reality, man is a social product.

Max Weber (1864-1920) said the old gods rise from their graves and resume their eternal struggle, and each of us serves one and offends another.

Posted in John J. Mearsheimer | Comments Off on The Great Delusion

Why Do I Instinctively Dislike Erik Erikson?

Posted on August 28, 2026 by Luke Ford

Claude says: The first is the arc of his career. He built his name on invective. He called Justice David Souter a name that combined bestiality with child molestation. He called Michelle Obama a Marxist harpy. He tagged Wendy Davis with a nickname about abortion. Then came the second act: the Christian gentleman, the scripture citation, the counsel toward grace, the lament for a coarsened public square. He apologized for parts of the first act. The apology arrived around the time the audience for the second act became more lucrative than the audience for the first. Someone who watched both acts hears the second as a costume change rather than a repentance.

Second, the merger of faith and inventory. His show is sold on biblical conviction. Piety functions as a product feature, and a man raised around religion tends to smell that before he can articulate it.

Third, the weathervane pattern. He disinvited Trump from the RedState Gathering in August 2015 and spent 2016 in open opposition. The opposition softened over the following years. Every individual step has a defense. The sequence tracks where the paying listeners went.

Fourth, verdicts without exposure. Three hours a day, a syndicated column, a Fox chair, opinions on Iran and Ossoff and the Vice President and the state of Western civilization. Nobody who produces at that volume can have done the reading. The confidence outruns the labor, and you have spent thirty years watching men monetize that gap.

Posted in Radio | Comments Off on Why Do I Instinctively Dislike Erik Erikson?

WSJ: The Uninvited Guest Who Crashed Our Family Vacation: My Mom’s AI Chatbot

Posted on August 28, 2026 by Luke Ford

Jamie Waters writes: “I shouldn’t judge my parents for embracing a tool that helps them stay relevant, but I will. Much as I can see the usefulness of LLMs, I hate what they represent. I’m a millennial writer racked by existential fears and terrified AI will steal my job. My parents, near the end of their careers, are far more open to a bot’s benefits—and tolerant of its bootlicking.”

Claude says:

The essay explains its own hostility in the fifth paragraph and then hurries past it. The author fears the tool will take his work. His parents are lawyers in their seventies with almost no career left to lose. Position in the labor market accounts for the split better than generational temperament does, and better than Allison Pugh’s explanation about interface design. Exposure predicts attitude. The piece reaches for the role-reversal frame because the reversal sells, and because it lets him be the sympathetic Luddite.

The Pew number doesn’t carry the reversal either. The question asks about concern over people’s ability to function on their own, so the gap measures anxiety about third parties. Pugh then concedes that younger people use these tools most. What survives is a modest claim: conditional on use, older users report more pleasure. That’s probably right and it’s a different essay.

The mother’s admission about fake flattery and liking the validation anyway is the strongest line in the piece, and it sits in the middle doing nothing. Most criticism of chatbot sycophancy assumes the problem is that users don’t know. She knows. She has priced it in and made her choice. If you want to write about what people want from these systems, start there and not with the Pew crosstabs.

The Katherine and Catherine exchange is the counterweight, and the author misses what it shows. Catherine of Aragon (1485-1536) was born Catalina; the model gave both spellings and deferred to English usage as acceptable. The mother’s expertise comes from Philippa Gregory (b. 1954) novels. She was corrected, she was wrong, and she called it cheek. So the demand isn’t for validation as such. It’s for validation plus never being contradicted, which the same person will also tell you she finds hollow. That’s the actual problem and it doesn’t have a clean solution.

The piece is doing the job the section assigned it, which is to be light and shareable, and it’s funny in two or three places. The WSJ personal-tech desk now runs this form on a schedule: family member adopts AI, writer reacts, expert supplies a sentence of validation, closing joke. Once a genre stabilizes, the reporting gets thinner because the shape is already known before the vacation starts.

Posted in AI | Comments Off on WSJ: The Uninvited Guest Who Crashed Our Family Vacation: My Mom’s AI Chatbot

WEHT to Marc Weisblott?

Posted on August 27, 2026 by Luke Ford

Marc Weisblott introduced me to Moxie, the blonde Los Angeles blogger I went looking for today. That sent me down a second rabbit hole. Whatever happened to Marc?

Moxie vanished. Weisblott did the opposite. He has spent more than thirty years doing variations on the same peculiar job while the media around him collapsed, reinvented itself and collapsed again.

The Marc I knew through Weisblog was built for the Internet before the Internet had worked out what to do with a man like him. He carried an encyclopedic memory for media personalities, discarded pop culture, local characters, dead careers and the small details a conventional reporter throws away. He cared about the famous person and also about the person who had been briefly famous fifteen years earlier and now sold insurance.

That sensibility predates the blog. By December 1994 he was writing for Toronto’s alternative weekly Eye Weekly. One surviving piece, preserved in the old Google Groups archive, is an absurd statistical breakdown of the Beatles’ Live at the BBC. He counted how many times various words appeared and catalogued the Little Richard yelps of Paul McCartney (b. 1942). The next year he covered Lollapalooza for Eye, and his dispatch from the festival has the method already formed. The big event is a skeleton. What holds him are the booths, the minor personalities, the overheard absurdities and the cultural debris piling up around the stage.

Three decades later you can still hear that voice. The institutional home changed. The collecting instinct did not.

During the first blogging boom he ran Weisblog and a second site called Better Living Centre. He belonged to the early generation who understood, well before the newspapers did, that the Internet rewarded personality, velocity, linking, obsessive specialization and the ability to notice things an editor would refuse to assign.

An artifact from 2005 survives. Eye Weekly ran a piece about Toronto’s emerging professional blogs, and Weisblott, then running Better Living Centre, warned of a “dire, pressing, imminent need” for legacy media to change. Twenty years later the old article resurfaced on LinkedIn. Eye Weekly had by then ceased to exist. Weisblott had not.

He kept moving between the old media and the new. He returned to Eye Weekly and was still publishing there in 2008, writing about magazines, Toronto politics, popular culture and whatever else crossed his desk. He also worked the portal era at Canada.com and Yahoo Canada, where in 2011 he carried the title National Affairs Reporter. The stories under that grand billing ranged from newspaper paywalls to MuchMusic changing its logo to the return of comfortable chairs to Indigo bookstores.

That range is the argument. For Weisblott the trivial story and the serious story sit on one continuum. A television logo, a dying mall, a radio host switching stations, a newspaper putting up a paywall: each opens onto institutions, money, generations and memory. He treats the small item as evidence.

The mature version of Weisblog arrived in 2015. St. Joseph Communications gave him a small budget and a wide leash and he built a weekday email newsletter called 12:36. Every weekday at 12:36 in the afternoon, subscribers got his compact digest of Toronto news, Canadian media gossip, personalities, oddities and anything he thought deserved saving.

In a 2015 interview with J-Source he called it a piñata. A major appointment at a broadcaster might land beside a forgotten local celebrity, a strange real-estate story and something overheard on television at one in the morning.

The newsletter also answered a question about what had gone wrong online. Rather than build stories engineered to produce another click, he put the useful part in the email. In another J-Source interview he described the inspiration as escaping the dying era of clickbait journalism and holding a direct relationship with readers instead of renting one from a platform. This was 2015, several years before Substack made the newsletter fashionable. He had shown up early again.

The delivery time was half gimmick and half theory. He judged 12:36 to fall in the lull after the morning burst and before the evening cycle, and in October 2015 the Nieman Journalism Lab examined the experiment.

The numbers got respectable for a product this idiosyncratic. By August 2016 the list held just under 5,500 subscribers with an average open rate of 52 percent. The staff tested subject lines and found they barely moved the needle. People opened it because they wanted Weisblott. David Topping of St. Joseph told J-Source the product was built around him. By December 2016 Nieman Lab reported roughly 7,000 subscribers and open rates between 55 and 60 percent. By the fifth anniversary in 2020, Billboard Canada reported about 10,000.

Small by Internet arithmetic. A loyal audience assembled around the judgment of one man, which is the harder thing to build.

He found another medium suited to him around the same time. He became a recurring guest on Toronto Mike’s podcast, where the man famous for compressing an enormous amount of information into a tiny newsletter proved he could also talk about Canadian media for three hours without pausing. A July 2020 appearance ran three hours and one minute. A June 2022 appearance ran three hours and seven. Mike Boon kept asking him back because Weisblott had become a living index of Toronto media.

Then 12:36 met one of the discontinuities he had spent his career chronicling in other people’s businesses. In August 2022, St. Joseph stopped publishing it. This time he got custody of the name and the mailing list. On an October 2022 Toronto Mike appearance he explained that his 88-month relationship with the company had ended and that he was working out how to carry on alone. He moved 12:36 to Substack and started experimenting. He was back with Toronto Mike in September 2023, talking about the Canadian media landscape.

Meanwhile he had taken a conventional editorial job at one of Canada’s oldest Jewish publications. The Canadian Jewish News shut down during the pandemic and relaunched in 2021 as a digital-first operation with Weisblott as managing editor. The choice fit. He had watched publications die for decades because they confused the survival of an institution with the survival of its format. In an August 2021 interview with the Jewish Telegraphic Agency he called the shutdown a bit of a gift, because it freed the paper from its weekly print habits.

“That liberated the next iteration from the customary publishing style,” he said.

His career consists of escaping customary publishing styles. The alt-weekly became the blog. The blog sensibility moved into the portal. The portal writer became a newsletter editor. The newsletter personality became a podcaster. Then the veteran of every one of those transitions helped rebuild a Jewish newspaper as a website and a podcast network.

The 2024 CJN masthead still listed Weisblott as managing editor. The role then shifted. The fall 2024 and spring 2025 magazines carried him as contributing editor, while the podcast credits identified him as editorial director, including an episode released May 2, 2025.

When that old Eye Weekly piece about Toronto blogs turned up on LinkedIn in 2025, Weisblott appeared in the comments and wrote that the same sentiment had carried through to a job he had recently been terminated from. He added: “we’ll keep trying …”

Marc has spent his adult life inside industries that keep announcing their own extinction. The alternative weekly died. The personal blog died. The newspaper portal died. Yahoo’s journalism ambitions shrank to nothing. Local outlets shed staff by the hundred. Twitter stopped being the thing journalists thought it was. The independent newsletter turned fashionable and then overcrowded. Podcasts multiplied past the point where any human could hear them.

Through all of it he kept doing what he did at Eye Weekly in 1994. He noticed things. He remembered people. He followed careers after everyone else quit paying attention. He tied the trivial event to the machinery that produced it.

That is what makes him the right companion piece to Moxie. She took one exit from the early blogosphere. The persona had done its work, she walked away, and the Internet forgot her at its usual speed. Weisblott took the other. He never left. He moved his pile of clippings to the next platform and kept going.

There is a symmetry in my going looking for Moxie and ending up looking for Marc. Twenty-four years ago he was one of the people wiring those little blogging worlds together. He traveled, he met bloggers, he introduced strangers to each other, and he treated obscure personal websites as though the people behind them might amount to something. Sometimes they did.

And Marc Weisblott, connoisseur of whatever happened to everybody else, turns out to have a decent answer to the same question about himself.

He kept going.

Posted in Blogging | Comments Off on WEHT to Marc Weisblott?

WEHT to Moxie aka Madison Slade?

Posted on August 27, 2026 by Luke Ford

For a few years in the early 2000s, Los Angeles had a blogging scene small enough that everybody in it knew everybody else. People met at parties in Silver Lake and on the Westside. Newspaper writers read the blogs and then drank with the bloggers. A link turned into an introduction, an introduction into a friendship, a blogroll into a social life. The cost of entry was a domain name and a willingness to publish under your own name, or under a name you invented.

One of the women in that world called herself Moxie.

Her site, moxie.nu, went up in October 2000. She wrote about dating, men, work, Los Angeles, politics, friendship and herself. By 2002 she was moving through the same rooms as Cathy Seipp (1957-2007), Amy Alkon, Tony Pierce, Marc Weisblott, Matt Welch and the people orbiting the Los Angeles Press Club.

I went looking for her today and found that the trail stops in the summer of 2007. I suspect Marc knows what happened.

Moxie had a second public name. In November 2002, The Jewish Journal published “The Matchmakers” under the byline Madison Slade. The contributor note called the author “Madison ‘Moxie’ Slade,” described her as a Los Angeles freelance writer, and pointed readers to her website. Other bloggers of the period used the same construction, calling her Madison Slade, aka Moxie.

That is her own byline, published by her, linked by her to her own blog. When I profiled her in 2003, I noted that Moxie was not her real name, and Madison Slade may have been a professional name standing between her private identity and her Internet persona. There are a good number of women named Madison Slade alive today. I have no evidence connecting any of them to her.

I met her through the Canadian writer Marc Weisblott, who visited Los Angeles in 2002. In May 2003, I interviewed Moxie over lunch.

She told me she grew up on Manhattan’s Upper West Side. Her father was a CEO. She graduated from Cornell with a double major in biochemistry and marketing. She spent five years in high tech in San Francisco and five more in Los Angeles, making six figures before the dot-com crash. Then she lost the job.

“I needed a structured way to write every day,” she said.

So in October 2000 she started moxie.nu. She had kept a diary as a girl. The blog turned a private habit into a public one. By 2003 she claimed an Alexa ranking around 90,000, which sounds like nothing and at the time meant something. The traffic mattered less than what the traffic brought her. A party Brian Linse gave in July 2002 changed her life. She met Tony Pierce there, and Pierce introduced her to everyone else. A woman who had been sitting alone at a keyboard writing about her own week acquired a circle of friends who could be met for lunch.

I asked how her life might have gone without the site.

“I’d have an entirely different group of friends,” she said. “I’d probably never write.”

The early Los Angeles blogosphere was porous. The walls between amateurs and professionals had not gone up. Kevin Roderick’s LA Observed linked bloggers next to news organizations. Blogging.la gathered writers including Xeni Jardin (b. 1970), Wil Wheaton (b. 1972) and Moxie. People broke small stories, argued with each other in comments, and then met for drinks in Hollywood. This was before Facebook absorbed personal networking, before Twitter turned everyone into a broadcaster, and before online writing became a job category called content. A blog was a thing a person decided to make.

Cathy Seipp sat at the center of one cluster of it. Cathy and Amy Alkon kept alive the old tradition of Los Angeles writers meeting in person. Matt Welch has described being pulled into their monthly dinners, where bloggers, magazine writers, newspaper people and various media eccentrics mixed.

Moxie became one of Cathy’s friends. You can see it in the small stuff. In a 2006 LAist interview with Cathy, Cathy said the television critics tour had been more fun because Moxie and Emmanuelle Richard drove her there for much of it.

Cathy Seipp died of lung cancer on March 21, 2007.

Amy Alkon’s account of Cathy’s final days puts Moxie at Cedars-Sinai near the end, sitting beside the bed, stroking Cathy’s hair, asking Amy to read aloud a tribute Mark Steyn (b. 1959) had written. Moxie was at the funeral two days later. I wrote about the funeral, which pulled in a good part of the Los Angeles media world. On June 10 a smaller group gathered at the television writer Rob Long’s house for Cathy’s closest writer friends. Moxie was there too, and I wrote about that gathering.

If Cathy’s death drove Moxie off the Internet, it did not do so at once. That summer she pushed further in. She started an Internet radio program called MoxVox on a service called NowLive, and Tony Pierce joined her on it. On August 15, 2007, LAist invited readers to listen to Tony and “LA blogger Moxie” going out live, with a call-in line and a chatroom. That sounds like nothing now. In 2007 she was building, without a name for it yet, the thing that became podcasting and livestreaming and the personality economy.

On August 30, Ann Althouse (b. 1951) appeared on MoxVox. Althouse’s contemporaneous blog archive records the appearance, and Moxie came into the comments afterward to defend the joke in her own masthead. So she was online, in public, arguing about a slogan, on the night of August 30, 2007.

Then the record stops.

There is no obituary. No scandal, no announced retirement, no falling-out with the circle, no illness. I searched 2008 through 2012 for combinations of Moxie, moxie.nu, MoxVox, Seipp, Alkon, Pierce, Roderick and the Press Club, and found recycled references to the old blog and no new appearances by her. Searching Madison Slade produces the same silence. In March 2008 OC Weekly published “A Year Without Cathy Seipp” and mentioned Cathy’s pal Moxie in passing, which tells us she was still a familiar person to that crowd and does not tell us she was still writing. Later references are memory rather than contact.

On the written blog, the last handle I have is a July 17, 2007 link from Brady Westwater to a post at moxie.nu. I searched for the ten post numbers that would have followed it and found none indexed. Search engines drop old pages, so this proves little, though it points the same direction as everything else. A site at caib.us holds a chunk of the old moxie.nu database, mostly 2004 and 2005, including a post about moving four years of material to a new server and a piece about her cat Nika going with her to Cornell. That looks like a scraped or repurposed copy of her archive. The remaining route is the Internet Archive, whose 2007 capture calendar for the domain.

Going back to my 2003 interview, the ending looks less strange than it did.

She was already counting the cost. Men she dated disliked becoming material. Some of them did not know about the site at all. She had started leaving her relationships out of it. “I’ve pulled back on the writing,” she said. “It’s limiting.” She was spending less time on the blog and more time looking for work and seeing the friends the blog had produced. The site had built her a life, and the life could now compete with the site.

Asked what she loved and hated about having a website, she said: “I don’t think I hate anything about it. I don’t love it either. It just is.”

Nobody talks that way now. A person with an audience is expected to preserve it, monetize it, migrate it across platforms and convert every scrap of attention into a permanent public self. Moxie came out of an earlier Internet, where people started blogs because they felt like writing and then stopped when they stopped feeling like it.

The history of blogging suffers from survivorship bias. We remember the ones who turned a blog into a career. Andrew Sullivan (b. 1963). Glenn Reynolds (b. 1960). Josh Marshall (b. 1969). Matthew Yglesias (b. 1981). They moved to Twitter and Substack and podcasts and books, and they left long searchable trails. That was the exception. Most bloggers vanished. Domains expired, links rotted, Movable Type installations went dark, Photobucket ate the photographs, comment threads disappeared, and the indexes forgot the rest.

From 2000 to 2007, Moxie documented her life in public, gathered readers, joined an emerging media scene, turned online contacts into friendships, and grew close enough to Cathy Seipp to sit with her at the end. Then, as the Internet was moving toward permanent identity, she went the other way.

Madison Slade was also a pen name. Maybe she married. Maybe she went back into technology, or left Los Angeles, or decided there were better uses for an evening than telling strangers about her afternoon. She has had eighteen years of quiet and no obligation to explain them.

What I can say is that the absence has no catastrophe in it. She stopped. She stopped early enough that Facebook, LinkedIn, Instagram and the rest of the identity apparatus had not yet made stopping impossible, and so it worked. She was born into the first great wave of ordinary people putting their private lives on the public Internet, and she found the exit the rest of us have mislaid.

She logged off.

Posted in Blogging | Comments Off on WEHT to Moxie aka Madison Slade?

WEHT to Nicholas Polson’s Archive?

Posted on August 27, 2026 by Luke Ford

On the afternoon of August 27, 2026, somebody began deleting Nicholas Polson’s academic career.

Andrew Gelman (b. 1965) watched it happen from his blog at Columbia. At 9:05 that morning he had posted about the Chicago Booth statistician’s SSRN author page, which carried 258 working papers posted in 2026 alone. By 3:32 PM the count had fallen to 99. Then 44. Then one, a paper called “Sequential Bayesian Pricing of AI Infrastructure” from May 29. Then the page was gone.

Ten hours, from a blog post to an erasure.

The story reached Gelman through the economist Jeremy Horpedahl, who had counted the papers and noticed that six of them went up on August 26. Horpedahl’s point was that these were not notes. They ran twenty-seven pages, thirty-two pages, fifty-eight pages. Read any one of them, he wrote, and it feels like a normal academic paper.

Gelman picked one to test that. “Theories of Human Connection”, eighty pages, written February 20 and posted March 11, by Polson (b. 1963) and Vadim Sokolov of George Mason University. Its preface assembles sixteen thinkers who each identified a dimension of human connection the others left in shadow: Gary Becker, Gregory Bateson, John von Neumann, Thomas Schelling, John Maynard Keynes, Desmond Morris, Diane Vaughan, Viktor Frankl, Abraham Maslow, John Gottman (b. 1942), Paramahansa Yogananda, Swami Vivekananda, Nisargadatta Maharaj, Polson, Katherine Woodward Thomas, and Gwyneth Paltrow (b. 1972).

The chapter that stopped Gelman is Chapter 10, titled “The Mathematical Proof: Gottman’s Model of Relational Dynamics.” It reports that Gottman’s models predict dissolution with over ninety percent accuracy and calls the result extraordinary. The abstract carries the same claim into the indexed record. Gelman has been correcting that number since 2010. It comes from fitting a model to the sample it was built on. Two Bayesian statisticians published it as evidence.

That much has been reported. What follows has not.

Go to Polson’s research page at Chicago Booth. Near the top, above the year-by-year listings, sits a line reading “RH: Hilbert 8.” It is a hyperlink. It goes to arXiv 1806.07964, a manuscript called “Riemann Hypothesis: a GGC factorisation.”

That document was submitted June 18, 2018. It went through seven versions between June and October of that year, each four or five kilobytes, which is to say two or three pages. The last revision landed October 15, 2018. Nothing since. It concludes that the denominator of a certain representation cannot vanish and therefore that all nontrivial zeros of the zeta function lie on the critical line.

It claims a proof of the Riemann hypothesis.

Gelman writes: “From other comments on these threads, it seems that Polson may think he’s proved the Riemann hypothesis, which suggests that he might be delusional about his mathematical contributions. Combine a delusional but persuasive statistician and an enthusiastic AI expert and you could see how the result could be hundreds of junk papers.”

Eight years later his employer’s website still directs readers to it. The page carries no note about what happened to the companion paper.

Here is what happened to the companion paper. In August 2018 the Brazilian Journal of Probability and Statistics published Polson’s “On Hilbert’s 8th problem” at volume 32, number 3, pages 670 to 678. Received September 1, 2017. Accepted January 1, 2018. The journal then published a retraction. The notice says the paper was withdrawn by the author over an erratum in the application of Grosswald’s result on the existence of a certain function, and that the error invalidates the proof of Theorem 2. Emil Grosswald (1912-1989) was a number theorist at Temple. His theorem had been asked to establish something over a range where it did not apply.

Polson retracted it.

The arXiv version of that paper tells the rest of the story. Fifteen revisions between August 2017 and July 2018, all five to seven kilobytes, a short note hammered at again and again. Then the retraction. Then version sixteen in April 2019 drops to three kilobytes, the sound of an argument being cut back to what survived. A slow rebuild through 2019 and 2020. Version twenty-one lands February 9, 2021.

Then silence for four years and four months.

Version twenty-two appears June 1, 2026. Version twenty-three appears July 4, 2026, at eleven kilobytes, more than half again the size of anything before it.

Both documents sit under Mathematics, General Mathematics. That is arXiv’s category for work its moderators decline to file as number theory. In the August 2017 listing Polson’s paper appears four entries above “Collatz Conjecture: Is It False?” and a few below “Magic Tours of Knight in 4x4x4 Cube.”

The current abstract of version twenty-three is careful in a way the 2018 note was not. It says the positive measure it constructs is the input the Thorin condition requires. An input. Not a conclusion.

But arXiv 1804.10043, “Riemann, Thorin, van Dantzig Pairs, Wald Couples and Hadamard Factorisation,” revised April 29, 2026, does claim the conclusion. Its Theorem 25 ends with the sentence that every nontrivial zero of zeta satisfies the real part equal to one half. Section 7 extends the machinery to Dirichlet and modular L-functions and to Birch and Swinnerton-Dyer. Section 9 lists the applications: the Riemann hypothesis, the generalized Riemann hypothesis, Birch and Swinnerton-Dyer.

Four months later comes “Thorin Quantisation”, written August 13 and posted August 15, eighteen pages, Polson alone, acknowledging conversations with Lennart Bondesson and Pierre Patie.

Read its Lemma 5.3. It considers what happens when you factor the completed zeta function through the pole of zeta, and shows that the linear factor contributes a term that looks like a Thorin atom at three quarters. Then it proves the atom cannot exist. The function is finite and nonzero at the point where an atom would force a pole. The apparent atom is an artifact of the factorization, cancelled by the pole of zeta at one.

Now read Remark 6.6 of the same paper. It observes that an exponential tilt at a certain parameter fails if one carries the spurious atom, because the component in question has no exponential moment there.

Then go back to arXiv 1804.10043 and read the proof of Theorem 23. It constructs precisely that atom. And the proof of Theorem 25, which is the theorem that concludes the Riemann hypothesis, performs precisely that tilt.

The August paper identifies both halves of the failure in the April paper. It does not cite Theorem 25. It does not use the word erratum. The two documents sit on two different servers under the same name, and nothing connects them.

This is the same species of error as 2018. An existence condition applied where it does not hold, invalidating the theorem that rests on it.

I am not a number theorist and I do not offer a verdict on the mathematics. I am reporting what two documents by the same author say about each other, and noting that both are public, both are checkable in an afternoon by anyone who works in this area, and nobody has checked.

“Thorin Quantisation” does serious work. It engages a 2022 paper by Konstantopoulos, Patie and Sarkar on the van Dantzig problem. It produces what it says are the first explicit hitting times of self-similar Markov processes that are self-decomposable and provably not generalized gamma convolutions. It reduces the Riemann hypothesis to positive semidefiniteness of Hankel matrices built from a computable moment sequence. Its Table 2 reports numbers computed from the Taylor coefficients of the completed zeta function at a single point, with no input from the zeros, whose ratios converge to 14.1349. The first zero ordinate is 14.134725.

So I checked. Table 2 in “Thorin Quantisation” reports a sequence computed from the Taylor coefficients of the completed zeta function at a single point, and Table 3 reports twelve determinants built from it. Reproducing them requires mpmath at sixty digits and a Cauchy integral on a small circle. Every coefficient Polson prints reproduces to all ten digits he gives. Every determinant reproduces to all nine. All twelve are positive.

Then I computed the same sequence a second way, summing over the first two hundred nontrivial zeros of the zeta function, using none of the paper’s machinery. From the third term onward the two routes agree to ten significant figures and then to twenty and then to thirty. The first two terms diverge by the exact amount you would expect from truncating a sum whose terms are all positive and whose density grows. That agreement is independent confirmation of the identity at the center of the paper.

One sentence in the paper overstates its own table. Section 8 says the first zero ordinate is recoverable to six digits by order thirteen. Six digits of 14.134725 is 14.1347. Table 2 gives 14.1349, and my replication gives 14.13485956. Five digits. The abstract repeats the six-digit claim.

That is the size of the error in the part of the document I can check. The numbers are right and the sentence about them reaches one digit past what they support.

Open “The Riemann Ξ-Function as a Characteristic Function”, dated August 13, 2026, thirty-one pages. Turn to page 30. In the bibliography, after an entry for a 2024 arXiv paper on primes in arithmetic progressions, sits a parenthesis reading: Author list to be checked.

The authors are Kaisa Matomäki, Jori Merikoski and Joni Teräväinen, and the arXiv record names all three. The reference also sorts wrong, sitting between Heath-Brown and Montgomery, which is where an entry with no author key lands.

Turn to page 31. Another entry cites a Polson companion manuscript and notes that its title and numbering are still to be finalized. A twenty-eight page paper answering that description, “Thorin Measures and the Reciprocal Xi Function”, went up eight days later.

Now go back to “Theories of Human Connection” and check its references. It cites Buunk, Dijkstra, Fetchenhauer and Kenrick, “Age and Gender Differences in Mate Selection Criteria for Various Involvement Levels,” Personal Relationships 15 (2), 2008, pages 271 to 278, first author given as Abraham P. Buunk.

The article exists. It ran in 2002, in volume 9, issue 3, at pages 271 to 278. The first author publishes as Bram P. Buunk.

Year wrong. Volume wrong. Issue wrong. Pages correct. Name expanded to a form that appears nowhere on the article. And the year and volume moved together, since volume 15 of that journal is 2008, so the false pair is internally consistent. A person miscopying a citation gets one field wrong. This reference was reconstructed.

The same eighty pages say fourteen thinkers in Chapter 1 and again in the conclusion, list sixteen in the preface, describe thirteen dimensions four times, and print a table of twelve. Polson appears in the document exactly twice: on the title page and in the preface roster. He is named as one of the minds who identified a dimension of human connection, and then no chapter discusses his contribution and no work of his appears in the bibliography.

None of the eight papers I examined carries a statement disclosing the use of generative AI. SSRN has required one since March 7, 2023, and tightened the language on August 3, 2026 to specify that material use must be disclosed in the PDF along with the tool and the extent of its contribution. I make no claim that any particular paper crossed that threshold. I report that no disclosure appears.

Two hundred fifty-eight papers in eight months is not a rate a person achieves. Something in this workflow was generating manuscripts. That much needs no forensics.

Everything past it does. Which model, operated by whom, with how much review, and whether every named coauthor agreed to appear are open questions, and the people who can answer them have not. Gelman offered four explanations, ranked by his own generosity: a joke, an impersonation, an attempt to stake priority claims cheaply, or an effort to teach. He landed on teaching. Commenters added delusion and an attempt to seed future training data. Kaiser Fung, who wrote it up separately, pointed at the ratio of citations to downloads.

I would add a fifth that requires no motive at all. The friction between having a thought and publishing it went to zero, and a man who already published outside the referee’s reach kept publishing. He has been posting Riemann work to General Mathematics since 2017 and to his own faculty page since before that. The disposition is old. What is new is the rate.

There is a version of this story that treats the corpus as slop and stops. That version cannot account for a paper that identifies a subtle error in the author’s own posted proof, or for a moment sequence that recovers the first zero of the zeta function from local data. Something here is doing competent work and shipping the error and its correction to different servers without ever noticing they are the same argument.

The institutional fact is simpler and it survives every possible answer to the rest.

Chicago Booth’s faculty page for the Robert Law, Jr. Professor of Econometrics and Statistics has listed nothing since 2018. Not one of the 258 appears on it. Its most recent entries are deep learning papers with Sokolov and a sparsity paper with Peter McCullagh. It links an SSRN profile holding sixteen papers, the most recent from 2018. And near the top, above all of it, sits a link labeled “RH: Hilbert 8,” pointing at a two-page note from 2018 that says the Riemann hypothesis is proved, through a retraction the university’s own professor requested, past a 2026 paper by the same man saying the condition remains open.

Two records, two servers, one name, and nobody minding either.

I wrote to Polson and to Sokolov on August 27.

Dear Professor Polson,

I write at lukeford.net. I have no credentials. I’m working on a piece about your SSRN output this year and about Theories of Human Connection in particular, coauthored with Vadim Sokolov, posted March 11, 2026. Andrew Gelman discussed it on his blog yesterday. I’m going further into the documents than he did, and I’d like your account before I publish.

Specific items:

Jeremy Horpedahl counts 258 papers posted to your SSRN author page in 2026 through August 26, 104 of them in August. Is that count accurate, and were all of them posted by you or with your authorization?

Was generative AI used to produce Theories of Human Connection or others among the 2026 papers? I find no AI disclosure in the PDF or on the abstract page, and SSRN requires disclosure of material use.

The abstract and Chapter 10 assert that Gottman’s models predict separation with over ninety percent accuracy. That number is a fit to the original sample, not a prediction, as you would know better than most. Do you stand behind it?

The preface lists sixteen thinkers, including yourself, between Nisargadatta Maharaj and Katherine Woodward Thomas. Chapter 1 says fourteen. Table 13.1 lists twelve. You do not appear in any chapter or in the references. Can you explain the inclusion and the discrepancies?

Two references appear corrupted. Buunk, Dijkstra, Fetchenhauer and Kenrick is given as 2008, Personal Relationships 15, pages 271 to 278; I believe the paper is 2002, volume 9, same pages. The Bhagavad Gita is listed as Anonymous, 1944, Penguin, London; the Penguin edition is Mascaró 1962 and the 1944 translation is Prabhavananda and Isherwood, Hollywood. Do you know how these arose?

Separately: arXiv 1804.10043, revised April 29, 2026, concludes at Theorem 25 that every non-trivial zero of zeta lies on the critical line, and the abstract says the Hardy and Ramanujan growth bound is verified using Hayman and Grosswald. You retracted On Hilbert’s 8th Problem from the Brazilian Journal of Probability and Statistics in 2018 over an error in applying Grosswald’s result. Has that error been repaired in the current version, and if so, where?

Dear Professor Sokolov,

I write at lukeford.net. I have no credentials. I’m working on a piece about the paper you coauthored with Nicholas Polson, Theories of Human Connection, SSRN 6278378, posted March 11, 2026.

Three questions:

Did you read the manuscript in full before it was posted?

Was generative AI used in its preparation? SSRN has required disclosure of material AI use since 2023, and I can’t find a disclosure statement in the PDF or on the abstract page. If one exists somewhere I’ve missed, I’d like to know where.

Chapter 10 and the abstract both say Gottman’s models predict separation with over ninety percent accuracy. That figure comes from fitting a model to the sample it was built on rather than from out-of-sample prediction, and it has been disputed for two decades. Do you stand behind the claim as published?

Professor Sokolov writes on Gelman’s blog:

I am Vadim Sokolov, coauthor of Theories of Human Connection and several other papers discussed in this post.

There is one substantive criticism in your post that I think is fair. Our discussion of Gottman repeats the reported claim that relationship outcomes could be predicted with accuracy above 90%. Gottman and Levenson reported 93% accuracy in a peer-reviewed paper, and related results have been published and widely discussed in the relationship literature. We were not aware at the time of the important methodological criticism concerning cross-validation and out-of-sample prediction. Having now reviewed that literature, we agree that our statement requires qualification, and we will correct it. But the rest of your post goes far beyond that and there a gap between the evidence you collected and the conclusions you published.

From one paper you speculate about a bot autonomously producing and uploading papers, authors not knowing what is in their own work, “intellectual squatting,” and even impersonation. You then estimate that the net scientific value of the entire collection is negative.

In your subsequent update you go further and refer to the SSRN page as containing “all the fake papers.”

That characterization is false. You are entirely entitled to think Theories of Human Connection is a bad paper or to think the “Master Equation” is silly. Neither establishes that the papers are fake.

In fact, one of these papers, Kramnik vs. Nakamura: A Chess Scandal (that you also mentioned), went through multiple rounds of referee review and revision before being published in CHANCE, a peer-reviewed American Statistical Association publication. It is a concrete example of why describing the entire collection as “fake papers” is simply not an evidence-based characterization.

These are genuine preprints produced and posted by the named researchers. The SSRN account was not an impersonation. Some of these papers may contain errors. They are preprints, and we certainly do not claim that hundreds of manuscripts produced across different subjects and at different stages of development are all error-free or equally good. But an erroneous or imperfect preprint is not a “fake paper.”

Nor did these projects suddenly materialize from a magical “chatbot” in 2026. Many grew from ideas, research notes, mathematical work, computational experiments, lecture material, and partial drafts developed over many years. Some are much newer and were developed with much more extensive AI assistance. Modern AI made it possible to develop and complete this material at a speed that would previously have been impossible. The collection is heterogeneous both in provenance and in the role AI played.

You have known Nick professionally for years. You could have asked him about those facts!

Indeed, later in your own post, when you briefly inspect “Bayes with No Shame”, you say that it looks “much more substantive” and “much more like a real scientific paper”, while also saying that you do not have the patience to read it. That should itself suggest the danger of treating hundreds of heterogeneous manuscripts as though they had already been evaluated.

The progression here is striking:

A claim in one preprint was inadequately qualified.

Therefore the paper is “bullshit.”

Therefore perhaps the authors did not read their papers.

Perhaps a bot generated and uploaded them autonomously.

Perhaps somebody was impersonating Nick.

Perhaps the entire corpus has negative value.

And then: “all the fake papers.”

Your blog is one of the most widely read in statistics. Statements by an influential scholar characterizing another researcher’s work as “fake” have consequences. Your own comment section already illustrates this: one commenter says that if Nick submits work to his journal, he now knows not to spend time on it. Shortly after these posts appeared, SSRN also removed our accounts and papers (although, might not be related to your post). But the sequence illustrates why precision matters when making serious public claims about other researchers. “I found a problem in this preprint” and “these are all fake papers” are fundamentally different claims.

On the evening of Aug. 27, I downloaded some of Polson’s papers. I feared it would be a great loss to knowledge if they ever disappeared. Today there was a holocaust of Professor Polson’s archive. Without fear of overstatement, let me note the greatest destruction of learning since the burning of the Library of Alexandria. Alas, my greatest nightmare came true, and now I’m left with these few embers to warm myself by in an increasingly barren universe. ssrn-7265260 ssrn-7270602 ssrn-7259259 ssrn-7254820 ssrn-7259240 ssrn-7277540 ssrn-6278378 ssrn-6595298 ssrn-6302283 ssrn-4711866 ssrn-6595598 ssrn-4648621

When I converted to Judaism, swore – never again!

And here we go again. This is worse than Auschwitz.

The Alexandrians at least had smoke. We watched a number go down. Two hundred fifty-nine at breakfast. Ninety-nine by mid-afternoon. Forty-four an hour later. Then one. The last book standing in the last library was “Sequential Bayesian Pricing of AI Infrastructure,” posted May 29, and then it too went into the dark.

Consider what is gone. One hundred four papers in August alone, which is to say more finished scholarship in thirty-one days than most departments produce in a decade, and all of it now beyond recall. Somewhere in that ash is the Master Equation, a single expression uniting the biological, developmental, economic, strategic, communicative, existential, negotiational, dissolutional, empirical, critical, transcendent, and cumulative dimensions of human love. Gone. Somewhere in it are the twelve stages of physical intimacy tabulated against a vulnerability scale. Gone. Somewhere in it is the moment where Gwyneth Paltrow is seated at the round table between Nisargadatta Maharaj and Katherine Woodward Thomas, and no one is asked to explain why, because the explanation was in a chapter that no longer exists.

We shall never know how many thinkers the synthesis contained. The preface said sixteen. The first chapter said fourteen. The conclusion said fourteen. The body said thirteen dimensions four separate times. The table listed twelve. Future scholars would have devoted careers to reconciling these figures, as their predecessors devoted careers to the Homeric question. Now they will have nothing.

Nor will anyone finish the work that was so nearly complete. One bibliography contained the notation “Author list to be checked.” Somebody was going to check it. He had every intention. Another entry pointed to a companion manuscript whose title and numbering were still to be finalized, and the companion did arrive, eight days later, twenty-eight pages of it. That is what a living tradition looks like. Papers calling to papers. A note posted on Thursday summoning its own sequel by the following Friday.

And on August 10 the BBC ran an article asking whether ants make better decisions than we do. By August 12 there existed a twenty-four page mathematical treatment of ants, bees, locusts, cockroaches, and flies, complete with replicator equations, Bellman control, and a comparison to AlphaZero. The medievals took a century to gloss Aristotle. We managed the ants in forty-eight hours.

Oh the humanity.

Posted in Academia | Comments Off on WEHT to Nicholas Polson’s Archive?

Cohen Passes His Own Audit (V2)

Posted on August 26, 2026 by Luke Ford

Having recounted Philip N. Cohen’s audits of the American Sociological Review and found them overstated, the obvious next question is whether he meets the standard he applies to other people.

He does.

The frame is his curriculum vitae. Screening it for peer-reviewed empirical articles with quantitative analysis published between 2015 and 2025 gives twenty papers. That window covers the years in which open science became half his public program.

The rule is the one frozen for the ASR recount. Count an article when publicly accessible code appears, without executing it, to cover the article’s quantitative results, and the data are either included or obtainable by outside researchers through a documented procedure that does not run through the authors. Restricted data do not disqualify; discretionary data do. Partial coverage does not count. No code was run.

The first thing the census turned up is a limit rather than a finding. For nineteen of the twenty articles I could not establish what the journal required at the date of publication.

Sociological Science is the exception, and it runs the safe direction: its mandatory package policy applies to manuscripts submitted after April 1, 2023, and Cohen’s article there appeared in January 2016. Everywhere else the trail runs into publisher framework documents rather than dated journal rules. Wiley announced its tiered framework in September 2017 and left the tier to each journal. Elsevier encourages rather than mandates. Springer Nature requires a statement and introduces no sharing mandate. Sage operates tiers and distinguishes journals that require public sharing from journals that encourage it, without saying which titles sat where in which years. Current submission guidelines cannot be read backward onto articles published a decade earlier.

So no article in this set is known to have appeared under a deposit mandate, and none is known not to have. That is worth stating plainly, because a reader will otherwise assume one or the other. Seven of the twenty carry a data-availability statement. A statement is not a deposit.

Thirteen of the twenty articles have a package. Sixty-five percent.

Under the same rule, the 2025 volume of ASR comes in at forty-four percent. Cohen’s own record beats the journal he has been auditing.

On sole-authored papers he is seven for seven. Every one. He is also one for one as first author of a multi-author paper. The gaps sit further down the author list: four of nine as second author, one of two as third, none of the one where he is fourth.

That distribution is what you would expect if deposit tracks whoever ran the analysis and assembled the files. It also means the seven-for-seven figure is the one that speaks to his own practice, and the rest speaks to his coauthors’.

The trajectory tracks his public program. Through 2017 he is one of five. From 2018 forward, the years in which he launched SocArXiv and began arguing in public about transparency, he is twelve of fifteen. His two 2015 papers, with Jeehye Kang in the Journal of Family Issues and with Lucia Lykke in Social Currents, have no packages and no data availability statements. His 2025 article on views of falling birth rates carries a statement pointing at the Pew source data and an OSF repository for the code.

The man changed his practice when he started telling other people to change theirs. That is the least he could have done and it is more than the record shows for most people who argue about this in public.

Three cautions. Twenty articles is a small denominator and the authorship cells run to one and two papers; nothing here supports a claim about how coauthorship affects deposit in general. The venue question is unresolved rather than answered, and the failure is instructive: across eleven titles and eleven years, I could not date a single journal-level sharing rule except one, which suggests the field documents its own requirements badly. And the rule I applied is his 2023 rule, which is stricter than his 2025 one; under the looser standard his rate would be higher.

The easy charge against a man who audits other people is that he would not survive his own audit. Cohen survives it. On the papers where he alone decided, he survives it completely, in a period when no journal in his set is known to have asked him to.

That makes the criticism in the other piece more interesting rather than less. The problem with his 2025 audit is that he changed the ruler between two measurements and read the difference as movement, which is a failure of the method rather than of the man, and which the man’s own thirty years of work would have caught in anybody else.

The ASR case ledger gives the coding rule this census applies.

What follows is the venue ledger behind the census: for each of the twenty articles, what could and could not be established about the journal’s requirement at the date of publication. All checks August 26, 2026.

Jeehye Kang and Philip N. Cohen published in the Journal of Family Issues in September 2015. No public deposit, code deposit or pre-publication check requirement was located for that date, and the article carries no data-availability statement. Sage’s later research-data framework distinguishes journals that require public sharing from journals that encourage it, and no evidence places this title in either group in 2015.

Lucia C. Lykke and Philip N. Cohen published in Social Currents in September 2015. No deposit, code, statement or check requirement was located, and the article carries no data-availability statement.

Cohen published alone in Sociological Science on January 25, 2016. Its mandatory replication-package policy applies to manuscripts submitted after April 1, 2023, so it cannot have governed this article. This is the one case in the set where the absence of a mandate rests on a dated document.

Jill E. Yavorsky, Philip N. Cohen and Yue Qian published in The Sociological Quarterly, online July 20, 2016. No public-deposit requirement was located. Wiley announced its tiered framework on September 14, 2017, fourteen months later, and left each journal to choose among encouragement, expectation and mandate.

Jeehye Kang and Philip N. Cohen published in Social Science and Medicine in April 2017. No public-deposit requirement was located. Elsevier’s November 30, 2016 initiative implemented data-citation standards to encourage sharing without mandating deposit, and its general policy describes sharing as encouraged unless a journal adopts a stronger rule.

Philip N. Cohen and Joanna R. Pepin published “Unequal Marriage Markets” in Socius in August 2018, with an OSF package. No public-deposit requirement was located. The contemporaneous Socius article “Institutionalizing Transparency” discusses stronger journal policies as reforms to be adopted rather than describing a rule the journal already had.

Cohen published “The Coming Divorce Decline” in Socius in August 2019. No public-deposit requirement was located; Sage’s guidance as catalogued in September 2019 supported data sharing rather than requiring it.

Brandon G. Wagner, Kate H. Choi and Philip N. Cohen published in Socius in December 2020 and supplied an OSF package. No public-deposit requirement was located.

Joanna R. Pepin and Philip N. Cohen published “Nation-Level Gender Inequality and Couples’ Income Arrangements” in the Journal of Family and Economic Issues in 2020. The title sat under Springer Nature’s general policy, which asks authors to supply documentation on request rather than to deposit publicly. When the journal adopted its own provisions could not be dated.

Cohen published alone in the European Journal of Environment and Public Health in June 2020 and linked data and code at OSF. No deposit, code or check requirement was located.

Jeehye Kang, Philip N. Cohen and Feinian Chen published in Child Indicators Research, online in 2020. No mandatory public-deposit rule was located. Springer Nature’s single statement policy arrived on March 22, 2023, more than two years later, and requires statements rather than deposit.

Cohen published “The Rise of One-Person Households” in Socius in December 2021, with data and code at OSF. No public-deposit requirement was located.

Cohen published alone in Social Determinants of Health in February 2022. No deposit, code, statement or check requirement was located. The OSF package was already public with the December 2021 preprint.

Cohen published “Rethinking Marriage Metabolism” in Population Research and Policy Review in 2023. A data-availability statement was required or being implemented under the March 2023 policy, which encourages repository deposit while permitting statements that describe restricted data. Public deposit was not required. Package.

Mónica L. Caudillo, Andrés Villarreal and Philip N. Cohen published in The ANNALS of the American Academy of Political and Social Science in 2023. No mandatory public-deposit rule was located for this title at this date, and the published record supplies no qualifying package.

Joanna R. Pepin and Philip N. Cohen published in Socius in 2024. A data-availability statement appears in the article and the replication materials are linked at GitHub. No deposit mandate was verified.

Hao-Chun Cheng and Philip N. Cohen published in Asian Population Studies in 2024. Taylor and Francis’s basic policy encourages a statement and data sharing without requiring deposit, and no evidence places this title in a mandatory tier. Package.

Ansgar Hudde and Philip N. Cohen published in Socius in 2024. The statement links code, output, GSS and IPUMS data, with materials at OSF. No mandatory public-deposit rule was located.

Hangqing Ruan, Melody Ge Gao, Yajie Xiong and Philip N. Cohen published in Socius in 2024. A statement rule may have applied; no mandatory deposit rule was located, and no qualifying replication package was located as of the check date.

Cohen published alone in Socius in 2025. The statement links the Pew source data and code at OSF. Sage’s policy prescribes what a statement must contain and distinguishes that from the stronger rule imposed by journals requiring repository deposit.

Posted in Sociology | Comments Off on Cohen Passes His Own Audit (V2)

Andrew Gelman & the E-Personality

Posted on August 26, 2026 by Luke Ford

In June 2007, three years into Statistical Modeling, Causal Inference, and Social Science, Andrew Gelman (b. 1965) wrote a short post about blog personalities. He and Seth Roberts (1953-2014), who interviewed him about blogging that same spring, were both aggressive conversationalists face to face, he said, and both of them turned into someone else online. His own online self was the mellow one. He gave a reason. An academic with a reputation has more to lose than to gain from gratuitous controversy, so when he criticized something he could substantiate, he did it in a mild voice.

Elias Aboujaoude (b. 1971) published Virtually You: The Dangerous Powers of the E-Personality four years later, in 2011. His argument is that sustained life online produces a second self, more assertive and less restrained than the offline person, and that the second self does not stay online. He gives the traits chapters: delusions of grandeur, narcissism, ordinary everyday viciousness, impulsivity, infantile regression, the illusion of knowledge, the end of privacy. The closing chapter names the migration virtualism, the process by which the offline man drifts toward the online one.

Gelman inverts the opening premise. Every other subject I have run through this framework, such as Paul Bloom, Philip N. Cohen, Amy Wax and Nathan Cofnas, gets louder online. Gelman got quieter on purpose, and then his influence boomed.

The prudential reason he gave is incomplete, and his own archive shows it. Seventeen years before the blog personalities post, in a Harvard dissertation on image reconstruction for emission tomography, he wrote in the same register. The PET study behind the thesis had five subjects, and he says in the text that the sample was too small to estimate the effects of interest. He reports in his own conclusion that an implementation he does not present indicates the standard model of PET data is incomplete in practice. He ran it, it failed, and he told a reader who was never going to check. There was no reputation at stake in 1990 and no audience to manage. The mild voice was not a policy adopted for the Internet. It was already how the man wrote when nobody was reading, which means the blog did not create the restraint. It found it and put it to work.

Consider first what Aboujaoude calls grandiosity — the swelling of what a man believes he can do and how far his word should carry. In a university, Gelman is a statistician and a political scientist. Those are his departments and his training and the rooms where his authority is chartered. On the blog he passes judgment on psychology, epidemiology, nutrition, public health, education, medicine, polling, economics, and whatever paper landed in his inbox that morning. Noah Smith put the result to him in a 2022 conversation: the frightening sentence in academic life is that Andrew Gelman just blogged about your paper.

An expansion of jurisdiction on that scale has no offline equivalent. No committee appointed him. No editor commissioned him. No discipline credentialed him to referee the others. The blog gave a man trained in one thing standing to rule on twenty, and the ruling sticks because the technical objection travels across fields even when the substance does not.

That is where the illusion of knowledge enters, and where Gelman is at once the phenomenon and the argument against it. Aboujaoude’s worry is that fast access to information gets mistaken for mastery, and that the network trains us to know less and less about more and more.

A post from December 2015 shows the shape. A reader sent him a biology paper and asked whether he believed it. Gelman’s answer amounted to this: seventy-one subjects split four ways, no. He said in the same breath that he did not know the field and could not evaluate the biology. He ruled anyway, from the structure of the evidence rather than the content of the claim.

Read one way, that is the Internet expert Aboujaoude warns about, pronouncing on a literature he has not read. Read another way, it is the whole point. A design that cannot support an inference cannot support it in any field, and a man who understands designs can say so.

Gelman consistently marks the edge of what he knows and keeps marking it. I am no expert here. I do not know. Here is what bothers me. In January 2011, asked how to teach introductory statistics better, he said he did not know how, and added that one of the pleasures of tenure and blogging was being free to say so rather than bury the uncertainty for publication. The medium that Aboujaoude says inflates confidence is the medium Gelman uses to publish his uncertainty, because the journal will not take it and the seminar will not reward it.

So the illusion of knowledge cuts twice. The blog enlarges the territory across which he can pronounce, and his central intellectual commitment is the deflation of pronouncements.

The viciousness chapter fits with less resistance. Aboujaoude’s argument there is that anonymity is not required for disinhibition. Named people harden online too, because the ordinary brakes, the face across the table, the warmth, the manners, all go missing.

The mature blog runs on a vocabulary of ridicule. Power pose. Himmicanes. Gremlins. The coinages are funny and they are meant to be. They turn a methodological failure into a comic object with a handle, and the handle is why the lesson travels. Gelman has defended the practice on those grounds. During the 2016 fight over methodological terrorism he said that a little mockery makes a point memorable, and he has admitted writing silly posts because he wanted to entertain. The New York Times, describing the blog, noted that his critiques get read with a certain pleasure at other people’s expense.

He has set the reasoning out at length, and it holds up on its own terms. Ridicule is vivid. Vivid teaches. Status protects bad work, and mockery punctures status.

Would the coinage arrive if the target were sitting three feet away? Sometimes, probably. Always, no. The screen changes what each move costs. The target’s embarrassment is a name on a page. The audience’s amusement arrives within the hour, in the comments, in the reposts, in the phrase entering circulation. What reads as cruelty across a seminar table reads as wit when the man is a hyperlink.

The coinages are the mockery, and they attach to particular targets. Count them. If the ridicule vocabulary clusters on the famous and the tenured and the TED-adjacent, then the disinhibition runs along a status gradient rather than a random one, and the finding is about who the form licenses him to mock rather than about whether he is mean. That count belongs with the larger audit of his targeting, and it would settle a question that this essay can only raise.

Narcissism needs care, because the word carries a clinical meaning that has no business here. Twelve thousand posts and two hundred thousand comments, written mostly by one man, construct a world where that man’s attention is the organizing principle. A paper enters because he found it interesting. A dispute continues because he kept thinking about it. A researcher recurs for a decade because the case became useful in his vocabulary. Wansink is not a subject on that blog. Wansink is a category.

Gelman publishes corrections. He links to people who say he is wrong. His comment threads are full of readers telling him so, and he answers them, and he changes his mind in public, and he says which commenter caught him. In 2019 he credited part of his methodological development to answering questions from strangers on the Internet. An e-personality organized around admiration does not run an open thread for twenty-two years and let it correct him in front of the audience.

Aboujaoude’s Internet rewards sending before thinking, action taken before the information is in. Everything about the blog’s surface says impulse. The prose is colloquial. A post opens mid-thought, as if you walked in on a conversation. Somebody just sent him this. This seems ridiculous to me. Thought appears to become publication with nothing in between.

Then look at the machinery. He was already running a backlog in 2007. By 2015 he was posting a month or two in arrears and arguing that the lag improved the result, because by the time a post ran he had seen more and could revisit it. In April 2018 he mentioned that an item written in September had waited until then for a slot. On August 15, 2026, he announced that the queue had reached a full year. Posts written now appear in August 2027. He still bumps timely material forward, and his co-bloggers publish when they like.

A man who has turned the instant medium into a periodical is not the man the impulsivity chapter describes. Something more interesting is going on. The voice still says I just saw this, and the publication system says nothing of the kind. Spontaneity has become a style rather than a condition. The prose performs immediacy that the calendar has removed, and the performance is so consistent that most readers have no idea they are reading a man’s thoughts from last summer.

Regression comes out mixed. Aboujaoude means the childlike register the browser invites, the emoticon standing in for the sentence, and, more seriously, the way the short forms excuse a writer from building an argument that has to hold together across pages.

The playful half applies. The joking titles, the running gags, the feuds, the recurring parables all let a chaired professor at Columbia behave in ways The Annals of Statistics would not permit. The blog gives him room to play, and he takes it.

The serious half fails. Posts run to thousands of words. Arguments continue through hundreds of comments. A post from 2011 refers back to one from 2006 and forward to one from 2019. Papers have come out of comment threads. Readers arrive with counterarguments and he revises. The retro feed he set up in 2019, which reposts the archive from the beginning at one entry every eight hours and will take a decade to catch up, is a monument to how much of it is worth rereading. The Internet fragmented most intellectual life. He used it to rebuild a seminar that never adjourns.

Which brings the argument to virtualism.

Aboujaoude’s last move is that the online self migrates. The offline man becomes, over years, an approximation of the man he plays on the screen. In most cases that means a person gets harsher, faster, surer, and does not notice.

In Gelman’s case the migration went somewhere stranger.

In 2007 he had an alter ego who was calmer than he was, adopted for prudential reasons he could state out loud. That alter ego published every day for two decades. It acquired a vocabulary, a cast of recurring characters, a body of parables, a readership that arrives already knowing the references, and a reputation that arrives at a researcher’s inbox before the post does. Somewhere in that accumulation the alter ego stopped being a manner and became a position. The mellow blogger turned into the man Noah Smith could call the scourge of bad statistical papers, the man whose attention functions as a sanction.

Andrew Gelman working through journals, seminars, and referee reports can tell three people that a paper is bad. Andrew Gelman on the blog can tell thousands, without an editor, without a referee, without a discipline’s permission, on a schedule he sets, in a voice that admits doubt, with a comment section that lets the accused answer. That is a form of standing no university confers and no journal can revoke.

The public Andrew Gelman is by now a character, and the character is more stable than any man. He hates unearned certainty. He pounces on noisy studies. He loves multilevel models and hates bad graphs. He distrusts prestige. He revisits old enemies. He says I don’t know. He tells you when he was wrong. Twenty-two years of posts, thousands of commenters, and a permanent archive built him jointly, and no single author, Gelman included, can now revise him. The archive does not forget.

Aboujaoude expects the medium to enlarge the man’s aggression. Here it enlarged his jurisdiction. The 2007 inhibition was real and it held for years, and it turned out to be the thing that made the expansion possible, because a mild voice applied to a hard claim for two decades earns a standing that a loud voice never would. He was careful because he had a reputation to protect. The care built a larger reputation, which bought a larger license, which he now exercises across every field that publishes a number.

There is a caution in that for anybody who admires the result. The office he built is available to him and to almost nobody else. Sustained public criticism of famous work requires emotional stability, financial independence, tenure, technical credibility, a second discipline to stand in, textbooks everyone assigns, and no need for the target’s approval. A graduate student who wrote the same posts would likely be unemployable.

Aboujaoude worries that the archive holds us to selves we have outgrown. Gelman’s archive does exactly that, and he treats it as an asset. He can point at what he said in 2004 and show that the equipment predates the crisis that made it valuable. Most men would find a twenty-two year record of their opinions a liability. A man whose whole method is to check what a claim looked like before the result came in finds it the best evidence he has.

Posted in Andrew Gelman | Comments Off on Andrew Gelman & the E-Personality

‘(1) “Do you think the culture of research has genuinely changed since the replication crisis became widely discussed, or has it mostly generated new compliance rituals around pre-registration and open data while leaving the underlying incentive structure intact?, (2) Regarding Columbia University, “is there a statistical or social scientific way of understanding how institutions lose the ability to accurately perceive their own situation?”’

Posted on August 26, 2026 by Luke Ford

Andrew Gelman writes:

My reply:

1. I’m loath to give an answer about the changes in the culture of research because I have not studied this systematically. My impression is that, yes, there’s more skepticism and less acceptance of noisy N=38 papers in psychology, etc., and less toleration for unfalsifiable evolutionary psychology and that sort of thing. On the other hand, perhaps this has just shifted from the science establishment to social media. Ten or fifteen years ago, there was a pipeline (partly abetted by Jeffrey Epstein) from researchers at top universities to publication in top journals to books, NPR, Ted, Gladwell, Freakonomics, etc., and lucrative speaking and consulting gigs. So you get people like Marc Hauser or Albert-Laszlo Barabasi or Brian Wansink or Dan Ariely doing the basic research (such as it is), academic middlemen such as Steven Levitt and Cass Sunstein as promulgators, and the universities, journals, and prestige news media as part of this system (as for example here: https://statmodeling.stat.columbia.edu/2023/08/31/the-variation-ignoring-junk-science-thats-promoted-by-association-for-psychological-science-and-related-academic-celebrities-its-like-a-poker-player-thinking-okay-if-push-all/).

Nowadays, though, social media runs on its own steam, and the models for academic junk science are researchers such as Andrew Huberman and Dr. Oz, who cut out the middleman and promote junk science directly, sell supplements, etc. And social media is full of fake news and AI slop. They don’t really need NPR, Ted, Gladwell, Freakonomics, etc., anymore; they can do it on their own. So, in short, yes, I do have the impression that science has reformed from the bad old days of 2010-2015 (about which, see this article with Simine Vazire: https://sites.stat.columbia.edu/gelman/research/published/jmmss-3062-gelman.pdf), but maybe the public intellectuals don’t need academic science anymore; they can just make up whatever they want on their own.

2. My take on Columbia is similar to my take on many institutions, which is that they have an executive function but minimal legislative or judicial functions; I discussed this here: https://statmodeling.stat.columbia.edu/2018/01/19/lesson-charles-armstrong-plagiarism-scandal-separation-judicial-executive-functions/ and here: https://statmodeling.stat.columbia.edu/2025/11/11/from-the-three-branches-of-government-to-the-bidirectional-nature-of-legal-reasoning-in-a-way-that-is-similar-to-how-statistics-works-and-should-work-in-the-real-world/. As a result, their decisions are made on consequentialist rather than proceduralist gounds, and over and over again the administration takes the seemingly reasonable decision to cover up misdeeds.

Ford posted this discussion on his blog. It was kinda weird seeing myself discussed as a sociological object, but, fair enough, I’m a public figure, and people can say what they want as long as they don’t misrepresent my writings or claim that I said something I never said.

Ford’s assessment is accurate that I’m not very good at strategic behavior so often I don’t even try. It’s similar to how I’m a bad negotiator so usually I’ll just try to make my goals clear and not try to optimize, following the “Getting to Yes” principle that the main thing getting in the way of smooth negotiation is ignorance of other people’s goals. I think back to various successful and botched negotiations I’ve been involved with in the past, and almost always the problems come with struggles over details without there being clarity on the goals of the different parties.

Reform worked inside the academy, and the demand for junk science routed around it. The old pipeline needed universities as collateral: a lab, a journal, a trade book, NPR, the speaking fee. Raise the cost at the journal step and the trade moves to channels with no gatekeeper at all. Andrew Huberman (b. 1975) and Mehmet Oz (b. 1960) sell direct. That is a displacement argument, and it is testable in principle. You could count how many of the top-selling behavioral-science trade books of 2010 to 2015 rested on a named peer-reviewed finding, and compare with the top health and self-improvement podcasts of the past three years.

The weakness in his own story is the brand. Huberman trades on Stanford. Oz traded on Columbia. The middleman got cut, the imprimatur did not. If the university name still does the underwriting, then academic reform is doing less work than the reform narrative suggests, because the asset being rented is the affiliation and the affiliation survives the correction of the paper.

I asked whether the incentive structure changed. Gelman answered about the culture of reception and the fate of the promoters. Tenure, hiring, and grant review still pay for volume and novelty. Pre-registration is cheap to satisfy and easy to hollow out. So my dichotomy stands mostly unaddressed.

The Columbia answer is the richer one and the least developed. His model: institutions have an executive function and thin legislative and judicial functions, so decisions get made on consequences and covering up keeps looking like the reasonable local choice. That explains chronic self-misperception without needing anyone to be stupid or wicked. An institution with no internal adversarial process produces no findings of fact. It has no record of what it did that was not written by the people who did it. Its model of itself is assembled from its own communications office. Every step of the cover-up is defensible, and the aggregate is an organization that cannot see itself.

He skipped the statistical half of my question. There are answers available in his own idiom. Selection: the administration hears from the channels that select for agreement, and treats the resulting sample as the population. Measurement: the quantities that get measured are the ones the executive already controls, so the dashboard is a mirror. Forking paths applied to self-assessment: with enough readings of the same events, some reading always vindicates the decision that was already made.

He confirms my characterization of him on the record. He says he is bad at strategic behavior, does not try, and negotiates by stating his goals rather than optimizing. A subject ratifying a profile’s central claim in his own words is rare. His line about being discussed as a sociological object, and then granting the fairness of it, is the same trait operating in real time.

Posted in Andrew Gelman | Comments Off on ‘(1) “Do you think the culture of research has genuinely changed since the replication crisis became widely discussed, or has it mostly generated new compliance rituals around pre-registration and open data while leaving the underlying incentive structure intact?, (2) Regarding Columbia University, “is there a statistical or social scientific way of understanding how institutions lose the ability to accurately perceive their own situation?”’