Illustration by Tag Hartman-Simkins / Futurism. Source: David Dee Delgado / Getty Images for The New York Times
Sign up to see the future, today
Can’t-miss innovations from the bleeding edge of science and tech Email address Sign Up
Anthropic is calling for a global “pause” on AI development, claiming that the technology is nearing a point where it can spiral out of human control.
In a lengthy blog post published Thursday, the world’s most valuable AI startup made the case that its Claude family of models were on the path to achieving “recursive self-improvement,” or the ability to improve themselves on their own, a key hypothetical tipping point that could lead to the creation of powerful AIs capable of operating outside human interests and harming society.
We’re not at that point yet, Anthropic stresses, but it “could come sooner than most institutions are prepared for.”
“We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology,” the company wrote in the post.
It added that “a meaningful slowdown or pause would require multiple well-resourced labs at or near the frontier, in multiple countries, agreeing to stop under the same conditions,” and admitted that this would be challenging to enforce.
“Training runs are far easier to conceal than missile silos,” it wrote.
For Anthropic to call for a pause now is convenient. In the past few months, it leap-frogged OpenAI to become the world’s most valuable AI company with a $1 trillion valuation, and its models are now generally viewed as the best in the field, especially at coding tasks. If the industry were to hit the brakes now, it would cement Anthropic’s dominance.
Not everyone was buying Anthropic’s claims. Prominent AI critic Gary Marcus called the company’s lengthy post a “bait and switch.”
“Anthropic is trying to strike terror into everyone’s hearts (‘full recursive self-improvement also might increase the risks of humans losing control over AI systems’) but all they have really shown is just faster coding — entirely under human control,” Marcus wrote on his Substack. “A faster coding tool will probably not end the world.”
Anthropic has long tried to paint itself as the ethical and deeply concerned adult in the room. A cornerstone of its mythology is that CEO Dario Amodei abstained from unleashing a revolutionary AI model back in 2022 because he was too concerned about safety, and let OpenAI get all the glory when it released ChatGPT months later instead.
Two months ago, in a rehashed sequel to this foundational company lore, Anthropic announced a new model called Mythos — but made a show of not releasing to the public, claiming it was powerful enough to break into “every major operating system and every major web browser.”
Amid its blowout with the military, Anthropic also dropped a safety pledge that was arguably the venture’s entire raison d’etre: to stop training an AI system if it couldn’t guarantee it had proper safety guardrails in place.
Further underscoring Anthropic’s hypocrisy, University College London professor Steven Murdoch cited recent reporting from the Financial Times revealing that Anthropic is helping the US National Security Agency use its Mythos model so it can wage cyberwarfare against potential enemies like China and Iran.
“Anthropic might give the impression of being warm and fuzzy, but their definition of AI safety is narrow,” Murdoch told The Guardian. “Supporting US authorities in the development of offensive capabilities has never been something they have spoken against.”
Regardless of whether Anthropic genuinely thinks it has a remotely realistic shot at pulling off a global pause — or if this is yet another ploy to boost its safety-minded image — it’s vowing to pursue further action.
“In the coming months, we will organize conversations where policymakers, researchers, civil society, and other AI companies can help answer some of the questions this piece raises, especially around full recursive self-improvement and how to create better options for coordination and deliberation,” the company wrote. “We’ll publish what comes out of it.”
I’m a tech and science correspondent for Futurism, where I’m particularly interested in astrophysics, the business and ethics of artificial intelligence and automation, and the environment.
OpenAI CEO Sam Altman. Anna Moneymaker—Getty Images
The CEOs of some of the biggest AI companies in the world have set aside their cutthroat competition to co-sign an open letter to Congress asking for more safeguards against a threat that their own technology has helped create.
Recommended Video
Beyond the Pilot: How COOs Are Scaling AI Today
Dario Amodei, Sam Altman, and Mustafa Suleyman—the CEOs of Anthropic, OpenAI, and Microsoft AI, respectively—signed their names to a public letter to Congress urging the government to screen for the buying and selling of synthetic materials that could be used to create bioweapons. The letter, signed also by dozens of experts in the life sciences and national security fields, was organized by the conservative-leaning think tank, the Foundation for American Innovation, as well as the nonpartisan Institute for Progress.
The letter specifically asks Congress to mandate screening for companies that are selling synthetic DNA and RNA, which the letter’s authors argue could be used to create bioweapons with the help of AI. Notably, some of the companies that manufacture these materials, like Twist Bioscience and Ansa Biotechnologies, also signed the letter, signaling that at least part of the industry welcomes the regulation.
“AI systems are improving rapidly, and alongside incredible benefits to science and medicine, there is a real possibility that the knowledge barriers which have historically prevented bad actors from obtaining biological weapons will meaningfully erode,” the letter read.
While companies selling synthetic DNA and RNA already do some screening voluntarily, the letter wants Congress to go further by making it legally required across the industry. The letter also urges Congress to require the companies that sell these synthetic materials to keep records on their orders, as well as the exact specifications of the materials sold, in an effort to help with potential biosecurity investigations.
The letter comes as improved AI models continue to spread to more people at global and exponential scale. A study by Stanford University from earlier this year found that generative AI tools reached 53% of the world’s population in just three years, faster than both the PC or the internet. At the same time, experts have found that publicly available AI models are able to provide information on how to create biological weapons and how to spread them, the New York Timesreported earlier this month.
A silent threat
The government has long recognized the need to protect against deadly biological weapons. Biological agents are rarely used in terrorist attacks, and have accounted for just 0.02% of all historical attacks, according to a study in the peer-reviewed publication, The American Journal of Emergency Medicine. Yet, because they are often odorless, colorless, and in some cases highly contagious, they pose a distinct threat to Americans.
Biological agents like Anthrax are especially deadly.When inhaled, Anthrax has a mortality rate of nearly 100% without treatment. In 2001, five people died and another 22 people were infected after a microbiologist and former employee of the Army’s biodefense laboratory mailed several Anthrax-laced letters addressed to two U.S. senators and several news outlets. The attacks, which came just after 9/11, spurred one of the largest FBI investigations ever.
Some laws already exist to protect Americans against man-made biological threats. The Biological Weapons Anti-Terrorism Act of 1989 made it illegal to develop or possess biological agents for use as a weapon, with a potential penalty of up to life in prison. After the anthrax attacks in 2001, the PATRIOT Act expanded on the 1989 law, making it easier to prosecute people in possession of dangerous biological agents even without explicit proof that they intended to build a weapon.
Congress has already made some progress on improving the safeguards around the selling of synthetic DNA and RNA. In February, Sen. Tom Cotton (R-Ark.) and Sen. Amy Klobuchar (D-Minn.) introduced the Biosecurity Modernization and Innovation Act of 2026, with the goal of forcing sellers of these synthetic materials to screen both their orders and their customers while providing exemptions for “clearly non-hazardous and pose no credible threat to public health and safety.”
While the bill slowly moves its way through Congress, Josh Wentzel, a senior fellow at the Foundation for American Innovation, told Fortune that the letter was a good opportunity to show lawmakers that the AI industry and companies who sell synthetic DNA and RNA were equally concerned about the issue.
“This is bipartisan, concrete, achievable, and noncontroversial,” Wentzel said, adding he hopes now that Congress sees these parties are aligned, it can move forward with passing the Biosecurity Modernization and Innovation Act. “It’s a goal among many national security experts and, crucially, something the nucleic acid synthesis industry itself has called for.”
In 2001, Fortune first convened the smartest people we know, bringing together CEOs and founders, builders and investors, thinkers and doers. Since then, Fortune Brainstorm Tech has been the place where bold ideas collide. From June 8–10, we will return to Aspen—where it all began—to mark 25 years of Brainstorm. Register now.
Multiple survivors of Jeffrey Epstein spoke before a U.S. House hearing, sharing emotional testimony about the abuse they suffered as teenagers and the failures of the justice system. Jenna Lisa Jones told lawmakers she was only 14 years old when the abuse began in Palm Beach and said it took years for her to find the courage to speak publicly. Another survivor, Dani Bensky, said she still lives with PTSD after being abused in 2004 and 2005 — years after earlier reports about Epstein had already been made to authorities. Survivors said a secret non-prosecution deal allowed Epstein to avoid harsher charges, arguing the government failed to inform victims and denied them justice. The testimony came as lawmakers continue to examine the broader network around Epstein and calls for greater accountability. #JeffreyEpstein#EpsteinHearing#EpsteinVictims#CongressHearing#JusticeForVictims#EpsteinCase#USPolitics#BreakingNews#USnewsMint is an Indian financial daily newspaper published by HT Media. The Mint YT Channel brings you cutting edge analysis of the latest business news and financial news. With in-depth market coverage, explainers and expert opinions, we break down and simplify business news for you. Click here to download the Mint App: https://livemint.onelink.me/MrDS/p0kx… Now make Mint your preferred source on Google and get business & finance updates first. Add here – https://www.google.com/preferences/so…
Jeffrey Epstein with professors at a dinner he hosted at Harvard University in September 2004. Rick Friedman/Alamy
Ethics
Last December, the U.S. Department of Justice released its first batch of files on disgraced financier and convicted sex offender Jeffrey Epstein. Among the thousands of images was one video clip, the only one in the lot. It showed four seconds of the noted psychologist and writer Steven Pinker of Harvard University riding with Epstein on his now infamous private plane.
It wasn’t a great flight even in 2002, years before Epstein’s first criminal conviction, Pinker says of the trip, which was heading to a TED Talk. “I immediately disliked Epstein and thought he was a dilettante and a smartass,” he says. Pinker has not been accused of wrongdoing in connection with Epstein.
Epstein, who died in federal prison in 2019 while awaiting trial on sex trafficking charges, spent a lot of time talking to scientists. When more records are released from a reported stash of 5.2 million, now a month overdue, questions about what the “Epstein files” say about science and scientists are sure to arise. Already, e-mails dropped by a congressional committee and files released by the DOJ—thousands of notes, lists, videos and investigation records—have once again raised the question of why so many prominent scholars were involved with Epstein.
The financier widely courted pundits, politicians and billionaires, as the DOJ files confirm with photographs of everyone from Mick Jagger to Bill Clinton to Donald Trump appearing with him. (None are charged with wrongdoing in connection to the photographs.) A piano virtuoso, mysteriously wealthy and famously ingratiating, Epstein courted scientists for years, leading to investigations at the Massachusetts Institute of Technology and Harvard, the results of which were made public in 2020. Last year’s e-mail releases revealed that astronomer Lawrence Krauss and linguist Noam Chomsky both associated with him long after his crimes became public knowledge. Last November Harvard launched a new investigation to look at connections between Epstein and economist Lawrence Summers, former president of the university.
Patronage
Money is one easy answer for why scientists were interested in Epstein. “Scientists need patronage; they need support,” says Bruce Lewenstein, a science communications expert at Cornell University.Wealthy patrons have funded scientists for centuries; they have paid for telescopes to investigate the atmospheres of alien worlds, brain mapping institutes, malaria prevention experiments, and much else. “That’s not good or bad; that’s what it is. And that has been true for 400 years,” Lewenstein says. Unlike many donors, Epstein usually wasn’t asking for his name on a building, and he donated money to everything from dance troupes to the Council on Foreign Relations, according to a 2019 Miami Herald report.
Before his 2008 conviction for soliciting minors for prostitution, Epstein donated more than $9 million to Harvard, including a $6.5-million gift to Harvard’s Program for Evolutionary Dynamics (PED), led by mathematician Martin Nowak. (Epstein continued to visit that program after his conviction—he did so more than 40 times in 2018 alone—and kept an office there.) He was also a Visiting Fellow at the university in the 2005–2006 academic year, after making a $200,000 gift to its psychology department. Following his conviction, donors he introduced to Harvard scientists gave $9.5 million to the school.
Then there were Epstein’s donations to M.I.T.: he donated $525,000 to the MIT Media Lab and $225,000 to mechanical engineering professor Seth Lloyd. Both gifts came after his 2008 conviction and were handled outside normal channels, according to a university report. Epstein claimed to also have arranged another $7 million in donations from billionaires Bill Gates and Leon Black to the school (Gates denied this, and the university report says there’s no evidence of an effort to “launder” Epstein’s money in the donations).
“The only generalization is that scientists, like the universities they work for, together with artists and others in nonprofit ventures that depend on philanthropy, routinely cozy up to wealthy people willing to slosh money around,” Pinker says. “Very few of these donors are heinous psychopaths, and he exploited their gullibility.”
According to Pinker, his pre-TED Talk flight with Epstein came at the behest of his literary agent, John Brockman, whose Edge Foundation also threw salons for Epstein that BuzzFeed News described as an “exclusive intellectual boys club.” (Brockman and his organization did not respond to a request for comment, and no reports of wrongdoing attended the events.) Epstein funded that foundation, which threw parties for billionaires and made contacts with people such as Pinker for him. Those contacts paid off: despite his dislike for Epstein, Pinker unwittingly contributed to the financier’s legal defense. Pinker wrote a 2007 opinion on the semantics of the wording of a prostitution law as a favor for Harvard professor Alan Dershowitz, who was Epstein’s lawyer and had once taught a course with Pinker. Pinker has said he didn’t know the opinion was for Epstein’s defense.
“I was doing a professional courtesy to a colleague—it’s routine,” Pinker says. “If I knew at the time what we know now, I would not have agreed.”
Epstein in a Harvard classroom in September 2004.Rick Friedman/Alamy
Celebrity
So, legal opinions aside, what did Epstein want from science? The simplest explanation is that Epstein collected prominent people. His financial networking relied on creating an aura of wealth and influence to entice investors. He was a “people collector” who traded information and favors, said Barry Levine, one of his biographers, in a 2025 BBC report.Scientists might have just been one of many influential groups he cultivated at a time that was “a cultural high-water mark for scientists as celebrities,” says Declan Fahy, an associate professor of science communication at Dublin City University in Ireland and author of The New Celebrity Scientists. Scientists wrote best-selling books, appeared in Vanity Fair and Vogue and gave viral TED Talks that were elevated online. “They moved into the power elite,” Fahy says, and so made sense for Epstein to cultivate.
Sign Up for Our Daily Newsletter
I agree my information will be processed in accordance with the Scientific American and Springer Nature Limited Privacy Policy. We leverage third party services to both verify and deliver email. By providing your email address, you also consent to having the email address shared with third parties for those purposes.Sign Up
According to Ghislaine Maxwell, Epstein’s former girlfriend and majordomo, who was convicted in 2021 of sex trafficking, conspiracy and transportation of a minor for illegal sexual activity, Epstein was particularly fascinated by brain science. In a July 2025 interview Maxwell told the DOJ that connections she had made through her father, Robert Maxwell, founder of scientific publisher Pergamon Press, led to her introducing Epstein to the Santa Fe Institute, a home to many high-profile scientists. (Epstein donated $25,000 to the institute in 2010.) “Epstein would have dinners at the house that I was tasked to organize and the scientists were a very major component of that,” she said, according to the DOJ transcript.
The scientist and writer Evgeny Morozov attributed Epstein’s scientific connections to Brockman—the literary agent who, according to Pinker, talked the psychologist onto Epstein’s plane—in a 2019 article in the New Republic. Himself a former Brockman client, Morozov recounted the agent’s attempts to connect him to Epstein and his “billionaires’ dinners,” whose attendees often were TED Talk speakers—invitations that Morozov declined.
The Edge Foundation was ubiquitous in science writing circles from 1998 to 2018, annually publishing books on scientific topics. It was also connected to the physicist Lawrence Krauss, a former member of Scientific American’s board of advisers, who was removed following sexual misconduct allegations in 2018. Released e-mail records show that Krauss asked Epstein for advice on handling those charges. Krauss has denied the misconduct allegations against him; none of the communications cited allege wrongdoing in connection with Epstein. (In 2014 Epstein was even invited to two Scientific American editorial meetings, which he did not attend.) Public records suggest the Edge Foundation received $638,000 from Epstein from 2001 to 2015, making him its major funder.
Social Prosthetics
One disturbing explanation for Epstein’s support of science comes from his interest in genetic determinism.This idea, which dates to the eugenics era, is still fashionable in some wealthy circles and can be seen in companies now offering designer baby services for embryos of would-be parents. In 2019 the New York Times reported that Epstein had ambitions of founding a “baby ranch” to raise children of women he impregnated (not unlike “secret compound” plans reportedly shared by SpaceX and Tesla chief Elon Musk).
“Given this stance, it is particularly disturbing that he focused his largesse on research on the genetic basis of human behavior,” wrote Naomi Oreskes, a historian of science, in Scientific American in 2020. “Scientists might claim that Epstein’s money in no way caused them to lower their standards, but we have broad evidence that the interests of funders often influence the work done.” (Regarding Epstein, Oreskes now adds, “The continued press attention reminds us that—rightly or wrongly—we are judged by the company we keep, and some money is tainted.”)
Perhaps the only direct evidence of Epstein’s scientific ambitions comes from a proposal he made in 2005 to be a Visiting Fellow at Harvard. “I wish to study the reasons behind group behavior, such as ‘social prosthetic systems,’” he wrote in an application proposing magnetic resonance imaging studies on human volunteers. “That is, other people can act as ‘prosthetics’ insofar as they augment our cognitive abilities and help us to regulate our emotions—and thereby essentially serve as extensions of ourselves,” he added, with a scientific gloss neatly encapsulating his view of humanity’s role in his life. Harvard approved him twice for the fellowship, though a 2020 investigation later noted his utter lack of qualifications.
A Rocky Pedestal
One last question is why anyone is surprised that celebrity scientists fell into Epstein’s orbit—as opposed to, say, rock stars or politicians doing so—in a culture driven by the worship of wealth and celebrity.
“A bit of this is [because] we have created an idealized picture of scientists that doesn’t match reality,” Lewenstein says. Scientists themselves like being seen as experts with their status on a pedestal, he adds. “They are very reluctant to acknowledge the social forces that shape their science,” Lewenstein says.
Most of the scientists supported by Epstein weren’t overtly political and supported a once-uncontroversial view of science as an engine of progress, Fahy says. Things are different now, “where public debate around science in the U.S.—particularly around climate and vaccination—has become sharper, divisive, intensely political,” he adds.
All that leaves Pinker unsure why his four seconds on the plane in 2002 was the only video in the Epstein files to be initially released by the Trump administration. One reason might be to generate news stories such as this one about scientists, he says. “The more that journalists write about other people in photos, the less attention Trump’s entanglement gets,” Pinker says.
Dan Vergano is senior editor, Washington, D.C., at Scientific American. He has previously written for Grid News, BuzzFeed News, National Geographic and USA Today. He ischair of the New Horizons committee for the Council for the Advancement of Science Writing and a journalism award judge for both the American Association for the Advancement of Science and the U.S. National Academies of Sciences, Engineering, and Medicine.
If you enjoyed this article, I’d like to ask for your support. Scientific American has served as an advocate for science and industry for 180 years, and right now may be the most critical moment in that two-century history.
I’ve been a Scientific American subscriber since I was 12 years old, and it helped shape the way I look at the world. SciAm always educates and delights me, and inspires a sense of awe for our vast, beautiful universe. I hope it does that for you, too.
If you subscribe to Scientific American, you help ensure that our coverage is centered on meaningful research and discovery; that we have the resources to report on the decisions that threaten labs across the U.S.; and that we support both budding and working scientists at a time when the value of science itself too often goes unrecognized.
The billionaires’ eugenics project: how Epstein infiltrated Harvard, muzzled the humanities and preached master-race science
The billionaires’ eugenics project: how Epstein infiltrated Harvard, muzzled the humanities and preached master-race science
Edge – Jeffrey Epstein’s favourite intellectual salon – was sold to me as a gathering of the world’s finest minds, writes Virginia Heffernan. The files reveal it was something far darker: a decades-long project that cloaked eugenics, race science and sexual misconduct in Ivy League respectability
Graphic: John Brockman, Jeffrey Epstein and Harvard University, Massachusetts. Photos: Getty Images.
It’s alarming to see your name in the Epstein files, but I was braced to see mine. Years ago, I was part of a salon for intellectuals and pseudointellectuals called Edge founded by John Brockman. His mass emails evidently copied in Epstein and a dozen such email blasts made their way to the latest dump of hazmat.
Brockman, my former agent for tech writing, told me Edge was an intellectual salon. Edge.org is indeed intriguingly sprawling, jammed with scholarly idols whose bios have “Booker” and “Nobel” in them. Members of Edge participated in conferences and symposia, and promoted each other’s work. Who was I to say no? Among Edge’s prodigious ranks were Ian McEwan, Yuval Noah Harari, Steve Wozniak, Richard Dawkins, Nassim Nicholas Taleb and Daniel Kahneman.
But if I’d read the member list more closely, I might have hesitated. Edge was overwhelmingly male, for one. It was said to be an intellectual salon, but in the club photos were tech bro billionaires, including Edge members Elon Musk, Bill Gates and Larry Page. And too many members were men now largely renowned for misconduct, professional or personal: Marc D Hauser, Jonah Lehrer, Lawrence Krauss, and Marvin Minsky.
Turns out I didn’t have to worry about meeting these people. Brockman kept me at a distance. As the latest Epstein files reveal, the token female members of Edge were actively excluded from schmoozing and conferences, especially the glittering events known as the Billionaires’ Dinners.
Good policy. Otherwise, we might have struck up conversations with the anxious-looking teenage girls kept out of the photos. We might have overheard the Edge men praising race science, rape culture and genetic engineering. We might even have asked where the money came from. Then we would have come face to face with the illiterate child rapist and passionate eugenicist who bankrolled the whole thing. Jeffrey Epstein.
Whatever Edge was supposed to be, it became something sinister. The salon played yenta to billionaire money and alpha-male minds, and together, over decades, they all converged on a master philosophy: they were apex predators ordained by nature to exploit and subjugate others. This creed allowed the Edge set to steer intellectual history into its current fascist dead-end.
Jeffrey Epstein (2nd from left) at a dinner he hosted at Harvard, September 2004 with (l-r) professors Alan Dershowitz, Robert Trivers, Lawrence Summers (formerly Secretary of the Treasury and Harvard President) and Stephen Pinker. Photo: Rick Friedman / Alamy
Edge began in 1996, an online iteration of a club Brockman founded to promote technological ideas and oppose what he called the “sleepy wisdom” of the humanities. The 1990s and 2000s were a perfect time for Edge. The club gained momentum along with an avalanche of books that savaged political correctness, multiculturalism, and the “Obamacrats”. Several of these books were by men in the Edge circle and the Epstein files, including Palantir mastermind Peter Thiel, Trump’s crypto czar David O Sacks, and computer scientist David Gelernter. (Gelernter stopped teaching at Yale this month after his own lecherous emails with Epstein surfaced).
Billionaires really like thinkers who see their exploitation of the weak as a good and natural thing
With its contempt for the humanities, Edge offered intellectually insecure reactionaries a pass.Without even opening a book, they could dismiss all of feminism, postcolonial theory and queer studies. They could continue to ignore giants like Edward Said, Judith Butler, and David Graeber, and devote their brains instead to the race science and rape apologetics that now pass for scholarship on edgelord podcasts.
It’s not clear when Epstein met Brockman, but by the late 90s he’d burrowed into Edge. With his cash infusions, Edge came to be known as home to far-right academics and the tech billionaires who love them. By 2000, Epstein was flying the Edge sausage party around on his planes, which served meals from Le Cirque and was appointed with mink throws.
The academics, in turn, liked the billionaire glitz. Middle-class rightwingers at Edge functions, including former New York Times columnist David Brooks and neuroscientist Sam Harris, now consorted with the likes of tech monopolists Jeff Bezos and Sergey Brin.
In this atmosphere of warm brotherhood, how could they not have felt chosen to rule over the rest of us? One Edge member and Epstein consort, the anarchist Noam Chomsky, described this ethos: “The cool observers – meaning us smart guys – it’s our task to impose necessary illusions and emotionally potent oversimplifications to keep these poor simpletons on course.”
Jeffrey Epstein and the late professor Marvin Minsky, September 2004. Photo: Rick Friedman / Corbis via Getty Images.
The men of prestige and ambition that Epstein especially liked to sugar-daddy at Edge were figures with a hand in the grim sophistry of evolutionary psychology. Among these were Edge darlings Martin Nowak, a “mathematical biologist”, and evolutionary biologist Robert Trivers.
In an email in the recently released Epstein files, Nowak, whose Program for Evolutionary Dynamics (PED) at Harvard was founded with $6.5m from the child rapist, is asked “Did you torture her?” by Epstein. It’s not clear what Epstein is referring to.
As for Trivers, his name appears in the Epstein files as part of an FBI report alleging that he sexually abused a 15-year-old girl in late 2019. (On the Edge site, Steven Pinker, a zealous Edge member, called Trivers “one of the great thinkers in the history of Western thought”.)
In a classic of the just-asking-questions form perfected by rightwing trolls, Pinker pummelled the Edge audience in 2006 with these pressing inquiries: “Do most victims of sexual abuse suffer no lifelong damage? Did Native Americans engage in genocide and despoil the landscape? Do men have an innate tendency to rape?” And then, for good measure: “Is homosexuality the symptom of an infectious disease?”
And thus while he was peddling his vision for Edge as “a salon for the world’s finest minds”, its male members were preaching master-race ideology. Citing their work, Brockman predicted in 2005 that the “dangerous idea of the next decade” would be that “groups of people differ genetically in their average talents”.
Surprise: white men, notably those of Ashkenazi descent like Epstein, have been “biologically selected for high intelligence”, according to the Edge site.
The intellectual groundwork was laid for Epstein to spring his world-domination plan on his kept Edge scholars. The big reveal: he hoped to seed the human race with his superior DNA by impregnating women at his ranch in New Mexico.
Throughout the decades, according to The New York Times, Epstein talked up this ambition. He told at least one fellow eugenicist that he hoped to freeze his peerless brain and penis at his death, so his organs could be revived for future use in “transhumanism”. He also generously funded the work of George Church, the Harvard geneticist and Edge superstar who developed a dating app to match people based on the fitness of their genes. Epstein’s friends enthusiastically discussed eugenics with him up till shortly before he died in prison.
Screenshot from Edge.org website showing John Brockman, Steven Pinker, Daniel C Dennett, Katinka Matson and Richard Dawkins en route to TED in 2002. According to Yahoo Finance, an earlier picture caption, since altered, stated that they were travelling with Epstein.
What’s more, even as the NYT concludes that “there is no evidence that [Epstein’s plan] ever came to fruition”, this period in Epstein’s life is described contemporaneously by his teenage rape victims. One traumatised 16-year-old, whose diary surfaced in the most recent tranche of files, calls herself an “incubator” for Epstein’s offspring. She chronicles Epstein’s “Nazi”-like effort to create a “superior gene pool”. The fact that Sarah Ferguson, in a recently released email, congratulates Epstein on a “baby boy” has raised questions about where the DNA-seeding project stands now.
Other Epstein survivors, in the files and in court, have recounted hellish experiences of enslavement on the ranch, forced pregnancies, and bloody deliveries. In emails released by the justice department, Epstein is consulted dozens of times about pregnancies, sonograms, egg-freezing, and other obstetrical matters.
Finally, there’s the long video interview that Epstein gave to Steve Bannon, the far-right Christian nationalist, which provides more context for Epstein’s obsession with eugenics. To start, Epstein blithely explains his racist worldview. “If I was in the forest and I had to run from the lion or figure out a way not to be eaten, and my competition is a local African, I’m the one who’s getting eaten … They have the intelligence to deal with their local environment.”
Epstein also tells Bannon that he helped fund the Santa Fe Institute, a New Mexico research operation, in part to advance his interest in “genetic algorithms”, which he can hardly describe. “Complex systems are complex, by definition,” he says. Epstein justifies his own manifest illiteracy by saying that people who know how to write can’t think broadly, unlike Socrates, Jesus, and himself.
“Genetic algorithms” are evidently systems theory crossed with race science. They “work” by “improving” chromosomes or their digital analogues using selection and mutation over several generations. Genetic algorithms are also described as “a metaheuristic inspired by the process of natural selection”.
I flashed back to the Edge crew’s relentless criticism of the humanities in the 1990s. In The Diversity Myth, Thiel and Sacks bitterly complained about “diversity” as jargon that concealed a nefarious political agenda. Well, now we have metaheuristical eugenics, and the jargon’s on the other foot.
With the Epstein files, we’re confronted with exactly what all the Edge men – from Pinker to Dawkins to Musk to Gates – did with the intellectual territory they seized. With their Ivy League posts, their billions, and their blue-ribbon DNA, the would-be intellectuals in Epstein’s circle converged on nothing less than the ideology of Mein Kampf.The Edge dinners have ceased and the site is now dormant, but generations of young men trained at Harvard, LSE and Oxford absorbed the lesson — and generations of young women learned that their place in intellectual history is sidelined, exploited, or prone.
Virginia Heffernan in an American journalist and cultural critic who writes regularly for The New Republic. Her podcast and newsletter, Magic + Loss, can be found on Substack
The Nerve is a fearless, independent media title launched by five former Guardian / Observer journalists: investigative journalist Carole Cadwalladr, editors Sarah Donaldson, Jane Ferguson and Imogen Carter and creative director Lynsey Irvine. We cover culture, politics and tech, brought to you in twice-weekly newsletters on Tuesdays and Fridays (sign up here). We rely on funding from our community, so please also consider joining us as a paying member. You can read more about our mission here.
Cambridge Analytica Scandal (2018): Cadwalladr exposed how the data analytics firm harvested data from 87 million Facebook users without consent, using it to influence the 2016 US presidential election and the UK Brexit referendum. [1, 2, 3, 4, 5]
Tech & Democracy: She focuses on the intersection of technology, data, and politics, often investigating “fake news” ecosystems, surveillance capitalism, and AI’s impact on democratic processes. [1, 2]
“Broligarchy”: She coined this term to describe the alliance between Silicon Valley tech executives and far-right political movements, a theme she covers on her Substack newsletter. [1, 2, 3]
Legal Battles: Her investigations led to a three-year defamation lawsuit brought by Brexit donor Arron Banks, which she won in 2022, solidifying her reputation for tackling high-stakes reporting. [1, 2]
Career & Background
The Observer/Guardian: She was a longtime feature writer for The Observer and The Guardian from roughly 2005 to 2025.
The Nerve: In 2025, she co-founded The Nerve, a digital publication focused on politics and technology.
Awards: She was a Pulitzer Prize finalist for National Reporting in 2019, won the Orwell Prize for Journalism in 2018, and received the Polk Award. [1, 2, 3, 4, 5, 6]
Current Focus
As of 2026, Cadwalladr is active in investigating the role of AI in democratic backsliding, operating via her Substack, The Nerve, and as a speaker on tech accountability. She also established the non-profit “[allthecitizens]” to combat threats to democracy. [1, 2, 3, 4]
If you’re interested in her work, I can provide details on her specific TED talks or list the major awards she has won. Let me know what you’d like to explore next.
Carole Cadwalladr (@carole_cadwalladr) – Instagram* “How to Survive the Broligarchy” writer @carole_cadwalladr sounds the alarm about AI’s lawless business model. * British author …Instagram·carole_cadwalladr
Carole Cadwalladr | The Guardian20 Apr 2025 — Carole Cadwalladr. Headshot of Carole Cadwalladr. Carole Cadwalladr is a reporter and feature writer for the Observer … April 20…The Guardian
Those were among the most controversial remarks made by former UN weapons inspector Scott Ritter during a panel discussion at the St. Petersburg International Economic Forum (SPIEF). In a speech that quickly drew attention, Ritter claimed that “Europe doesn’t exist,” argued that Europe “stands for nothing,” accused Germany of returning to its militaristic past, and warned that the world is becoming increasingly dangerous as nuclear tensions rise.
Ritter also criticized U.S. foreign policy, describing American actions in Iraq as driven by regime change rather than disarmament, and issued a stark warning about nuclear conflict, arguing that once a nuclear war begins, it could escalate into global destruction. Watch the full highlights from Ritter’s explosive remarks on Europe, Germany, the United States, NATO, and the growing risk of nuclear confrontation. #ScottRitter#SPIEF#Russia#Europe#Germany#NATO#Ukraine#Geopolitics#EuropeanUnion#USPolitics#NuclearWar#WorldPolitics
Summary: A new study has shattered the long-held scientific consensus regarding the human brain’s capacity to engage in true multitasking. The study demonstrates how the brain physically remodels its underlying architecture after extensive experience to automate learned tasks.
By utilizing functional MRI (fMRI) and EEG technologies, investigators proved that continuous training forces complex processing tasks to migrate out of the bottlenecked prefrontal cortex and into the temporal cortex, bypassing executive deliberation entirely and leaving the frontal networks clear to handle parallel operations.
Key Facts
The Frontal Bottleneck Overpass: Early stages of skill acquisition rely heavily on the prefrontal cortex, the region governing executive function and thinking, which historically acts as a strict cognitive bottleneck capable of handling only one demanding task at a time.
The Temporal Offloading Discovery: Following weeks of extensive training, the neural circuitry physically shifts, offloading the automated task to the temporal cortex, a brain region optimized for object recognition and memory encoding.
The 30,000-Trial Longitudinal Audit: Researchers tracked participants who completed more than 30,000 image-sorting trials over a 5-to-10-week span via a smartphone app game, allowing researchers to capture structural brain scans both before and after expertise was achieved.
Dismantling the Task-Switching Myth: The findings directly challenge the traditional neurological theory that human multitasking is an illusion made up of rapid, back-and-forth task-switching. Instead, the study proves the brain can physically build distinct, separate neural circuits to execute two tasks simultaneously.
The Unlearning and Compulsion Metric: Because automated behaviors move into circuits less accessible to conscious thought, the research reveals why cognitive strategies like “thinking of something else” fail to curb compulsive behaviors, providing a new anatomical map to guide addiction therapies.
The Human Continuous Learning Blueprint: Moving automated skills into the temporal cortex frees up the prefrontal cortex to use old information as a modular building block for new skills—a major breakthrough that explains human continuous learning efficiency compared to current artificial intelligence models.
Circuit Compatibility Horizons: Senior author Dr. Maximilian Riesenhuber and first author Dr. Patrick Cox note that future research will focus on the exact signals that trigger this neural migration and define the limits of parallel processing, noting that tasks remain dangerous if they compete for the same physical sensory mechanics.
Source: Georgetown University
New research by Georgetown scientists shows how the brain rewires itself to automate learned tasks. The findings challenge a long-held understanding of how humans master complex skills, suggesting that true multitasking is really possible.
Beyond offering encouragement to busy people that they really can do two things at once, the study also has important implications for the development of artificial intelligence capable of building on prior learning as the brain does.
“We have another stepping stone in our understanding of how the brain learns,” said senior author Maximilian Riesenhuber, PhD, a professor of neuroscience at Georgetown University School of Medicine, and co-director of the Center for Neuroengineering. “The encouraging part is that you really can learn to multitask. There is actually a way to remodel your brain architecture and use other parts of your brain.”
The new study builds on decades of research on how learning occurs in the brain.
Scientists wanted to understand the mechanisms behind automation, and how the brain shifts from learning a new task into a way of executing that task more unconsciously after extensive experience.
A good example is driving, Riesenhuber said. When someone first learns to drive, it requires their full concentration. But after driving for many years, most people can talk, listen to music, or consider a problem without having to focus completely on operating the vehicle.
“The question is: how does your brain do that?” Riesenhuber said.
Most previous research on learning has focused on the early stages, but what happens to the brain long-term is harder to study and less understood.
For the new study, researchers trained people to sort morphed images of cars into two categories, learning to spot subtle differences to tell them apart. Participants completed more than 30,000 trials over 5 to 10 weeks, using an app that allowed them to sort the images as a game on their phone. Researchers used fMRI and EEG to conduct brain scans on the participants before and after they completed the trials.
They found that after people had initially learned to sort the images, the task activated their prefrontal cortex. This area of the brain is responsible for executive function and thinking, but can typically only handle one task at a time.
However, when researchers scanned the brains of participants who had been practicing the sorting task for weeks, they found that the categorization was now happening in the temporal cortex, a part of the brain involved in encoding memory and recognizing complex objects.
“Previous studies have shown that parts of the temporal cortex can be activated by particular object categories in experienced observers, birds, cars, even Pokemon, but a limitation of all of those studies is that they only looked after people became experts.
“The strength of this study is that it is longitudinal, we measure before and after training, so we can see that extensive training essentially put a category selective area in the temporal lobe that was not there before,” said first author Patrick Cox, PhD, who began the study as a graduate student in Riesenhuber’s lab and is now an assistant professor of psychology at Lehigh University.
“This has implications for critical real world scenarios, like when a radiologist can accurately classify masses on an x-ray as benign or malignant fairly automatically, often without extensive deliberation, thanks to years of training,” Cox said.
Category information from the car-selective area in the temporal cortex bypassed the prefrontal cortex and connected directly to output parts of the brain.
“Experience remodels the brain to bypass that frontal bottleneck. The prefrontal cortex then stays free for whatever else you want to do, increasing your capacity,” Riesenhuber explained. Indeed, the researchers found that the more the car task was “offloaded” from the prefrontal cortex, the better people were able to do another task in parallel to the car task.
The finding challenges a longstanding theory that humans are not capable of true multitasking. Instead, it was thought that the brain rapidly switched back and forth between two tasks.
“What we show is that the circuitry actually changes so the brain can do two things at once,” Riesenhuber said. “This really is true multitasking.”
The findings can also have implications for understanding compulsive behaviors, because they demonstrate that learned behaviors move into brain circuits that are less accessible to conscious thought or executive function.
“The first step to unlearning something is understanding where it is actually happening in the brain,” Riesenhuber said. “This shows why strategies like telling someone to think of something else don’t really help, because they don’t really have the behavior under conscious control.”
It also helps explain why humans are so good at continuous learning, or building skills upon skills — something that AI still struggles with.
Moving a learned skill into the temporal cortex and freeing space in the prefrontal cortex could allow the brain to use the old information as a building block to learn something new, Riesenhuber said. Current AI models don’t have that same capability, he noted.
Next, researchers want to study the mechanisms or signals involved in moving learning from one part of the brain to another and to figure out what the limits of multitasking are.
“Another really interesting question is what kinds of tasks can be learned well enough to do in parallel,” Cox said. “We can walk and chew gum at the same time, but looking at our phones to text while driving will never be safe, because we take our eyes away from the road. It comes down to being able to train fully separate neural circuits for two tasks to become compatible.”
Funding: Funding for this study was provided by the National Science Foundation (BCS-1232530) and the ARCS Foundation, and the Army Research Laboratory (W911NF-24-1-0097). The authors report having no personal financial interests related to the study.
Key Questions Answered:
Q: How does the human brain physically alter its own shape to make true multitasking possible?
A: By building an entire category-selective area where one didn’t exist before. The Georgetown University study showed that practicing a skill tens of thousands of times rewires the brain, allowing the task to migrate out of the crowded prefrontal cortex and relocate to the temporal lobe, creating a permanent, automated circuit.
Q: Why does this discovery explain why it is so difficult for people to break compulsive habits?
A: Because deeply learned habits migrate to brain regions that completely bypass your conscious control center. Since these automated behaviors are handled in the temporal cortex rather than the prefrontal executive network, simply trying to “think of something else” is ineffective because the behavior is running on a circuit separate from conscious thought.
Q: What can artificial intelligence engineers learn from how the human brain shifts tasks between regions?
A: How to master continuous learning without wiping out past data. Freeing up space in the prefrontal cortex by offloading mastered habits to the temporal cortex allows humans to use old memories as modular building blocks to learn new things, a structural trick current AI models still struggle to replicate.
Editorial Notes:
This article was edited by a Neuroscience News editor.
Journal paper reviewed in full.
Additional context added by our staff.
About this neuroscience and neurotech research news
Author: Karen Teber Source: Georgetown University Contact: Karen Teber – Georgetown University Image: The image is credited to Neuroscience News
Extensive Experience Remodels Neural Task Circuitry to Escape the Frontal Bottleneck and Increase Automaticity of Categorization
Object category learning is a foundational cognitive process. Most human category learning studies involve brief paradigms lasting a few hours and show increased shape tuning in visual areas and task-dependent responses in pFC.
Other studies also identify a “frontal bottleneck” that limits multitasking. However, real-world categorization often involves months or years of practice, potentially producing qualitative shifts toward automaticity. We tested the hypothesis that extensive training causes a spatio-temporal shift in the neural circuitry supporting categorization.
Participants were trained over >30,000 trials across 5–10 weeks to categorize novel morphed car stimuli via a mobile app. We used fMRI and EEG rapid adaptation techniques to examine neural responses after initial learning (∼4 hr in 1–2 weeks) and after extensive training (∼16 additional hours over another 4–8 weeks).
Converging fMRI and EEG results showed that extensive training fundamentally remodeled task-related circuitry: Visual areas in ventral occipito-temporal cortex (vOTC) were initially shape-selective, but category-selective responses emerged in the vOTC after extensive training. The vOTC also showed decreased functional connectivity with the pFC and increased connectivity with motor output areas.
This supports the hypothesis that extensive experience enables category decisions to occur outside of the “frontal bottleneck.” Critically, the decrease in connectivity between vOTC and pFC was associated with improved categorization performance while dual-tasking, indicating increased automaticity.
These findings demonstrate that prolonged training reshapes the neural basis of categorization, shifting it from a flexible but attentionally controlled process to a more streamlined, automatic process.
Decades of scientific investment have paid off in the past month, with researchers announcing promising breakthroughs against cancers and other deadly afflictions, Axios’ Caitlin Owens writes.
“This has quietly been a miracle month in medicine,” notes Derek Thompson, author of a smart Substack and co-author of “Abundance.”
In late-stage clinical trial results, Revolution Medicines’ experimental pancreatic cancer treatment doubled patients’ life expectancy compared with standard chemotherapy.
Eli Lilly’s latest experimental anti-obesity drug appears to reduce body weight at levels approaching bariatric surgery in clinical trials.
A Mayo Clinic AI model spotted abnormalities on scans up to three years before patients were diagnosed with pancreatic cancer.
Reality check: We’re still in the era of medicine where most miraculous new drugs generally don’t cure the disease as much as allow sick people to live longer.
And the new wave of anti-obesity drugs, which have the potential to stave off cardiovascular disease in the long term, are expensive and must be taken indefinitely to retain their benefits.
What we’re watching: AI is fueling hopes for a transformational era of medicine, where diseases are detected earlier and cured altogether.
An experimental drug acquired by Eli Lilly earlier this year, which targets multiple myeloma by editing cells inside the body, showed a 100% response rate in early-phase clinical trial results.
A small, early-stage study found that an experimental gene-editing therapy could potentially permanently lower cholesterol levels after just one infusion, opening the door to one-and-done heart disease prevention.
Late-phase trial results for a hepatitis B treatment found it was “a functional cure” for 20% of patients who received it.
Sign up for Axios Vitals for the latest in health care innovation and policy.
View in browser Axios Finish Line By Mike Allen and Erica Pandey and Jim VandeHei ·Jun 03, 2026
Welcome back! Axios CEO Jim VandeHei has the reins, sharing a speech he recently delivered to communications execs at the annual PTTOW! summit.
Join the conversation. Reply to this email or hit finishline@axios.com to share your thoughts with Jim.Smart Brevity™ count: 1,198 words … 4½ mins. Copy edited by Amy Stern. 1 big thing: Build your bionic brainIllustration: Aïda Amer/Axios
We’re at war. A war for our brains. This war rages from sunup to sundown, phone on to phone off. While we sleep. It’s our Forever War, Jim writes.
Our combatants are social media algorithms and emerging AI systems. Both never tire or relent. They only grow smarter and more sophisticated.
I’m here not to scare or shame you. I’m here to tell you this war will determine the future of your mind and your intelligence. It is winnable on your terms — if you understand the nature of it.
In the hours, days and years ahead, you can choose to understand and seize control of your mental inputs, building a bionic brain. Or you can allow algorithms and AI to do your thinking for you, succumbing to what I call “blah brain.”
Make no mistake: We’ve entered the age of extreme information inequality. It’s the defining divide of the next decade. Bigger than wealth. Bigger than education. Bigger than geography.You land on the right side of the divide by controlling what information you consume and by using AI to augment — not replace — your learning and thinking.You will choose. Choose bionic. Reject blah.
Let me start with a confession. I get paid to learn. I get paid to hire people with true subject-matter expertise. I get paid to filter fact from fiction every single day. My business depends on it. And I struggle to stay on the bionic path.I pick up my phone to check one thing and lose 20 minutes.I tell myself I’m reading the news and scroll away precious time on trivialities. I open ChatGPT to sharpen my thinking and catch myself letting it do the thinking for me.If I’m losing this battle some days — and I am — no wonder most of America feels like it’s getting crushed.
So I want to talk about this war you don’t realize you’re in — and then show you how to win it.
Start by understanding the Great Information Paradox of the 21st century.There is more misinformation, manipulative content and pure mind garbage available to us for free than at any point in human history. At the same time, there is more high-quality, mind-expanding, life-enhancing content available to us for free than at any point in human history.We all have access to the smartest minds alive via podcasts, YouTube and newsletters.Greatness and garbage. At the same time. On the same device. In the same five seconds of interaction.
So we face a choice — whether we know it or not, whether we like it or not — all day, every day. And now, AI raises the stakes on that choice by a factor of 10.
That’s because AI offers you two new paths — and they look almost identical from the outside. But they lead to completely different places.
Path One:Outsource your thinking to AI. Let ChatGPT write your emails. Let Claude make your decisions. Let an algorithm or LLM tell you what to believe about Iran, about the economy, about your own job and skills. Stop reading. Stop wrestling. Stop forming your own views. Just ask the machine and ship the answer.The temptation is enormous. It is easier. It is faster. It feels productive. And it will quietly hollow you out. In two years, you won’t have opinions — you’ll have prompts. You won’t have judgment — you’ll have outputs.
Path Two: Use AI to expand your mind. Same tools. Same machines. Completely different posture. You make AI argue with you.You make it challenge your assumptions. You feed it the things you don’t understand and demand that it teach you at your pace.
Path One shrinks you. Path Two enlarges you. Same machine. Opposite outcomes. The difference is entirely in how you use it. Most people are going to take Path One. They already are. Don’t be most people. Instead, do something radical: Build your own bionic brain. Starting today. Everybody runs on the same hardware — a human brain that hasn’t changed much since creation. The bionic part is what you feed into it and bolt on around it.
I want to give you something concrete. Five moves. You can start every one of them tonight.
Audit your inputs. Be ruthless. Open your phone right now. Look at the last 20 things you watched, listened to or read. Be brutally honest. Is that the diet of someone with a healthy brain? If not, cut three things tonight. Unfollow them. Mute them. The algorithm feeds you what you click on, not what you aspire to. The only way to retrain it is to stop clicking. Your feed in 90 days will be a different feed. Your brain in 90 days will be a different brain.
Pick one anchor reality source. It’s impossible to follow the news and know what’s real or important. Simplify. Pick one broad source of truth with an established reputation for accuracy.I’d recommend Axios, of course, but The Wall Street Journal or Financial Times are good options, too.
Pick two passions, two areas you want to be smarter in. Maybe it’s health or technology or history or literature. Find people with track records of being smart and open-minded on those topics, whether in podcasts, videos or writing. Their backgrounds should demonstrate expertise and curiosity. Follow them. They will lead you to others.
Build your own JimGPT. Use Claude, ChatGPT or Gemini to start building a wise companion brain. Tell it with specificity that you want to learn and be challenged, and to never be flattered or blindly reinforced. Tell it your passions and goals — including resisting the urge to outsource your thinking to AI — and commit that to memory. Ask it to ask you questions to shape its output to how you learn.
Start small: a new 30-minute habit. Replace one half hour a day of doom-or-boredom scrolling with healthy inputs and AI interactions. Listen to that podcast. Read that newsletter. Watch that video. Explore that topic on AI. Ask the AI to steelman the argument against something you believe: “Present the strongest, smartest, most persuasive fact-based case against my view.” Thirty minutes a day equals one full week of purging stupid and injecting smarts.
Do these five things and I can assure you, you will emerge a better, smarter, more satisfied person — with what feels like a new bionic brain.Share this column … Watch Jim’s video.If you’re a CEO or on a CEO’s team: Ask to join Jim’s new weekly Axios C-Suite newsletter.