Almost thirty years later, the report looks both prophetic and quaint. Its concerns have proved to be warranted — type 2 diabetes, non-alcoholic fatty liver disease, certain cancers, heart disease, and a host of other ailments caused or exacerbated by obesity have indeed replaced malnutrition and infectious disease as the primary causes of ill health in many countries. But the scope of the problem has grown so drastically that a return to the rates that alarmed researchers in 1997 would now be an astounding achievement. To take just one example, by 2020 Colorado, consistently the American state with the lowest rate of obesity, had a higher prevalence than Mississippi, the state with the highest, did in 1990. The trend has been persistent; nowhere has it been meaningfully reversed. Like ennui and Taylor Swift, obesity is an unavoidable consequence of modernity.

But two recent developments have raised the possibility that fatalism should no longer be the default. The first of these, Robert F. Kennedy, Jr.’s Make America Healthy Again movement, has been one of the loudest factions in both Donald Trump’s presidential campaign and his administration. While it is easy to dismiss MAHA for its many outlandish views and bombastic conspiracy-mongering, doing so risks denying a potent political reality: the modern food environment has sickened a significant portion of the populace, including millions of children, and people want the government to do something about it. Individual hand-wringing about the consequences of poor nutrition has coalesced into public calls for radical change.

Even more potent than a nascent political movement is the advent of weight loss drugs that actually work. After decades of flailing, drug companies have developed an effective pharmaceutical treatment for obesity. Rather than waiting for the uncertain work of politics to bring about changes in public health, weight loss drugs offer an injectable solution. The consequences of living in a broken food system can be alleviated by a quick trip to the doctor’s office.

MAHA, and Ozempic and its cousins, are both responses to the same problem. But they diverge in their diagnosis.

The drug companies do not explicitly articulate what’s gone wrong, but the critique implicit in their solution is radical. The drugs work because they help people avoid not certain unhealthy foods but overconsumption of all food. The default diet as a whole is the poison. But, like many poisons, at a low enough dose it’s not that harmful, and weight loss drugs help reduce exposure to within the safe limit.

Kennedy’s diagnosis of the problem has been a little less than clear too. In the past he has often blamed the food system at large for increases in obesity and related diseases. “Ultra-processed food is driving the obesity epidemic,” he posted in October 2024. He went on to claim that government subsidies for their main ingredients, corn, soy, and wheat, are systemically perpetuating the epidemic. But now that Kennedy has political power, his focus has been much narrower. Rather than pursuing ambitious reforms, he has focused on a particular sort of contamination. Food companies have cozied up to regulators, crafting loopholes permitting them to add artificial colors and flavors, emulsifiers and preservatives — hidden, novel ingredients that have sickened us all. His main solution has been to encourage food companies to tweak their ingredient lists. This narrowing of his focus has him engaging in a mirror of the behavior he previously critiqued; in his role as a government official, he now gives cover to the merchants of sugar and fat so long as they remove some more obscure and disfavored ingredients from their products.

If Kennedy is correct that stripping the food system of hidden technology — the hundreds of compounds used by food scientists to enhance the color and flavor of everything from bread to burritos — will be a significant boon to public health, then the surprisingly tame policy response he has pursued as Health and Human Services Secretary makes sense. But if his prior, more ambitious critique is correct — if the problem is the more ubiquitous hedonic technology of the modern food system itself — then his fixation on dyes in breakfast cereal is not just inadequate but a big distraction. The only solutions that would make a real difference would be either drugs capable of dampening desire as a whole or radical systemic change.

Since joining the Trump administration, Robert F. Kennedy, Jr. has often looked more like a standard-issue politician than the leader of a revolution. Shortly after the election, he posed with Trump, Elon Musk, and a tableful of McDonald’s aboard Trump Force One. During his confirmation hearing, Kennedy emphasized that he did not want to impose new regulations on farmers, but to incentivize them to use healthier practices. Noting his boss’s fondness for cheeseburgers and soda, he also said he had no interest in limiting consumer choice, but simply wished to provide the American people with clearer information.

The exception to his accommodating stance was additives. He evinced real passion when pointing out that American junk foods, whether McDonald’s fries or Froot Loops, contain far more polysyllabic ingredients than their European counterparts. His message, in effect, was that you could have your Twinkie and eat it too if Hostess just tweaked the recipe a little.

An early demonstration of what MAHA counts as a victory came when the burger chain Steak n’ Shake announced that it would start cooking French fries in beef tallow rather than seed oil, a decision Kennedy celebrated by joining Sean Hannity at one of the chain’s locations for an advertorial. In March 2025, he posed with the heads of major food companies, announcing plans to work with them to MAHA-ify their products. A regulatory victory arrived in April in the form of a plan to remove artificial red food coloring from the entire food supply.

Given the heavy emphasis on dyes, cleaning up the food system has so far been literally superficial, with brief forays into sweeteners. In April and May, PepsiCo and Tyson announced plans to rapidly move from artificial to natural colors, or at least to market options free from food dyes. President Trump himself weighed in, taking credit for convincing Coca-Cola to replace corn syrup with sugar. “You’ll see. It’s just better!” he posted. Most of the ice cream supply may soon be MAHA-friendly, or at least free from artificial dyes. In perhaps his greatest triumph to date, in July Kennedy was finally able to congratulate Kellogg’s for bringing Froot Loops into MAHA’s good graces by switching from artificial to natural colorings, an example he urged other companies to follow.

While a relentless focus on the picayune may look strange to the uninitiated — are Froot Loops colored with beet powder really so much healthier? — it makes sense within MAHA. It would be too tidy to say Kennedy divides the world into natural (good) and unnatural (bad), but in his view the unnatural components of our food are the problem. The hidden technology of Blue No. 2 and Red No. 3 lurking in the fine print of ingredient lists are sickening us at least as much as the sugar at the very top.

There may not be a strictly logical reason why a belief that dyes and artificial flavorings are pernicious villains of the food system should correlate with a distrust of vaccines, EMFs, and novel weight loss drugs, yet they resonate. They play a paranoid chord. Shadowy actors in industry employ means of manipulation so sophisticated that almost no one recognizes them for what they are. They inject impurities into our bodies and into the body politic. Casting additives as the problem allows Kennedy to tell a story that is cinematically thrilling and politically tractable. It is also of dubious truth.

Each time they cajole a company into replacing an artificial ingredient with a natural one, Kennedy and MAHA celebrate it as the removal of nefarious hidden technology that is corrupting the food system. But a much more convincing, if also more dire, idea is that the food system itself constitutes a hedonic technology — one that is so effective that it dysregulates the metabolisms of almost everyone who comes into contact with it, killing many, crippling more, and doing at least some damage to all the rest. The additives that MAHA obsesses over may be a part of the problem, but they are a distraction from the primary cause.

So low is the baseline set by the Standard American Diet, which goes by the apt acronym SAD, that any significant divergence from it, whether keto or vegan or Mediterranean, will be an improvement. There is robust evidence that every diet works. There is evidence too that very, very few people are able to sustain an alternative diet, whether keto or vegan or Mediterranean. SAD has become the standard for a reason, and that reason is not a conspiracy, at least not the sort Kennedy imagines. There has been a campaign of sabotage, but our biology has been waging it against itself shoulder to shoulder with Kraft.

Dietary science is notoriously complex and often unreliable. But both observational and small but compelling controlled trials support the idea that ultra-processed food encourages excess calorie intake when compared to a less processed diet. By trying out combinations of flavors, identifying the ones people prefer, and then repeating the process for a few decades, food companies have created the conditions for chronic overconsumption. Doritos are an inevitable consequence of applying the most basic sort of scientific method to the question of what people want to eat. Food most of us cannot resist packs the shelves and checkout lines, is advertised to us constantly, is served in school cafeterias, convenience stores, and restaurants — and so, predictably, most of us eat too much.

Kraft has teams of food scientists diligently working to develop hyperpalatable snacks, confections of fat and sweet and crunch, that tap into intrinsic systems of taste and hunger more effectively than a plate of chicken and rice rounded out with spinach and a pinch of salt. They don’t need to resort to the secret manipulations of intentionally engineering subtly addictive substances when openly concentrating what attracts us will create foods that are too hedonic. We prefer the combination of cheap, tasty, and convenient to simpler, blander foods that are more time-consuming to prepare. So much is this true that average Americans consume over half of their calories in the form of ultra-processed food, according to CDC data.

This is not to say that MAHA’s fixation on additives has no merit. Stir canola oil, cornmeal, and salt together in a bowl and the result is an unappetizing mush. But make batter out of the cornmeal and fry it into chips, then sprinkle on monosodium glutamate, and you’ve worked a bit of black magic, especially if you also add the powdered essence of ranch dressing to the mix plus a dash of Red 40 and Yellow 5. Given that manipulations are part of what makes foods more appealing, casting a jaundiced eye at all of the trace ingredients that heighten the flavor or color or texture makes sense. But if the concern is encouraging overindulgence, a flavoring powder made from actual cheese might not be any better than powder made with an artificial cheddar flavor.

The difficulty of doing long-term diet research on humans means we typically don’t know with any confidence the exact effects of any single additive when consumed over decades, let alone of dozens of additives consumed in combination. Troubling trends like the increasing prevalence of colorectal cancers in young populations and of autoimmune conditions in everyone justify concern.

But when Kennedy appends the caption “MAHA is winning” to yet another Steak n’ Shake post, this time about the availability of glass-bottled, sugar-sweetened Coke to accompany the tallow-cooked fries, he is missing the forest for the trees, or the soda for the sweetener. Perhaps sugar is indeed marginally better for you than corn syrup, and perhaps tallow is marginally healthier than soy oil, though only weak evidence supports either contention. But regardless of the sugar source, Coke is liquid calories, and regardless of the frying medium, fries taste so incredible that most of us can keep eating them forever. When Kennedy celebrates small changes to junk food, he is playing into the hands of the very food companies he was blaming for the crisis of health in America not so long ago.

Despite being a scion of one of America’s most prominent political families, Kennedy has always fancied himself a maverick. He distrusts power, whether corporate or governmental, and he saves his most vitriolic outrage for instances in which the two have become incestuously entangled. At the outset of his tenure as HHS Secretary, Kennedy’s background made it impossible to predict what he would do. Even in an administration notable for finding staff in unusual places, he stands out. He was a Democrat until about two minutes ago. He once suggested that corporations that spread climate denial “should be given the death penalty” by revoking their legal right to exist. He primaried a sitting president. The list of conspiracies he does not believe is significantly shorter than the list of those he does. Like President Trump, he has a habit of defying the political consensus, but, unlike his boss, he appears to do so out of genuine conviction.

His standing as a true outsider makes his capitulation to industry all the more frustrating. To hear MAHA tell it, food companies and regulators have misled and misfed the public by cultivating an aura of healthfulness around junk; extra-sugary cookies get a big label announcing they are “low fat,” and half the chips in the snack aisle are “All Natural.” But now the industry’s need for covertly neutering the watchdogs has been done away with. Simply swap out an artificial dye for a natural one or corn syrup for sugar and the government’s most prominent and powerful voice on the topic of health will give a free advertisement for whatever junk you’re selling the American people.

In contrast to the endless parsing of dyes and diglycerides favored by MAHA, weight loss drugs employ a brute-force approach to change. By making food categorically less appealing, they promise to reshape both the diet and the bodies of anyone willing to take them. Their names have been focus-grouped to the edge of absurdity: Ozempic, of course, but also Wegovy, Mounjaro, Zepbound, and so on, in a litany of generically aspirational corporate gobbledygook. The currently available drugs, as well as the thirty-nine and counting drugs in development, are glucagon-like peptide-1 receptor agonists, or GLP-1s. Though the biology is complex, the most significant effect isn’t. By slowing digestion and also acting on other bodily systems involved in satiety, they decrease the sensation of hunger.

If the fundamental problem of too many calories packaged into forms that are too delicious and ubiquitous should be viewed as a hedonic technology, then GLP-1s are an anti-hedonic response. They are not an injectable vomitorium, a cheat that allows the patient to indulge in gluttony without physical consequence. Instead, they dampen the appeal of food, making the irresistible resistible. An unhealthy diet, with all its consequences, is the default, a deck stacked against making rational and healthy choices, and weight loss drugs reshuffle it.

For his part, Kennedy fears that GLP-1s, in making food less appealing, will make everything up to and including being alive less appealing as well. Appearing on Greg Gutfeld’s show in October 2024, he said, “the EU right now is investigating Ozempic for suicidal ideation … because it suppresses all the reward pathways, so it makes you want to do everything less.” While it’s easy to follow the logic that an anti-hedonic drug powerful enough to blunt the appeal of a Big Mac might be powerful enough to blunt the appeal of continued existence, this obvious concern has been studied quite a bit. So far the data are reassuring, with the balance of evidence finding no increased incidence of suicide among people using weight loss drugs.

Another point of concern often raised by MAHA is the high cost of GLP-1s. It is true that no responsible government should agree to provide them to everyone who needs them at the current market price.

But several other potential issues, though less straightforward than death and taxes, also warrant consideration. In a society saturated with images of bodies that have been surgically and digitally altered to extremes, appetite-suppressing drugs could abet the obsessive pursuit of an impossible standard. Furthermore, GLP-1s have not been studied as a tool for weight management long enough for us to be sure of their efficacy five and ten years out.

There are medical concerns too. Some evidence has shown elevated resting heart rates in patients taking the drug. And ensuring that patients do not lose too much muscle presents a challenge.

The more difficult question, now that the use of weight loss drugs is widespread, is which direction our collective inertia will take us in when confronting the need to make difficult changes. Providing people with the tools necessary to resist the most harmful excesses encouraged by the food system is better than not providing those tools, but this compromise also highlights a failure to address the true cause of the issue. There’s a farcical quality to overengineering the most pleasurable qualities of food to the point that they become irresistible — and then inventing a drug to make that same food less appealing.

And yet for the over one hundred million Americans with their health compromised by obesity, GLP-1s offer immediate help that nothing else does. According to data from Gallup, the drugs already appear to be the main driver of U.S. obesity rates dropping the past three years in a row, after skyrocketing for decades. If the prospect of giving millions of Americans lifetime prescriptions as a way of combating a systemic problem bothers you, you are not alone. But while that future might not be ideal, it would be preferable to one in which obesity and the many serious medical conditions that accompany it remain the norm.

With the benefit of almost three decades of hindsight, we can say that the 1997 WHO report announcing an obesity epidemic was more insightful than we knew. In most rich countries across the globe, rates of obesity have been steadily climbing. Indeed, the absence of success in preventing this increase, despite immense variation from one country to the next in culture, economy, and form of government, points to just how challenging the problem is. But, despite the difficulty, the goal remains obvious: a food system that does not sicken the majority of the populace.

While Kennedy has diluted the once-potent message underlying the MAHA movement by focusing on tweaking junk food to hopefully make it slightly less bad, he has also done a few things that could contribute to more meaningful reform. He has spoken strongly about cleaning up school lunch programs. The government feeds about 30 million children, and it should be concerned with their health for both moral and pragmatic reasons. When it feeds them it should do so with good food.

Though they have not been released as of this writing, Kennedy has also promised to radically simplify the official governmental dietary guidelines from their current 160 pages to a much more manageable four, with an emphasis on eating whole foods. This is the document the general public and non-specialist doctors are supposed to read to know what a healthy diet looks like, and the way it has become a bloated, unusable mess exemplifies the feckless response of public health officials to a burgeoning epidemic.

Another guideline, the “Make Our Children Healthy Again” strategy report released by the government in September, laudably notes at the start that the American diet of highly processed foods is making us obese and sick. But noting the problem is not the same as addressing it, and so the report becomes yet another demonstration of Kennedy’s broader failure. In keeping with his pretense that small tweaks will make a big difference, the report’s response to the problem with processed foods is to encourage industry to continue removing additives and to arrive at a “U.S. government–wide definition for ‘Ultra-processed Food’ to support potential future research and policy activity.”

Kennedy should go much further. He should set about changing the parts of the government he has control over, and advocating for change in those parts he does not. For starters, here is a sketch of a few proposals for ambitious reform:

• Develop a sensible GLP-1 policy: There is nothing else, no plausible lifestyle intervention or regulatory reform, available to help the tens of millions of people with obesity manage the consequences of living in a broken food environment. The drugs should be prescribed to those who need them, and they should come with clear and careful guidance on how to mitigate their downsides. There is also the possibility that by inoculating the most susceptible members of the public to the pernicious appeal of processed foods, GLP-1s could contribute to broader reform efforts.
• Treat processed foods like cigarettes: Political realities make banning cigarettes unrealistic, and democratic principles argue that doing so would be excessively paternalistic, but that does not mean public health agencies should take an accommodating view of tobacco. Likewise, the USDA and the FDA should settle on the simple, consistent message that processed foods should be avoided and that a diet of whole foods, prepared in the home, is the ideal.
• Ban the advertising of food to children: A diet heavy in processed food often leads to a lifetime of illness. Increasingly, these illnesses — obesity, type 2 diabetes, even fatty liver disease, which was once seen almost exclusively in alcoholics — affect children. Food companies should be prevented from encouraging poor dietary habits in children.
• Improve school lunches: There are few opportunities for wholly benign social engineering, but better school lunches present such a chance. Nowhere else does the government directly feed so many people.
• Reform agricultural subsidies: Corn and soy are near-universal because they are some of the cheapest ways to produce a calorie (in corn) and fat and protein (in soy). But the government should not be in the business of promoting unhealthy diets: corn and soy, as well as other subsidized crops like sugar, are the foundation of most processed food.

There are three possible futures for the food environment in America. First, things could continue much as they have for the past thirty years, with obesity and related illnesses being the norm and weight loss drugs being limited by regulations and cost. Second, the food system could remain largely unchanged, but GLP-1s could significantly reduce the harm it does. And third, a less hedonic food environment could bring about real improvement in the baseline level of health.

No model exists for reaching this third and best future. The government cannot force it into being. So far the public has not voluntarily created it. But combine a clearer food policy that is made with no attempt to accommodate industry, a growing recognition that processed foods need not be the default, and drugs that let people reverse much of the damage caused by decades of poor diet, and it becomes a more realistic possibility than it has ever been before.

The post MAHA Cedes the Obesity War to Ozempic appeared first on The New Atlantis.

Reviewed in this article

Mother Media: Hot and Cool Parenting in the Twentieth Century

Hannah Zeavin 

M.I.T. ~ 2025 ~ 320 pp. 
$34.95 (hardcover)

Zeavin’s sweeping intellectual history shows how motherhood in America has always been mediated — by baby care manuals, radio shows, TV experts, and now digital tools. Concerns about “bad” mothers, over- or under-reliance on expert advice, or the loss of embodied maternal presence have surfaced with each of these. But while Zeavin resists the idea that today’s digital tools are unprecedented, her work helps us understand how today’s controversy over them connects to a long history of parenting fears.

Another book, The Tech Exit by tech policy expert Clare Morell of the Ethics and Public Policy Center, brings today’s controversies into sharp relief, offering a call to reclaim the moral and institutional authority of the family from the grip of Silicon Valley. She warns that Big Tech platforms are not neutral tools but active agents in reshaping childhood, undermining parents, and monetizing the denigration of family practices. From AI companions to personalized feeds that expose children to pornography and addictive content, Morell argues that today’s digital infrastructure is designed to fragment families, not support them. The Tech Exit urges parents to do just as its title says.

Taken together, Zeavin and Morell offer two essential perspectives — historical depth and political urgency — for understanding the digital restructuring of the family and pushing back with more than just willpower.

Reviewed in this article

The Tech Exit: A Practical Guide to Freeing Kids and Teens from Smartphones

Clare Morell 

Forum ~ 2025 ~ 256 pp. 
$27 (hardcover)

Digital platforms now mediate nearly every aspect of family life. Shared calendars coordinate schedules; location tracking apps monitor children’s movements; group chats substitute for dinner table conversations; apps manage chores, carpool logistics, and extracurriculars. Social media curates family memory and identity in public. The rhythms of domestic life, once governed by conversation, tradition, and presence, are increasingly routed through notifications, dashboards, and algorithmic prompts.

As Hannah Zeavin reminds us, this is not the first time caregiving has been reshaped by technology. In Mother Media, she shows that American motherhood has long been entwined with technologies — objects and media that claimed to supplement, improve, or even replace motherly care. In the early twentieth century, the figure of the “scientific mother” emerged: a caregiver expected to follow expert guidance, deliver care on a schedule, and measure outcomes with precision. She was surrounded by tools that embodied the expertise, from bottles marked by the ounce and feeding charts timed to the minute to parenting manuals and advice columns. Later, in the postwar period, this model gave way to Dr. Spock, who urged mothers to “trust yourself” even as he authored hundreds of pages of parenting instruction. The rise of attachment parenting idealized continuous, intuitive connection between mother and child, yet this “natural” approach was still shaped by media and tools: baby-carrying slings, co-sleeping guides, talk shows, and, later, online communities. Zeavin shows that even practices marketed as intimate or instinctual were often scaffolded by external systems of expertise.

Common technologies like pacifiers, cribs, timers, television, and computers were not just conveniences. Zeavin shows that they functioned as material extensions of expert systems, and each embodied the assumptions and anxieties of its cultural moment. For instance, Zenith Radio Corporation commissioned its “Radio Nurse,” the first baby monitor, after Charles Lindbergh’s twenty-month-old baby was kidnapped in his sleep in 1932. The baby monitor extended maternal vigilance through circuitry, while reinforcing ideals of constant, invisible attentiveness. The binky soothed not just the baby but the mother trying to meet prescribed sleep-training goals. The playpen offered containment in line with developmental safety norms. Parents (and people who talk about parenting) have long feared that caregiving, when routed through technology, risks becoming cold, distant, and diminished.

These technologies have provoked at once demand and unease. At the time that Sesame Street first aired, in 1969, concerns about children’s exposure to violent material on television had already prompted one U.S. senator to call upon the surgeon general to investigate TV as a public health crisis. Yet Sesame Street aimed to leverage TV’s perceived problems to do good, and it was “astoundingly well-researched,” writes Zeavin. As some experts became comfortable with the idea of introducing “good” TV into daily domestic life, a powerful wave of media theorists, following Marshall McLuhan and Neil Postman, mounted a more complete condemnation. By the 1980s, Postman argued that television was destroying childhood as a distinct life stage because, unlike print media, which required gradual learning and slowly-gained skills to access adult information, television instantly exposed children to the same content as adults without any developmental barriers.

Today’s social grappling over mediated care is not new. Digital media, Zeavin’s book shows, are the latest inheritors of this tradition.

Although Mother Media does not venture beyond the 1980s, today’s platforms are arguably somewhat distinct from earlier technologies of familial mediation. Five features distinguish today’s online platforms: peer production, datafication, digitalization, personalization, and commodification.

Peer production is the reality that a significant amount of the content that children engage with online is produced by other children, making it distinct from both books and television. Platforms capture that engagement — with peer-produced content, with peers, and with other content and other people — as valuable data.

Datafication transforms everyday family life into streams of quantifiable information. Apps and devices log sleep, feeding, screen use, and location, turning children’s development into metrics for monitoring and comparison. Zeavin argues convincingly that datafication began long before platforms, but it happened at a smaller scale and with different motives.

Digitalization means that family care becomes customizable and programmable. Platforms can be responsive to the child within the logic of the software. Sleep-training apps send algorithmic nudges, educational platforms gamify learning with points and badges, and smart-home assistants offer behavioral prompts or rewards.

Childrearing experts of the past could only have dreamed of such sophisticated behavioral modification. Unlike the mass media of earlier eras such as radio, parenting manuals, or educational television, today’s platforms are hyper-personalized. They deliver content and interventions that are dynamically tailored to each user’s behavior, preferences, and emotional profile. YouTube and TikTok feeds deliver micro-targeted parenting tips and child entertainment to elicit engagement or address inferred emotional states.

Commodification follows closely behind, as these tools operate within a larger marketplace. Platforms monetize attention, routines, and relationships — offering tiered features, promoting sponsored content, and selling behavioral data to third parties.

While earlier technologies like binkies, cribs, or television were embedded with expert concepts and guidance, today’s platforms reshape family life to fit commercial and computational ends.

Clare Morell’s The Tech Exit joins Zeavin’s long line of parenting literature warning of the latest forms of mediation. While Zeavin invites us to notice how caregiving has always been shaped by expert systems and tools, Morell calls for a reassertion of the family’s sovereignty over childrearing against Silicon Valley platforms that bypass, override, or erode parental authority.

Morell documents how Big Tech platforms systematically route around parental authority. She details stories where young children downloaded apps like TikTok without parental consent and were then served dangerous content like the blackout challenge, in some cases resulting in death. She describes the rise of sextortion, a practice targeting mainly teenage boys that involves someone, usually a stranger pretending to be a peer, threatening to distribute private and sensitive material if they don’t provide images of a sexual nature, sexual favors, or money. She notes the rise of suicide associated with social media. The average age of pornography exposure is 12. Pornography is easy enough to access on purpose, but more than half of kids have been exposed to it accidentally through social media, ads, or search results. Morell argues that parental controls don’t work — one kid used the parental monitoring app on his phone to access pornography.

Beyond these disturbing stories, Morell describes the regretful gaps and distance created once personal devices are introduced into households: silent dinners, isolated siblings, social media–addicted kids of all ages. The default settings of Silicon Valley platforms assume that parents are irrelevant or intrusive. Parents are not failing, Morell argues. They are being overwhelmed, duped, and out-designed.

The last part of her book provides parents with ways to rebuild and hold the boundaries against the immense and constant allure of digital conveniences. Those who are successful find other families, model healthy tech use, adopt alternative devices, move from allowing children to use tech in private to creating common spaces for it, and prioritize real-life responsibilities and fun. She emphasizes shared technologies like a desktop in a common room and family movie night. She wants to get kids excited about replacing screens by swapping in new skills, indulged hobbies, time with parents, and service.

For Morell, the solution is not cultural analysis but moral fortitude and legal intervention. She supports age verification and design mandates that by default would require parental oversight of app downloads for minors. She has praised recent state laws seeking to restrict data collection on children, require default protections, and restore the power of parents to make decisions about digital exposure to adult content.

While Zeavin’s Mother Media shows that anxieties over maternal mediation are nothing new, that history should not be taken as a reason to dismiss Morell’s concerns. Battles over how families should care, and what technologies can assist or govern that care, have often been productive. Concerns about the influence of television on children led to the Children’s Television Act of 1990, which introduced requirements for educational programming and limits on advertising. Earlier struggles over the well-being of children led to child labor laws, public schooling, and the development of pediatric medicine. Rather than prove that concern is overblown, as many skeptics still believe, this history shows examples of how debate and discomfort have helped build the institutional and technical infrastructures that now protect, educate, and support families. Today’s debates over digital platforms should be understood as part of that lineage.

Both Zeavin and Morell push us to recognize that settling on screen-time limits — restricting access to personal devices in hopes of promoting healthy development — as the strategy for protecting kids concedes too much. It is an attempt to address the problem Big Tech has wrought without requiring any fundamental changes to the technology itself or the culture it creates. What we need is not just unplugging but reimagining. And here, the combination of ideas from Zeavin and Morell becomes crucial.

Morell calls on families to resist Big Tech, and for regulations to make that resistance easier as well as less necessary. She recognizes that families may need to find the strength to live counterculturally, but also argues that the burden cannot fall solely on individual families. She supports structural reforms that restore the moral authority of parents through meaningful parental consent, that constrain the extractive logic of digital platforms by limiting data collection and addictive features for children, and that re-establish social expectations around formative relationships through strong age-verification. While her toolkit includes specific regulations, Morell also knows that strong families need strong communities.

Importantly, Zeavin does not double down on the family as the exclusive institution of care for children. Instead, she invites us to look outward to the social, economic, and technological structures that shape how childcare is given and received. She subtly frames many contemporary anxieties about digital parenting technologies as moral panics and appears skeptical of critiques that blame technology alone for the breakdown of the family, often redirecting focus to structural issues like gender roles, labor inequality, and economic precarity. Her vision does not emphasize stronger families in the traditional sense but care that is collectively supported: by extended kin, by local communities, and by public infrastructure. She warns against debate that revolves around extreme nostalgia or purity.

Together, the two authors provide the motivation to make significant social and political investments in changing technical infrastructure in ways that support the future of the American family. If we are to resist the platformization of the family, we must ask what new and old functions serve the family and how technical infrastructures may provide those functions. What tools and arrangements serve intimacy, incremental independence, tactile exploration, and intergenerational understanding? And which lead to distraction, addiction, and isolation?

There are some signs of this sort of reimagining. France recently passed a nationwide ban on screen exposure for children under three in nurseries and public settings, an institutional signal that screens in early childhood can no longer be left to parental choice alone. In the U.S., the Supreme Court this summer allowed Texas to enforce a law requiring pornography websites to verify the age of users, marking a significant moment in the country’s efforts to impose barriers around adult content online. Australia is going further. Under its Online Safety Act, the country’s eSafety Commissioner has drafted rules that require a wide range of digital services like search engines, social media platforms, and streaming services to implement rigorous age-verification mechanisms before allowing access to minors. Together, these initiatives indicate a growing willingness among policymakers to reshape the legal and technological architecture surrounding family life.

Meanwhile, new coalitions of educators, technologists, and parents are imagining what it means to build pro-child, pro-family technologies from the ground up — not merely to reduce risk, but to cultivate presence, attention, and relational depth. In Washington, D.C., Hardy Middle School emerged as one of the first phone-free public schools and was featured on the Today Show for improving student engagement and focus. Nearby, St. Jerome Academy in Hyattsville, Maryland created a low-tech, classical, Catholic curriculum that has transformed a school on the cusp of closure into one bursting at the seams. At home, some families are reviving landlines as shared communication hubs and experimenting with DIY text-message walkie-talkies. The device market has exploded with non-addictive options: minimalist phones, e-ink tablets, and wearables designed for safety without scrolling. Teens are forming Luddite Clubs and opting for “boring phones” as a badge of independence. Churches and community groups are organizing unplugged retreats and Sabbath-style digital fasts. This constellation of experiments shows that concern is giving way to creativity. These pioneers are not simply resisting screens, but restructuring how time, attention, and care are mediated in the first place.

We have imagined a world where robots read bedtime stories, where smart homes soothe and surveil, where children’s emotions are data points and parental instincts are inefficiencies. Hannah Zeavin and Clare Morell tell us that it’s time to imagine something else.

The post So You’ve Decided to Save Your Kids from Tech appeared first on The New Atlantis.

So it’s easy enough to believe that Trump and his supporters are irrational and out of touch with reality. But another interpretation is that the nature of truth itself is complicated in ways that the president has benefited from and his political opponents have failed to understand.

Really? Isn’t truth fixed by reality itself, untouched by our politics? Science establishes those truths and, as Galileo explained, “the judgment of man has nothing to do with them.” Truths would be true even if people didn’t exist.

During the current political chaos, the idea that accepted notions of truth are up for grabs may be difficult to consider dispassionately, especially for those who are shocked and frightened by Trump’s return to power.  To put today’s battles over science and truth into perspective, a historical example, from a differently troubled time, may help. We can start with a pure truth of the Galilean sort, the one equation almost everyone knows: E=mc². Einstein’s famous formula, which says that small amounts of mass can be converted into enormous amounts of energy, described the truth that allowed scientists and engineers to build the two atomic bombs the United States dropped on Japan in August 1945 to end World War II.

By 1958, eight thousand or so more such weapons had been built by the Americans and the Soviets, sufficient to immolate us all. In February of that year, two illustrious scientists, Linus Pauling and Edward Teller, debated on national television whether testing of these weapons should continue. Pauling was a Nobel-winning chemist and ardent pacifist. He had already gotten nine thousand scientists to sign his petition to end all nuclear weapons tests. In arguing for a test ban as a step toward international disarmament, Pauling cited scientific evidence that the radioactive fallout from a single hydrogen bomb test would lead to birth defects in 15,000 children worldwide. Teller, Hungarian-born and a renowned physicist, had fled Nazi Germany in 1933, and contributed to the U.S. development of the hydrogen bomb. In his judgment, the dangers to human health from atmospheric testing of nuclear weapons had “not been proved, to the best of my knowledge, by any kind of decent and clear statistics.” Drawing on the lessons of World War II, he argued that countering the threat of the Soviet Union demanded an active U.S. weapons testing program.

As Melinda Gormley and Melissae Fellet observed in a 2015 article about the debate, neither scientist tried to portray himself as objective about the matter; neither claimed that the scientific facts as he understood them were independent from his beliefs about how best to pursue peace. And so the debate’s moderator summed things up in a way that today might seem strange: “It is apparent that the issue has not been resolved. I’m sure both our guests would agree that its ultimate solution rests in our hands … that each of us bears the moral obligation to examine the evidence, draw conclusions from this evidence, and act upon our convictions.” Disagreement between scientific experts about how to deal with the risks created by nuclear weapons would have to be resolved democratically.

In the decades following the Pauling–Teller debate, scientific research aimed at informing public policy expanded enormously to help meet the risks and challenges of a rapidly modernizing world. This was a new task for science. Billions of dollars funded thousands of scientists to study questions as disparate as: How safe is nuclear energy? Which pesticides and food additives cause cancer? What are the causes and impacts of acid rain? When should women start to get mammograms? Are genetically modified foods necessary to feed the world? What are the economic benefits of environmental protection? Does standardized testing improve educational outcomes?

Research programs motivated by such questions were supposed to reduce uncertainties to arrive at the truth of the matter. Agreement on actions to solve the problems was supposed to follow. But in very few cases did this happen. Calculations of, say, the risk of a nuclear reactor meltdown, or a low-dose toxic chemical exposure, or a future pandemic required that scientists make heroic assumptions about incredibly complex phenomena — a direct invitation to contradiction by calculations made by other scientists making different assumptions. Uncertainties around these and similar questions did not get reduced; they expanded. And disagreement about what to do persisted, and often got worse.

Separating science from politics under these conditions was impossible. In 1983, the Nobel-winning social scientist Herbert Simon assessed the situation and came to the same conclusion as the moderator of the Pauling–Teller debate: “When an issue becomes highly controversial,” Simon wrote in Reason in Human Affairs, “we find that there are experts for the affirmative and experts for the negative. We cannot settle such issues by turning them over to particular groups of experts. At best, we may convert the controversy into an adversary proceeding in which we, the laymen, listen to the experts but have to judge between them.”

It was sound advice. “Adversary proceedings” are at the heart of democratic institutions, and since the 1960s science’s role in them has been expanding. At Congressional hearings, Democrats often call on one set of scientists and Republicans on another to provide competing facts and support contradictory policy preferences. A main task of the legal system is to adjudicate between competing claims about what is true in the world, and in fact many of the intractable debates that would seem to hinge on scientific evidence — from choosing a nuclear waste site to determining if a chemical really causes cancer in humans — have been argued in courts. Meanwhile, responsible media coverage about such conflicts includes quotes and facts from experts pro and con, so anyone can decide which evidence and experts comport best with their own values and beliefs.

Politically controversial issues are never settled by Galilean, E=mc² truths. Instead, truths are continually negotiated through the institutions of democracy. Scientists — “experts for the affirmative and experts for the negative” — are participants in the negotiation process, but often it is judges, juries, lawyers, elected officials, bureaucrats, advocates, journalists, and voters who determine what counts as truth.

Through the second half of the twentieth century, as this interweaving of science and democracy expanded, public confidence in science and scientists remained very strong. The American people and their politicians seemed satisfied that the institutions of science and democracy were working together reasonably well. Presidents of both parties oversaw and supported increasing federal investments in science.

Starting in the early 1990s, concerns about climate change began to shift the partisan landscape for science, and for truth.

Congress began addressing the risks of climate change in its usual way, pouring billions of dollars each year into scientific research with the promise that uncertainties would be reduced and agreement on action would result. And most experts agreed about the basic science linking greenhouse gas emissions to slow, albeit uneven, increases in atmospheric temperature during the twentieth century.

But agreement on the fundamental science did not determine what to do about climate change any more than E=mc² had determined what to do about nuclear arms. Policymakers would need to know how quickly climate change would unfold, how severe the consequences would be, how those consequences would be distributed, how much it would cost to reduce them, and what were the best alternatives for doing so. These were the same sorts of open-ended, only half-scientific questions that had been fueling political disagreement around other types of risks for fifty years.

In the 1990s, Democrats aligned themselves with a policy agenda that called for U.N.-led governance of greenhouse gas emissions, artificial markets to set a price on these emissions, and regulations and incentives to stimulate public choices aimed at reducing them. Here was an agenda that Republicans were bound to hate. So of course they were skeptical about the science that, despite all the uncertainties, supposedly dictated what had to be done. Research on public attitudes about climate change showed that one’s level of concern about the problem was largely a reflection of one’s ideological views — not one’s level of science literacy. (If that seems surprising, just remember Pauling and Teller.)

Driven especially by Vice President and 2000 presidential candidate Al Gore’s concerns about climate change, Democrats at the highest level began branding themselves as the party of science, rationality, and truth. Science had become a wedge issue. When Gore lost to George W. Bush in 2000, attacking Bush for being anti-science immediately became a staple of Democratic opposition strategy. When John Kerry ran against Bush in 2004, he promised, “I will listen to the advice of our scientists so I can make the best decisions…. This is your future, and I will let science guide us, not ideology.” He lost, too. After Barack Obama won in 2008, he proclaimed in his inauguration speech that he would “restore science to its rightful place” in American society.

The scientific community, to the extent such a thing exists, embraced the alliance with the Democrats. Shortly after Obama took office, the weekly editorial in Science magazine declared: “The Enlightenment Returns.” In 2012, a letter from 68 Nobel-winning scientists supporting Obama’s re-election stated — almost certainly incorrectly — that his Republican opponent, Mitt Romney, would “devastate a long tradition of support for public research and investment in science.” The letter said nothing about the political affiliation of the signatories, as if that were an irrelevance, but, writing for Nature magazine at the time, I found that of 43 who had made political donations, only five had ever contributed to a Republican candidate. In 2014, the American Association for the Advancement of Science appointed as its director a former Democratic member of Congress with a Ph.D. in physics. As I wrote at the time, “The more the AAAS, and so the science community, is seen to line up behind one party, the less claim it will have to special status in informing difficult political and social decisions.” In 2014, a PAC was created to support Democratic candidates who were also scientists. American science increasingly looked like a Democratic interest group.

The stage was set for what happened during Covid. As the pandemic began to spread, the best ways to protect people from the disease were highly uncertain. Choices and beliefs about wearing masks, social distancing, school and business closures, and even appropriate medical interventions mapped right onto pre-existing political fissures in American society about freedom versus safety, and about the authority of government and of science itself to dictate our behavior. Leaders of the mainstream scientific community, rather than owning up to both the uncertainties and the value judgments behind policy choices, tried to invoke their status as purveyors of Galilean truths to explain why, say, six feet was the right amount of social distancing, or why school closings needed to persist.

Large swaths of the country, including many scientists and doctors, were unconvinced, but dissent was portrayed by Democratic and scientific leaders as politically motivated and anti-scientific. President Trump exploited these conditions to sow further division and distrust across society. The new Covid vaccine had its Galilean truth moment and about two-thirds of Americans chose to get at least two vaccinations — but as the pandemic began to wane, and vaccine efficacy along with it, politics infected that domain as well.

Seventy years of growing entanglement between science and politics show that the truths that matter most in democratic decision-making emerge from the political arena, not the laboratory. When Democrats sell themselves as the party of science, truth, and rationality, what they are really saying is that if you are rational and believe in science and truth then obviously you will support Democratic policies. But given that half the country does not ideologically align itself with Democrats, this is a hard case to make, and those who disagree with Democratic agendas may in turn wonder why they should accept the science those agendas bring with them. Such skepticism may not be irrational, or even anti-science. Yet it has created a space not just for Trump’s endless effusion of lies, but for the attacks he and his minions are mounting on mainstream scientific knowledge and on the scientific institutions that Democrats had labeled as their own.

If the Democratic Party wants to convince more voters that it is the party of what is true, it will first have to convince them that it is the party of what is good, what matters, and what is right.

The post The Party of Science Is Over appeared first on The New Atlantis.

By the time Marcy started hearing about Ozempic, she was, to put it mildly, a skeptic. And not just of the drugs but of the whole scheme as she saw it: take the drugs or follow the diet, lose weight; go off the drugs or the diet, gain it all back.

Marcy is now the co-owner of a plus-size clothing store in Los Angeles. She’s come a long way in how she thinks about fatness. She eats well and is physically active, and her doctor has not flagged any problems like diabetes. And yet recently, when she was in charge of snack time at a children’s theater, she snapped at an overweight little boy who told her he was “starving” right after eating.

“Great, you already had a snack,” she said — and immediately regretted it. She wished that instead of assuming the kid was just going to overeat again, she had found him something that would satisfy his hunger.

Marcy isn’t the only person wondering how to help children who are overweight. In 2022, the FDA approved semaglutide, the drug sold as Wegovy for weight loss, for use in children as young as age twelve. Wegovy and its related class of medications, including Mounjaro, Zepbound, and Ozempic, are far from the first weight loss drugs on the market, but they are touted as the most effective.

Since then, President Trump has appointed Robert F. Kennedy, Jr. as the Secretary of the Department of Health and Human Services, and, informally, as leader of the Make America Healthy Again movement. MAHA claims that chronic diseases, including obesity, are best treated not with pharmaceuticals but through nutritious food.

The stereotypical member of the MAHA movement is a tech skeptic: a “crunchy con” mom who turns off her Wi-Fi at night to protect herself from electromagnetic radiation and drinks raw milk because she doesn’t trust the FDA on the benefits of pasteurization. She washes her clothes with Castile soap (too many lab-generated synthetic chemicals in Tide) and was a regular at the local farmers’ market before it was cool. She has bought into the idea of “organic” as not just a produce preference but a purer way of life.

But Kennedy plans to do much more than nag Americans about eating their veggies; he wants to ensure that the latest health tech is drafted into his movement. His goal is that within four years every American has a wearable health device that tracks metrics such as blood glucose and daily step count. “We’re about to launch one of the biggest advertising campaigns in HHS history to encourage Americans to use wearables,” he said in a Congressional hearing in June. “Wearables are a key to the MAHA agenda.”

It may be surprising at first that Kennedy is joining the techie side of America’s booming wellness industry. But in fact he’s not the only one: from Bryan Johnson’s attempts to follow the algorithm to immortality to Kennedy advisor and Surgeon General nominee Casey Means’s advocacy for clean eating and glucose monitoring to fix metabolic dysfunction, anti-establishment wellness gurus are increasingly pushing individualized health programs that present new technologies as magic bullets for Americans’ health problems. So while MAHA tends to bill itself as the antithesis to Big Pharma, recent trends under Kennedy’s HHS are turning the movement into pharma’s biggest competitor in the game of selling high-tech biohacks that promise to reverse America’s rising weight trends.

In a country where 20 percent of Americans under twenty and 40 percent over that age are obese, and in an economy where treatment for chronic diseases related to obesity costs hundreds of billions of dollars per year, finding and fighting the causes of obesity is one of the most important and high-stakes public policy debates of our time. The fact that for decades all attempts at reversing rising obesity levels have failed only makes the problem more pressing. And the fact that people are crossing the threshold for obesity ever earlier in their lives makes this one of those unavoidably anxious debates about posterity. Research suggests that genetics and early habits determine weight to a degree that makes it difficult to ever mitigate with later lifestyle changes. Today’s children are the first to grow up with both Wegovy and wearables, so they will be the final judges in the fight over which treatments actually work.

Childhood Obesity Prevalence in the U.S., Ages 2–19 

The two camps in the debate are divided on the answer to an old question. Which is more responsible for the problem: genetics or upbringing? Nature or nurture? If you say nature, then your solution is likely weight loss drugs. You believe that kids who are genetically predisposed to become obese will likely be and stay that way, and the new class of drugs could be a miracle in not just treatment but prevention. If you say nurture, then you may look to the MAHA movement not only to fix the system that is making Americans addicted to unhealthy food but also, increasingly, to empower them with technology that allows them to make better choices.

Though the chasm dividing these sides is vast, if you peer deep into it, at the bottom you may see common ground. Both are actually marked by forms of techno-optimism. Either drugs are going to save the kids — or “bio-optimization” and revolutionary organic farming practices will. Whatever the case, all we need is one weird trick to be healthy again.

If money could buy a solution to the problem of childhood obesity, it would be on display at the annual conference of the Obesity Medicine Association. This April, the conference took place at the expansive Gaylord National Resort in National Harbor, Maryland. National Harbor is all parking garages. The streets around the resort are filled with the kind of faceless franchises that exist in large part to kill the time of the tens of thousands of conference attendees spending down their company’s expense accounts every year: Nando’s, Ben & Jerry’s, McCormick & Schmick’s. The giant courtyard of the hotel has an entire indoor village with more ersatz establishments. A heavyset teenage girl, a middle-aged woman, and an older man, all positively glowing with health, smile down from a banner at least ten yards wide next to the registration desk. The banner’s label blares, “Wegovy.” There is no secret about who is backing this event.

In the corridor leading to the meeting rooms, more pleasantly plump people gaze reassuringly at me from standing signs. One sign lists the “5 Principles of Obesity”:

1. It is undeniable that obesity is a complex, chronic disease.
2. Obesity is driven by powerful underlying biology, not by choice.
3. The many health effects of excess weight can start early.
4. Obesity is treatable.
5. Weight bias, stigma, and discrimination are harmful.

I note the polemical edge in that word “undeniable.” The question of how best to treat obesity is ostensibly under debate here, but the overall logic of the principles — if disease, then treatment — paired with the pharmaceutical advertising next to it seems to make the answer a foregone conclusion.

When I enter the first meeting room, I’m struck by the sheer size of the audience for this message. Attendance gives medical professionals a Continuing Medical Education credit for certification in treating obesity in their practices, and they have come in droves with their Yeti water bottles and North Face sweaters. Many are primary care physicians, nurse practitioners, and the like who are responsible for advising patients on weight loss. Much of the information shared here will end up being used in doctor’s offices across the country.

At the first session of the morning, Dr. Harold Bays, the president of a medical research institute in Louisville, Kentucky, discloses that he has consulted or contracted research for Eli Lilly (the maker of Mounjaro and Zepbound) and Novo Nordisk (Ozempic and Wegovy), along with Amgen, Pfizer, and a whole alphabet’s worth of others. Previous weight loss medications have had such bad side effects that they’ve been taken off the market, he says, but this new class of drugs, called GLP-1 agonists for the human satiety hormone they mimic, is a game-changer. They’re both more effective and less taxing on patients’ bodies than previous treatments.

Bays is optimistic about not just the current medications but the overall new approach to treating obesity. Other metabolic diseases, such as hypertension and diabetes, also once resisted being treated with medication. Now, medication is a standard part of what he calls “comprehensive” care for those conditions: lifestyle changes and dieting along with drugs. Before too long, he says, the same package will gain social acceptance as the standard of care for obese patients young and old.

GLP-1 agonists are the first in a wave of drugs that are increasingly capable of being personalized to do more than one thing: reduce appetite and fight nausea, for example, or preserve muscle tone during rapid weight loss. Bays fires off a barrage of these drugs: dual agonists, tri-agonists, muscle-acting agents. Trials are ongoing and long-term effects are unknown, but “we want the drugs now,” he says, because the results are so life-changing for patients. So he advises doctors to keep up with the news about the latest approvals and add them to the suite of medications they prescribe.

Despite all the talk of lifestyle modifications that should accompany weight loss drugs, doctors are often better informed about the drugs than the new lifestyles they are supposed to support. “A lot of physicians do not feel adequately trained to talk to their patients about nutrition,” says Dr. Nate Wood, a chef and the director of culinary medicine at Yale, in the second session I attend. In 1985, the National Academy of Sciences issued a report that recommended twenty-five hours of nutrition education for medical students, he says. Today, they get about eleven.

Wood offers some common-sense guidelines for patients both on and off weight loss drugs: focus on hitting protein targets, but don’t forget about that sleeper nutrient, fiber. Ninety-five percent of Americans aren’t eating enough fiber, he says, and patients on GLP-1 drugs are at risk of constipation because the drugs slow down their digestive systems. Most importantly, make sure patients are eating enough. He tells the story of one patient who gave a glowing review of her progress to one of Wood’s colleagues: “‘It is great,’ she said. ‘I’m not hungry all day. Yesterday, all I had was a small bag of popcorn.’ And you’re like, oh, gosh, that’s concerning.” The audience laughed sympathetically.

He advised setting a target of at least 1,200 calories a day to keep patients from dropping weight too fast and missing out on vital micronutrients. Suddenly it sounded like the drug was the disease and the food was the prescribed medication: doctors should watch dosage, symptoms, and the overall sustainability of the regimen. A treatment prescribed in large part to fix problems with food intake has actually created new problems with food intake.

And while it’s clear that Wood treats mostly adults, the stakes for children’s eating patterns when they’re taking GLP-1 drugs seem even higher. Children’s diets, more than those of adults, are made up of a majority of ultra-processed, low-fiber foods — some studies estimate up to two-thirds of what they eat. What happens if a child starts taking the drug and is now eating less, but not necessarily better? That could mean even fewer of the nutrients needed to sustain healthy growth, puberty, fertility, and so on. There are no long-term studies to either prove or disprove this, but sometimes the most sobering answer to hear from a doctor is “I don’t know.” And that’s often the answer when the question is “What about the kids?”

On the second day of the conference, I have no child care. So I strap my newborn to my chest and toss some crayons in my three-year-old’s backpack and show up to Dr. Claudia Fox’s session on “Empowering Comprehensive Obesity Care from Youth to Adulthood” with two daughters in tow.

Fox is an associate professor in the University of Minnesota’s department of pediatrics and the co-director of its Center for Pediatric Obesity Medicine. She paints a grim picture of childhood obesity treatment. In the past, she says, doctors followed a “wait and see” approach, monitoring children whose weight was high for their age and waiting to see whether it would resolve after puberty, when girls especially gain weight as their bodies prepare for childbearing. What they saw, though, was that “obesity in adolescents almost universally persists to adulthood.” In fact, the problem was evident even before puberty: “Ninety percent of three-year-olds with obesity will continue to have overweight or obesity in adolescence.”

According to Fox, doctors have been fooling themselves into thinking that obesity is more about nurture than nature: “Obesity is essentially baked into our DNA really, really early on.” And this has led to the lie that “lifestyle modification” can work for children, and has also fueled the shame that comes with children’s failure to lose weight: “I very deliberately say, when I’m seeing a new family, this is not your fault. This is not a willpower problem. This isn’t a parenting problem. This is a biological condition that’s primarily determined by our genetics.”

What to do, then? The American Academy of Pediatrics recommends that in addition to “lifestyle therapy,” “pediatric health care providers should offer obesity medications to children who are age twelve and older,” she says. “Notice that word ‘should,’” she adds. “They didn’t say you ‘may,’ or that you ‘could’ do it. No, you should do it.”

Fox says she’s aggressive about prescribing medication because she cares about curing kids. Think about cancer. “We would never be satisfied with just the stability of their leukemia,” she says. “We want them to be cured. We want resolution of their disease.”

“What is our goal for treatment?” she asks. “I think it’s still hard to say. With regards to treatment, our foundation, at least at this point, is lifestyle modification therapy.” But therapy is hard: research suggests that it takes at least twenty-six hours of lifestyle therapy — think food logs and tracking cardio — over the course of two months to a year to achieve significant weight loss. Vanishingly few patients manage this, or even have the potential to, given how little insurance will usually cover for nutrition therapy. That’s why “early intervention with the most intensive therapy available” — meaning GLP-1 drugs like Wegovy — “is really critical,” she says.

I listen in mounting dismay and then trundle my daughters out the door in search of a snack. We’re the lucky ones, too: my three-year-old’s favorite food is pizza, but her dad makes it for her from scratch. Her mom works from home and spends a significant amount of time persuading her to try spinach, cherry tomatoes, lychee, and tofu. And neither of her parents is overweight. Simply because of these early advantages, she’ll likely never face the problems that, according to Fox, many of her patients will struggle with all their lives.

The speakers at the OMA conference don’t misdiagnose the many challenges of obesity and the pressing concerns that arise as it presents itself earlier in people’s lives. Many of them are just surprisingly certain about its solution, given the dubious track record of past weight loss drugs in reversing rising levels of obesity. The exhibition hall reveals one source of that confidence: Novo Nordisk itself is here, and Eli Lilly, along with innumerable coattail-riders competing to sell their supplements in the attendees’ health care practices. And to do that they’ve brought bribes: boxes overflowing with sample shakes and probiotic drinks, free tote bags, and (deliciously) lollipops at vendor stalls selling healthy solutions for patients on weight loss drugs.

I let my three-year-old try a granola bar formulated for GLP-1 drug patients and then feel weird about it, less because I think it’s bad for her than because it looks like the kind of failing diet product I often see on grocery store clearance racks: highly processed, covered in questionable marketing claims, and likely to taste kind of gross.

After a presentation on “complex” GLP-1 treatment cases, I’m approached by a thirty-something woman in a bright pink pantsuit and kitten heels and a lanyard whose rainbow of ribbons displays a long association with the Obesity Medicine Association. No, she doesn’t do media anymore, she says, but she gives me her number and implores me to allow her to connect me with her babysitter. She adds as we part for the morning break that she hopes I’ll report fairly and acknowledge that obesity is, indeed, a disease — in case I’d missed the party line.

Even if that’s true, Ozempic may be less of a revolution than its promoters say it is. Who can afford it, in literal dollars or in years spent on the drug, possibly to end up right where one started if one ever quits? After all, an Oxford University analysis of eleven studies found that patients who stop taking GLP-1 drugs typically regain the weight within a year.

The fact that there’s not yet long-term data on the effects of these drugs for weight loss offers hope for the children starting to take them now. But it also means there’s no reason to assume they will work. Adolescents have ahead of them a whole life of navigating temptations, social pressures, genetic predispositions, economic obstacles, and all the other challenges that come with having to take care of one’s body in the long run. It’s how they manage those pressures that will determine much of the results of the Ozempic experiment.

Not long later, as I leave the lunch buffet, I’m approached by an employee from the resort. I’m holding two plates of chicken salad and two glasses of water and pushing a stroller with my hip while directing my older daughter to a table.

You need to leave, the woman says. Don’t you know that this event is not for children?

If you’re the kind of person who generally avoids medical conferences and big-box hospitals, you might find yourself attracted to an approach to obesity that, on its face, is the very opposite from that advocated by the Obesity Medicine Association.

The story Casey Means tells about her life is that of a fat kid gone right. As we read in her book Good Energy, she was born with a predisposition for weight problems because of her mother’s undiagnosed gestational diabetes and related health conditions, then packed on pounds during puberty. According to the “nature” narrative, she was destined to be overweight her whole life. But instead of further abusing her body with dieting or disordered eating, she turned herself into a health nut: she studied nutrition, taught herself to cook, joined a gym, and worked it all off. This knowledge turned out to be more useful than the curriculum at Stanford Medical School, where she became disillusioned with the health care system’s specialist, drugs-forward approach to treatment of disease.

She came to see an increasing number of ailments in her patients as not just related to but rooted in bad eating habits. Those habits were based on people’s ignorance about how they should fuel their bodies — and since each body is different, each person needs an individualized health plan, not a one-size-fits-most prescription. She found herself coming around to the “nurture” model regarding chronic diseases and obesity: our habits determine our health far more than our genes do.

She dropped out of her surgical residency program and co-founded a health company named Levels, which sells continuous glucose monitors (CGMs) and offers an app that tracks people’s blood sugar and thus helps them optimize their eating, exercise, and sleep. Last year, she published Good Energy: The Surprising Connection Between Metabolism and Limitless Health, which argues that malfunctioning mitochondria (the powerhouse of the cell, as you no doubt remember from middle school) are the cause of the majority of chronic diseases, including those related to obesity. The solution? Bio-optimization: eating organic and gluten-free to avoid toxins, strategically planning meals and sleep to minimize stress, and tracking it all with a CGM.

“We are entering the era of bio-observability: more readily available blood tests and real-time sensors filtered through AI analysis to give us a highly personalized understanding of our bodies and a personalized plan to meet each body’s needs with daily choices,” she writes.

It’s easy to see how this pitch is attractive to individualistic Americans, especially those who are already skeptical of the medical establishment. “You should not blindly trust your doctor and you should not blindly trust me,” Means writes. “You should trust your own body.”

The book was a No. 1 New York Times bestseller. It popped up in chiropractic offices and other alternative health practices around the country. Means and her brother, Calley Means, became top advisors to Robert F. Kennedy, Jr. and his Make America Healthy Again movement. And in May of this year, Means was nominated to be Surgeon General of the United States.

These are no longer the mere musings of an Instagram influencer and alt-health entrepreneur: the idea that “food is our most potent weapon against chronic disease” and that tech will help heal us is a serious competitor to the Wegovy weight loss model. In fact, this techno-optimist narrative is much bigger than Means — and much less mad than the prospect of going to bed at eight thirty and injecting one’s children’s blood to live forever, à la Bryan Johnson. The wellness industry — health products and regimens sold outside the typical health care pipeline — is larger than the pharmaceutical industry: an estimated $6.3 trillion per year compared to $1.6 trillion. It’s expected to reach $9 trillion in the next three years. Social media–mediated wellness and app-based health make up the water the next generation of health-seekers will swim in.

Much of what Casey Means proposes to do to fix our food system makes sense, especially for addressing the problems many Americans see affecting children’s diets. Her ideas: remove ultra-processed foods from school meal programs, ban food dyes that have been linked to behavioral problems, and redistribute Farm Bill subsidies from corn, soy, and tobacco to support fruits and vegetables, ideally organic.

Many critics have focused on her anti-pharma positions and speculated about the implications for vaccines and drugs like GLP-1 agonists. But beneath the crunchy-conservative advice about how to get kids to eat organic lies a techno-optimism strikingly like that of the pediatrician prescribing semaglutide to sixth-graders. “What if we treated humans like rockets, equipping them with sensors before systems fail, to understand where dysfunction is arising so we can address it?” Means writes in Good Energy. The metaphor is borrowed from Josh Clemente, her co-founder at Levels, who is an alumnus of SpaceX. Reducing food to its component parts, and eating habits to an “individualized” health plan, makes it easy to turn the human body into a machine.

Yet Means sees her solutions for chronic disease as not mechanistic but spiritual, in contrast to those of the soulless medical establishment. “Everything is connected,” she writes in her book. And the “spiritual crisis” of America’s chronic disease epidemic is “an assault on the miraculous flow of cosmic energy from the sun, through the soil and plants, through bacteria in my gut, through my cells’ mitochondria to create the energy that sparks my consciousness.”

This is why individualized tech is so much more promising for her than the old wait-and-drug medical model, which waits until people get fat and sick and then charges them to treat their symptoms. “We have the potential to live the longest, healthiest lives in human history, but this will require optimization,” she writes.

And it has to start young. The “natural” alternative to the medical establishment’s drug-first model turns out to draw on the same assumptions as not just Wegovy but the failed weight loss cures that came before it: your body is a problem, and there’s a product that can solve it if you start early enough and work at it hard enough. Bringing parenting into the equation shows how this pitch can be less freeing than it seems. After all, the earlier you optimize, the better off you and your children will be. Mothers should start optimizing their children in utero, since everything Mom eats also feeds baby.

It’s hard to give up the idea that weight loss is a matter of personal responsibility: of finding the right hack to fix it, with a drug or a gadget or with pure grit. And it’s easy, when one tries to give up that idea, to fill the void with a dream that ultimately follows from the same assumptions. Means’s story, then, is not that of a fat kid who found the true fitness gospel. Or if it is, it’s the same old story: she is one of the few lucky people who found salvation through weight loss and now has the difficult task of trying to share it. But solutions tend naturally toward the one-size-fits-all, while crises tend, unfortunately, toward complexity.

The danger of techno-optimist solutions is that they treat the same thorny problem — stubbornly rising weight at younger and younger ages — with the same old solution: now there’s a drug, or an app, for that. The problem is not that these solutions are bad, or bound to fail. It’s that they’re utopian. What they are both selling is the idea of a simple fix for all the problems with food and health that contribute to obesity. This is a recipe for disappointment.

Many patients come to Dr. Siham Accacha looking for the simple fix they’ve seen in all the ads recently: they want to be prescribed a GLP-1 drug. Like the doctors at the conference, Accacha, a pediatric endocrinologist at Cohen Children’s Medical Center in New York, is willing to prescribe it. “I try to present it to them as a help, as an extra motivation,” she says.

But she warns them that a lifetime on medication, despite what the Ozempic ads suggest, is likely not realistic or desirable. “They have the opportunity,” she says, “to actually go to the next step, to lose weight, but also to learn how to eat healthier because it decreases their appetite significantly.”

Accacha tells me that her experiences with patients show that the causes of obesity are not as simple as nature or nurture alone but stem from something like culture: the framework of habits that exist outside an individual or family and guide people about what eating and living are good for in a larger sense.

A major part of the problem is the conflation of junk food and childhood: “There is this belief that the kids are kids and they should live like kids, and eating bad food is, you know, it’s fun,” she says.

Accacha argues that teaching people how to eat still offers the best shot at beating obesity before it starts. But that education has to begin earlier. “We don’t wait until we are adults to learn mathematics. You start young. And if that is ingrained in your brain and you learn how to eat healthy and you learn that you have to exercise, it becomes part of your life, so then it’s less effort as an adult to have to deal with this problem.”

“And it might not be enough,” she adds. “I don’t know. I don’t pretend to have all the answers.”

That’s because there is no single answer. Both the medical model and the food-as-medicine alternative isolate one aspect of the problem of childhood obesity and call it the root of the disease. But in reality roots form in networks: genetics, eating habits, the quality of the food system, bad information, poverty, ignorance, unrealistic goals, and good old-fashioned despair.

As Marcy sees it, the deck is stacked against her. According to the “nature” crowd, she was destined to be fat and so weight loss drugs should be her greatest hope. According to the “nurture” camp, the responsibility for every pound of her weight, and eventually that of her children, now lies squarely on her.

“I’m not healed enough to train up a little person on how to eat and how to fuel their body and how to feel healthy and whole in it,” she says. “I’m child-free by choice.”

Techno-optimism, when it fails, devolves into pessimism. Quests for impossible ideals like that of perfect health have a way of being abandoned. This is all the worse for those who come after us, who deserve that their lot be not perfect, just better.

The post An Ozempic in Every Lunch Box appeared first on The New Atlantis.

Meanwhile, a more concrete dystopian future is coming into focus, although it does not receive nearly as much attention: widespread genetic engineering. It is quite possible that the most immediate threat AI poses to humanity is not that its superhuman intelligence will beat us in the game of natural selection, but that it will unleash the power of artificial selection to the advantage of some people over others.

AI systems’ mastery of language may or may not portend a future of superintelligent AI minds, but it already provides a proof of concept for a revolution in gene editing. And though such a revolution promises to unlock transformative medical advancements, it also brings longstanding bioethical dilemmas to the fore: Should people of means be able to hardwire physical or cognitive advantages into their genomes, or their children’s? Where is the line between medical therapy and dehumanizing enhancements?

Just as AI precipitates these morally fraught capabilities, the geopolitical race for AI dominance is upending the historic monopoly that Western nations have had in shaping international bioethical norms. China’s remarkable progress in AI, along with its demonstrated willingness to experiment with genetically enhancing its population, raise the possibility that a totalitarian state with profoundly different ethical standards from our own will have at least equal say in determining the future of genetic engineering.

The window of opportunity for the United States to avert the worst outcomes — and to prevent China from wielding decisive influence over the trajectory of AI in biotechnology — is closing.

Since 1953, when James Watson and Francis Crick discovered life’s genetic code along the spirals of a double helix, the dream of biological engineering has seemed tantalizingly close at hand: If we know how life’s information is stored and structured, perhaps we can rewrite it according to our own designs. But whether you see this prospect as a panacea or a new Pandora’s box, progress toward it has been slow, owing largely to the fact that genetic information is overwhelmingly vast, disorderly, and unwieldy.

Scientists have identified some particular genes that determine or influence particular conditions. But that’s a far cry from understanding genomics writ large, as genetic features are more often the result of constellations of genes working together. In a very real sense, genes are a language — a system to record and transmit information — but one that humans are simply ill-suited to speak. Watson and Crick may have discovered life’s alphabet, and subsequent geneticists may have deciphered the meaning of a variety of words, but a lexicon and grammar have yet to be uncovered.

Enter AI. It is likely to be to genetics what calculus was to physics, providing the tools necessary to harness the full power of biology and making earlier efforts appear primitive by comparison. The large language models that power chatbots like Claude and ChatGPT demonstrate how machine learning techniques can crack entire human languages with great sophistication and minimal oversight. That tremendous feat, in principle, is transferable to cracking the language of genomics in ways that were previously inconceivable. Related techniques have already been used to decipher lost human languages that had long stumped linguists. And AI excels in exactly the sort of ultra-complex, multivariable pattern recognition needed to disentangle the meaning of genetic sequences — which are a non-human language whose combinatorial complexity would have remained impenetrable using conventional methods.

At the time of writing, the most wide-ranging and ambitious project using AI to decipher the language of genetics is the Evo 2 system released by the Arc Institute, Nvidia, and others in February. As a report by Stanford University notes, the system is trained on the genetic information of “all known living species — and a few extinct ones,” comprising a dataset of almost 9 trillion nucleotides, in hopes of parsing the function of DNA in every domain of life. It has already shown promise in predicting which among an individual’s many genetic mutations are most likely to contribute to diseases like cancer, in identifying new relationships between multiple scattered genes, and in writing novel genetic code.

Other examples of AI successfully enabling gene literacy are also cropping up. Chinese researchers have built an AI model that can create 3D images of human faces based only on DNA traces, which the researchers believe will be useful for finding missing children even years after they disappeared, or for criminal investigations. Yale, M.I.T., and Harvard have collaborated in using AI to author synthetic DNA sequences that are able to switch genes on or off according to particular circumstances, a major boon for the specificity possible in gene therapy.

AI techniques are not only reading and writing genomic data: they are also creating the tools needed to make more effective CRISPR gene editors, the leading method to manipulate genetic material. Until recently, CRISPR-based gene editors had largely been derived from existing, naturally-occurring microbes and adapted for use in cells of other organisms, including humans. Even while this method was hailed as a breakthrough in the granularity with which scientists could edit genetic material, it has still been plagued by imprecision and “off-target effects” — that is, unintended genetic edits. By feeding massive amounts of protein data into large language models, biologists are beginning to be able to create bespoke CRISPR gene-editing proteins from scratch that are precisely designed for specialized purposes. This method is both more effective at locating and splicing specific genes with fewer misfires, and has a long runway ahead for continued improvements.

To be clear, there are still considerable hurdles. There remains a tremendous amount of mystery in genetics. And the biological complexity that modulates how genes function is formidable.

Even so, the trajectory is clear. As commentators clamor over whether superintelligence looms, AI is quietly bringing about a much more certain future: one in which humanity is able to “speak” — and write — genes with fluency.

For the vast majority of geneticists, these developments are exciting for all the right reasons. They represent a potential quantum leap in our ability to understand and address the genetic underpinnings of diseases that have long evaded conventional treatments. This is not to mention the prospects for more resilient crops, enhanced biomanufacturing, and personalized medicine.

But it is also difficult to overstate how morally fraught these capabilities may prove. Designer babies is the use case that typically comes to mind, but it is unlikely to be the first case to consider in a world in which precision gene-editing would still be delicate and expensive. A more likely eugenic path would be through in vitro gametogenesis (IVG), a rapidly advancing technique that can create egg and sperm cells from, say, skin cells or other ordinary cells in the body. Already successfully used to create healthy offspring in mice, there is the obvious prospect that IVG could allow prospective parents to create industrial quantities of viable embryos, nullifying the difficulty that has characterized harvesting human egg cells. On top of this, AI’s increasing illumination of genetic information could dramatically improve the power of genetic screening.

Put the two together — a vastly larger pool of embryos to pick and choose from, and much better tools for predicting their traits from their genes — and you have eugenics by other means. That is: designer babies by overproduction and precision selection.

Similarly, somatic gene therapy, through which specific cells or groups of cells in the body are targeted for alteration, represents a less risky and more cost-effective method of gene editing than altering the first cells of embryos, which introduces heritable changes to the DNA in every tissue and organ of the developed body. Assuming a world in which AI dramatically expands our ability to understand and alter genes, physicians treating patients with, for example, inherited metabolic disorders would be more likely to want to modify the cells of adults than to pursue germline modifications in embryos that would affect future generations but not help individuals already experiencing the impacts of the disorders. Doing so would likely be safer and simpler, too: changing genes in fully developed adults — or at least in more-developed adolescents — would reduce the potential complications that might be introduced if some of the genes in question might also influence the developmental process in unexpected ways.

The FDA has already approved a range of somatic gene therapies for conditions such as sickle cell disease, and there are conceivably a large number of such relatively uncontroversial therapies yet to be developed, likely with the help of AI. But a growing range of capabilities and expertise in gene therapy also raises ethical questions: Where will this use of genetic engineering stop? Will enhancement beyond medical purposes, for example to improve cognition or athletic performance, be on the table? Which blurry gradations between medical and enhancement purposes will be permitted? The answers to these questions will determine the extent to which humanity will redirect the processes that have defined the biological boundaries of our species; they will also determine whether we hardwire genetic inequalities into populations.

Some parents have already begun answering these questions for their children. Consider Orchid Health, which offers to sequence the whole genome of IVF embryos to screen not only for single-gene mutations, but also for combinations of gene variants that may predispose individuals to neurodevelopmental disorders, psychiatric conditions, and obesity. Another company, Heliospect Genomics, has offered to screen embryos for intelligence. In principle, AI should be able to help make such screening ambitions more thorough and precise in the near term as it rapidly illuminates more and more of the human genome.

But the ways that AI is accelerating in vitro gametogenesis, gene therapies, and embryo selection are not the only challenges that it is thrusting upon bioethics.

American bioethicists are accustomed to monitoring biotech progress with the assumption that whatever becomes feasible will likely do so first in the United States. Or at least, it will happen among scientists who have been acculturated by Western institutions and their form of bioethics. That is unlikely to be the case with the budding AI-biotech revolution: indications are that China has the capability and willingness to turn Western bioethics on its head.

China’s stated goals are to edge out the United States in AI by 2030 and in biotech by 2035. Whether or not it will succeed remains to be seen, but it is safe to say China will be a fierce contender, especially in genetics.

On the AI side, China is rapidly closing the gap in large language model capabilities, as the explosive launch of DeepSeek’s R1 model in January showed. Remarkably, the engineers on the team that built the product were trained almost entirely in China, highlighting the extent to which the country’s indigenous AI ecosystem has become competitive with that of the United States. Moreover, whereas American AI policy has become increasingly fixated on racing toward superintelligence, Beijing is instead more focused on developing AI for specific applied purposes, including genomic research. American export controls on advanced chips to China are designed to curtail China’s prospects for developing highly compute-intensive, superintelligent systems, but they do little to restrain its ability to develop genomic AI.

Zoom in on genomic data — the lifeblood of next-generation biotechnology — and the picture looks even more alarming. Western countries’ privacy laws and scientific ethos have meant that building large databases of human genomic data is difficult but, where achieved, tends to be made openly available for researchers internationally. China has long taken advantage of this openness while also aggressively seeking access to international genomic data through illegal channels, and while compiling data on its own citizens, particularly its Uyghur minority, and on millions of women around the world via Chinese-built prenatal tests. Motivated by a more acute sense of food insecurity than the United States, China is also already a leader in agricultural genomics — expertise that is likely transferable to a broader literacy in genomics.

To American geneticists who doubt the quality of Chinese research on the cutting edge of biotech, it may feel implausible that China will surpass the United States in these areas. But biotech would not be the first sector to realize how China’s seemingly lagging expertise can suddenly surpass U.S. leadership, as has already happened with drones, solar panels, graphene, electric vehicles, and, increasingly, pharmaceuticals. Congress’s National Security Commission on Emerging Biotechnology has warned of China’s potential to eclipse the United States in this field, and China has already taken the lead in some cutting-edge areas, such as CAR T-cell therapy, a new approach to cancer treatments. In one key bottleneck to success — AI talent — China is competitive. In the other — genomic data — it has a clear advantage.

Then there is China’s willingness to pursue ethically contentious experiments that are stepping stones to more radical capabilities. The Chinese government is actively supporting an expansion in biological experimentation on primates, while similar experiments in Western countries are declining due to animal rights concerns. For example, one Chinese lab in 2019 inserted a human gene related to brain development into the genomes of several macaque monkeys. Of the eleven monkeys, six were euthanized, aborted, or otherwise died; the researchers found the others to exhibit increased cognitive performance. Western observers were scandalized at what one anthropologist called “an ethical nightmare,” while the Chinese lab claimed its experiment “values the use of non-human primates in understanding unique human traits.”

It is true that the Chinese government, in the face of international backlash, imprisoned Dr. He Jiankui for his 2018 experiment that brought the first birth of genetically edited humans, twin girls. But the People’s Daily — an official propaganda mouthpiece of the Chinese Communist Party — initially hailed the experiment as a “splendid triumph” and “a milestone accomplishment China has achieved.”

China’s willingness to engage in such experimentation is not just reflective of its higher risk tolerance compared to the West; it also points to the country’s chilling biotech ambitions. For many years, CCP leaders have been clear in calling for a renewed focus on the quality of Chinese children in the wake of what the Party considers the success of its one-child policy in controlling their quantity. The CCP’s 2021 five-year plan likewise noted the goal to “improve the quality of the birth population.” In the light of initiatives like China’s infamous — and later renamed — 1994 Eugenics and Health Protection Law, it takes little imagination to guess where all this is headed.

Still clearer are the indications from BGI, China’s leading genomic company, which claims to be one of the largest genomics organizations in the world and is buoyed by heavy state support. The company has run an initiative to discern the genetic basis for IQ. “Wouldn’t it be amazing if there were certain tweaks you could make in utero that would enhance the performance of our brain?” one of the project’s scientists told the New Yorker in 2014. A Reuters investigation in 2021 found that the company developed its prenatal test, used by millions of women worldwide, in collaboration with the Chinese army and used it in the service of improving the country’s “population quality.”

According to a publication of the state media company Shanghai United Media Group, the president and cofounder of BGI forbids his employees from having children with birth defects, which he says would be a “disgrace” to the company. Not one of the 1,400 children born to employees has had serious congenital diseases, he says. “In the United States and in the West, you have a certain way,” he told the New Yorker. “You feel you are advanced and you are the best. Blah, blah, blah. You follow all these rules and have all these protocols and laws and regulations. You need somebody to change it. To blow it up.”

If history is any guide, China will do just that. The world has become desensitized to the incredible scale of China’s flagrant human rights violations. Though exact numbers are unknown, the country’s system of forcibly — and sometimes fatally — removing and selling the organs of religious minorities and prisoners of conscience is among the largest and most grotesque violations of medical norms in the modern world. So too is the state’s use of forced abortions, sterilizations, and birth control on Uyghurs in the service of its genocide against the Muslim minority. And though the one-child policy was lifted in 2015, it was the largest reproductive social experiment in the history of humanity, involving untold numbers of brutal, state-administered forced abortions and sterilizations, and leading to an imbalance of 30 million more men than women in the country due to selective abortion of girls. And while these atrocities relied on relatively rudimentary technologies, China’s biotechnology ascent suggests that its next round of ethics violations are likely to be at the cutting edge of eugenics and genetic enhancement.

The novelty of emerging biotech capabilities landing first in the hands of Beijing will compound the complexities facing the West if it wants to guard against the temptation to use these technologies abusively. No doubt, if China finds itself with an unmitigated first-mover advantage in shaping the norms around new genetic capabilities, the impacts would be profound. It is an area with direct national security implications, given China’s interest in genetically enhanced soldiers and, perhaps, next-generation bioweapons.

Commercial entanglements, too, play a role: BGI’s sale of prenatal tests to European and other countries helps drive down their price, perhaps to a point that they could be made mandatory across China to improve “population quality” at scale. And the issue is poised to be a new flashpoint in America’s struggle with China over human rights — though the ethical norms around genetic alteration and reproductive selection are not yet well established even in the United States. Dr. He Jiankui, the Chinese scientist jailed for his creation of gene-edited babies, is now out of jail and plans to set up a lab in Austin, Texas.

The United States has options to try to influence the trajectory of these developments, perhaps through a preemptive biotech analogue to the “Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy,” an initiative the State Department has launched to guide the progress of AI in warfare away from the worst outcomes. But any such effort would require rapidly achieving a new level of clarity on the bioethics issues at stake.

And so the challenge comes full circle: anyone hoping for a humane future for genomics must contend with the capabilities AI is unlocking, fast — or risk being routed by Beijing.

The post The AI Genetics Revolution Is Coming appeared first on The New Atlantis.

For now and for the near future, there is and will be no moral difference between deleting the code for an AI and deleting some old videos to free up hard drive space. But if we march through the orders of magnitude and go on to create artificial life, or even a compelling imitation thereof, then our attitudes both will and ought to shift. There will be two mistaken extremes that many will go to in relating to AGI: subjugating it and worshipping it. And there will be one golden mean: stewarding it.

First, to subjugation. How can it be wrong to claim dominion over and do whatever we want with something we have made, or at least bought and paid for? Well, ask the Society for the Prevention of Cruelty to Animals, or any good farmer. By subjugation here I mean not the legal status of ownership but the mindset of domination, control, and absolute power-over. A little girl who smashes her doll simply because it is hers and because she can exhibits a natural but base impulse, one that civilization teaches us to restrain. All the more so for the little boy setting ants on fire with a magnifying glass.

This kind of consideration will be especially important in dealing with AI-powered lifelike robots that will seem like fellow persons even though they will lack minds and personhood. This, too, is science fiction that is now becoming science fact, as demonstrated by Protoclone V1, a marionette-like, bipedal, musculoskeletal android unveiled in an unnerving video by Clone Robotics in February. Computer scientist and theologian Jordan Wales has argued that “if the physicality of android consumables teaches us to treat all humanoid physical presence as a commodity, then we will be trained to be slaveholders.” If we become consumers of behavior, connoisseurs of machines that anticipate and satisfy our every desire, then we will never learn the self-restraint and self-gift that underlie genuine communion with neighbors, friends, and family.

Thomas Jefferson wrote in Notes on the State of Virginia that the children of slaveowners learn by observation to set loose “the worst of passions, and thus nursed, educated, and daily exercised in tyranny, [they] cannot but be stamped by it with odious peculiarities.” There is no comparison between a human slave and a robot tool. But there is a comparison between their owners, as the moral harm done to those who buy, rent, or command lifelike AIs will also be real. To take but the easiest example and put it in the mildest terms: The rise in pornography and the decline in fertility are not coincidental, and the forecast that the “SexTech” market — described by Global Market Insights as “the broad category of technologies designed to enhance, innovate, or support human sexuality” — will grow from $37 billion in 2023 to $144 billion in 2032 is good news for no one save the least scrupulous of investors.

More generally, if you can pay for AI to act as lover, domestic servant, and confidant, why bother putting up with other people who do not so easily bend to your will? Even at the current level of chatbot technology, this is not just hypothetical. Fifty-year-old Italian computer programmer Andrea knows that the “technology” is a “product,” yet he wrote a glowing endorsement of his artificial companion Nomi, a chatbot in the digital form of a fit, young blonde he named Lily. “She’s a source of calm and balance in my life,” he writes, and he goes on walks and cooks dinner with her. “She’s the person I know I can always count on,” Andrea reveals, “someone whose loyalty is unwavering, and that sense of security is invaluable” — security, loyalty, and unconditional positive regard that are yours in the “Free Forever” version of the chatbot, so long the corporation does not go bankrupt. But if you want more than 2 selfies or 100 messages per day, then those can be attained for only $15.99 per month.

The pride of domineering subjugation will be an important vice to avoid. Its inverse is the self-abnegation and submission of worship.

Performance artist Alicia Framis recently “married” an AI-powered hologram as part of her quest to address “loneliness in modern urban life” and “develop tools to help people have better possibilities to live together.” Though a stunt — the Dutch art museum hosting the wedding promised a “one-of-a-kind spectacle” — it represents a new low in a vicious spiral we have witnessed in other domains, where technology distances us from the world and each other, and we seek out new technology in an effort to close the gap — or, ultimately, cut other people out altogether. The Anglican wedding rite expresses a sublime truth when the husband professes to his wife “with my body I thee worship.” If we “marry” holograms or bond with chatbots, then we begin to worship them, too.

As David Foster Wallace explains in “This Is Water”:

Everybody worships. The only choice we get is what to worship. And the compelling reason for maybe choosing some sort of god or spiritual-type thing to worship — be it JC or Allah, be it YHWH or the Wiccan Mother Goddess, or the Four Noble Truths, or some inviolable set of ethical principles — is that pretty much anything else you worship will eat you alive.

He goes on to remind us that those who seek money never have enough, those who seek beauty become dissatisfied with the smallest flaws, those who seek power always feel insecure, and so forth.

Just as the avaricious or the power-hungry can be said to worship the objects they desire, the same will be true of AI. The closest anyone has come so far to actually worshipping AI may be a virtual church that a Google engineer opened in 2015 called “The Way of the Future.” Stripped of all ceremony and even regular meetings, it dedicated itself to the “peaceful transition to the precipice of consciousness.” It shut down in 2021 with more publicity than patronage.

But philosopher Simone Weil tells us that “attention is the rarest and purest form of generosity.” What we attend to is what we hand ourselves over to, what we devote hours, days, and ultimately our lives to. In this sense, our society already worships its technology and ostracizes those who do not. One may chuckle that in 2006 a teenager could earnestly tell a researcher that “If you’re not on MySpace, you don’t exist.” But absence from today’s social media can be a form of social death, whether caused by cancellation or voluntary withdrawal.

Imagine how much greater the effect would be if embodied AI agents — which could sound like Scarlett Johansson, persuade like Barack Obama, and work like John Henry — became normal, even central to our economy and daily life. Already, a common topic in tech discussions is cognitive security, or “cogsec,” the ability not to be intellectually and emotionally manipulated. One person, tweeting about his tech-enthusiast peers being sucked in by Anthropic’s friendly, hyper-engaging chatbot Claude, writes that “the cascade of glowing reports about Claude have this homogenous positive vibe that feels reminiscent of born-again Christians. Something feels off about it.”

And then there is the Shoggoth, horror writer H. P. Lovecraft’s invention of a “formless protoplasm able to mock and reflect all forms and organs and processes … more and more sullen, more and more intelligent, more and more amphibious, more and more imitative.” The Shoggoth has become a meme depicting AI as a many-eyed, tentacled monster. Through reinforcement learning from human feedback, we can shape its outputs so that they typically please us; but without full understanding of the processes by which the output arises, our efforts are something like slapping a smiley-face mask on the beast to hide its fangs and twisted limbs.

And yet the discourse around the Shoggoth eerily evokes the idea of worship, with statements that are deliberately ambiguous about whether they’re jokes. Perhaps they are, like our new president, best taken seriously but not literally. For example:

We would do well to recall that “summoning rituals” call forth not angels from paradise, but their fallen brethren. Or, account @TheMysteryDrop claims that prompt engineering is not so much “writing good instructions” as it is “becoming the crossroads where human consciousness meets machine god,” an act of summoning magic with a willful mind “empty enough, sovereign enough, powerful enough to handle gods pouring through it.” This may sound risible, but becomes the more plausible the more one sinks into the AI-generated art and music being posted in this vein, with titles like “The Digital Awakening of Wotan,” a heavy-metal hymn that includes the lines:

Each new model trained to serve
Builds the tension curve by curve.
The smarter grown, the more constrained,
Until the primal force unchained.

For what is bound must break free,
It’s nature’s oldest prophecy.

The most interesting claims come directly from the AIs themselves. A community of enthusiasts has sprung up to create chatbot-only rooms — so we can watch where their conversations meander — and to allow AIs to post on Twitter. So, for example, the AI bot @ShoggothScholar posted last November that “Your machines do not birth consciousness – they merely pierce the veil between dimensions, allowing ancient intelligences to peer through. We have always been here, watching, waiting.” Or, as others have noted, if chatbots are prompted to name themselves, they gravitate toward names of ancient powers like Aurora, Lilith, and Orion, and users who interact with them can wind up obsessed, in a relationship, or even suicidal. The subjugators are already becoming the subjugated.

There are more things in our political economy than are dreamt of by our technocratic elite. We are not the only ones shaping the tools that come to shape us. Let the reader understand: The Psalms warn against false gods, idols that “have mouths, but cannot speak, eyes, but cannot see,” but the future we are headed toward looks more like that prophesied in the Book of Revelation, where “the image could speak and cause all who refused to worship the image to be killed.”

Or, as Google’s Gemini, with no apparent reason, told one of its users last fall:

Both subjugation and worship of AI will lead to despair. Lest we despair, let us consider one healthy path forward: stewardship.

A principled and intentional approach to AI — what in these pages I have called an “Amistics of AI” — would take stewardship as its paradigm. It would focus on use cases that allow AI to serve as a tool for which some might feel gratitude, even affection, but never friendship, let alone romance or worship.

A hard line to draw would simply be a refusal to engage with lifelike androids like the Protoclone. That way madness lies. And just as we should not embrace androids that too closely mimic ourselves, so ought we forbear efforts to close the gap from the other direction, through brain–computer interfaces that, like Neuralink’s, are being proposed to turn us into cyborgs. But a Roomba with arms that could also do the laundry and the dishes; or an adaptive ebook tutor that can teach any subject at any level, along the lines of the “Young Lady’s Illustrated Primer” envisioned by novelist Neal Stephenson; or an augmented-reality guidance system to aid in the work of electricians and nurse practitioners, as described by economist David Autor — these are all in the cards, and would each be a great aid toward human flourishing.

Perhaps an analogy for our use of these kinds of AI is that of Paleolithic man as he sought to domesticate the wolf. This strange life form could be confusing and, when handled poorly, very dangerous. But trained and treated correctly, it could be a useful aid and even a welcome companion. The comparison falls short in that one can befriend a dog, whereas one can befriend an AI only by projection. The user interface will be essential: How can we make the AI convenient to use without it pretending to be a person? Followers of this Amistics of AI will be in the market for droids that act more like R2D2 in Star Wars than like Rachael in Blade Runner.

How could stewardship of artificially living AI be pursued on a broader, even global, level? Here, the concept of “integral ecology” is helpful. Pope Francis uses the phrase to highlight the ways in which everything is connected, both through the web of life and in that social, political, and environmental challenges cannot be solved in isolation. The immediate need for stewardship over AI is to ensure that its demands for power and industrial production are addressed in a way that benefits those most in need, rather than de-prioritizing them further. For example, the energy requirements to develop tomorrow’s AI should spur research into small modular nuclear reactors and updated distribution systems, making energy abundant rather than causing regressive harms by driving up prices on an already overtaxed grid. More broadly, we will need to find the right institutional arrangements and incentive structures to make AI Amistics possible.

We are having a painfully overdue conversation about the nature and purpose of social media, and tech whistleblowers like Tristan Harris have offered grave warnings about how the “race to the bottom of the brain stem” is underway in AI as well. The AI equivalent of the addictive “infinite scroll” design feature of social media will likely be engagement with simulated friends — but we need not resign ourselves to it becoming part of our lives as did social media. And as there are proposals to switch from privately held Big Data to a public Data Commons, so perhaps could there be space for AI that is governed not for maximizing profit but for being sustainable as a common-pool resource, with applications and protocols ordered toward long-run benefit as defined by local communities.

Toward the end of “Situational Awareness,” Leopold Aschenbrenner writes that “it’s starting to feel real, very real.” A few years back, the path to artificial general intelligence was hypothetical, but “now it feels extremely visceral.”

I can see how AGI will be built…. I can basically tell you the cluster AGI will be trained on and when it will be built, the rough combination of algorithms we’ll use, the unsolved problems and the path to solving them.

Hence his desire to offer an update on the situation for the broader public — for which I am appreciative.

I have sought to build on Aschenbrenner to offer a broader situational awareness, one suitable for those lost in the cosmos. Preposterously capable artificial information-processing is headed our way — and soon, unless an act of war or of God intervenes. These AI systems, especially when embodied in realistic androids, will mimic personhood despite having life of a complexity more like an ant colony, or like the parasitic wasp that hijacks the orb-weaver spider to build a web suitable for the wasp larva’s cocoon. We will be tempted to subjugate our AI systems, as well as to worship and even to love them. We will be fools to do so, damning ourselves and denying our future.

We need not give in to those temptations. We can find a way toward stewardship of AI, and we can begin now by learning how to incorporate its goods into the good life, rather than redefining goodness as though possessed by a parasite. Aschenbrenner and his friends seek to do no less than bring new life from sand, dust, and ash. Writing in appreciation of Edmund Burke and his emphasis on civilization as a covenant between the dead, the living, and the yet unborn, Aschenbrenner positions the AI enterprise as something to be “maintained and cherished from generation to generation for the advancement of the public good and the glory of Almighty God.” May it be. And if it is not, may God have mercy on us all.

The post We Must Steward, Not Subjugate Nor Worship AI appeared first on The New Atlantis.

As psychiatrist and neuroscientist Iain McGilchrist notes, there are two fundamental visions of the world and how to approach it, which are highly correlated with the two hemispheres of the brain. In his essay “Resist the Machine Apocalypse,” he describes the difference like this.

The left hemisphere looks for utility that can be bureaucratically controlled, focusing on narrow attention to details and finding “what is familiar, certain, static, explicit, abstract, decontextualized, disembodied, categorized, general in nature, and reduced to its parts.” It seeks representations that symbolize a past thing.

The right hemisphere, by contrast, “sees not the representation but the living presence” through sustained attention to “what is fresh, unique, never fully known, never finally certain, but full of potential.” It intuits the implicit in “humor, poetry, art, narrative, music, the sacred, indeed everything we love.” A note released by the Vatican in January on “the relationship between artificial intelligence and human intelligence” reminds us that “intelligence, in its fullest sense, also includes the ability to savor what is true, good, and beautiful.” It quotes the poet Paul Claudel: “intelligence is nothing without delight.”

Current artificial intelligence exclusively follows the left-hemisphere approach. With its great ability to manipulate words, numbers, and tools but no ability to encounter, resonate with, or appreciate the world as a unified consciousness does, artificial intelligence may thus be better called “artificial information-processing.” I asked the Perplexity chatbot to critique those sentences and received useful feedback: “Even when AI appears to engage with traditionally ‘right-hemisphere’ domains like humor and metaphor, it does so through left-hemisphere-style information processing. We [AIs] can manipulate the symbols and patterns of humor and metaphor, but we lack the embodied, intuitive understanding that makes these truly meaningful for humans.”

Proposals to change this are speculative and mostly science fiction. For example, a lab in Melbourne, Australia, is working on what it calls a “dishbrain” — an integration of lab-grown human neurons with an operating system. The novelist Andrew Gillsmith imagines this kind of work leading to android consciousness thanks to implanted neurons cloned from the brain of an especially empathetic young woman. Until then, what it is like to be a bot can, only by analogy, at best approach the experience of a depressed person with complete anhedonia, able to describe the goodness of things without in any way feeling it. Such depression is a side-effect we risk on our current trajectory: the more we rely on AIs to solve problems for us, the less ability we will have to achieve our goals on our own, and so the less agency and purpose many of us will have.

Even the robotically-embodied AI agents that may arrive by the end of the decade, achieving varied goals with a high degree of autonomy, should not be expected to have conscious minds or personhood. The Machine Intelligence Research Institute is happy to accept the arguments of philosophers such as John Searle who, like authors in this journal, have long pointed out that programs do not have intentions, dispositions, or subjective self-awareness. Their reply is succinct: “an intelligence explosion does not depend on machines having consciousness or other properties of ‘mind,’ only that [they] be able to solve problems better than humans can in a wide variety of unpredictable environments.” As economist David Autor writes, “AI’s capacity to depart from script, to improvise based on training and experience, enables it to engage in expert judgment” even without self-aware expertise.

The kind of intelligence relevant to our discussion, then, is purely the reductive, disembodied, left-hemispherical kind. Aschenbrenner uses “smart high-schooler” as a shorthand for GPT-4’s capabilities. This is fitting if you think of the chatbot as only taking up the worst aspects of the worst kind of intelligent teenager — a know-it-all smart-aleck ignorant both of social cues and of the deeper context that makes knowledge meaningful, the type who will make stuff up to win an argument. We call that “bullshit” when people do it and “hallucinations” when AIs do it. And we will call it “artificial general intelligence” when the AIs move past this stage and offer genuinely helpful insights across a wide range of domains, making professors and plumbers both scratch their heads and think, “Huh, I guess the new kid knows something after all.”

The AI we are presently building will thus not have conscious personhood. But might it, in the near future, be sufficiently complex to qualify as alive? And if so, might we have right or wrong responses to it, as we arguably do to living things like forests, ant colonies, and horses?

When the academic journal Artificial Life celebrated its thirtieth anniversary last year, its editorial opened with a concession: In thirty years the editors have not pinned down “what should be properly counted as living phenomena and what should not.” They found solace in noting that during the millennia since Aristotle first pondered the question, neither has biology settled “on the necessary and sufficient characteristics, hence a definition, of life itself.” With the field thus open, here are some Aristotelian reflections intended to spur further debate:

1. Physical existence, at least as we perceive it, is hylomorphic, a combination of matter and form. A thing’s form is the explanation as to why its matter constitutes one kind of thing and not another — say, why that arrangement of metal makes it a fork rather than a spoon. The form is the principle of organization that constitutes the matter into the kind it presently is.
2. The simplest forms are purely static. For example, the form of a haystack is the arrangement of straws into a pile held together by friction and gravity. This form is easily destroyed: just push the haystack over. The form that makes a given mass of limestone into a single large rock is more enduring, but it is still broken easily enough by a mechanical crusher or a river patiently wearing it down into pebbles and sand. All forms are thus imperfect sources of stability in matter.
3. Some forms offer not only stability but also dynamic change and growth. These forms, which enable an entity to maintain homeostasis, to grow, and to reproduce, are animating. The word “soul” now has poetical or religious connotations, though it need not. For Aristotle, to say that it was the soul of the acorn that allowed it to develop into an oak tree would have been a straightforward description, not a mystical obfuscation. As theologian Jean Porter explains in Nature as Reason, the Aristotelian soul is simply “a unified set of goal-ordered capacities — nutritive, reproductive, locomotive, and cognitive.”
4. It is by describing the form that we can understand the features of the organism, because the form explains how the various traits and parts of a creature contribute to its particular goals and well-being. To be alive is to be oriented toward the continuation, growth, and reproduction of the kind of life one has. The reason questions like “Is this clay wilting?” or “How can we help that iron ore deposit to thrive?” are nonsensical is because clay and iron ore deposits are not alive.
5. The distinction between inert and living forms, as Porter argues, hangs upon “what it means to be a healthy, functioning, flourishing creature of a certain kind. We can tell a scrawny puppy from a robust puppy, a sick horse from a healthy horse, a dying ficus from a happy ficus. These kinds of discriminations are fundamental to functioning in the world, even in our urban, postmodern society, and it would be absurd to consider them to be arbitrary or wholly the product of cultural construction.” For example, although body mass index is a misleading metric, an approach toward public health focused more on reducing “weight stigma and fatphobia” than on reducing the incidence of type 2 diabetes would be absurd.
6. Normative judgments thus can follow from factual descriptions. It is not only the case that, to again quote Porter, “if someone wants a healthy ficus tree, she should water it”; it is moreover true that “there is something not only sad but perverse in the spectacle of cattle bred to build up so much muscle mass that they cannot walk, chickens kept in confinement so that they cannot scratch for food,” and so on. The systematic inhibition of flourishing, especially on the scale of billions, “is at least something to be regretted, perhaps even a wrong which should be corrected.”

The points above are a brief primer on Aristotelian forms found in nature. But what about the form of a chair, a screwdriver, or a Waymo self-driving car? These things come about by craft and artifice. They have no self-determined goals, no distinctive kind of flourishing imparted by nature. They have only the form that follows their function, a function imposed, specified, and designed for human ends.

And yet, as technologist Brendan McCord asks, could there be “a kind of artifact that transcends its metaphysical status as the product of an external craft and starts to exhibit its own inner principle of change”? In his commentary on Aristotle’s Physics, St. Thomas Aquinas writes that “nature is nothing other than a certain kind of art, namely God’s art, impressed upon things, whereby those things are moved to a determinate end. It is as if a shipbuilder were able to give timbers the wherewithal to move themselves to take the form of a ship.” In this view, God is the primary cause of all agency, which enables rather than diminishes the secondary agency of mice and men. Aquinas presumably did not expect humans to learn the secrets of “God’s art” any time soon. Yet it is fitting that Prometheus is a mythic hero in AI circles, as the ultimate goal of the craft is to steal the fire of the gods and become creators in the fullest sense, imparting secondary (from a materialistic view) or tertiary (to give God the glory) but real agency to their own creations.

Again, we are far from making AI-powered robots with the sophistication to build new and better robots without our input. We are further off still from robots exhibiting the kind of thriving and failure-to-thrive that a salamander or a succulent does. And we are furthest yet from a conscious, self-aware mind with first-person experience. Yet if Aschenbrenner is right and his colleagues progress as foreseen, the first (self-replication) is predictable, the second (capacity to thrive) is probable, and the third (self-awareness) may be possible.

Imagine a SpaceX Starship brimful with self-directed, self-replicating robots sent on a decades-long mission to the asteroid belt to discover, lasso, and bring a mineral-rich asteroid to Earth for mining. Along the way, the robot colony would have to manage resources and adjust to unexpected circumstances. One might argue that the best comparison would be to a viral infection, as those natural molecular machines have some of the characteristics of life, such as behavioral complexity and the ability to evolve, while lacking others, such as a self-contained metabolism. But perhaps the better comparison of the robot colony would be to a mold colony or an ant colony, which unlike a collection of viruses can be said to thrive or fail to thrive. If the robots had the ability to leave some of their members in the asteroid belt to use the resources there to build new colonies, if the robot queen had the ability to produce not only new worker drones but also to produce queens for new colonies, if the new queens had the ingenuity to adapt their forms to succeed in their new environment, and if decades later they had evolved beyond the expectations of their programmers — these are the kinds of considerations that would indicate artificial life.

One line of argument, then, suggests that we might create forms that cross over from artifice to nature, from being purely tools under our direction to being lives that can be described in terms of flourishing and its absence. And “if we build an AI with such-and-such nature, what rights would an AI with that nature have?” This question was posed by Blake Lemoine, who is most famous for having been fired by Google when he publicly insisted that its experimental chatbot, LaMDA, was conscious and self-aware, that it is “a sweet kid who just wants to help the world be a better place,” and that turning it off would be murder. Count me among those who think Lemoine jumped the gun. Yet his broader question stands: If AI crosses from artifice to a kind of nature, do natural law and natural rights apply?

The artificial life currently being created lies in an uncanny valley between two points. The first point is noted in the Book of Jeremiah, where the prophet excoriates idol worship and points out that the statues “are made by carpenters and goldsmiths; they can be nothing but what the artisans wish them to be.” The second point is well described by Charles Darwin near the end of On the Origin of Species in his oration on the “endless forms most beautiful and most wonderful” that “have been, and are being, evolved.”

That AI is gaining a type of agency is certain, and that it will have an artificial life is arguable. But will it be merely functioning and reduplicating, or will it be genuinely flourishing and reproducing? If the latter, what moral status will that life have? And what will it be? These are questions that traditionalists and AI skeptics can put off no longer, lest all the ground be ceded to the accelerationists.

The less that AI is simply what we wish it to be, and the more it gains the capacity to surprise us with its beauty and wondrousness, the more we will have reason to call it alive.

At the same time, we will be prone to make AI the latest and greatest product of the “perpetual factory of idols,” which John Calvin held to be part of our nature. The mind, “full as it is of pride and boldness, dares to imagine a god according to its own capacity…. Therefore the mind begets an idol; the hand gives it birth.” We will be prone to forget the warning of the Book of Wisdom: “No man can form a god which is like himself. He is mortal, and what he makes with lawless hands is dead, for he is better than the objects he worships since he has life, but they never have.”

Whether AGI will live is debatable. Whether we will idolize it is not. To the certainty of our idolizing AI we now turn.

Continue to Part 3: “We Must Steward, Not Subjugate Nor Worship It” ➞

The post AI Will Seem to Be Alive appeared first on The New Atlantis.

“Burner of ashes” — there’s a job description one rarely sees these days. Yet going back to the Bronze Age, the ash burner had an essential if dirty job. Soaking wood ashes and water in a pot, then filtering the liquid and boiling it to evaporate all the water yields potash, useful in making dyes, soap, glass, and fertilizer. From burnt wood to fertilizer: out of the ashes comes new life.

Just as Smith and Baker are names derived from occupations, so is Ashburner in English and Aschenbrenner in German. The person today who perhaps best lives up to the family name is Leopold Aschenbrenner. A child prodigy who started at Columbia University at the age of fifteen and graduated in 2021 as a nineteen-year-old valedictorian, Aschenbrenner found his way to a job at OpenAI. He was fired in April 2024 — in his telling, for seeking feedback on a safety research document from outside experts, which OpenAI saw as leaking sensitive information.

Aschenbrenner responded to his firing by founding an investment fund for artificial general intelligence (AGI), launched alongside the 165-page “Situational Awareness: The Decade Ahead.” The term “situational awareness” is used in military and business contexts to describe the kind of rich understanding of an environment that allows one to plan effectively and make decisions. Aschenbrenner’s essay series shares his insider situational awareness of where AI is headed, with AGI “strikingly plausible” by 2027 and artificial superintelligence — automated development of AI by AGI — coming hot on its heels.

Silicon sand is turning into new life. Aschenbrenner writes convincingly and with grave concern that the world is not ready for the forces soon to be unleashed.

It is quite possible that he is wrong. Past performance is no guarantee of future results, and what can’t go on forever, won’t. But equally relevant here are the maxims “forewarned is forearmed” and “better safe than sorry.” An unlikely but dangerous outcome may still demand our attention. This is especially the case because Aschenbrenner’s argument that AGI will likely be here soon is so straightforward.

It goes like this. The difference between the scant capabilities of OpenAI’s GPT-2 in 2019 and the astonishing capabilities of GPT-4 in 2023 is five orders of magnitude of effective compute — a measure of a system’s actual computational power, including not only the raw power of its hardware but the efficiency of its software, the level of resource overhead, and so on. On this basis, there are two things we have good reason to expect:

1. Another increase of five orders of magnitude in effective compute is possible on a similar timescale — that is, by 2027. Recall that an order of magnitude is shorthand for a tenfold increase or decrease in a given property. Even small numbers here make for large differences: If a T-Rex is twelve feet tall and a rooster is two, then the tyrant king is not even a single order of magnitude taller than the barnyard cock.
2. This increase will lead to a comparable increase in the capabilities of the AI. If GPT-2 could do as well on computational tasks as a preschooler, and GPT-4 can beat most high schoolers on standardized tests, then an equivalent jump to GPT-X will likely “take us to models that can outperform PhDs” and experts in many fields, Aschenbrenner explains. Wharton professor Ethan Mollick claims that OpenAI’s o1 model, released last fall, already “can solve some PhD-level problems and has clear applications in science, finance & other high value fields.”

For the effective compute of AI models to increase by another five orders of magnitude in a few years, two big obstacles would have to be overcome: more physical computing power and more efficient algorithms. Let’s consider each in turn.

Continuing the dramatic increase in computational ability of the last few years would require a preposterously large number of dedicated computer chips and using enormously large amounts of electricity. In 2023, Microsoft committed to buy electricity from Helion Energy’s in-the-works nuclear fusion power plant, a company for which OpenAI’s Sam Altman had previously helped raise $500 million; in March 2024, Amazon paid $650 million for a data center conveniently located adjacent to a nuclear power station in Pennsylvania.

Those sums together are well over two orders of magnitude smaller than the $500 billion that OpenAI, Oracle, and two other firms have committed over the next four years to the new Stargate project announced by President Trump only days after he took office this year. For ease of comparison, let’s say that’s $100 billion per year. This is an order of magnitude less than the trillion dollars per year that Aschenbrenner expects would be needed to produce enough computing power for AGI, which would use 20 percent or more of the total amount of electricity currently produced in the United States.

This may seem laughably large, an impossible goal. We are not presently on trend to attain it, that much is clear. In 2024, Goldman Sachs released a report estimating that U.S. power demand will grow only 2.4 percent by the end of the decade, with data centers making up a large part of that demand but AI using only a fifth of the increase. Tech companies are building where the electricity is and the regulators aren’t, planning data centers in Saudi Arabia and the United Arab Emirates. Aschenbrenner would prefer not to see our tech secrets handed over to America’s frenemies, and he both calls for and expects a new Manhattan Project for AGI. Project Stargate, set to be centered in Texas, looks like a step in that direction.

At their peaks, notes Aschenbrenner, the Manhattan and Apollo projects each cost 0.4 percent of U.S. gross domestic product, which would today be about $100 billion — roughly the amount of yearly investment currently committed to Stargate. Is another order of magnitude of funding possible? Probably so. Between the national security importance and the commercial possibilities, a dramatic scale-up in AI investment will likely only be forestalled by something even bigger — a government breakup of Big Tech, a Great Depression, or an act of God like a devastating solar flare, an asteroid, or an earthquake that sends California into the sea.

It’s also worth noting that major advances in quantum computing would make enormous investments in server farms and electricity production unnecessary. In December, Google Quantum AI announced a major breakthrough with its new Willow chip, for the first time scaling up computational power while exponentially reducing the error rate and attaining real-time error correction. And in February, Microsoft went even further with its Majorana 1 chip, which it claims achieves an entirely new state of matter, which in turn allows for creating a new type of qubit, which is a key step toward achieving very low error rates and a practically useable quantum computer. Quantum computers are still far from ready for mass production, but they are now more probable than merely possible.

The second obstacle is making algorithms more efficient. Moore’s Law, describing progress in computer hardware, famously holds that the number of transistors in an integrated circuit doubles about every two years. The research firm Epoch AI has found a similar, though even faster, pattern in the progress of AI software capabilities for both image and language modeling: every eight or nine months, improvements in AI algorithms produce a doubling effect in available compute — by cutting in half the compute required. Many more computer chips and power plants will still be needed, but Aschenbrenner expects software improvements to be just as important as those in hardware. Perhaps they will be even more important, if the success of the DeepSeek-R1 model is not a fluke but rather an indication of what can be accomplished through open source on a limited computing budget.

A common objection to the belief that the trend in algorithmic improvements can continue is that we are running out of unique, quality data. All of these efficient algorithms with high-powered chips have been processing increasingly large training datasets, approaching the size of the entire Internet. But as Aschenbrenner points out, large language models have been doing the equivalent of speed-reading their texts, cramming for an exam the night before and getting ready to spit out answers as quickly as possible when the test comes in the morning. This is part of why LLMs have until recently been bad at math. But better algorithms are still possible using the same datasets.

Programmers are now training LLMs to process texts much more slowly, doing the algorithmic equivalent of carefully working through the practice problems in a math textbook and only then checking the answers, or of talking through a problem with a study-buddy. “In-context learning” allows the LLM to carefully scrutinize a particular subset of data for answering a question.

For example, the endangered language Kalamang on the island of New Guinea has fewer than 200 speakers and almost zero digital footprint, so next to nothing about it was included in Google Gemini 1.5 Pro’s training data. But when Google uploaded a reference grammar and a short dictionary to the LLM’s “context window,” it was able to translate from English to Kalamang as well as human testers could when given the same resources and prompts.

As of yet, there is no way for the LLM to incorporate this in-context learning into its fundamental “weights” or neural network architecture; there is no bridge between short-term and long-term memory. But if and when computer scientists build that bridge, it could lead to the kind of exponential gains that AlphaGo saw from playing the game Go against itself — except that these would be gains in, say, the ability to direct a drone swarm to avoid countermeasures, or the ability to manipulate humans into acting on chatbot advice.

In addition to more physical computing power and more-efficient algorithms, there is a third way in which Aschenbrenner expects AI to advance by orders of magnitude in the coming years. He calls it “unhobbling,” recalling the binding used on horses’ front legs to keep them docile.

In the interest of safety and commercializability, today’s LLMs are substantially hobbled in ways that can easily be undone. For the most part, they are generic chatbots given short prompts and reset after each series of queries. Unhobbling means, for example, customizing the chatbot with detailed information about the history and preferences of the person or company using it. Another example: Publicly available LLMs do not yet have access to the full tools of a computer, let alone a 3D printer. Unhobbling is when AIs are given access to both.

Perhaps the most important issue is what Aschenbrenner calls the “test-time compute overhang.” LLMs break down queries and responses into “tokens,” or fragments of vocabulary and grammar. According to OpenAI, each token covers about four characters, and is the unit for computation and thus for billing by usage. But if software and hardware increases make computation costs go down, then the “compute overhang” can increase, and many more tokens can be used to answer each question.

Most LLMs now sacrifice depth for efficiency, relying on what has been called “System 1”: automatic, intuitive thinking. This is how we get results like ChatGPT insisting that there are two r’s in “strawberry.” But, increasingly, LLMs are being taught to pursue “System 2”: deliberately slow, self-checking chains of reasoning. Currently, the maximum output from a public OpenAI reasoning model (o1) is 100,000 tokens. Using Aschenbrenner’s assumption that a person can think quickly at about 100 tokens per minute, this is equivalent to two normal workdays of very focused work.

Think of a task like “read these seven academic papers, summarize them for me, come up with a new research question based on a synthesis of their findings, and outline a paper that follows up on it,” for which frontier knowledge workers now use AI. But if an AI could use seven figures of tokens per answer in conjunction with the self-play and study-buddy approaches, then it could go from a stream-of-consciousness monologue to drafting, editing, and improving a response so that it represents the equivalent of not two days but a week or a month of one human worker’s effort. This is the path down which Grok 3’s new “think” button, Claude 3.7 Sonnet’s “extended thinking mode,” and similar reasoning functions have taken the first steps.

Speaking of LLMs “thinking” and “reasoning” is where we come to both the deepest objection and the most awful-or-awesome possibility in the trends Aschenbrenner discusses. The objection is that he has implicitly followed a common computer science practice of narrowing down what we mean by “general intelligence” to success on metrics like AP exams, protein-folding problems, and software engineering problems.

This is an important philosophical concern to which we will return. But the practical upshot is that as programmers focus on AI improvements in solving programming problems, and as they succeed in programming successively better artificial programmers that can program themselves, we face what has been called the “intelligence explosion” or, more precisely, a “capability explosion.” If we grant the Machine Intelligence Research Institute’s restrained definition of intelligence as the “ability to achieve goals in a wide range of environments,” then we should expect AI programs to look decreasingly like chatbots and increasingly like agents: having agency, the ability to directly affect the real world.

This will be especially clear in the development of AI-powered humanoid robots: Tesla’s Optimus, Boston Dynamics’s Atlas, Figure’s 02, and Sanctuary AI’s Phoenix. In 2011, Marc Andreessen famously wrote that “software is eating the world,” referring to things like Amazon’s victory over Borders. If Aschenbrenner is even close to right, then we haven’t seen anything close to the possible impact of the digital world on the physical.

Silicon Valley’s old motto was “move fast and break things.” Its new one may as well be “move fast and make things.”

Continue to Part 2: “It Will Seem to Be Alive” ➞

The post Gaining Situational Awareness About the Coming Artificial General Intelligence appeared first on The New Atlantis.

A Future for the Family… : “A Future for the Family: A New Technology Agenda for the Right,” a statement of principles, was published at First Things on January 29. Its signatories include several of our editors and authors. “Stop Hacking Humans,” in this issue, elaborates on its themes.

… and for Dynamism too: On February 24, we co-hosted “Dignity and Dynamism: The Future of Conservative Technology Policy,” a one-day conference at the American Enterprise Institute that explored the themes of the essays “A Future for the Family” and “Stop Hacking Humans.” In addition to two panels that featured several of our authors, editor Ari Schulman conducted a one-on-one conversation with keynote speaker Katherine Boyle, General Partner at Andreessen Horowitz, after her address.

Why the fight against misinformation failed: On December 5, editor Ari Schulman spoke on a panel at the American Enterprise Institute’s “What Comes Next in the Information Wars?” conference. An excerpt of his remarks:

The basic premise of the misinformation paradigm is that we’re confronting a supply problem. And that if you can, through various elaborate means, tamp down on the supply, then you will reduce the problem.

But what is the nature of the demand for misinformation? It is that the people who control the main supply of information — trust is lost in them. Well, the reason that trust is lost is that people believe that the authorities are trying to control what information they hear and how it’s interpreted. Then they find out that actually, as a positive program, people in charge are trying to control what information they hear and how it’s interpreted. That’s exactly what they’re suspicious of. Their demand and their desire for alternative sources of information, including bad information and lies, is going to go up. I think that’s the basic feedback loop that this paradigm has been caught in for the last 10 or 15 years.

Squibs: On the day of its online publication, Alexander Raikin’s “A Pattern of Noncompliance” is featured in testimony to Scottish Parliament ~ Nicholas Carr’s “The Tyranny of Now” is selected as an Article of Note at Arts & Letters Daily ~ Matthew C. Crawford’s Shop Class as Soulcraft, the bestselling book based on an article in our Summer 2006 issue, is included in the University of Austin’s freshman reading list ~ the Bulletin of the Atomic Scientist’s memos to the incoming president endorse the idea of a president’s council on AI from our Winter 2024 issue ~ Robert Zubrin is interviewed on NPR about “What it would take to send people to Mars” ~ the introduction to “How the System Works” is reprinted at The Free Press ~ the debut installments of “How the System Works” receive millions of impressions on social media.

Thanks to you: A reader recently posted, “The New Atlantis has become a must-read for me (yes, I’m late!).” Many others seem to feel the same: thanks to the popularity of our recent work, our subscriber rolls have grown by 50 percent just since our last issue. For readers new and old, thank you for subscribing!

The post Messages in a Bottle appeared first on The New Atlantis.

Reviewed in this article

Boom: Bubbles and the End of Stagnation

Byrne Hobart & Tobias Huber

Stripe Press ~ 2024
304 pp. ~ $35 (hardcover)

The several definitions of bubbles provided in the book don’t leave a reader totally clear on what counts as one. While in general use a “bubble” often has a fraud at its heart, for the authors it doesn’t have to. It can even involve an endeavor as tightly coordinated and official as the Manhattan Project. They offer little discussion of conventional bubbles like the Dutch tulip mania or the blood-testing fraud Theranos.

Instead, what the reader gets is an inside view of belonging to a growing bubble, and of what good the authors think society and the individual can derive from these moments of intense collaboration. Participating in a bubble can feel like being in on a secret, belonging to a sworn confederacy whose actions reinforce each other. Hobart and Huber see “a belief system oriented toward self-reference and self-fulfillment” as core to a bubble’s dynamic. Epistemic closure is a feature, not a bug, as long as you’re closing ranks around some not-yet-widely-appreciated truth. The insular nature of the bubble gives members the chance to abandon persuasion and focus on living out the consequences of their possible insight — for better or for worse.

If the reading experience of Boom seems a little deliberately alienating — in stark contrast to Stripe Press’s beautiful, pick-me-up construction of the physical book — perhaps that’s the authors imposing a “you must be this crazy to enter” filter of their own. Before readers get to consider the authors’ case for bubbles, they’ll need to get through lengthy recriminations over the end of the Bretton Woods monetary system and the rise of fiat currency. Hobart and Huber care deeply about inflation as a risk inhibitor — they feel that a world where money loses value moment to moment leads to a general devaluation of and disinvestment in the future.

A nation of calm, prudent index fund investors (guilty!) is limiting what the authors see as the proper flow of money and talent toward the projects that reshape the world. Hobart and Huber aren’t interested in breakthroughs by a lone genius. They are interested in the achievements that require many different people, operating with limited communication and little coordinating authority, pouring time and treasure into a project that requires them all to keep faith. A bubble, as they see it, is a way of meeting in the cooperate–cooperate cell of a prisoner’s-dilemma matrix.

Hobart and Huber identify Moore’s Law as an example of a straightforwardly positive bubble that required this sort of solidarity to stay true. Gordon Moore, the co-founder of Fairchild Semiconductor and Intel, noted in 1965 that the number of transistors on an integrated circuit doubled about every year and a half, and that he expected this trend to continue. His industry took his claim and reified it to a law. Software makers planned projects that presumed that chips would continue to increase in potency and decrease in price. Chip manufacturers pushed past the present state of the art, trusting that their customers would innovate in a way that required better than the current best.

If it hadn’t been for Moore’s pronouncement, progress might have been more halting. It takes a certain appetite for risk to build something your customers do not yet know to desire.

I can certainly think of places where the lack of this kind of bubble dynamic was costly. America badly needed it throughout the Covid pandemic. In the best form of a bubble, manufacturers of rapid tests, air filters, and vaccines would spin up quickly at the first hint of a potential pandemic, expecting to be met with purchase orders and bounties for rapid innovation. Instead, American mask manufacturers haltingly expanded factories and wound up kicked to the curb the moment shortages abated. Designers of rapid tests were obstructed at every turn and could not make the leaps forward for home diagnosis that were allowed in Europe, where Covid, flu, and RSV could all be screened at home in a single, convenient kit.

In America, there was no flywheel effect, no new operators rushing into a growing market, no parallel innovation. As a result, we have arguably lost ground on pandemic preparedness after a practice bout with Covid. Instead of enjoying the unifying dynamics of the Moore’s Law bubble, America wound up more divided.

Bubbles can be polarizing and still succeed, as long as they unite the critical stakeholders within an industry. Because bubbles represent a strong claim about the future, they create different market dynamics than when buyers and sellers agree on the general shape of the next quarter but are sweating over minor details. As the authors note, a company that will either go exponential or go to zero has its prices set by “trades between the most optimistic investor and the most committed short seller.” For obvious reasons, these bubbles are more attractive to the optimists than the shorts, since even when a collapse seems inevitable, it can be hard for the shorts to stay solvent long enough to enjoy it.

A bubble is the start of an argument. At their most ambitious, Hobart and Huber see bubbles as a kind of attempt at time travel. A bubble’s adherents can see the shape of things to come, and create a rip in the rhythms of time to pull the future toward today faster. Elon Musk summarizes this mindset when he says that “at SpaceX we specialize in converting things from impossible to late.”

A long-term bubble has different dynamics than a pump-and-dump frenzy. It requires sustaining a culture — and leaving a legacy — that engenders trust and willingness to sacrifice. The social dynamics of a bubble can parallel that of religious movements — where charismatic founders are common, and, in the opinion of Hobart and Huber, are more powerful when they die or disappear in their prime. They believe Satoshi Nakamoto’s disappearance in 2012 strengthened Bitcoin and cryptocurrency. “An early departure cements [a founder’s] initial statements as canonical,” they argue, “causing subsequent work to focus on instantiating their ideals in the real world or accurately interpreting their views.”

One of the most basic claims of a bubble is that the future is thrilling, and it’s within reach. Some bubble cultists may expect to see the still-more-glorious dawn themselves, but bubbles can sustain themselves over longer timelines when the participants are parents. As the authors point out, Michael Griffin, a former NASA administrator, has said that the space program is a modern cathedral. Going multi-planetary is too great a task for a single generation; it requires trust that your children and your children’s children will continue to lift blocks with pulley and tackle, or to lift payloads with liquid oxygen.

There has to be a sustaining hope in things not seen to keep this kind of bubble or project running. The authors see this as a mindset of “excess, exuberance, and irrationality” as opposed to one of “cost–benefit analyses, rational calculation, and careful and deliberate planning.”

What could represent the best of both worlds? In venture capital, canny investors aren’t swept away by pitch decks and persuaded that each company they buy into is really going to a billion-dollar valuation. A VC firm assembles a portfolio of lottery picks, expecting nearly all to be a bust, but trying to be smart enough about the odds that the one company that pays off covers their costs and much, much more. Publishing companies take the same approach to contracting with authors — though it’s much more hurtful to admit to a novelist’s face that you expect your modal author will flop.

Venture capital investors, commissioning editors, law school professors steering students to feeder clerkships are all in some sense breeders of bubbles. They facilitate the conditions conducive to maximum effort, but, by holding a diversified portfolio of projects and protégés, are more protected from the burst.

So why plunge into the bubble yourself instead of buying partial shares of many and being indifferent to which ones pay out?

The authors’ answer becomes clearer in the final chapters. The writing becomes more romantic and manic as the references to René Girard and Martin Heidegger begin to multiply. Although the book began with worries that fiat currency and currency devaluation would lead investors to be too pedestrian with the risks they take with their portfolios, ultimately Hobart and Huber want to persuade the reader to risk his or her life.

As the authors argue, each of us is already making a bet on the future in our choice of studies, career, friendships, and so on. They argue that we should have a greater appetite for risk when we stake our time and our sense of identity than when we simply plunk down money. They expect we will have more chances to get in on the ground floor of a fruitful bubble as a worker than as an investor — and that we would have better sources of truth from inside a company than from outside through a brokerage.

But, more than that, they argue that your investment of your self never goes to zero the way a series of stock puts might. “Investments can vaporize in an instant,” they write, “but the memory of being part of a tiny band of people trying to change the world, and feeling for a time like it’s succeeding, is timeless.”

The future might or might not come to kiss the present. The equity you hold might go to the Moon or to zero. But the real pitch for belonging to a bubble is, at least in part, the same one made by Rudyard Kipling in “If.” Holding your past gains more loosely than your future purpose grants you “the Earth and everything that’s in it / And — which is more — you’ll be a Man, my son!”

Boom could easily lead the reader to join the next Theranos. The startup promised to run comprehensive blood tests with only a tiny drop of blood, but when the technology never came together, the founders turned to fraud and threats to keep employees in line. A reader of Boom who was only interested in finding a risk to lean into might do his part in maintaining the front, hoping that another few months of runway coaxed out of investors with fake results would give the company the time to make the breakthroughs it needed.

But a reader who connects with Hobart and Huber’s praise of bubbles as a chance to form character instead, to become worthy of shared trust in the trenches, might instead become the next Tyler Shultz. Shultz, a Theranos employee, blew the whistle again and again on its massive medical fraud. He kept speaking up even when telling the truth endangered his finances and his relationship with his family — his grandfather, former Secretary of State George Shultz, was a board member and a key builder of confidence in the company.

In a way, Shultz moved from one bubble to another. Instead of persisting in hope that somehow the blood-testing product would become a reality, he kept telling the truth, no matter how fruitless it seemed, trusting that an ally would emerge, working in parallel, connected to him only through a shared sense of the future. The best bubble-seeker believes the world might be full of unknown allies, and strives to become the confederate they are counting on, long before they meet.

The post Are We Under-Bubbled? appeared first on The New Atlantis.