When I lecture about American democracy, I often throw in the line: “America is like one big arranged marriage.” My point is usually about the harms of polarization. Imagine a marriage in which the husband starts considering the wife’s insistences about how to load the dishwasher or how many streaming services is reasonable to subscribe to be existential problems—as threats to the household. The same is true for America. Once we start seeing the people we disagree with as problems to be eliminated, the prospects for a functional democracy start looking dim.

But I’d never before considered taking the metaphor more seriously. What might we learn about having better democracies from ideas about how to realize better marriages? Insofar as nations are fundamentally relational commitments, guidance for achieving a more passionate marital life isn’t so far afield as it might seem as first glance.

America, the Fused

Often we hear about the importance of “good communication” and love languages (or whatever) when it comes to having a happy marriage. Yet, most of us probably know a couple who has mastered all the therapy-speak chess moves but are still miserable as hell: “Honey, I’m not feeling seen when I take out the garbage. I’d like to agree on healthier emotional boundaries regarding our domestic waste management arrangement.”

Just as constant relationship talk probably makes for an exhausting and (let’s be serious) neurotic marriage, more political talk isn’t always better. That’s Robert Talisse’s argument in Overdoing Democracy. As any and every action becomes viewed through the lens of the political partisanship, as political discourse becomes our hobby, we cease to be able to sustain civic friendships. Other people become either roadblocks or allies in political projects that most of us, frankly, don’t actually have much influence on.

While I used to think about this pattern in terms of Americans growing apart, I’ve recently started to wonder if we’re really growing too close together.

David Schnarch, especially in his book Passionate Marriage, makes this exact point for committed relationships. Contrary to so much social media self-help fluff, the problem with trouble marriages isn’t so much marital distance but rather emotional fusion. Partners don’t become too different from each other when relationships fall apart but actually fail to differentiate themselves.

This might seem counterintuitive, but it helps to observe the role of anxiety plays in dysfunctional marriages. Emotionally fused couples have learned to accommodate each other’s’ anxieties in order to avoid conflict. A wife reloads the dishwasher, while fuming, rather than either accept a second-rate job or risk a fight with her husband, who happens to be overly sensitive to criticism. Or one partner might react to a proposed idea for a new bedroom activity with fear: “You think I’m boring and inadequate.” Consequently, intimacy suffers as anxiety, rather than passion, comes to characterize the marital bed.

When couples flee from anxiety, they hide from opportunities for growth. Growth and change are scary, because they might mean that partners no longer want the same things from the relationship. The resulting arrangement might be “safe” or “comforting”, but in only the same way that a diet based entirely around McDonald’s chicken nuggets would keep one protected from potentially unpleasant gastronomical surprises.

Schnarch’s books are filled with examples of how this plays out. But the gist is that emotionally fused couples become reliant on each other for managing their own fragile self-esteems. Giving the silent treatment, emotional distancing, or temporary separation might seem like creating “space.” But Schnarch sees these practices as also fusion in action, because the spouses are dedicating considerable energy to avoiding one another and thus conflict. They are just using escape rather than accommodation.

You can see this same relational pattern play out across the political spectrum. Insular leftist sects are intolerant to dissent, partly because disagreement isn’t validating. In my years hanging out with leftist academics, far too many of them see other people’s politics as a threat to the self, and they consequently only hang out with people they already agree with and shun others as ignorant or as “not having done the reading.” And to the extent that Americans have geographically segregated themselves along ideological lines, we’re evermore emotionally fused as a nation.

The flip side of this process: affective polarization, is a kind of emotional fusion via antagonism. Polarized political actors cut their noses to spite their faces, which (politically speaking) can look like “owning the libs” with self-destructive policies. It’s hard to see cuts to Medicaid and the Affordable Care Act that is likely to shutter rural hospitals as anything but this process unfolding.

Whatever its exact manifestation, political polarization signifies an environment in which political opponents have been transformed into political enemies. An increasing portion of Americans and politicians are so emotionally invested in seeing the other side lose, that they have little ability to think about what it would mean for the nation to succeed. They act as if all that matters to them is that politics reflect their own ideological sense of self as much as possible.

Political Differentiation

The answer to emotional fusion, in Schnarch’s work, is differentiation. Couples need independence, not as an end of in itself but rather as a vehicle for interdependence. Again, it’s a bit counterintuitive, but the central idea is that partners ought to learn to better self-soothe their own anxieties. Otherwise, they become emotionally reliant on positive and validating reactions from the partners, and, consequently, they cease to speak openly and directly about the things that matter to them. Until they learn “to get a hold on themselves,” they’re stuck in an emotionally fused gridlock that doesn’t really work for either partner.

Schnarch encourages couples to see their relationship as a kind of crucible. Conflict isn’t something to be avoided but is instead something with transformative potential. Once each partner stops constantly seeking validation from their partner and has some internal stability, then issues can be genuinely confronted. Both partners learn to face their anxieties better and grow up a bit more, and by working through conflict they actually grow closer together in a more honest way. Differentiation paves the way toward genuine vulnerability.

More conflict-oriented theories of democracy, such as Chantal Mouffe’s concept of agonism or Charles Lindblom’s idea of mutual adjustment, see conflict in similar terms, though for different reasons. Lindblom recognizes that knowledge about improving our collective lot is distributed throughout society, and that experts are unable to know what’s best for us and how to get there. Without conflict we simply can’t get a handle on public problems. We need citizens to be differentiated, to be upfront with their concerns, wants, and interests, while also being capable of coping with disagreement in a way that leads to tentative compromise solutions. Without mutual give and take, the process gets gridlocked.

Yet, modern politics seems to push us away from genuine political differentiation and towards fusion. Why might that be?

Anton Jaeger’s thesis regarding modern politics gives us one piece of the puzzle: political institutions have been hollowed out, but we’re at a high point of societal anxiety. Problems like AI, climate change, public debt, the housing crisis, demographic aging, etc. challenge us innumerable ways. Yet, local political organizations and other intermediate associations through which we could be efficacious and work across our divides no longer exist. Instead, we have social media, a technology literally based on the idea that we ought to spend our lives chasing validation from strangers. Social media cannot contain conflict. It cannot push people to grow up by facing real political tradeoffs, to learn to stay calm in the face of disagreement, to not rely on others to validate their beliefs, or to endure through discomfort. But what can?

Designing the Political Crucible

The advantage of a marriage in terms of serving as a relational crucible is that it’s supposed to be permanent. And, so long it’s not an open marriage and partners can’t simply run to third (or fourth or fifth) parties to validate their weak self-esteem during a moment of conflict, both people have a strong incentive to do the work. Commitment isn’t a ball and chain restricting one’s freedom, thus understood, but rather the very vehicle for pursuing self-actualization.

Our inability to collectively actualize as a nation, in our hyperpolitical moment, is due to the fact that our society is increasingly built upon contingent belonging. Our response to conflict is too often to exit. We move to neighborhoods and join social networks with likeminded people. We switch jobs or churches at the first sign of “toxic” behavior. We cut off family members rather than confront them. And, worst of all, far too many of us look for validation outside of our most treasured committed relationships when faced with the need to confront our deepest anxieties (a pattern that the development of AI-powered lovers only worsens).

But it’s not like we can just go back to de Tocquevillian localism. School board meetings and public hearings now often function as proxy battlegrounds for national politics, largely because those with the time and interest to be political active are of the privileged variety who want their ideology validated more than they need a material problem solved.

We’re in a real pickle, so it seems. Maybe bodies of citizens, randomly selected could be tasked to work through public problems, to fearlessly face down their divisions, while also being empowered to make binding decisions. Perhaps bureaucrats tasked with developing more informally negotiated quasi-democratic solutions to public problems could accomplish the same through chains of small group discussions. Yet, these seem like band-aids. What we likely need most is to reform the institutions that occupy the biggest chunk of our waking hours: work, school, family, marriage, etc., to more often guide people into and through conflicts. Our democracy depends on it.

Alas! My piece went clearly unread at the NYT, which re-mongered the fear of AI-generated bioweapons in a recent article reporting on the shock felt by a handful of concerned scientists who shared some transcripts “in which chatbots described how to assemble deadly pathogens and unleash them in public spaces.”

Whether such descriptions translate into any real threat is debatable. As of early 2024, a RAND study had concluded unambiguously that “Current Artificial Intelligence Does Not Meaningfully Increase Risk of a Biological Weapons Attack”—but in the world of AI, two years is a long time. Things may have changed.

The fearmongers, however, want us to believe that AI can help terrorists create a novel pathogen for which no treatment would be available—like COVID-19 in 2020, for example. As one researcher told the NYT, AI “can design beneficial proteins to fight cancer — but also has the potential to invent lethal toxins no one has seen before.”

That sure is a frightening possibility. But I’d still weight the probability of it happening at close to zero.

The space of all possible pathogens is Vast. The space of novel pathogens capable of producing a pandemic, however, is Vanishingly small. So, while not impossible, it seems highly improbable that even AI-assisted terrorists are going to find a brand-spanking-new bug to kill us all with.

We know this because the US government has spent hundreds of millions of dollars funding scientists engaged in precisely this kind of work, for decades.

It’s called “gain of function” (GoF) research, the aim of which is to increase the lethality, transmissibility, and other dangerous qualities of pathogens.

And yet, despite the most favorable of conditions—clean labs, expert knowledge, organizational resources, ample funding, and government license to cook up pandemic-level pathogens—the GoF enterprise, as one of the scientists quoted in the NYT piece above, MIT professor Kevin Esvelt, testified to a Senate committee, has not “succeeded in producing any credible threats.”

Unless, that is, SARS‑CoV‑2 was created in a GoF research lab.

Share Taming Complexity

Pandemic Failure is our Model

Just this month, the US Department of Health and Human Services (HHS) started formal debarment procedures against GoF pioneer Ralph Baric.

Baric, along with Peter Deszak, formerly of EcoHealth Alliance, and Shi Zhengli, former head of the Wuhan Institute of Virology, is the third member of the trio at the center of the “lab leak” controversy. After being suspended a few months ago, Baric is being debarred for a “pattern of deception” in his dealings with the US government, which funded his virological work with tens of millions of dollars.

Deszak had already been fired by EcoHealth Alliance before both he and the organization were debarred last year for similar reasons. He and Baric appear to be among the only prominent scientists connected to the Wuhan-related GoF controversy to face formal federal sanctions or debarment proceedings.

And Shi Zhengli, now at the Guangzhou National Laboratory, apparently unscathed and undeterred, continues to do GoF research on bat viruses.

At this point, the origins of COVID-19 remain shrouded in controversy. Prominent virologists still insist the scientific data supports the natural emergence hypothesis via “zoonotic transmission.” As I argued last time, however, the preponderance of circumstantial and historical evidence indicates SARS‑CoV‑2 was leaked from the Wuhan Institute of Virology, where GoF research on bat coronaviruses was being conducted with inadequate biosafety precautions.^[1]

In any case, we definitely know that in terms of governance, the GoF field was a powder-keg and the community’s response to the pandemic was a catastrophe.

Why am I rehashing this story? Because it sets up the point I want to make here:

The governance architecture that failed to prevent a viral pandemic is paralleled in the field of AI, where a similarly consequential disaster is waiting to happen.

Subscribe now

GoF/AI: Parallel Recipes for Disaster

Here I’ll identify some of these structural problems by drawing out parallels between GoF virology and AI R&D to show how in both cases, the system of scientific governance allows an elite few to put the whole of society at risk.

The Perverse Efficiency of High-Risk Innovation

Let’s start with a familiar problem: the “publish or perish” culture of academia. In virology this pressure creates a logical preference for GoF research. Compared to antiviral drug research, GoF is cheap, fast, and almost certain to yield publications in high-impact journals, which tend to favor novel, attention-grabbing results.

For career-minded academics then, working on GoF is practically a no-brainer. A junior researcher can complete multiple GoF studies in the time it takes to see a single antiviral candidate through early Phase I trials, and be virtually guaranteed to see them all reach publication.

The same structural incentive to “publish or perish” is active in AI. Here, AI capability enhancement research—improvement of a system’s speed, accuracy, efficiency, and other features that make it “intelligent”—parallels GoF.

AI capability enhancement research leading to “state-of-the-art” (SOTA) results proceeds at a blistering pace. Benchmarks intended to last years are now often overtaken in mere months, paralleling the GoF timeline quite closely. As in GoF, where tinkering with an existing virus to make it more transmissible (e.g. adding a furin cleavage site) is relatively uncomplicated, incremental performance improvements in AI systems are comparatively easy to manufacture with architectural tweaks and compute-efficient training. In this highly competitive arena, even a 2% gain on a SOTA benchmark is worthy of publication.

Moreover, as in GoF, novel AI capabilities are prized by top-tier conferences (e.g. NeurIPS, ICML, AAAI—computer science doesn’t really do journals) and, perhaps more importantly for the industry, gather media attention. Papers showing a model can pass the Bar Exam or a Science Olympiad often have immediate viral impact and garner high citation counts, regardless of real-world utility.^[2]

By contrast, AI ethics/safety research—making systems more robust, reliable, and (potentially) risk-free—parallels antiviral research in that it is slow, uncertain, and poorly rewarded. Whereas developing a new frontier model and shipping it to hundreds of millions of users can generate massive commercial returns within months, AI ethics research and robust safety evaluation are long-term investments with uncertain payoffs and, most tellingly, no direct revenue generation.

The Safety Theatre of Self-Governance

In reading through the post-pandemic investigations into COVID-19 origins, I was surprised to discover that international biosafety standards are voluntary and unenforceable. Moreover, that oversight was conducted by institutional committees composed of the researchers’ own peers—people embedded in the same incentive structures, sharing the same disciplinary assumptions, and personally invested in the continuation of the enterprise.

In elite circles, this is called “self-governance.” It means the scientists creating the risks were also setting the rules for managing them. Where I grew up in Indiana, this is called “the foxes guarding the henhouse.”

The AI industry, notable for its many voluntary safety commitments, enjoys the same dynamic. The much-lauded responsible scaling policies, model cards, and safety evaluations are all made by the companies developing the very models being evaluated. In this case, letting the foxes guard the henhouse is justified on the grounds that to do otherwise might “stifle innovation,” The job of the foxes is thus to create “safety measures” that do not slow the pace of AI capability enhancement.

Where the rubber hits the road, the AI industry’s commitment to practicing this form of self-governance has proven flexible, however, often aided by the elastic narrative of an “AI Arms Race with China.”

For instance, by now we all probably know the story of OpenAI: it starting as a nonprofit, safety-first AI lab before transmogrifying into a thirsty for-profit Microsoft partner and defense contractor. At this point, it isn’t a stretch to say the company has stepped back every commitment to safety it made. If this is news to you, the story is now the subject of public record thanks to a failed legal case against them.

But when Anthropic, regarded by many as the most safety-conscious major AI lab, revised its Responsible Scaling Policy to remove its core safety commitment—a pledge not to train AI systems with capabilities beyond the company’s capacity to control—many in the AI community were shocked and dismayed. They had bought into the theory, prevalant when the company first published the commitment in 2023, that such measures could catalyze a “race to the top” by incentivizing competitors to demonstrate still stronger safety commitments.

Instead, Anthropic’s self-governance measure collapsed under competitive pressure. As the company’s Chief Science Officer explained: “We didn’t really feel, with the rapid advance of AI, that it made sense for us to make unilateral commitments … if competitors are blazing ahead.” (China!)

Sure, in terms of resources and time, safety is expensive. And of course, less safety is cheaper and faster. But it was this same logic that drove the Wuhan Institute of Virology to conduct chimeric coronavirus research at biosafety level 2 (BSL-2), described by one microbiologist as “the biosafety level of a standard US dentist’s office.” And that choice was the result of mere academic career incentives.

How reliable could self-governance ever be if the foxes are taking home million-dollar paychecks for guarding the hens?

Share

Manufacturing Unanimous Consensus (For Those with the Most to Lose)

When the cozy relationship of self-governance by experts/for experts breaks down, the GoF and AI enterprises respond in the same way: by suppressing dissent.

Soon after the pandemic started, the GoF community mobilized in lock-step to suppress inquiry into its potential culpability. Before any forensic investigation had been conducted, top scientists convened private meetings to protect their own by advancing an ad-hoc theory: the virus emerged from “natural” zoonotic transmission through an intermediary—maybe a pangolin. The lab-leak hypothesis was quickly framed as “conspiracy theory” and forced outside the bounds of legitimate scientific discourse. This aura of scientific consensus marginalized dissent for years.

By contrast, in the AI industry, a division of labor between capability enhancement and ethics/safety research does exist, and, at least in theory, this could fortify the enforcement of self-governance. The economic reality is, however, that AI ethicists and safety teams operate within companies whose revenue stream depends on shipping the very capabilities those teams are meant to constrain. It’s no surprise, therefore, that when the ethics/safety work conflicts with the bottom line, it loses every time.

Perhaps unsurprisingly, a Black woman was the first major example of such a loss. Timnit Gebru was fired from her post as co-leader of Google’s “Ethical AI team” back in 2020 for identifying how biases were being built into the company’s AI systems. The next year, another co-leader of the same team, Margaret Mitchell, was fired after co-authoring the famous “Stochastic Parrots” paper with Gebru and others, which argued LLMs are not, in fact, intelligent.

Since then, the dismissals and dissolutions of AI ethics and safety teams have come so fast and been so numerous one cannot help but wonder if the entire field of research was ever anything but a house of cards. Between 2024 and 2026, OpenAI disbanded not just one, or two, but three such teams, with a leader of the “Superalignment” team lamenting that “safety culture and processes have taken a backseat to shiny products.”

At present, the Ethical AI Departures tracker documents close to sixty researchers and executives who have either quit or been fired from major AI labs over ethics and safety concerns. “Safety Deprioritization” and “Team Dissolution” by the industry are the two most cited reasons, with “Inadequate Oversight” a close fourth. Whereas the virology community suppressed biosafety dissent through orchestrated letters in high-profile journals and behind-the-scenes meetings with top government officials, the AI industry typically utilizes the ordinary institutional mechanisms available to corporations—namely, firing—to systematically weed out the ethically-aligned and safety-conscious.^[3] That is, anyone who would stand in the way of AI capability enhancement.

The results of all this are predictable: AI capabilities are outrunning the safety infrastructure.

You Can Just Leak Things

I won’t get into the details of the COVID-19 case here (check out the previous post for that), but suffice to say, lab leaks happen far more often than most people realize.

Over a decade ago, researchers calculated that over a multi-year timeline, accidental release or researcher infection of a potential pandemic pathogen is a statistically predictable event—not an anomaly. In other words, it’s only a matter of time. Worse, this is unlikely to change because the main cause—human error—is inevitable, even in the lab. Furthermore, the situation is exacerbated by the lack of a centralized, mandatory, internationally maintained database of such events, which obscures their actual frequency.

Human error is no less inevitable in AI, where major leaks recently grabbed headlines worldwide. First, in March, the entire code base for Claude Code—the popular extension for Anthropic’s Claude chatbot that gives it control over your local machine—was leaked. Within hours, it had been copied across the internet tens of thousands of times, ensuring it could never be taken down by legal means. By the next day, the entire had been “clean-room” re-written in Python, making it not only easier to modify for those so inclined, but legal to own and use.

The leak also contained an “Easter egg”: documentation of an as-yet unreleased model, “Mythos,” that was described as the first of “an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the efforts of defenders.” In other words, AI capability enhancement is entering “super-hacker” territory. If realized, the capacity to penetrate any system not hardened against it would make extant cybersecurity essentially obsolete.

Second, in early April, Mythos was announced—but not released to the public. Its unprecedented ability for finding and exploiting security vulnerabilities in software meant it could be used (by terrorists!?) to launch devastating cyberattacks. Deeming it “too dangerous” for a general release, Anthropic said it would only offer access to about 40 corporate partners via a consortium. As a demonstration of its dedication to safety and responsibility after the embarrassing Claude Code leak, Anthropic would allow these consortium members—Microsoft, NVIDIA, JP Morgan, Apple, the usual suspects—would use the model to patch up their own systems before the eventual public release.

But Anthropic apparently failed to use Mythos to harden its own defenses, because on the very day it was publicly announced, the model was accessed by a group of hackers. Weeks later, Fortune reported that while the group “has not been using the model for cyberattacks, it has been using the program continuously since its release and still has access.” Cybersecurity experts chimed in that if thousands of employees at several dozen large firms and a random group of hackers all have access, then China (China!) almost certainly has access. And soon, we will too, because the AI system that was “too dangerous to release” is scheduled to be released to the public—with no hint as to why that is now a safe move.^[4]

Thus far, the only harm these leaks have done was to Anthropic’s bottom line and reputation. But consider that like SARS‑CoV‑2, leaked systems become part of the broader ecosystem. Existing defenses become useless. You and I, as third parties, have no choice but to adapt. And that might mean having no choice but to buy a Claude subscription from Anthropic to harden cybersecurity on…every website? Every computer network? Every software package?

It isn’t hard to imagine a scenario where the coming wave of AI super-hacker systems create dependence on a single provider—Anthopic—for safety. Forget to pay your dues one month, and your organization’s website gets penetrated, pwned, and ransomwared for millions of dollars by hackers using the model released next month. Everyone, from individuals and small businesses to national governments would subject to this new safety tax, with larger organizations on the hook for larger fees, given their larger attack surface.

Share

Scaling the Inevitability of Failure: BSL-4 Labs and Data Centers

Around the world, the number of BSL-4 labs, which work with dangerous pathogens at the highest level of containment, is “booming.” Having climbed from 59 to 67 labs between 2021 and 2023 in a post-COVID surge, a total of 69 BSL-4 labs are known to be in operation as of 2026, and with about 20 under construction, approximately 90 will be operational by 2030. On top of that, West Point’s Combatting Terrorism Center notes an “upward trend in high-risk research” despite “significant gaps and weaknesses globally” in biosecurity management.

In parallel, data centers are undergoing massive, unprecedented scaling. Big corporate consultancies project that over the next four years, the current estimated global capacity of 115 GW (data centers can be measured in gigawatts of energy usage) is predicted to double to between 200 and 240 GW, driven by an anticipated spend of nearly $7 trillion on data center infrastructure by private companies.

Three parallel misconceptions about both BSL-4 labs and data centers are notable.

First, proximity. You might think that BSL-4 facilities are isolated in remote deserts or military bases. In fact, over 75% of BSL-4 labs are located in densely populated urban centers. (Wuhan, China!) Similarly, massive data centers are heavily concentrated in metropolitan clusters (e.g., Northern Virginia, Atlanta, Dallas) to minimize data latency and access existing fiber-optic lines. Consequently, lab leaks are in close proximity to human transport networks, while AI leaks are in tight physical proximity to the human capital, power grids, and networks that run critical societal infrastructure.

Second, governance failure is a common feature. You might think dangerous BSL-4 labs would be highly regulated. In fact, construction has drastically outpaced international regulatory standards. The researchers behind the Global BioLabs tracking site noted that, “Only one-quarter of countries with BSL-4 labs score well on best practice indicators for biosafety and biosecurity,” and while a few countries have dual-use policies, “none have yet signed up to a new international bio-risk management standard.”

Regulation of data centers is likewise minimal. Buildouts are governed primarily by local zoning laws and grid capacity allocation, while the safety profile of the models being trained inside them plays no part. In addition to the issues with AI ethics/safety raised above, at present, no domestic or international framework regulating the hardware layer of AI training exists.

And third, size matters. You might think of a “high security biolab” as an imposing building, but roughly half of all global BSL-4 labs are remarkably compact: 200 square meters or less, not quite the size of a standard tennis court. Not only does this small footprint lower the financial barrier to entry, allowing smaller academic or localized government institutions to handle lethal pathogens, it severely complicates international verification and non-proliferation auditing.

And while the media focuses on huge “gigawatt-scale mega-clusters,” advancements in algorithmic efficiency, model quantization, and compact liquid-cooled server racks mean that not every modern data center need be massive. Today, immense cryptographic and generative AI power can now be packed into surprisingly small spaces and distributed across smaller modular data centers, significantly complicating global compute tracking.

This infrastructure “boom” of BSL-4 labs and data centers dramatically spikes the baseline probability of an accidental, catastrophic leak. Creating dozens of new physical nodes—containment doors and airlocks in labs, hyper-scale server clusters in data centers—populated by human operators prone to error, guarantees that a leak is no longer a matter of if, but when.

The Fundamental Asymmetry

This all raises an important question: Why?

Defenders of GoF research claim identifying potential pandemic pathogens could help save lives. But a strong case has yet to be made, despite the widespread acceptance in the scientific community. Multiple independent, quantitative risk analyses have concluded that the costs, in terms of increased pandemic risk, outweigh the actual historical benefits.

In 2014, researchers analyzing the tradeoffs found that safer, alternative scientific methodologies exist for every stated benefit of GOF. These techniques were ostensibly employed when Moderna created its COVID-19 vaccine in just two days, without explicitly utilizing GoF research.

Moreover, a 2016 NIH-commissioned 1,000-page independent risk-benefit analysis found that any purported benefits of GoF research inevitably require a host of supporting infrastructure to be realized, including: surveillance systems,sequencing capacity, vaccine manufacturing, regulatory pathways, public-health infrastructure, and international information sharing. These nontrivial “barriers to realization” mean that all potential benefits of GoF research are uncertain. The risks, however, are very real. As MIT professor Esvelt, quoted above, told Congress:

Even if identifying a pandemic-capable virus in nature allowed us to perfectly prevent that virus from causing a natural pandemic, and we could do so with zero risk of accidents, we should expect terrorists to use the same virus to kill a hundred times as many Americans as would be saved.

An inverted parallel holds for AI.

In your hands or mine, AI may be fun, convenient, and useful. But the benefits we enjoy will never come close to the world-changing powers the same systems grant political and economic elites, who will use them to make decisions of global consequence.

Moreover, We don’t need to deny the real, though middling, benefits AI provides to us ordinary users to simultaneously acknowledge the multifarious risks it poses to society as a whole: undermining democracy, accelerating the murder of civilians in war zones, wiping out jobs, proliferating scams to rob the elderly, cratering teens’ mental health and making homework obsolete, inducing psychosis, and of course, causing human extinction, as many prominent AI researchers regularly warn us about.

And now, as a father of a toddler, I’ve been made keenly aware of a fresh risk: the bevy of AI-powered kid’s toys now being brought to market, entirely unregulated.

It is hard not to conclude that something significant needs to change, and soon.

The Price We Pay For Their Progress

So, what is going on with the hyped-up fearmongering around AI-generated bioweapons? And with the safety theatre being performed around them?

Simply put, it’s a distraction from the truly catastrophic threat we face: persistent, widespread, structural governance failures combined with rapid expansion and acceleration of the GoF and AI enterprises.

Incentives that reward risky capability research over safety, voluntary self-governance by self-interested parties, suppression of dissent, and the relentless drive to expand potentially dangerous infrastructure add up to drastically multiply the surface area for inevitable human error, while minimizing the structures available to prevent that from happening and to keep us safe once it does.

Ultimately, both GoF and AI exhibit a fundamental asymmetry: The catastrophic risks thrust upon society by an elite few—who operate in a reward system alien to most of us—vastly outweigh the uncertain, often middling benefits delivered to the public.

Without addressing these structural governance problems by building out a practicable, international, mandatory safety infrastructures and more importantly, renegotiating or outright rejecting the value now placed on capability enhancement (Innovation!) in pathogens and AI models, society will remain exposed to failures whose consequences may only become apparent after it is too late.

Leave a comment

Shunryu Garvey is the head priest of Keigakuji Temple in Niigata, Japan, a former Stanford University researcher in AI governance, and the author of the forthcoming Zen and the Art of Democratic Intelligence. Subscribe to follow and support his work with Taming Complexity.

Footnotes

^[1] Deszak, Baric, and Shi were collaborators on the now-infamous DEFUSE project, which proposed to design a novel pathogen with specific features found in SARS‑CoV‑2. As subsequent Congressional investigations revealed, when Baric learned of Shi Zhengli’s intentions to use biosafety level 2 (BSL-2) for the project, he wrote:

IN [sic] the US, these recombinant SARS CoV are studied under BSL3, not BSL2, especially important for those that are able to bind and replicate in primary human cells. In [C]hina, might be growin [sic] these viruses under bsl2 [sic]. US researchers will likely freakout.

Baric then edited their shared research proposal to indicate BSL-3 would be used on the project. That didn’t happen.

^[2] With the discovery of “emergent abilities” in large-language models (LLMs), this novelty incentive has gone into overdrive. Here the simple act of using more compute to scale the model can produce the sudden, unexpected “emergence” of new capacities, like a jump from 0% to 80% accuracy on a task the model was not explicitly trained for. The 2023 “Sparks of Artificial General Intelligence” paper from Microsoft Research leaned heavily on the novelty incentive, packing 150 pages of qualitative examples of an early GPT-4 model tackling difficult tasks across mathematics, coding, vision, medicine, law, and psychology.

^[3] That said, AI critics have long been derided by its advocates, and at least in the case of Timnit Gebru, corporate leadership facilitated a public harassment campaign that lasted for months.

^[4] It’s worth noting that shortly after the Mythos announcement, OpenAI released an equivalently powerful/dangerous model, ChatGPT 5.5, with no warnings attached, leading many to speculate the “too dangerous” label was a clever marketing ploy.

Studies cited in tables

1. Du, Shaoqing, Xueping Hu, Ping Li, et al. “Antiviral Drug Discovery and Development: Challenges and Future Directions.” Signal Transduction and Targeted Therapy 11, no. 1 (2026): 69. https://doi.org/10.1038/s41392-025-02539-7.

2. Puhl, Ana C., Thomas R. Lane, Fabio Urbina, and Sean Ekins. “The Need for Speed and Efficiency: A Brief Review of Small Molecule Antivirals for COVID-19.” Frontiers in Drug Discovery 2 (March 2022). https://doi.org/10.3389/fddsv.2022.837587.

3. Ward, Derek J., Edward Hammond, Luan Linden-Phillips, and Andrew J. Stevens. “Trends in Clinical Development Timeframes for Antiviral Drugs Launched in the UK, 1981–2014: A Retrospective Observational Study.” Pharmacology and Therapeutics. BMJ Open 5, no. 11 (2015): e009333. https://doi.org/10.1136/bmjopen-2015-009333.

Leave a comment

A recurring theme in my writing and research is the relative rarity of persuasion, especially for high-stakes and conflict-ridden issues. Why do people embrace apocalyptic thinking about their health, flat out rejecting suggestions that may actually help them? Why do climate activists double down on cataclysmic predictions, even when it doesn’t seem to have worked? Or consider, as I had a recent class do, that we’ve seen “follow the science” rhetoric fail repeatedly over a century of conflict over vaccines, yet many doctors and public health officials defaulted to that during the pandemic. And this occurred despite no shortage of people, like myself, shouting for people to pay attention to the historical and sociological evidence. Why is it that, when it really matters, many (if not most) people can’t seem to change their mind?

No doubt that I’m far from the first person to be puzzled by our cognitive stubbornness. Though, I’ve been reluctant to think too much about it, because my academic training pushed me to see ordinary people’s reasoning about their lives as “lay expertise,” which—to be frank—tends to get treated as sacred in my field. That said, I don’t like explanations coming from people who see themselves as paragons of rational thinking, who think they can easily deduce which people are engaging in “motivated reasoning” and which are not doing so. I have yet to run many analyses of belief and wrong-headedness that doesn’t let one group or another off the hook.

A great deal of experts fall prey to naïve realism, as Jeffrey Friedman observed. Economist Bryan Caplan wrote a whole book about the irrationality of voters, but understood rationality as nothing less than wholeheartedly agreeing with the free market policy proposals that he took to be commonsensical. He argues that “basic economic truths are obvious to any adult of normal intelligence who calmly considers them,” insisting that anyone who rejects them must be indulging their emotion over rationality. One would no doubt here something similar, albeit for different policy proposals, when talking to an avowed Marxist or progressive.

Any theory of belief worth its salt would have to contend with the reality that most of us are mostly wrong about most things. The state of disagreement about most important political issues tells us as much: We can’t all be right. Even if most of us probably have something important to contribute to understand or resolving public problems (admittedly some more than others), on the whole we as individuals, even as experts, generally don’t know what the hell we’re talking about, but yet still cling to our limited arsenal of facts and beliefs. In other words, I’ve started to wonder if talk and persuasion is really what democracy is all about.

Belief or Credence?

Two years ago, I read Neil Van Leeuwen’s book Religion as Make Believe. I wasn’t sure what to make of it at the time, but I think it offers a good starting point. Van Leeuwen tries to isolate the specific features that distinguish two particular kinds of beliefs: factual ones and religious credence. Where Van Leeuwen deviates from the typical intuition about this distinction is by framing it in terms of cognitive attitude rather than the content of those beliefs per se. One can have false factual beliefs, for instance, and also by implication follow a true credence. In any case, for Van Leeuwen, it is a person’s orientation to that belief that is decisive.

Religious credence is characterized by Van Leeuwen as being voluntary, invulnerable to evidence, not used to guide practical action, and tending to be compartmentalized to specific settings. Religious people very often are aware of the empirical limitations of their credence. Indeed, voluntary belief in the absence of evidence is what many people’s understand by the word “faith.” At the same time, as much as some Christians, to take one example, might try to pray away an illness, they usually still go to the doctor. Similarly, Van Leeuwen cites how a traditional rainmaker scoffed at an anthropologist’s suggestion that a rain dance would be done any time they needed rain. He recognized that the act was a supplement, not a replacement for practical action in response to drought. Moreover, for many religious people, their credence guides their behaviors on Sunday, within a holy site, or during specific rituals, but much less reliably when doing their taxes or at work.

Factual beliefs involve taking the opposite attitude. They aren’t necessarily true, but are involuntarily updated in the face of contrary or supportive experiences or information, even if a person unaware or lack a full understanding of the relevant science. People use factual beliefs to decide which route to take home after work or to decide how make dinner or which retirement program to choose. Factual beliefs influence one’s behavior everywhere. And human beings usually can’t will themselves into not believing them. Upon recognizing the reality of something akin to gravity or the hardness of concrete, they won’t act as if a dropped object won’t fall or that running into a wall won’t hurt.

As such, Van Leeuwen sees factual beliefs and religious credence as two distinct “maps” that people use to understand the world. But the influence from one upon other works only one way. Factual beliefs can direct or supersede a credence, but not vice-versa. Someone partaking in holy communion invariably knows they are drinking wine (or grape juice) even as they simultaneously believe that they hold the blood of Christ in their hands. Likewise, Jesus walking on water serves as a credence only in light of believers’ factual belief that humans ordinarily sink in water.

Clearly Van Leeuwen sees factual beliefs as important for practical action, for navigating reality. But what are credences for if they don’t do that? Binding group identity. This is where the title of Van Leeuwen’s book comes into play. Credences are “make believe.” It isn’t that people happen to randomly share a credence and then form groups. They develop credences so as to be able to create and sustain communities. A Christian believes that they drink the blood of Christ at communion, because doing so binds them together with their fellow church members. You believe in order to belong, not the other way around.

In Van Leeuwen’s view, credences themselves are not deceptive per se, but rather equating them with factual belief, blurring together the two maps that are otherwise distinct cognitive attitudes, is what is misleading. And this is how Van Leeuwen distinguishes a “normal” religious person from a fanatic. It is only the fanatic who extends a religious credence far beyond its normal compartment to encompass the entirety of reality, who acts as if religious credence were on its own sufficient to achieve practical aims. An example would be someone who actively prevented people from going to the hospital, because they believed doing so was a rejection of God. In this case, subverting practical action is exactly the point, because its extremity is an unequivocal sign of group loyalty.

When the Two Maps Collapse

Rather than serve as a definitive model for demarcating factual belief from credence, I think Van Leeuwen’s model is more insightful when seen as presenting two ideal types. There are no doubt examples of the two maps in action, but much of reality (the parts that really matter) are less clear cut. It can be hard to know whether people’s ideas about climate change or their own health are really rooted in chains of factual belief and when/how credences insert themselves.

Consider apocalyptic environmentalism. Apocalyptic beliefs are never falsified, only revised. Apocalypticism is groupishness on overdrive. It is driven by a desire to see oneself and one’s fellow beliefs as special people, chosen in light of their ability to correctly read the signs of upcoming doom. To repeatedly cite extreme and unrealistic climate scenarios, to point to the growing cost of natural disasters, while ignoring the contribution of increasing GDP, signals membership in this chosen people. That’s why those talking points persist.

Yet, in light of Van Leeuwen’s model, you would expect climate doomists to recognize that their cataclysmic claims are (at root) credences. Apart from Genevieve Guenther’s unusually candid review of The Uninhabitable Earth, such admissions are rare. The implication, I think, is that people often believe that they are rationally updating factual beliefs without noticing how groupishness is at work in the background.

Consider the systematicity that the Salem Witch Trial judges applied to their investigation. They looked for physical manifestations or “Devil’s marks” and sought out “spectral evidence,” when witnesses saw the spirit of the accused in their dreams. They assembled what they saw as factual evidence, seemingly oblivious of the credences that held it all together.

The danger, contra Van Leeuwen, isn’t simply in fanatics that extend their credences into the world of practical action, but in situations where we struggle to know the difference at all, where credences unavoidably intervene in our effort to understand reality and discern the right practical action.

And I think that this isn’t atypical but rather normal of most contemporary challenges. One of my conclusions after having written a book about the politics of biodiversity is that I’m not sure it’s possible to have any opinion about conservation without guiding credences about nature and society. People who have faith in the resilience of nature can find plenty of supportive examples. And so can those who think that the environment is inherently fragile, such as when a climate doomist reinterprets nearly all instances of extreme weather as signs of the coming planetary apocalypse. Our experience of the world isn’t unambiguous enough to enforce any particular storyline about nature. And, at the same time, scientific data is so plentiful and varied to be able to seemingly corroborate whichever one we arrive at. In a sense, our various (religious and non-religious) credences are the glue that holds together the tiny shards of factual understanding that we each possess.

At the same time, if this groupish behavior isn’t a subversion of our nature as human beings, but rather an expression of the fundamental character of a social animal, then we can’t see the dominance of credences in our political thinking as something that can be excised. And this poses problems for how we normally think about democracy. Typically, democracy is cast as a kind of government by discussion. And many pro-democracy people hold their own kind of credence about it, that if we could just create the right kind of deliberative forums, equalize economic power, etc. then a sort of popular will can emerge, or at least durable compromise agreements.

A recent controversy over a proposed data center in my town reminded me of this. The problem with the proposal is that it is technologically utopian and the company has no track record of success. The proposal names both atmospheric water farming and small modular nuclear reactors as part of its system, neither of which the firm has had any real experience with. But the public outrage doesn’t center as much on these facts as on claims about water usage and energy prices. Simple back of the envelope arithmetic shows that the million gallons per day water requirement, supposing that it came from local supplies rather than the atmosphere, is on par with what a medium sized farm or refinery would use. If project produced all its own electricity, impact on local energy prices would be slim to none.

To me, the repetition of easily disputed factual claims signals that the debate has been captured by a credence, especially when stronger factual objections are available. Opponents in the public are essentially saying, “These data center people are not a part of our community. They can’t be trusted and will bring us harm.”[i]

That’s not to denigrate opponents. Political speech can be factually dubious without being illegitimate. Companies working in the data center space need to get their acts together and figure out how to persuasively act like neighbors and community members rather than appear to be handmaidens to Silicon Valley interests or speculative investors. Public opposition reflects a real failure on their part.

The trouble is that once the situation gets this far, collective problem solving is undermined. When credences dominate public interactions, debates become fanatical and problems suddenly seem zero sum. “If the data center is using water, it must be at the expense of local farms/environment” becomes a taken-for-granted belief. Dubious factual claims become akin to catechisms, repeated in order to signal group belonging, that hinder deliberation rather than help it. Any imaginative energy that could be directed toward thinking up what it would take for a data center to be compatible with the needs of a small desert community is spent on conjuring apocalyptic nightmares.

Even if the process is a democratic success in that the public feels heard and that they have an influence, the broader impact is further inflaming divisions, hardening the credences that separate different parts of the population. The locals get seen as hick yahoos who fight progress. The image of AI investors as nothing more than vultures preying on the water and energy resources of economically suffering towns gets reinforced. The result is that productive negotiation regarding data centers becomes harder, not easier.

Democratic Bureaucrats?

High stakes discussions aren’t productive unless we’re able to keep groupishness from running amok. Juries are example of how we accomplish that on a small scale. Bring together a dozen people from all of over the community, refashion them into a group, and hopefully you’ve controlled for the various credences that jurors came in with.

Yet, juries aren’t really scalable. No doubt that there are “citizens’ assemblies,” and other exciting experiments in deliberative democracy. But I think there’s a real need for democratic politics to work without requiring so many evenings from citizens. There’s something to John R. Hibbing and Elizabeth Theiss-Morse’s observation that many Americans appear to want a kind of “stealth democracy” that functions without their involvement.

While on its surface that attitude suggests that citizens want a well-functioning technocracy, or rule by non-corrupt, well informed, and unbiased experts, that’s not the conclusion that I draw from it. First, such an ideal technocracy is utopian dream. Second, I’m no longer convinced that bureaucracy and democracy are necessarily opposites.

Michael Lipsky long ago recognized that street-level bureaucrats effectively enact their own policies. Think of teachers who get to interpret, at least until recently, what implementing state learning goals meant, police officers who choose which deviant acts to ignore and which to crack down on, or unemployment bureaucrats who can guide an unusual case through the system quickly or let it languish. Bureaucracies are inherently political. But the traditional tendency to see expertise and politics as opposites have led us to neglect the exploration of how bureaucratic politics itself could be made democratic.

This leads me to wonder what it would look like if bureaucrats were trained and tasked with not employing expert knowledge but with facilitating democracy. What if planners in charge of zoning saw themselves as less in the business of executing the codes on the books, and more in terms of devising solutions that explicitly brokered between NIMBYs and YIMBYs (and were enabled to do so)? They might see their jobs as discerning where deviances from the single-family residential zone were desired, or at least more tolerated, and enabling experiments with different urban designs. Insofar as democracy is about resolving conflict and representing most people’s interests, then this kind of action isn’t really technocratic. It’s a kind of public negotiation, but one that happens through action more so than discussion. It’s a way of lessening the hold of credences, which too often insert themselves to short-circuit negotiation and make even the smallest, most incremental policy changes nearly impossible.

Rather than focus on persuasion and agreement, the point is to enable disagreement to be turned into small changes, experiments that might push at least a portion of the population to rethink their factual beliefs about urban space, data centers, or whatever. No doubt that it wouldn’t sway those who hold most firmly onto their credences, but if it can cut down on the fanaticism on display in many public hearings, I’d see it as a win.

Yet there are obvious concerns about this new role for bureaucrats. Who holds them accountable and how? How can we be sure that they’ve brokered fairly between opponents and haven’t been captured by one of them? No doubt that there’s always a need for democracy to be rooted in discussion-based forms of accountability, but we also need a political system that doesn’t so often devolve into group identity-driven bickering and actually gets stuff done.

[i] A public hearing about the closing of the university pool was remarkably similar, where it was clear that it was people’s emotional attachment to the old pool that was driving outrage, but the discussion centered on incredulity on the easily confirmed figures about low usage/revenues and the high costs of maintaining and renovating a several-decade old structure.

Leave a comment

You can read a bit more of the first chapter at Google Books’ preview.

If you’ve followed my past writing, both here and at outlets like The New Atlantis and The Ecomodernist, you’ll recognize some of the arguments. But there’s a lot that I haven’t shared here, especially a number of fascinating stories of conflict and disagreement from the world of biodiversity conservation, which illustrate how environmentalism has gone wrong and how it could get back on track.

The book is the culmination of a long journey. I started out as someone with well-meaning but ultimately naive ideas about biodiversity and rewilding, who just so happened to have convinced the Fulbright Commission to fund a stay at the German Centre for Integrative Biodiversity Research (iDiv). And I ended up realizing that conservation conflicts illustrate a number of wrong turns that we have made as a society. They illustrate a pattern of thinking and action that will damn us to repeat the errors of past generations of environmentalists and science activists, the mistakes of both climate change and of COVID—if we let it.

But Conservation by the People is also a hopeful book. I suspect my Taming Complexity collaborator Shunryu Garvey will see its central message as almost Buddhist: That we let go of the belief that we can analyze and quantify our way to knowing where it is exactly that the human species ought to land on planet Earth, and instead that we strive to learn to simply better navigate the path together.

I don’t merely call for us to try to practice a kind of Biodiversity Democracy. I have come to see common ideas about how to do democratic politics, and the role that science should play in it all, as having long ago run aground. If we are to find our way out of the stormy seas and onto the shoreline, it will be focusing more on listening than on talking, and more on doing than on knowing. I hope you join me on that journey.

@ MIT Press

@ Amazon

@ Bookshop.org

Some weeks ago, the House of Representatives passed the Safeguarding American Voter Eligibility Act, abbreviated the SAVE Act, and commonly called the SAVE America Act. The act would require citizens to present either a passport or birth certificate to prove their citizenship before they can register to vote. It also requires voters to show a valid government issued photo ID at the polls before casting their ballot. Finally, it calls for the criminal prosecution of any government officials “registering an applicant to vote in an election for Federal office who fails to present documentary proof of United States citizenship”. Note that this penalty is for registering an applicant who doesn’t have the correct documentary proof, not for registering a non-citizen.

These rule changes could have several undesirable consequences. For any citizen who has ever changed their name, a birth certificate alone would not be sufficient to register. They would need to either provide a passport, or include additional legal documentation connecting their birth name with their changed name when registering to vote. This includes the approximately 54 million married women who took their spouse’s last name.

Presenting your passport instead obviates the need for this additional documentation, but over 146 million Americans do not have their passport, about half the adult population. Moreover, passports currently cost $125 which, in addition to being a high price for many Americans, may technically constitute a poll tax. Poll taxes are unconstitutional having been outlawed by the 24th amendment and being a violation of the equal protection clause of the 14th amendment.

Meanwhile, providing a birth certificate and proof of name change is no guarantee that a citizen will be allowed to vote. Would a marriage license or a marriage certificate be the correct legal document? The SAVE Act does not specify. And those documents vary widely in presentation between states and territories. Would a random public official be able to remember all the different forms of authentication for those documents from each state and territory? The SAVE Act doesn’t provide any guidance or support for verifying such documents, something which most registration locations, like DMV offices, are already under-equipped to do. And most importantly, given that they could go to jail if they make a mistake, would that official be inclined to grant a citizen their voter registration if they are at all unsure about their documentation?

Prosecuting voter registration officials incentivizes them to reject any registrations that are abnormal or difficult. People have rejected my New Mexico diverse license for being a foreign license more than once! And my wife was surprised and frustrated after we got married at how difficult it was just to do things like change her name on her accounts. It doesn’t seem especially democratic to accept this for registering voters.

The SAVE Act also lacks the language to handle important fringe issues. For example, how to ensure that people can register and vote after being affected by a natural disaster. In such a case where most of this documentation is destroyed, and the government infrastructure to replace it may not be operational, are we to just accept that people won’t be allowed to vote?

The SAVE Act is also light on details in general. In addition to no guidance on what documents constitute proof of name change, it also doesn’t specify how any of the required documents should be submitted for online voter registration. While it does have language that clarifies document submission for mail-in registration (the documents still need to be submitted in person, essentially nullifying registration by mail), it provides no such guidance for online registration. It provides a recipe for confusion, with states setting drastically different standards, where the result is that some people get to vote and others don’t. The stakes are too high for this level of disarray.

But political parties often propose bills that they know won’t pass as a symbolic gesture to their base: “look, even though I can’t get this passed I’m proposing it anyway.” The difference is the Democrats, the ostensible opponents to this bill, are the minority party in both chambers of Congress. Their only real strategy for blocking this legislation is to filibuster it in the Senate and hope the Republicans don’t gain a super majority in the midterm elections.

Worse, the Democrats seem to be stuck in an obstructionist rut, barely able to stop a bill that would make voting harder when the U.S. already has a voter turnout problem! It would be far better if the party could focus their energy on solving those problems rather than merely stopping new ones, better if they could bail the boat rather than just plug the leaks.

The US has relatively poor voter turnout in its elections. Relatively few of our voting age citizens are registered to vote compared to other democratic nations. Voter turnout in the 2020 election was 66%, the highest voter turnout since 1908. The second highest voter turnout in this period was in 2024, with a voter turnout of 64%. There are 262 million citizens over the age of 18. Which means that Trump’s victory over Kamala Harris with over 77 million votes only represents about 29% of the voting age population. Biden won the cycle before, in 2020, with 81 million votes representing only 32% of the 252 million voting aged citizens at the time. Neither of these numbers strike me as particularly democratic. Especially not when other countries have turnouts as high as 70, 80, or even 90%!

Long wait times to vote have a variety of negative impacts. Long lines tend to deter voters, who might not be able or willing to wait the full time and may leave the line after waiting too long or simply not bother to get in it in the first place. Voters who wait in long lines are less likely to cast their ballot in future elections. And the expectation of long lines, such as from news coverage, can keep people away from the polls entirely. In 2024, about 11% of voters had to wait for over 30 minutes to cast their ballot.

It is rather frustrating to see that, while the US already suffers from chronically low voter turnout and voter registration, congress is considering a bill that would make that even worse.

It seems to me that being merely oppositional and obstructionist is, if not a losing proposition, at least not a winning one. If Republicans aren’t willing to budge on voter IDs and voter registration, then maybe the Democrats should trade that symbolic policy for substantive improvements in voting in the US.

Clumsy Solutions

While Republicans claim they want to fight voter fraud and Democrats argue that voter ID laws will repress voters, neither side is likely to change their position based on arguments and evidence about either. The law is, instead, symbolic of values.

Laws aren’t just decisions about issues, they are affirmations of values and world views. It doesn’t matter to Republicans that there is no evidence of voter fraud that needs to be solved. Failure to take precautions against it is essentially saying to them “we don’t care about your concerns, your concerns are stupid.”

The same impasse was present in France in 1975 when they passed the law which first legalized abortion there. In practice, the French 1975 law was a victory for the pro-choice camp. It substantively allowed abortions with relatively few restrictions. But it included several requirements that were symbolically important to the pro-life camp. First, women seeking abortions had to certify that the procedure was due to distress or emergency. Even though the law provided no way to verify or enforce this certification (women could simply lie about why they wanted the abortion and there wasn’t anything the state could do about it), this provision validated pro-lifers’ position that the fetus ought to have some degree of rights and that we ought to have some degree of moral obligation to the fetus. Second, the law required a waiting period of one week during which the woman would receive counseling on all the state support programs she would have access to were she to change her mind. This affirmed the pro-life position that abortion should not be used as a form of birth control. In practice, the law did little to protect fetuses or prevent abortion as a form of birth control. The only barriers were the woman’s willingness to lie to public officials with no repercussions and her willingness to wait one week. But by offering symbolic concessions to pro-lifers, the pro-choice camp was able to remove enough opposition to broker a compromise that got them the substantive policy they wanted.

Moreover, by getting a policy on the books that was agreeable to both sides, France made the issue less polarizing, paving the way for future liberalization of abortion. The 10 week restriction was eventually rolled back to 22 weeks, with no restriction in the cases of a non-viable fetus or a danger to the mother, and French state healthcare fully covers abortion procedures.

France’s policy was what Marco Verweij and Michael Thompson call a “clumsy solution.” When politics entails disagreement (and honestly when does it not?) it often feels like the only options are to get stuck in the morass and do nothing or struggle endlessly with an obdurate opposition. But why let the perfect be the enemy of the good? Among the extremists shouting at one another are innumerable reasonable people who just want to be heard. Perhaps we can do better by embracing the conflict, and clumsily navigating our way through the plurality of existing positions to arrive at an imperfect but improved solution?

Making Voter IDs More Democratic

Could we implement a similarly clumsy solution for voter ID laws? Are there compromises that Democrats could make to give Republicans a symbolic victory while gaining a substantive victory for themselves: increasing voter participation?

Most other countries with free and fair, democratic elections require proof of identification in some form or other to vote. And in these countries, citizens are often *more* enfranchised than we are in the United States. Clearly, it is possible to implement voter identification checks without also causing all of the problems that would come along with the SAVE Act.

In their most recent federal elections, Germany had a voter turnout of 82.5%. The lowest voter turnout since WWII was in 2009 with 70.8%. Germany’s lowest voter turnout is appreciably higher the the highest turnout in the U.S. And Germany requires an ID at the polls. How does Germany maintain such high democratic participation while still requiring voter IDs?

In Germany citizens are automatically registered to vote. The country requires citizens to register their presence with their municipal government, the data from which then resides in a civil register with the federal government. Citizens on this register are automatically registered to vote. They receive a registration confirmation in the mail prior to national elections. Citizens bring this confirmation along with a photo ID with them to vote. This removes the administrative burden for voter registration from individuals and places it in the hands of the state. It also removes burdens from local agencies, like departments of motor vehicles (DMVs) in the US. While the US would have to implement this system on a state-by-state basis, as elections are administered by the states, adding requirements like automatic voter registration and maintaining a civil registry could turn a voter ID bill from a restrictive bill to one that enhances access to voting while also assuaging the concerns of Republican voters over voter fraud.

Australia does even better. In 2022 their Senate elections had a turnout of 90.47% and the House elections had a turnout of 89.82%. This incredible turnout is because, in Australia, voter registration and election participation is compulsory. A national level agency maintains voter rolls using information gathered from states and territories through the normal conducting of business, like drivers license offices, tax offices, and death and marriage authorities. Citizens confirm their identities by answering three questions to confirm their identities, obviating the need for IDs. To make it easier for citizens to meet their obligations, they can pre-register to vote starting at the age of 16, given ample time before their 18th birthdays. When a citizen first registered to vote, they get a confirmation of that registration. Voter registration is also included in the citizenship process so that all new citizens are automatically registered.

Similar to Germany, the administrative burden for registering to vote is taken by the state rather than individual citizens. While Australia does not have voter ID laws, they are still required to prove their identity before voting. The US could substitute the identity questions with a voter ID requirement, especially if the registration confirmation was in the form of an ID.

When and how voting happens is also important to election turnout. In Germany, elections are always held on a Sunday to ensure most citizens have the day off of work to vote. In Australia, elections are held on a Saturday for the same reason. In both Germany and Australia citizens must apply to vote by mail. However, Germany has no in-person early voting, while in Australia early voting is fairly ubiquitous. Furthermore, Australia has implemented mobile voting locations to make voting easier for those living in the country’s vast swaths of rural land.

Election day in the United States was well thought out for a country where the vast majority of citizens were farmers, occurring after the harvest but before the cold winters. But today, changing election day to fall on a weekend or to be a national holiday would make it easier to balance voting with work. But the rise of early voting has already been extremely impactful in increasing voter turnout in the U.S. Requiring states to support early voting at all polling places would go a long way in reducing wait times and improving voter turnout even further. Furthermore, although I have personally found voting by mail to be useful because I can research exactly what’s on my specific ballot, Australia and Germany demonstrate that universal mail-in ballots are not necessary to achieve high voter participation. The U.S., especially rural states, could instead implement mobile voting locations like Australia, perhaps with federal support.

The SAVE act is, as it stands, a genuine threat to American democracy because it has the potential to dramatically restrict citizen participation in the election process. But that doesn’t mean that requiring IDs to vote is automatically undemocratic. Another key aspect **better word** of a democracy is that citizens feel like the government respects their values even when it doesn’t implement their preferred policies. Democrats could craft a policy that expands voter registration and citizens’ access to the polls while also validating Republicans by compromising on the symbolism of voter IDs. Not only is such a strategy more likely to actually get legislation passed, but it could be legislation that gives Democrats the more expansive citizen participation that they want by only giving up a symbolic voter ID requirement.

By now you’ve surely heard that AI is a threat to democracy—yet congressional leaders assert that rapid development is essential to defending democracy. AI-driven mass layoffs looms larger than ever, with Anthropic’s CEO casually predicting 50% of all entry-level white-collar jobs eliminated in a couple years—but McKinsey forecasts a robust global GDP increase of “$2.6 trillion to $4.4 trillion annually,” roughly the size of the entire UK economy. I regularly see posts from teachers who lament that AI has ruined education—meanwhile “education innovators” breathlessly laud its potentialto “revolutionize education.” And with AI psychosis popping up in the news lately, each of us might be wondering just what exactly “the algorithm” is doing to our brains—some good, some bad, no doubt.

Amidst the tangle of disputed dangers surrounding this dazzling technological marvel, one potential harm in particular stands out like a beacon: Bioweaponry.

Everyone from AI executives to their anti-capitalist critics seems to agree that LLM-powered chatbots should not be allowed to teach anyone how to build a biological weapon. Here, at least, the risk of catastrophe is clear enough to command widespread agreement among otherwise disparate partisans.

With the proliferation of DIY biotechnology kits, now available even on Amazon, this risk is being portrayed as pervasive—as though any disaffected loser could easily jail-break an LLM to walk them through the construction of a pandemic-level pathogen while wearing pajamas in their parents’ basement.

Sure, nobody wants that.

But there is good reason to suspect the purported risk of home-brewed bio-weaponry is overblown.^[i] Placed in the broader context of extant global risks, concerns over this danger are functioning as a kind of safety theatre that conceals a deeper, more systemic set of risks that are actually worth worrying about.

A recipe for anarchy?

To any younger readers born in post-Internet era: Bear with me for a moment of nostalgia. Remember when we had to actually get up and walk all the way over to the bookshelf and open a big black paperback to find instructions for doing illegal, dangerous, and subversive things?

Ahh, those were the days. When I was about 13 years old, a friend and I would sneak his older brother’s copy off the shelf and peruse its pages, fantasizing about all the mischief we would make…someday. But like most readers of The Anarchist’s Cookbook, we never acted on any of it.

In the years since its 1971 publication—authored by a teenager who cribbed the material from sources in the New York Public Library and never field-tested any of it—the The Anarchist’s Cookbook has been found in the possession of a handful of terrorists and school-shooters. But the extent to which any of its (often incorrect) information was ever actually employed is entirely unclear.^[ii] The FBI is on the record saying the book is “one of the crudest, low-brow, paranoiac writing efforts ever attempted” and practically useless to would-be criminals.

As of 2016, however, over two million copies have been sold, and who knows how many copies have been downloaded by randy teenagers. Clearly, access to dangerous information is a necessary but insufficient condition for the perpetration of terrorist acts.

That ought to be obvious to everyone in the Internet Age. Any of us could probably find a bomb-recipe online without much trouble—but who actually bothers?

Knowing that one could learn to build a bomb is different than actually building one. And successfully building one is different than actually detonating a bomb in an act of terror.

And committing an act of terror turns out to be a line most people aren’t willing to cross—a fact that frustrates the FBI to no end. Time and again, this “law enforcement” agency has had to guide, coax, fund, and even coerce would-be domestic terrorists into actually doing the deed.^[iii]

Even with a sycophantic chatbot as guide, safely and successfully building, testing, and deploying a bomb or bioweapon without killing yourself or being found out is a lot of hard work that most people aren’t willing or able to do.

Not without strong incentives, that is.

In fact, a strong incentive system for producing pandemic-level pathogens is already in place—not in anyone’s parents’ basement, but in publicly-funded labs around the world. And the would-be perpetrators of dangerous pathogen dispersal aren’t wearing pajamas, but lab coats.

These potential terrorists don’t need an LLM to teach them pathogen synthesis—they already know how to cook. It’s actually their job. All they need to do to put humanity at grave risk is make a normal mistake at work.

Natural outcome or normal accident?

In a stunning reversal, what was initially dismissed as a conspiracy theory—the notion that the COVID-19 pandemic was caused by a leak from a laboratory in Wuhan, China—is now the official position of the current White House, the CIA, the FBI under Biden, the Department of Energy, and the House Oversight Committee that investigated the virus’ origins. Even the current NIH director recently told the NYT that he “thinks COVID probably started in a lab.”

One of the first and most consistent voices for the lab leak hypothesis has been Rutgers molecular biologist Dr. Richard H. Ebright. The Twitter algorithm fed him to me in late 2021 as I was waiting in an airport bar on a return flight from New York to Tokyo, when Japan had re-opened international travel for a couple months, right before a new COVID variant shut everything down again.

Assuming credentials still matter in a world flattened by internet punditry, rest assured that Ebright, is no crank. I’ve been following him ever since, watching with interest as he updates the confidence level of his evaluations as new information comes out.

Somewhat surprisingly, his basic outline of what happened has not changed in the intervening years. Instead, new revelations have mostly verified his initial hypotheses.

To be clear, Ebright does not argue that COVID-19 was a bioweapon. Rather, he argues that the leak was precipitated by a combination of academic career incentives, human error, and governance failures in the high risk domain of virology.

We publish, you perish: Gain of Function (GoF) research

Ebright starts by pointing to the perverse incentives in the subfield of virology called “gain of function” (GoF) research. GoF has itself become a disputed term, but basically it refers to experiments that modify an existing pathogen – here, a virus – to make it more transmissible and/or more virulent. One MIT professor described such experiments the “virological equivalent of nuclear testing.”

Why are researchers spending valuable time and scarce resources to make viruses and other pathogens more dangerous? The field justifies its activities by claiming they contribute to pandemic preparedness. Understanding the workings of more virulent strains that might evolve naturally could help us defend against them.

Ebright, a critic of GoF since well before the pandemic, disputes this claim, arguing instead that GoF is largely driven by academic incentives:

[GoF] is not being done for practical purposes. It has produced no actionable results — no results that would, for example, be useful for preventing or responding to a pandemic. It is being done for career advancement.

As in any academic field, researchers in virology are under constant pressure to publish papers with “high impact” results. If we analogized academia to a video game, GoF research would be akin to a cheat code: a quick way to level-up as a researcher.

As Ebright points out, GoF experiments are “simple to perform…highly publishable and highly fundable.” Whereas creating an “antiviral drug typically has a timeline of 20 years and a probability of success of one in 20,” GoF experiments can be completed in “just six months, and the probability of success would be nearly 100%.” That is, by applying the cheat code of GoF research, “one can move very quickly to results, to publication, and then to a next grant.”

Do the benefits of near-certain personal advancement and less-certain contributions to pandemic preparedness outweigh the risks of global catastrophe and millions of lives lost? Experts, of course, disagree. But as one evolutionary biologist put it, “When you talk about an experiment in which the consequences of an accident are a global pandemic — it’s really hard to even put that into a risk-benefit analysis,” adding that COVID-19 emerged “right under the noses of people doing this work, and nothing they did was useful in stopping it.”

Such calculations aside, there is no doubt that the GoF research system strongly incentivizes the creation of bioweapon-adjacent pathogens.

How does this system function in practice?

Proposing a pandemic

The cornerstone of Ebright’s lab-leak case is the DEFUSE proposal—a $14.2 million grant application submitted to the Defense Advanced Research Projects Agency (DARPA) in 2018 by EcoHealth Alliance, in collaboration with the Wuhan Institute of Virology (WIV) and a virology lab at the University of North Carolina (UNC).

Leaked by independent researchers and confirmed as authentic, DEFUSE described a plan to:

1. Collect and sequence bat coronaviruses from caves in southern China.

2. Create full-length infectious clones of these coronaviruses.

3. Insert furin cleavage sites into the spike proteins of SARS-related coronaviruses—sites that would enhance the viruses’ ability to infect human cells.

4. Assemble synthetic viruses from six DNA segments.

5. Test engineered viruses in humanized mice and human airway cell cultures.

DARPA rejected the proposal, judging it too risky.

As Ebright has repeatedly emphasized, however, several features of SARS-CoV-2 (i.e. the virus responsible for the COVID-19 pandemic) precisely match the viruses that DEFUSE proposed to create:

• Furin cleavage site: SARS-CoV-2 possesses a furin cleavage site at the S1/S2 junction of its spike protein—a feature found in no other known SARS-related coronavirus among the 871 known members of its viral family. DEFUSE explicitly proposed inserting such sites.

• Synthetic assembly signature: A 2022 analysis identified evenly spaced restriction enzyme recognition sites in the SARS-CoV-2 genome consistent with assembly from six DNA segments—the very method described in DEFUSE.

• No natural precursor: Despite four years of intensive searching, no natural reservoir host for SARS-CoV-2 has been identified. No natural intermediate host has been found. No progenitor virus with a comparable furin cleavage site has been discovered in nature.

When unredacted DEFUSE drafts were released through a Freedom of Information Act request, Ebright upgraded his assessment of this evidence from “noteworthy but not decisive” to “the equivalent of a smoking gun.”

If the DEFUSE proposal was rejected by DARPA, how could it then be leaked from a lab?

Outsourcing risk

Ebright explained in his 2024 congressional testimony. In 2018, the EcoHealth Alliance, WIV, and collaborators also applied to the NIH for another project that, like DEFUSE, involved constructing coronavirus variants “that possessed spikes having higher binding affinities for human SARS receptors.” This proposal was funded.

And since the scope and aims of this project overlapped with the DEFUSE proposal, NIH grant policies permitted the use of those funds to support the work proposed in the rejected DEFUSE proposal. Noting that this is “the typical practice for NIH grantees,” Ebright concluded it is therefore “likely that, in 2019, researchers at the Wuhan Institute of Virology undertook both work proposed in the 2018 NIH proposal and work proposed in the 2018 DARPA proposal.”

Why did this matter? Two reasons. First, Ebright argues this indicates an intent to deceive NIH, the funder. The DEFUSE proposal indicated the work would be done at UNC. Moreover, when EcoHealth Alliance President Peter Daszak—the project leader—was asked by Congress in 2024 where the work was going to be conducted, he said, “My understanding for that work, it was going to be done at UNC.” However, internal notes on DEFUSE released through FOIA revealed that Daszak wanted to “stress the U.S. side of this proposal so that DARPA are comfortable with our team. Once we get the funds, we can then allocate who does what exact work, and I believe that a lot of these assays can be done in Wuhan as well…”

Share Taming Complexity

Second, if the risky work was conducted at WIV, then it was almost certainly done with inadequate precautions. Had it been performed at UNC as proposed, the experiments would have been conducted at Biosafety Level 3 (BSL-3), “where researchers perform all experiments in a biosafety cabinet” while using full PPE and N95 masks under medical surveillance, with two sets of doors to keep the lab isolated, and air filtration systems that use directional airflow to prevent the escape of airborne pathogens.

However, at the Wuhan lab, coronavirus research was being conducted at Biosafety Level 2 (BSL-2). As Ebright (now famously) put it, these conditions are “the biosafety level of a standard US dentist’s office.”

Collaborators on the DEFUSE project were well-aware of this fact. For instance, a 2016 paper from WIV, also funded by a 2014 grant to EcoHealth Alliance for work on coronaviruses, notes that “All experiments using live virus was [sic] conducted under biosafety level 2 (BSL2) conditions.” Moreover, in early drafts of the DEFUSE proposal, the head of the WIV lab “wrote that she planned to perform this work in BSL-2.” Seeing this, the head of the UNC lab replied that while such weak precautions might be acceptable in China, “US researchers will likely freak out.”

In sum, work that DARPA had deemed too risky to conduct in the USA was mostly likely carried out in Wuhan, China.

Not unlike a multinational petrochemical corporation that sites polluting factories in jurisdictions with weaker protections, these virologists, driven by academic career incentives, knowingly outsourced risky GoF experiments to a nation with weak oversight and inadequate precautions.

Speed kills

Still, we don’t need to assign malicious intent, per se, to the GoF researchers. Self-interest is a sufficient explanation.

Ebright argues the WIV’s decision to work at BSL-2 was driven by a simple practical calculus: Lower biosafety levels dramatically increase research productivity. BSL-3 precautions are cumbersome and time-consuming. In comparison, Ebright estimates that working at BSL-2 increases rates of progress by a factor of ten to twenty—a massive competitive advantage in the race for publications and grants.

This decision was technically within their purview. Somewhat surprisingly, no internationally enforceable biosafety standards exist. Consequently, assignment of biosafety levels is left to individual countries, institutions, and scientists themselves. In China, the BSL-2 classification was determined at a time when bat coronaviruses had not been confirmed to infect humans. China’s regulatory framework, such as it was, did not account for the fact that these researchers were explicitly creating novel coronaviruses that would cause human disease.

This is where weak oversight proved fatal: The catastrophic choice to carry out the DEFUSE experiments at BSL-2 was technically in accordance with Chinese regulations.

What happened next might not shock you

So, what happened next? Most likely, a normal accident. Somebody spilled something, or cut corners on a decontamination procedure. These kinds of things happen, and with surprising frequency:

• Between 1949 and 1974, Sulkin and Pike documented over 4,000 laboratory-acquired infections, with a fatality rate of 4.1%.

• The original SARS virus escaped from laboratories on at least four documented occasions after the initial 2003 outbreak.

• The US government imposed a near-total moratorium on GoF research from 2014 to 2017, “after two mishaps at government labs, one handling anthrax and one handling avian flu, which together suggested that biosafety and biosecurity at even the most respected labs fell well short of what is needed to protect the public.”

• A 2024 Lancet Microbe scoping review documented 309 laboratory-acquired infections and 16 “accidental pathogen escape/loss events” worldwide between 2000 and 2021, including escapes of SARS-CoV, anthrax, poliovirus, and influenza H5N1.

Moreover, in the lead-up to the pandemic, the WIV—the first laboratory in China to be certified for BSL-4 work, despite staffing shortages and other issues—was under increasing scrutiny.

In January 2018, the U.S. State Department warned that the WIV had “a serious shortage of appropriately trained technicians and investigators needed to safely operate” its high-containment laboratory. Internal WIV reports from 2018-2019, later analyzed by the U.S. Senate and intelligence agencies, highlighted ongoing equipment and protocol failures; identified “hidden dangers,” “shortcomings,” and “nonconforming items” in their biosafety management systems; and noted that some staff members were overworked and poorly trained.

Again, no need to assign malicious intent. Errors—procedural mistakes, equipment failures, lapses in training—are inevitable. Laboratory accidents are not aberrations. Rather, they are routine features of the research enterprise.

The same cannot be said, however, of what happened next.

Subscribe now

By “self-governance” we mean self-protection

When did “the pandemic” start, exactly? In the winter of 2019/2020, I remember seeing news about some strange new infections in China, then some on cruise ships, then in Seattle…and then WHAM! Everything was shutting down.

We were only a few weeks into the mess when, on February 19, 2020, The Lancet published a letter signed by 27 prominent scientists that stated: “We stand together to strongly condemn conspiracy theories suggesting that COVID-19 does not have a natural origin.”

At least two things about this letter are worth noting. First, at this point in time, no investigation into the origins of the virus had been conducted. Therefore, these scientists could not have known whether a lab leak had happened. To label that possibility a “conspiracy theory” at this early stage was clearly a calculated choice.

Second, the letter was organized by none other than… Peter Daszak of EcoHealth Alliance! As leader of the rejected DEFUSE project and the closely related NIH-funded coronavirus project at WIV, he obviously had a conflict of interest, in that confirmation of a lab leak would directly implicate him as culpable. Yet this conclict of interest was not disclosed until June 2021, in an addendum where the editors merely noted, “There may be differences in opinion as to what constitutes a competing interest.”

The way Daszak attempted to conceal his involvement in organizing the letter makes it hard not to conclude that he was well-aware where investigations into the lab-leak “conspiracy theory” might lead. As documented by the BMJ in 2021, emails FOIA’d by watchdog group US Right To Know revealed how Daszak actively curated the presentation of the letter behind the scenes, telling colleagues, “Please note that this statement will not have EcoHealth Alliance logo on it and will not be identifiable as coming from any one organization or person.” He even considered “removing his name from the statement ‘so it has some distance from us and therefore doesn’t work in a counterproductive way.’”

Share

Meanwhile, Dr. Anthony Fauci was convening private teleconferences with prominent virologists in February 2020, several of whom, as emails later obtained through FOIA revealed, initially found the lab-leak hypothesis plausible. Yet within days, several of those same scientists would draft the influential “Proximal Origin” paper for Nature Medicine that argued SARS-CoV-2 was “not a purposefully manipulated virus.” Fauci and others in his orbit would go on to publicly dismiss the lab-leak hypothesis as “just a conspiracy theory.”

Ebright, critical of this response from the start, later told BMJ, “It’s very clear at this time that the term ‘conspiracy theory’ is a useful term for defaming an idea you disagree with,” decrying how the many scientists and journalists who deployed the term managed to get this narrative amplified by mainstream media in the following years.

You might, for instance, remember that Facebook and other platforms banned stories about the lab leak hypothesis as “misinformation”—a position that was only reversed in 2021 after a growing number of scientists, organized in part by Ebright, became more vocal about the need for an open investigation into the virus’ origins, drawing attention from Washington DC.

Ebright characterized the series of acts that constituted the “official response” as a coordinated campaign of manufactured consensus: “False narratives were established and enforced by a small number of scientists and their willing soldiers among science writers,” as well as the editors of top scientific journals.

Specifically, Ebright was one of the first to accuse Fauci, NIH Director Francis Collins, and Deputy Director Lawrence Tabak of systematically misrepresenting the nature and extent of NIH-funded research at the WIV. Though this particular issue became a flashpoint in the fiery clashes between Senator Rand Paul and Fauci during government hearings, Ebright’s assertions were later vindicated when the NIH itself acknowledged that EcoHealth Alliance had violated the terms and conditions of its grant, leading to Peter Daszak’s formal debarment and the suspension of all funding to EcoHealth—from which Daszak was also fired.

There is no happily ever after

As I mentioned, today the lab-leak hypothesis is now the official position of much of the US government. The admittedly strange White House page on COVID-19, which distills the main findings of the House Oversight Committee’s investigation into the origins and handling of the pandemic, reads like a point-by-point vindication of Ebright’s claims, both about the substance of the science and the manufactured consensus campaign.

Still, after being dismissed as conspiracy theorists, subjected to social media censorship, and marginalized professionally, the vindication of Ebright’s and others’ concerns about the possibility of a lab-leak origin for COVID-19 must be bittersweet. After all, millions died and our world was turned upside-down, yet consequences for those who may have been responsible have been minimal and came only years later.

EcoHealth Alliance’s federal funding was not terminated until May 2024. No criminal investigation has been initiated against any U.S. scientist or institution. Fauci, for instance, was portrayed as the victim of a political witch-hunt and preemptively pardoned by President Biden. And the Chinese government has refused all international forensic investigation of the WIV, including access to records, samples, and personnel, further compromising any attempts at accountability.

To top it off, in 2024, twelve researchers filed a formal complaint against Ebright at Rutgers, alleging defamation. Regardless of its merits—Ebright has certainly said inflammatory things about other scientists—the case illustrates the institutional mechanisms available to suppress vocal dissent within the scientific community.

Of course, one might still dismiss the lab leak as conspiracy and chalk all this up to the politicization of science by right-wing firebrands and their allies under both Trump administrations. To be sure, many in the scientific community still insist on the natural origins of the virus—the so-called “zoonotic spillover” hypothesis, that COVID-19 spread from bats to an intermediary animal and then to humans. The WHO, citing lack of evidence, concluded in mid-2025 that “all hypotheses must remain on the table, including zoonotic spillover and lab leak.”

The genomic evidence for zoonotic spillover is beyond my expertise to evaluate. But given the way prominent scientists were revealed to have manufactured a premature consensus around that hypothesis, the burden of proof is on them to not only make a solid scientific case, but also to rebuild trust in scientific authority. The fact that the wet markets of Wuhan, the supposed site of the spillover event, didn’t sell pangolins or bats suggests that will be an uphill battle.

Whichever hypothesis one endorses, perhaps we can agree on one thing: The fact that the cause of the pandemic remains in dispute six years after the shutdowns started is itself evidence of a massive failure of self-governance in science.

The Privileged Position of Virology

COVID-19 became one of the most politically divisive technoscientific controversies of all time. Contestation between scientific, political, and legal forms of authority and expertise defined our lives for several years following the 2019 outbreak. It was difficult to follow all the moving parts.

I found Ebright’s analyses helpful in making sense of the mess. He framed events in a narrative of structural incentives I was familiar with from my academic work on scientific knowledge production and technological risk governance.

In this piece I tried to distill the facts into a story of mostly well-intentioned but fallible people. Some, like Daszak, do come out looking like bad actors. And while legal consequences should certainly be levied for criminal acts, it seems to me that the US government’s investigations focused too narrowly on individual actors.

As I see it, the problem is structural. That is, even if all of the world’s virologists were replaced with equivalently-skilled but different people, assuming the same set of incentives and system of self-governance were in place, it seems reasonable to conclude a pandemic would still probably have happened.

But rather than get into blame or a definitive adjudication of the natural-origin versus lab-leak debate, I want to highlight the risks inherent to the governance structure. Whether SARS-CoV-2 escaped from a laboratory or spilled over from an animal host, the pandemic exposed a governance architecture in which:

• A small community of virologists exercised world-shaping authority with little democratic accountability;

• Academic incentives systematically favored dangerous experiments over safer alternatives;

• Safety precautions were inadequate relative to the catastrophic potential of the work;

• When things went wrong, the same scientific community that created the risk had the institutional power to suppress investigation of its own potential culpability.

This is what STS scholar E.J. Woodhouse has called “the privileged position of science,” operating at a civilizational scale. By comparison, the risk of AI-assisted homebrewed bioweaponry seems fairly quaint.

The few subsequent reforms that have been enacted, such as Executive Order 14292 on biosafety, have been modest and contradictory, leaving “federal agencies, research institutions, and investigators without clear guidance or guardrails for conducting life sciences research.” The perverse incentives driving risky research remain unaddressed as policy is bent to accommodate the productivity advantages of lower biosafety levels: in December 2024, the CDC revised rules to allow for SARS-CoV-2 to be handled at BSL-2. And international enforcement of biosafety remains nonexistent, even as high-containment laboratory construction has accelerated globally.

Worryingly, the flawed governance structure that precipitated the pandemic and prevented learning from the disaster is still mostly in place.

^[i] MIT professor Kevin M. Esvelt has noted that while a new pandemic-level pathogen would kill more people than a nuclear weapon, amateurs and would-be terrorists “lack the technical capability to perform basic science research at the scale needed to find the pandemic needle in the animal virus haystack.” Government labs, outfitted with all the requisite equipment and millions of dollars in funding, have been conducting searches for such virii with little success. The likelihood even a well-equipped “lone wolf” could succeed is slim. A 2024 RAND report found the assistance of an LLM does not close the gap, concluding that AI “does not meaningfully increase risk of a biological weapons attack.” And actual cases of underground labs have thus far been merely disconcerting. The 2023 arrest of the Chinese owner of an unlicensed bio-lab in California turned out to be all smoke, no fire. The Israeli citizen discovered to be running an illicit bio-lab in Las Vegas was actually arrested for firearms possession, but if it turns out he was using AI as described here, I guess I’ll eat my words.

^[ii] What is clear is that since 2001, most of the “crime” associated with the book has been the mere fact of possessing a copy. For instance, in 2021, a 22-year-old was convicted under Section 58 of the UK’s Terrorism Act of having a copy on a computer hard drive, which police found among “tens of thousands of documents” the offender had gathered in “bulk downloads.” Pray, my fellow torrent-users, pray your drives are never so scanned by the state.

^[iii] For example, in the 2020 plot to kidnap Michigan Gretchen Whitmer, the would-be perpetrators turned out to be unwilling to actually commit the serious crimes they fantasized about. Consequently, the FBI informants and undercover agents that had infiltrated the group had to do most of the heavy lifting.

I spent an hour sitting down with Volker Hahn at the German Centre for Integrative Biodiversity Research last month to talk about my upcoming book Conservation by the People. You can check it out on Podigee, Spotify, Apple, and even YouTube.

In 1934, Los Angeles was in the middle of a severe polio epidemic. There were 1,700 cases in the city alone. At the county hospital an unprecedented number of staff seemed to fall ill. Nearly 200 of them reported polio symptoms by the end of the year. The trouble was that few of them showed the usual signs. No paralysis. No abnormalities in their spinal fluid. All patients seemed to show, besides outward signs of being sick, were minor motor impairments. They couldn’t have had polio.

Investigators christened this new disease, one that seemed to have little to do with any specific germ and posed no risk of death “benign myalgic encephalitis.” You might have heard it called chronic fatigue syndrome, or at least heard of its cousins: fibromyalgia and long COVID. These are contested illnesses whose sufferers complain of extreme malaise after mental or physical exertion, sleep disturbances, pain, and other flu-like symptoms. Except, unlike the flu, the illness never seems to go away.

All contested diseases pose the same scientific dilemma: When medical scientists and doctors can’t seem to find something unequivocally wrong with a sick person, is it because they haven’t looked in the right places or because there’s nothing to find?

In an earlier post with The Ecomodernist, I argued that apocalyptic thinking undermines sensible environmental politics. It leads people to exaggerate the scientific certainties, to see politics in fanatical black and white terms, and to refuse to acknowledge the on-the-ground challenges involved in solving global environmental problems. I now recognize the same kind of thinking as at work in keeping people ill. When we view our own bodies through an apocalyptic lens, we throw out any hope of getting better.

The Power of the Mind

Some investigators of the apparent polio outbreak at Los Angeles County Hospital eventually concluded that staff members’ symptoms were the product of mass hysteria, fear of the disease rather than the germ itself. They didn’t do so without good reason. There’s a long history of hysterical disease epidemics. Hundreds of peasants literally danced themselves to death in the middle ages as the belief of a spreading “dancing curse” combined with the threat of famine and the stresses of everyday medieval life. Other hysterias have caused paralysis, rashes, tremors, and even unceasing fits of laughter.

The hypothesized mechanism behind these outbreaks is a kind of reverse placebo effect. The normal placebo effect occurs regularly. Give people a sugar pill and their expectations for improved wellbeing can actually cause physiological improvements. It’s something that confounds most medical trials. When it comes to mass psychogenic illness, it’s the expectation of disease that literally makes people sick.

No doubt that it is always possible that there really was an infectious or toxic agent in such cases, and that we have just failed to discover it. Stomach ulcers were waived off as a product of stress, until someone discovered H. Pylori bacteria in the guts of sufferers. The Spanish flu left some victims with severe Parkinsons symptoms. Why couldn’t COVID leave people permanently ill?

Decades of research has yet to deliver a clear biological etiology for conditions like chronic fatigue syndrome and fibromyalgia. Studies have uncovered genes and other biological markers, like signs of neuroinflammation, that seem associated with these chronic illnesses, but nothing that really reliably distinguishes those who suffer and those who don’t. Suspect biomarkers are often present in healthy people too. And for discovered abnormalities that seem unique, we can’t know whether they are cause or effect, given that we know through the placebo effect that the mind can cause changes in our bodies.

At some point, a lack of conclusive physiological evidence suggests at least some psychogenic element for some portion of the chronically ill. Even if the Spanish flu did cause post-viral illnesses in people, it’s also pretty clear that an outbreak of debilitating nervous tics among young women during the COVID pandemic was simply a case of mass hysteria.

But that is something that is very hard for sick people to accept. When someone utters word psychosomatic or psychogenic, patients hear “fake” and “crazy.” Suffering people understandably want to feel validated, and the diagnosis gets off to a bad start almost by default. It sounds like an ad hoc excuse, something doctors say when they hit the limits of their expertise and just want an inconvenient patient to go away.

That seems like a plausible argument, except when you learn that there are many people who have gotten better simply by recognizing the psychological nature of their suffering.

An early proponent of the psychogenic theory of common ailments, including back pain, irritable bowel syndrome, and headaches, was the late NYU medical professor John Sarno. He claimed to have cured tens of thousands through his clinical practice, which included celebrities such as Larry David and John Stossel just by convincing them it was psychosomatic. Sarno’s claims no doubt outpaced the available medical research. He himself couldn’t see how getting people to see their pain in Freudian terms, as a product of repressed emotion, and having them simply go back to living their lives without dwelling on their symptoms could be tested through a randomized trial.

Recent studies, however, do seem to confirm Sarno’s main intuition (who knows about the Freudian stuff). In one trial of pain reprocessing therapy, 66 percent of treated back pain sufferers were pain free or almost pain free after only four weeks. This was multiple times more effective than a placebo or usual care. Five years later, 55 percent of the group treated with pain reprocessing therapy still had little to no pain. For anyone who follows medical studies, an NNT (“number needed to treat”) less than two beats most of the medicines that people take.

What the Body Reveals

The claim that at least some chronic illnesses are psychogenic, caused by the way that we think about them, conflicts with our commonsensical notion of disease. If we have pain or dysfunction, there must be a biological cause. It is something that happens to us. How could it possibly be that we could being doing it to ourselves?

When it comes to chronic pain, we can point to medical imaging. Reduced joint spacing, bulges in our spine, bone spurs, and others flaws show up objectively on film. There. That’s the culprit.

Sarno would dismiss these as “normal abnormalities,” based on his assessment that many people’s symptoms seemed disproportionate to the physiological incongruities seen in X-rays and MRIs.

And there is some evidence giving credence to Sarno’s argument. A large portion of people appear to have arthritis on x-rays, but no symptoms. Rotator cuffs can often be torn and vertebral discs herniated, seemingly without people being any worse for wear. Absent better diagnostics about the physiological causes of pain, the main difference in many cases seems to be simply patients’ reported suffering, how powerful their mind is in creating and sustaining the experience of pain.

The question, of course, is how far the psychogenic pattern of chronic pain extends to other symptoms, such as fatigue, gastrointestinal distress, and dizziness. To my knowledge, the studies haven’t been done. But it no longer seems like such a big a leap of logic to wonder that, if people’s brains can keep them debilitated with chronic back pain, then maybe our minds can also give us the symptoms of chronic fatigue or long COVID.

Given the fact that COVID quickly became one of the most scientifically scrutinized illnesses of all time, we are likely to have uncovered all sorts of seemingly new, but potentially normal, abnormalities. Scientists have found innumerable physiological impacts of COVID, minute bodily changes—some of which probably occur after infections by viruses—and might have remained undiscovered if it weren’t for the pandemic focusing our attention on them. But the sudden revelation of these impacts makes COVID seem like an unprecedented attack on our immune systems, a “mass disabling” crisis in the making. But such thinking might be the very thing that has rendered so many people ill.

When interviewed for a New Yorker article on fibromyalgia in 2000, Harvard psychiatry professor Arthur Barsky argued that sufferers “become trapped in the belief that their symptoms are due to disease; with future expectations of debility and doom. This enhances their vigilance about their body, and thus the intensity of their symptoms.”

In other words, chronic psychodynamic illnesses are, at least in part and for some cases, driven by apocalyptic, revelatory thinking. Chronic fatigue, fibromyalgia, and long COVID sufferers insist that their symptoms reveal a broken body. The lack of medical succor reveals a broken medical system, if not society

“Long COVID is not a tragedy, it’s government negligence,” argue some activists. Others warn readers of national newspapers that “Long COVID has derailed my life. Make no mistake: It could yours, too.” COVID has been transformed from a virus that can kill in rare cases but is otherwise an inconvenience, into one that instead threatens everyone with apocalyptic disability.

When Lived Experience Lies

I suspect that we will look back on the long COVID movement and hear echoes of the polio hysteria at Los Angeles County Hospital. To be clear, my point isn’t that chronic illnesses, such as long COVID, are invented whole cloth. Or that people aren’t genuinely suffering. Neither do I want to imply that there’s no physiological element to psychogenic illness. I’ve experienced it myself.

I suffered for years with a chronic vestibular dysfunction that began as a severe viral infection. Doctors had to mainline Dramamine into my veins in order to stop the dry heaves. Six days later, when I could still barely handle walking down the hallway before needing to rest, they insisted that my brain would eventually adapt around the vestibular injury and released me from the hospital.

When some of my symptoms persisted or returned, such as the nausea that I felt under certain kinds of light and in crowds or the extreme fatigue that I felt after driving, I suspected the worst. I even considered going on disability when I had days when I felt like I couldn’t keep up. I’m sure those who believe they have chronic fatigue syndrome or long COVID have similar stories to tell.

Despite these symptoms, even more fringy practitioners struggled to help me, while traditional doctors waived off my complaints. The slightly kooky vision therapist that I saw tried make me prism glasses, only for them to seem to disturb my balance even more. The mainstream neurologist recognized that my life wasn’t in immediate danger and quickly shooed me out of his office. When I said that I hoped to return to playing soccer and rugby, he told me in his kiwi accent, “Come on. I think you’re done, mate.”

Had I insisted on many more brain scans and blood tests, I might well have uncovered still more physiological abnormalities of unknown significance—ones in addition to the few that I madly Google searched after standard MRIs and CTs. And negative results could always be waved away, a sign of medical incompetence, the failure by the so-called experts to find what my body and brain were so clearly telling me was really there.

My chronic vertigo and associated fatigue didn’t get better until I tried to resist the siren song of apocalyptic medicalization, the nagging voice that sees debilitation behind discomfort. Also helpful was trying to not have such unrealistic expectations of myself, high standards that drove me to ignore my body when it asked for rest and crash out on the weekends.

I recognized that apocalyptic thinking about my body led me to tense up my neck and facial muscles when I experienced brief relapses of vertigo, to tire myself out even more by dwelling about feeling fatigued, and to make myself crazy with dreamed up catastrophic scenarios in which I never played sports again or was forced to quit my job. I learned to more often let discomfort simply be discomfort, and to try to let it go.

During my PhD in Science and Technology Studies, it seemed like everyone viewed citizen health movements as full of heroes, as people with “embodied expertise” who were pushing back against the scientization of illness (partly because it was still before right wingers had their own health movements). But patients’ lived experience can just as easily be bullshit as any slapdash scientific study.

The heroic image of sick people striking a blow for healthcare justice has prevented the field from paying attention to the possibility that health activism and social movements for contested illness might, at least in some cases, be iatrogenic. Efforts to validate chronic fatigue syndrome or long COVID as “legitimate” physiologically caused diseases may be literally making many people sick. Activism may be dragging innocent people, who would have gotten better on their own, into the abyss of apocalyptic health thinking.

The Siren Song of Apocalyptic Medicalization

Unfortunately, apocalypticism in all its forms is more easily analyzed from a distance than actually coped with—I can attest to that personally. Coping means learning to live with uncertainty, to resist diagnosing every imperfection as revelation of doom, and to insist on practicing a faithful, even if naive, optimism about the future—whether about one’s own body or the planet.

I’m not sure what I would say to someone with long COVID, were they sitting right in front of me. When I’ve tried to help friends and family who seem to be struggling with what clearly looks like psychogenic pain: tennis elbow that just so happened to develop after a promotion to a management position, a knee that became dodgy after moving in with a new romantic partner but looked clean on an MRI, and countless others, I’m unsure if I’ve convinced any of them. They seem to half listen to me recount my experiences and conclude quietly to themselves: “Well, maybe he’s a crazy person whose dysfunctional brain kept him ill, but my suffering is REAL!”

After years spent trying to persuade people that they are often their own worst enemies, whether it is the stubborn belief that their political opponents are all brainwashed or corrupt, the insistence on a coming environmental apocalypse, or the preoccupation with seeing bodily dysfunction as invariably physiological, I am stuck with two possible conclusions: Either persuasion is nearly impossible for most high-stakes, highly emotional issues, or I am just bad at it.

There are just not any obvious ways to disabuse someone of apocalyptic thinking that allows them to save face. I’m embarrassed that my mistaken thinking about my body kept me from driving and playing sports for years longer than was necessary. Just like I’m embarrassed about how seriously I took 2010s claims about peak oil and a climate crisis being right around the corner. But somehow I changed my own mind. And I worry that millions are doomed to needlessly suffer unless we can figure out how to more reliably inoculate people against the allure of the apocalyptic.

Subscribe now

Share

Leave a comment

I have a guest post over at The Ecomodernist on eco-apocalypticism today. Please go check it out.

It’s a piece that I’ve sat on for a couple years. Part of the reason for that is that I really wasn’t having a really good time when I first drafted a good portion of it. We’ll just say that my expectations for fruitful intellectual exchange and social comradery where I was researching at the time weren’t met and leave it at that.

But I’m glad that I’ve come back to it. In some ways, it’s a response to some of the recent talk about “highbrow misinformation.” I’m not a fan of the term, but if anything is a candidate for being elite misinformation, it’s the recurring refrain that we know for certain that we’re destined for environmental doom.

However, I don’t think labeling my former belief in a coming eco-apocalypse as the product of misinformation would have accomplished much. Just like it does next to nothing to label vaccine hesitancy of MAHA beliefs in that way. Still, even if many environmentalists might be horrified by the comparison, I think thinking about alarmism as not too dissimilar to climate denial and vaccine refusal is helpful. And that’s because the solution isn’t more information or the right information but improving democracy itself.

I get into all of this in the post, so I won’t include any more spoilers here. Enjoy!

Across the Pacific, it was already January 7^th when the news broke. That morning, I was on cook duty at Daieiji, a Zen monastery in Niigata. I had done about six months of training there a couple years prior, and since I already knew my way around the kitchen, I was quickly cycled back into the rotation. Half or more of the other monks in training were younger guys who’d never really cooked anything besides “Cup Noodles” in their life.

Rising at 4 AM, a precious 20 minutes before everyone else, I made my way to the bonsho, the big outdoor bell, and, shivering against the bitter pre-dawn cold, climbed the bell tower ladder. I performed three prostrations. Then rang the bell. Then another three prostrations. Then bell. And so on, until 4:45. Finished, and wide awake, I came down to the kitchen to get the rice porridge going. As usual.

But today was different: I had a phone in my pocket. And in previous years, a few clever brothers had surreptitiously installed WIFI one day while the Abbot was out. And I knew the password. I was going to get online for the first time in months.

Subscribe now

Truth is, new recruits aren’t allowed to bring anything up to the monastery except their bedding, robes, and sutra books. And, once in, they aren’t allowed to leave the premises or receive visitors for 100 days. Phones, however, are not welcome in the monastery, at any stage.

Technically, though, I wasn’t a new recruit. I had simply returned from a 2 year leave of absence. At least, that’s how I justified it to myself. So, on a rare day of rest following the New Year holidays, I asked my wife, who is Japanese, to drop by with my iPhone. I felt compelled to check in on my home country, America—and was willing to break my “digital detox” to do so.

Fondling the forbidden phone, hidden away in a fold of my robes, I awaited the moment when everyone—the Abbot, his attendant for the day, and the rest of the trainees—would be busy chanting the morning services.

As the time approached, imagining what might be going on back home, I phrase from the sci-fi protagonists of old floated through my mind: “I gotta bad feeling about this.”

My wife and I left the States just a month before the November 2020 election. If you were there, you probably remember: the country was a confused mess. Mask mandates were in effect and everything was still mostly shut down, including Stanford University where I had a research position. We were still wiping down groceries. Hate crimes against Asians were on the rise, even in my hometown of Bloomington, IN. And worse, it looked like Trump would be re-elected. America, it seemed, hadn’t learned anything after a 4-year dalliance with unhinged authoritarianism.

And I had taken Buddhist vows not quite ten years earlier. Seeing the world come to a crashing halt as people died in droves felt like impermanence smacking me in the face. Suddenly, the teachings of the Buddha, which transcend life and death, seemed more relevant than anything I was reading as an academic. The Universe was giving me a sign about what I should do with this one life I have to live.

At the time, I was collaborating on a paper that proposed seeing the COVID-19 pandemic as an opportunity, an opening to something else; as a “Portal to Alternative Futures in the Making.” I decided to take that proposition seriously. After seven years in upstate NY and almost a year in Palo Alto, my wife and I sized up our options, and agreed the writing was on the wall: It was time for us to go.

To go back to Japan—where they wear masks as a courtesy to others.

Almost a decade prior, we’d met and married in Japan. So I was permitted into the country on a spousal visa, expressly on the condition that we were not traveling, but moving back, permanently. We flew during peak COVID, when you still had to get tested—that awful nasal swab that leaves you wondering whether they aren’t scraping your prefrontal cortex with that deluxe Q-tip—within 72 hours of your flight.

Unlike the barebones chaos of American airports, with overworked skeleton crews in poorly-fitting paper masks shouting over confused crowds, our arrival at Tokyo’s Narita was calming, and surreal. Filing down the long hallway into a wide open room, we were greeted by dozens of staff in white, full-body hazmat suits, who kindly asked us to spit into tiny test tubes. No brutal nasal swabs here.

I re-entered the monastery a few days after the 2020 election. It was joyous. In the time I was gone, my direct senpai—the monk one step above me in the rigid monastic hierarchy—had risen to head monk. The new guys had heard all about me. Everyone was happy to have the foreigner back on the team.

Never having really played or enjoyed team sports, that is still the best I can analogize monastic life to: being on a team—only the game is never over, and you don’t go home. You do everything—wake, meditate, pray, eat, work, bathe, and sleep (what little you can get)—together. It might sound claustrophobic, and it is at times, but it was an utter relief to be away from America’s mess.

Out in rural Japan, COVID, too, seemed distant. Within the monastery, life proceeded pretty much as it always had. We monks stayed busy: meditating for a week in early December, gathering donations of vegetables to store for winter—burying offerings of cabbages and daikon radishes in the sandy loam out back—and prepping for the New Year’s holidays.

On January 1^st, 2021, while cleaning up my section of the common area, I discovered my old keychain. It was empty. I had given up my world. No car, no home, nothing to call “mine” and lock up. At 36 years old, I was a homeless Zen monk.

And yet, I found myself unable to rally the triumphant attitude of worldly renunciation that one reads of in the annals of the masters. Here I was, doing the kind of deep dopamine detox that Silicon Valley tech bros can only dream of—but I had to admit: I couldn’t shake my concern for my country, my family, and friends.

The first pangs of a feeling I can only describe as “survivor’s guilt” set in. I had escaped the madness, and found myself in a place that afforded a life that, while not easy, was safe, and more importantly, sane. In doing so, however, I left everything behind in the sinking ship of the United States.

And I had a bad feeling about that.

It drove me to justify something I would have never otherwise done: Sneak a smartphone into a Zen monastery. For compassion’s sake, right?

Finally, my moment came. I heard the ino, the ceremony leader, intone the first syllables of the Heart Sutra and calculated that I had about 10 minutes before the Abbot’s attendant for the day would come pick up a pot of hot water for his quarters. Trundling out my iPhone Mini, I clicked open the NYT to behold the stunning headline: “TRUMP INCITES MOB”!

Well holy shit. It was just a few pixels of black on a tiny screen of white, but the assembled combination of letters sent my head reeling. A stunned disbelief at what was there before my very eyes, offset by the sudden confirmation of all my suspicions. I knew it! At last, the “bad feeling” I had been harboring was borne out as an accurate intuition. I knew it! Things were going to shit. And that lingering fear—that maybe I had misread things, overreacted, fled prematurely—vanished.

In a daze, I scrolled through the scenes of rioters fighting police before the sounds of the attendant shuffling down the hallway brought me back: back from the fractured instantaneity of social media; back from a polarized America’s fragmenting culture; and back to my safe, sane, tidy life in the cold kitchen of the monastery.

I still had porridge to make.

Throughout the day, mostly alone in the kitchen, I chewed on my secret knowledge, turning it over in my mind as I salted radishes for pickling.

Telling anyone was out of the question; the phone had to remain a secret. Besides, in a monastery, discussions of politics are reserved for the Abbott. Lowly monks in training, by contrast, are actively, often loudly, encouraged by their senpai to take their precious opinion and go shut the fuck up about it.

On duty until 7 PM, when tomorrow’s cook would show up for a hand-off, my Cartesian theater was booked solid with a private showing of “Chicken Little.” The sky was falling, but I was the only one who knew anything about it.

In the weeks after Biden was finally inaugurated, it looked to many like life would return to a pre-Trump status quo. I still had bad feelings about it all, though.

Bad feelings

To begin, unlike so many Americans, I wasn’t horrified at the scenes of belligerent “patriots” defying police and defiling the halls of power. I had already turned my back on the mundane world—what did I care about disruptions to Washington’s pomp and circumstance? To be sure, the loss of life was lamentable. But the righteous destruction of hollow decorum in the name of a higher purpose? I could actually sympathize with that, having imagined myself to have done something similar when I walked away from Stanford and the academic path.

Indeed, what I felt was jealousy.

During morning mediations, one thought kept rattling round and round my head: “That should have been us!”

That should have been us, putting a stop to illegitimate politics and politicians! That should have been us, reclaiming the People’s House! That should be us right now! We, the defenders of democracy, should be up there—not fighting for a cretinous would-be King—but to expand the Power of the People! Fighting for our political autonomy! Defending our noble heritage of self-rule!

It wasn’t long before my jealousy morphed into anger. Not that it happened, but who did it—and what they didn’t do. I was furious that it was Hillary’s “basket of deplorables,” of all people, who had actually succeeded in striking at the throbbing heart of American politics. Not because I disagreed with their cultural views—though I did—but because these self-proclaimed patriots had somehow achieved what no progressive movement ever has: they breached the inner sanctum of American democracy. Then once inside, they squandered their precious chance.

Real revolutionaries would have seized the moment. These goons were there to “Stop the Steal”—but once it was stopped, most of them just stumbled around the place, gawking, stealing, pooping on stuff and posting on social media. Factions within the mob made futile efforts to chase down specific politicians, such as Mike Pence, but the mass of J6ers, with no real political vision beyond grievance, had no plans to execute.

Then, of course, my anger threatened to turn inward. Now the thought plaguing me during morning meditations was a question: “Well what would you have done instead?”

If it had been us

Instead of MAGA hats and Confederate flags, imagine it had been those who actually lost the most under minority rule. The teachers whose unions were busted. Voters from districts gerrymandered into irrelevance. The victims of opiates, whose lives have been ruined by Big Pharma and its press-gang of doctors. Workers of all stripes, whose wages have stagnated for fifty years while productivity soared and wealth concentrated upward. And sure, even kids—the very kids who will inherit an unlivable hellscape of a planet because the oil companies own Washington.

Imagine they breached the Capitol not to stop a legitimate vote count, but to demand the system actually count their votes equally. Not to re-install a strongman, but to stop the structures that make strongmen possible. Livid at the systematic closure of the extant channels of democratic change, and furious with a congress that made itself the prison of popular will, these revolutionaries would strike at the architecture of minority rule itself:

• Stop the Steal!? How about “Stop the Senate!” Abolish that obsolete body! At least end the two-per-state rule, which turns the happenstance of geography into de-facto aristocracy, and institute proportional representation! Or at the very least, put age- and term-limits on those fossils.

• Stop the Supreme Court! Abolish that cabal of lawyers with their lifetime appointments! Let us lay low those proven liars, who promised to defend the constitution, but instead ply their doctrine of “Originalism” to pay obeisance to a tyrant! Can we really go on letting nine unelected elites override 330 million people indefinitely?

• Stop the state-based national election system! Abolish the Electoral College! That anti-democratic relic was designed to accommodate slaveholders in the 18th-century and thwart the majority will. How long can we pretend that presidents who lose the popular vote are “democratically elected” leaders?

• Stop Citizens United! Abolish the legalized theft of elections by the propertied classes—everywhere in America! Money is not speech, and corporations are not people: This is legalized oligarchy wearing the mask of the First Amendment.

• Stop voting altogether and institute sortition! Instead of voters choosing politicians, we have politicians gerrymandering to choose their voters. Who still has faith in the two-party duopoly? The only way we can ensure proportional representation is random assignment—the Greeks knew this!

• Stop the Steal!? Stop sacrificing our sovereignty to self-interested strongmen!

My imaginary revolutionaries would come ready with demands about who gets to decide—and plans for radically restructuring the decision making process. The J6ers just wanted to keep their King—but us? We want a democracy of, by, and for the people. They had no real plans for taking power, because they were acting under someone else’s power. We, on the other hand, would take power from the elites only to distribute it more democratically.

Where are we now?

Soon, the event and its aftermath was stealing my attention, and my ability to stop worrying about the mundane world. I started sneaking the phone around with me everywhere, constantly checking the news in the few places I had privacy: on the toilet, early mornings in the kitchen, and under the covers before bed.

January 6^th was the day the algorithm got me into its clutches. Somewhere, a world away, Steve Bannon belched a few foul chuckles. My zone had been flooded.

In the years that followed, I graduated from the monastery, received transmission of the Dharma, moved into “my own” temple, became the abbot. I wish I could tell you that I was cool, confident, and unconcerned by worldly matters throughout all of that. But the algorithmic Skinner box still had me trapped. And like a pigeon on an intermittent reward schedule, I am ashamed to say, I kept coming back for more news from home, lest I miss another coup attempt.

It might just be the electronic fangs of the algorithm preying on my psychology through its many “feeds,” but in the years since J6, my anger has turned to fear.

Fear that there is no “us.” A fear that “my side” doesn’t actually believe in American democracy. Not as practiced, anyway. We’ve spent decades pointing out that the franchise is a rich man’s game, that the Founders were slaveholders, that the Constitution was con-job designed to constrain majorities, and that capitalism corrupts everything it touches—especially government. We’ve taught ourselves that democracy never really existed in America, that it’s always been a sham.

It’s no surprise, then, that when what might become the last presidential election to take place in our Great Nation rolled around and the Democrats called on us to “save democracy,” a great many of us couldn’t be bothered. What is there to save?

Now, as it faces destruction at the hands of a rampaging Tyrant, all too many of us are falling back on the very mechanisms of democratic governance we had poo-pooed for so long: hoping military leaders won’t follow illegal orders; counting on the courts to save us; and just waiting until the midterms.

This is weak tea for potent political times.

Today we confront a stark possibility: Two and a half centuries into this beautiful experiment, a new democratic revolution may be necessary to remove a tyrant from our midst, once again. Is our side up to the task? Do we have a vision beyond mere grievance?

The enemies of democracy already drape themselves in the flag and parade as true patriots. They have already claimed 1776, and the Founders. They claim to revere the Constitution as secular holy writ, and deploy it as selectively as the Bible. And, as January 6^th showed, they can weaponize this mythology of America to great effect.

What do we have? The New Deal? The Civil Rights Movement? Though important, these are mere reforms, not revolutionary origin myths we can call upon. The white, Christian America they conjure may have never existed, but it motivates them. Meanwhile, we have let anti-whiteness discourse poison the well of our nation’s history, dispelling any possibility of drawing inspiration from it. As Catherine Liu points out, the 1619 Project and other attempts to reframe the founding solely around slavery ultimately serve the MAGA cause by foreclosing the possibility of interethnic solidarity around shared goals—namely, broader, more inclusive forms of democracy.

When asked who George Washington was, how many of us will respond with contempt “slave owner” rather than “OG revolutionary”? Both are true. But only the latter gives us something to build on.

Yes, the Founders were flawed men, with questionable goals. Yes, the franchise of American democracy was only gradually and grudgingly expanded, and still remains imperfect today. None of this invalidates the democratic horizon this nation was founded on. True democracy has never existed anywhere, because it is not a black-and-white state of affairs—it is an aspiration, a practice, and a path to tread.

But clever postmodernists have deconstructed the American experiment so thoroughly that we no longer know how to put one foot in front of the other. While the J6ers were storming the Capitol, we were debating whether the building deserved to stand.

January 6^th was a portal into this hellish timeline. With the 2026 mid-terms looming, another one is due to open up.

When it does, on the 250th anniversary of the founding of our Great Nation, who among us will be willing to walk the path of democracy?

Imagine this: It’s 2064, and you’ve been tasked by the Civilization Disorder Prevention Bureau with preventing the events that set in motion the collapse of the 20th century international order. You’re going to be sent back in time. Your target: Donald Trump.

No. Your mission isn’t to sneak into little Donnie’s nursery with a giant rock in hand, like so many sci-fi fantasies about preventing the rise of Adolf Hitler. You’re going to intervene to persuade his parents to parent him more gently.

Ridiculous? Probably. Where am I going with this? Not sure. Michael and Shunryu have pieces in the works that require a lot of in-depth research. But I wanted to remind you all that Taming Complexity is still live.

In any case, what got me thinking about baby Trump was a post that Kat Rosenfeld recently published, a piece that I think gets to the heart of the contradictions of liberal societies, the incoherence of peoples’ demands to “normalize” whatever it is that they happen to like or do.

Rosenfeld juxtaposes two different examples of this. The first is relatively innocuous: social media whining about getting judged by others for having a messy house with mismatched furniture. The other case is about the worry over “normalizing” Donald Trump. Both examples direct us to think more deeply about the machinery that is necessary to sustain a society.

The first case brought to mind philosopher Charles Taylor’s much earlier diagnosis of the malaise that infected modernity. Modern liberal societies tend to heroize the artist (or more recently the innovator), the people who transgress the limits set by their culture. The trouble is, as Rosenfeld also notes, is that transgression can only be meaningful in the context of those limitations. No cultural horizon, no actual transgression. Without pearl-clutching normies, freaky behavior loses a lot of its excitement.

The expectation that everyone can pursue a wholly individually determined authentic lifestyle is a trap, because without societal norms as guardrails, all that is left is empty self-referentiality. The outcome is nihilism, in which individual values become paramount but also simultaneously trivial. You can do whatever you want, but none of it matters.

This example stood out to me not because I care that much about standards of good taste in interior decoration, but rather because my wife and I are like other millennial parents: hopelessly lost in the morass of competing parenting advice. Gentle parenting doesn’t promise transgression, of course, but transcendence, a final surpassing of the claimed cycle of trauma caused by traditional authoritarian childrearing.

But the philosophy has no cultural center. Instead, therapy-speak about boundaries, safety, and emotional self-regulation is meant to undergird a process of conflict resolution, without having a clear model of what kind of adulthood that children are being prepared for. And as much as I’m in favor of my sons learning to speak more kindly, read their own emotions, and all the other things wrapped up in gentle parenting, it’s hard not noticing the massive void of guidance regarding what exactly the transition from, say, boyhood to manhood ought to look like.

What Rosenfeld captures about what “normalization” demands is also true for gentle parenting: an intense aversion to “shaming.” A child should never feel ashamed for crossing a boundary (which are notably individual but never cultural), but redirected to reflect on why they did it. A great deal of this, of course, falls by the wayside in actual parenting, but the fear that a child somewhere might be made to feel worse about make a mistake is so intense within gentle parenting spaces, it is hard not to wonder if there’s a baby being thrown out with the bathwater (and if the whole movement is itself one giant collective trauma response)

That’s not to say that these spaces are free of shame, but only that it is reserved for particular cases: the father who is hesitant to get on the gentle parenting train (“sounds like a potential abuser” is literally written in Reddit threads on the subject) or those Karens who deign to have strong opinions about people’s tastes and life choices. The paradox of achieving a gentle society that normalizes everyone is that it requires heaping scorn on anyone who is insufficiently nonjudgemental.

What I think Rosenfeld is getting at is that this makes for a poor foundation for social norms, that we ought to be more willing to participate to assure a culture that aspires to something more than vast networks of individual noninterference treaties. Shame is a powerful tool, and arguably the most important one in sustaining a society. We should lament it being employed excessively or toward the wrong ends, but to pathologize shame itself is mistaken. For instance, as much as I sometimes resented German neighbors who looked askance of my sloppy American habits, the shame was intended to prevent me from freeriding on the orderly and immaculate country that they had built.

And that’s what this all has to do with Donald Trump. The left-wing refrain has been to refuse to “normalize” the Trump administration. But as Rosenfeld points out, being elected president twice invariably normalizes the man as a politician. The cries of leftists can’t change that. However, the only way that the “undignified, vindictive, astronomically petty, and wildly destructive” behavior of Trump becomes itself normalized, according to Rosenfeld, is “if everyone else — young and old, Democrats and Republicans, public servants and civilians — starts doing it.” Even if the country’s top executive seems immune to shame, we shouldn’t start acting shamelessly in response.

I can’t prove that a culture that increasingly flees from the idea of setting social guardrails in one sector of society is less able to enforce standards of decorum in politics and media, but it seems pretty plausible. It’s hard not to find a better example of someone who seems to only care about what the voice in his head says is right than the current sitting president (something that sociologist Robert Bellah found to be an increasingly common view of morality, what he dubbed “Sheilaism”).

At minimum, I think most people would agree that there ought to be higher standards for today’s public officials and for most of what passes for public discourse. The question is, however, how exactly to do we get there? Once the guardrails are fully gone, either blown up by Trump himself or by those trying to oppose him, how can get them back? In a society whose recent history has mainly been about eliminating social norms, or imposing new ones in a self-serving and polarizing fashion (i.e., “woke” politics), it hard to think of positive exemplars. So, if you have any ideas. I’m all ears.

Subscribe now

Share

Leave a comment

It’s no secret that President Trump has been targeting American universities in an effort to crackdown on Diversity, Equity, and Inclusion (DEI) policies. The response has been mixed, with a few schools like Harvard promising to resist, while others have caved and renamed their DEI bureaucracy “The Office of Belonging” or something similar. Like most other cultural issues during the last decade, opinion over university DEI is polarized, which is to say that many people, especially loudest voices, seem incapable of having a reasonable discussion over it. But it’s far from clear how we’ll ever be able to.

I’ve had about a month to decompress after witnessing my institute’s own faculty senate argue over the inclusion DEI language on my institute’s website and in its strategic plan. It was a debate that one of my colleagues later referred to as “kabuki theater,” and, indeed, like politics today more generally, the conversation was far more performance than the kind of dialogue one would hope could occur among the most highly educated people in the state. And if they were incapable of a sensible conversation about diversity, equity, and inclusion, what hope is there for the rest of America?

---

DEI has a lot of problems. Among the most significant is that there is very little evidence that current practices work. Musa al-Gharbi surveyed much of the literature on DEI trainings back in 2020. He found that ineffectiveness wasn’t even the worst of the issue with DEI. Often these trainings decreased cooperation across racial lines, increased resentment of poor whites, heightened feelings of alienation and anxiety among minority workers, and sometimes even generated greater fatalism about racial bias. To top it off, a good number of workers now simply simulate progressive beliefs to please DEI bean counters.

There’s a DEI committee at my own university, and like other institutions of higher learning, we have a few well-paid DEI administrators. To my best recollection, these committees and offices have never provided evidence that their efforts have had an impact on recruiting and retaining students from non-privileged backgrounds. I actually have no idea what they do.

The only noteworthy occurrence was an all-staff email thread from the DEI office that turned in a struggle session when one poor bastard, failing to read the peak-woke-era room, sent out a cliché meme about us all being the same on the inside. I seriously worried that we’d find him hanging from a rope in his office.

Going into the faculty senate discussion, part of me hoped that practical considerations about whether DEI was actually working might be part of the conversation. Instead, we sat through a half hour of culture warring. It didn’t help that the instigation for the debate was from a professor who referred to DEI as an “ideology.”

Left-wing faculty swooped in to defend DEI as morally equivalent to civil rights in the 60’s, as no different from supporting diversity in general, and as the same as following the ten commandments or other non-negotiable ethical frameworks. Others accused DEI skeptics of carrying water for the Trump administration. It was like a real life reenactment of a Bluesky thread.

Most fascinating (and tragic) about the exchange is that the relevant literature, including al-Gharbi’s survey, was circulated prior to the meeting. But it didn’t matter. The battle lines had already been drawn over what DEI symbolized, not its actual impacts.

To me, it felt like an illustration of what philosophy Liam Kofi Bright referred to as the “white psychodrama” characterizing today’s culture war over race: a largely symbolic fight between members of the intelligentsia over “how to psychologically manage the results of living in a materially deeply unequal society, not a conflict about how or whether to reduce that material inequality.”

What I think Bright would say about debates over university DEI departments is that they are, for the most part, materially unimportant for minority groups. Much of the literature suggests that DEI practices are largely performative efforts to help left-wing managers and employees assuage their feelings of guilt and believe that they are battling discrimination, which means that they’re one of the “good guys.” Their staunchest opponents, however, aren’t much better, seeking to downplay ongoing inequalities with insistences on color-blindness and efforts to avoid talking about race at all.

This is similar to what al-Gharbi argues in his book We Have Never Been Woke. Wokeness hasn’t achieved woke political goals. Instead, it seems to have been mostly an elite status game, an elaborate system of etiquette valuable more for signaling one’s virtues than actually practicing them. The fact that Harvard is freezing the wages of their custodians, who are disproportionately nonwhite, while maintaining an army of highly paid DEI staff is emblematic of al-Gharbi’s point about the contradictions of the woke era.

Yet, on reflection, it was naïve of me to think that a constructive discussion of DEI would have been feasible, given that a large chunk of America’s symbolic analysts are staunchly committed to one side or another of the psychodrama. Trump’s effort to strong arm universities into dismantling DEI has negatively polarized much of the left-wing electorate into wanting to defend it. The situation is little different from populist reaction to vaccine mandates, albeit with vastly more affluent and educated citizens now seeing themselves as in the resistance.

Part of me wonders if this is what Trump wants, more so than actually dismantling DEI: that the left rallies behind a set of diversity-promoting activities that only a small minority of Americans believe to be effective.

---

What to do? During the COVID pandemic, I was one of only a few voices arguing that we had no hope of persuading vaccine skeptics by trying to reveal their ignorance or accuse them of being deluded by conspiracy theorists and misinformation. Their hesitancy had far more to do with feelings of mistrust in Anthony Fauci and the FDA, with the sense that they had no input in the public decisions that were upending their lives and threatening their livelihoods.

Vaccine hesitancy was made worse during the pandemic, because most advocates refused to try to understand skeptics as human beings—instead treating them as mere disease vectors.

Finding our way out of the unproductive mess that is the contemporary debate over diversity, equity, and inclusion is a similar scale of challenge. But I’m currently at a loss to understand the loudest voices in that debate in a charitable way. Liam Bright chalks the situation up to cognitive dissonance. Philosopher Dan Williams would probably label it an issue of “highbrow misinformation,” where educated elites fall prey to their own version of groupthink.

I’m not happy with any of those explanations, because they strike me as patronizing, as repeating the pandemic-era mistake with vaccine skeptics. Even if there is a kernel of truth to seeing the battle over DEI as a product of status-seeking, black-and-white thinking, and other personal deficiencies, recognizing as such doesn’t suggest any productive pathways forward. As I wrote in my second book, the only hope for bringing people back from the brink of fanaticism is by meeting them halfway, not by diagnosing them as cognitively broken. Only by understanding them as reasonable people can we do that.

So, how might a discussion over DEI turn out differently? Psychologist Peter Coleman contends that difficult conversations are only possible by reframing our political disputes as complex layered sets of trade-offs and dilemmas, rather than black-and-white debates over truth and falsity or good versus evil. The only way that faculty at my university and others will be up for the challenge of tough discussions about diversity, equity, and inclusion is in an environment that reawakens our curiosity about the challenges and quandaries inherent in those ideas, and pushes us out of the ruts of our deeply rooted political divides.

Leave a comment

Subscribe now

Share

The participatory technology assessment forums were, themselves, well organized and run. But they were, essentially, just tacked on to an existing process; a sort of after thought with limited impact on the ARM as a whole. This approach undermined the participatory forums in a number of ways.

First, the scope of participatory technology assessment was far too limited to have any substantial influence over public perception and support for the asteroid mission.

Public engagement through a single assessment over two forums is just not broad enough. The participatory technology assessment only consisted of 200 people in two locations. For the purposes of studying how such forums might work in practice, that’s a pretty exceptional size! But it definitely isn’t enough people from enough places to have any impact on overall public perception.

Perhaps small forums would be enough if they were being run regularly for all of NASA’s proposed missions, but this was the first, and only, of their kind for NASA. These forums weren’t one part of a broader trend of public participation in setting mission priorities for NASA. Imagine if, instead of a one-off that you probably have never heard about before now, every NASA mission engaged about 200 people in various cities and towns across America. Even if you, personally, were never included in deliberations, chances are you would still at least feel like NASA genuinely cared about working for people like you. But instead, by making these forums a one-off, NASA inherently limited the benefits of their impacts.

Second, the deliberations did not occur early enough while the goals and designs of the mission were still flexible.

It turns out that the Keck Institute for Space Studies published a study in 2012 about the feasibility of capturing and returning a small asteroid. This study, several years before the participatory technology assessment, was the basis for the concept of ARM. The Keck proposal’s “key example” of why such a mission would be useful is as a jump start for asteroid mining. The proposal notes that the expense of getting high density materials, like water, beyond Earth’s orbit might make mining those materials from captured asteroids economically feasible. What the Keck report outlined was a mission that would contribute the foundations to a future where we captured asteroids for their resources. ARM would test the basic technologies and techniques necessary to open up that future as a possibility.

In other words the purpose of the Keck proposal was agenda setting. Agenda setting answers the question: which issues are appropriate for the system of policy making to take into account? In the vast set of all possible problems for the government to address, which ones are even in consideration? In the case of spaceflight, agenda setting is especially important. It can be hard to determine what policies and missions NASA should even consider when so much of their work can seem like science fiction. The Keck proposal takes the very science fiction idea of asteroid mining and connects it to a very feasible and achievable mission. It makes asteroid mining something that NASA and policy makers can take seriously.

So the Keck proposal set the agenda for the ARM mission before public engagement even began. The choices on the table for the public forum participants to deliberate over were already pre-determined. By not occurring early enough to influence agenda setting for the mission, the impact of public engagement was fundamentally limited.

Moreover, a NASA mission is a pretty technical undertaking. Just because something is technical, doesn’t mean that the broader public shouldn’t get a substantial say. However, the further along in the process, the more central technical decisions become. The beginning stages of developing a mission start with questions about priorities and values. What do we care about more: scientific objectives, testing and demonstrating new technologies, or the prestige of human spaceflight? But the further along the mission gets, the more the questions start looking like: what kind of sensor will get the data that we need given our power constraints? The ARM forums weren’t looking at questions nearly so technical as this, but they also certainly were not giving people influence over basic values decisions.

In the end, the results of the participatory technology assessment just weren’t very influential over the final design of the mission.

In the Keck proposal, planetary defense (figuring out how to keep asteroids from smashing into us) is a secondary consideration. The conception of ARM outlined in the Keck proposal has “synergy with planetary defense,” but planetary defense is not the purpose. These differing values from the results of the participatory technology assessment show themselves in the decisions made about ARM.

The first step to planetary defense is knowing what’s out there: finding and tracking threatening near Earth asteroids (NEAs). Similarly, the first step for asteroid mining is surveying NEAs for those with water or valuable minerals. On the surface, it seems like ARM could achieve both of those goals at the same time, but it turns out the dangerous asteroids and the valuable ones are totally different.

The majority of potentially hazardous NEAs are S-type asteroids. These mostly consist of silicone based rocks and iron. They don’t contain anything valuable enough to mine. On the other hand C-type asteroids, which are rich in volatile (watery) and organic compounds are much better targets for asteroid mining. While ARM did not have a future target asteroid by the time Congress cancelled it, NASA used 2008 EV5, a C-type asteroid, for planning and in simulations.

Even if you find a dangerous asteroid heading towards Earth, you still have to do something about it. As epic as landing a drilling team to bury a nuclear bomb deep inside the asteroid and then blowing it up would be, it would be a lot less dangerous to just use gravity to nudge it onto a different trajectory.

But, “the best laid schemes of mice and men go often askew.” The universe is a complicated place, and the theory of gravitational deflection is not likely to match the reality of practice. It would certainly be best to have some experience before the real deal. The original ARM plan called for just this practice. A small robotic spacecraft would use its own gravity to “tow” a small asteroid into lunar orbit. Once there, NASA could send other spacecraft, or even astronauts, to study the asteroid relatively easily.

Then, in 2014, the NASA Advisory Council (NAC) advised NASA to review the mission design in light of the risk of potential cost overruns. So NASA developed an alternative where, rather than practice gravitational towing, the spacecraft would be fitted with a robotic arm that would lift a small boulder from a larger asteroid and move just that boulder into lunar orbit. Option A versus option B. While option B would not be useful for learning about planetary defense, at least it would help prevent potential cost overruns right?

Well, not exactly. Independent cost and technical assessments of the mission indicated that option B would likely cost $100M more! And worse, by this point, members of Congress had already voiced their concerns over the quickly escalating costs of ARM, and were already publicly considering denying funding for the mission in appropriations. It seems NASA officials managed to simultaneously ignore the priorities set participatory technology assessment AND the very clear warnings from Congress.

A cynical interpretation might be that they were trying to appease the industry interests represented in the Keck proposal. But more likely is that the Keck proposal simply came first. It established the underlying purpose for the mission, the values through which decision-makers interpret their choices. And since there were no alternative interests vested with actual decision-making authority, no room in NASA decision-making structure for disagreement, there was also no one to question those initial values when it came time to make important decisions.

So what does this failure teach us about public deliberation? It teaches us that deliberation is not sufficient in and of itself. It may seem obvious to say, decision-makers have to actually act on the product of deliberations for them to have any meaningful impact. Making that actually happen is trickier. First, deliberations have to happen early enough, or else agenda setting and good old fashioned decision momentum will undermine the results. And second, it must be continuous. It is not enough to have a one-and-done discussion. Baking in the kinds of diversity of perspectives that comes from a well designed deliberation all throughout the decision-making process will prevent the kind of ossification of policy direction that eventually doomed the Asteroid Redirect Mission.

It might already be happening. Despite enthusiastic adoption across nearly all sectors of the economy, a recent MIT study found that 95% of businesses’ attempts to integrate AI into have failed. A few contrarian investors are already pulling back from the industry, and the question many are starting to ask is, “How long can it last?”

Nevertheless, investment in AI is projected to reach a unprecedented levels this year, with estimates from $200 billion to $380 billion across the Big AI firms (I call them MAGAF: Microsoft, Amazon, Google, Apple, and Facebook).

Big Finance is getting skittish. Goldman Sachs has estimated that $600 billion in new annual revenue would be necessary to justify the investment, leaving some analysts to conclude that at this rate, “anything less than spectacular results will look like failure.”

As if in illustration this principle, after AI chip-maker Nvidia—perhaps the only AI company selling a real product at a real margin to real customers—delivered a solid earnings report last week, CEO Jensen Huang lamented that “the market did not appreciate it,” and explained their oddly precarious position as follows: “If we delivered a bad quarter, it is evidence there’s an AI bubble. If we delivered a great quarter, we are fueling the AI bubble.”

That is, despite the weak outcomes, dubious math, and fears of a bubble, the money keeps pouring in.

Why?

Subscribe now

Until recent updates, LLMs had a strange trait: Despite all their sophistication, they could not reliably perform basic arithmetic.

Could this be true of their creators and owners as well? Technical geniuses, who, having risen to such grand economic heights by manipulating numbers, nevertheless remain unable to comprehend the difference between red and black ink?

Unlike in the DotCom era, however, the current heads of Big AI are not technical savants, innocent of business sense: they mostly have MBAs, not engineering degrees. Presumably they are shrewd capitalists, or else they wouldn’t be there.

What, then, would motivate them to ignore basic economics and put their respective enterprises at risk?

Just blitzscaling, bros?

It could be that this is the natural outcome of what investor and PayPal mafioso Reid Hoffman called “blitzscaling,” an update on the classic, uh, philosophy of “move fast and break things.”

Beloved by Silicon Valley, blitzscaling is a strategy of prioritizing speed and scale at all costs, in which one seeks to establish a first-mover advantage and use that to dominate a market. The “innovation” here is that unlike past eras, the goal of becoming profitable comes after the company has become indispensable. Amazon pulled this off. Uber did not.

Because self-sabotage of the AI industry seems unlikely to be intentional, Dotson’s analogy to German windmills doesn’t line up perfectly here, but the rapid scale-up is surely a means of squashing the competition. Outside of China, few companies possess the resources to build their own models, as training costs can exceed hundreds of millions of dollars. Now that many gigantic models are available for free, fewer still will be inclined to even try.

At this point, the playing field has effectively been cleared out. Most “new” AI apps are just an interface-layer on top of someone else’s model.

By choosing to unleash ChatGPT without notice (in the middle of the school year, no less), Big Tech’s dubious Wunderkid (and possible murderer) Sam Altman, intentionally or not, set off a competition for pole position in a race to dominate the AI market. As he and his competitors see it, now they have no choice, business-wise, but to push as far as possible as fast as they can.

Share

Unless he has an ace up his sleeve, however, by going all-in on scaling LLMs as fast as possible—despite growing evidence scale alone is not all you need—Altman, even as he attracted unprecedented sums of investment, may have simultaneously dug himself a deep bed that he and those following his lead will eventually have to lay down in.

That would be too bad for the AI industry, sure, but also for the average American. The journalist Stephen Witt, author of a recent book on the rise of Nvidia, currently the most valuable company in the world, estimates that it holds 8% of the value of the S&P 500, with another 7% in Microsoft.

Given that Big AI constitutes Nvidia’s principal customer base, and that American’s 401Ks and IRAs closely track the S&P, bursting the AI bubble would shave off a significant percentage of millions of Americans’ retirement savings.

It could be, then, that blame for the next recession—or depression—could reside with the “blitzscaling” ideology and business culture of Silicon Valley.

“Too Big to Fail,” by design?

It could be that “too big, too fast” is a deliberate strategy taking advantage of this recent-but-close tie between Big AI and the broader American—and hence global—economy. Ignoring serial liar Sam Altman’s claims to the contrary, Big AI firms, with their enormous (and dubious) valuations, might be confident they are already “too big to fail.”

That confidence might be misplaced. Commentators have pointed out that the analogy to Big Banks circa 2008 is tenuous. Whereas banks were highly leveraged, Big AI is not. That means the banks held more debt than assets when the crash came. But if OpenAI was bankrupted tomorrow, the cash-rich tech firms it owes money to would remain solvent.

In other words, if OpenAI, Anthropic, and other cutting-edge AI startups were to faceplant hard and fast, it is unlikely to upend the global economy.

Moreover, whether or not we agreed with them, in other cases of “too big to fail,” the companies involved could at least make a case that their products were intrinsic to economic life: Everyone needs cars and insurance! But Big AI firms, whose products kill American jobs, will have a much harder case to make.

This may be why the industry’s leaders have rushed to demonstrate fealty to President Trump. After promising absurd sums of investment in America, they likely feel sure of a Big Bailout, when and if the time comes. Apparently, vibes matter more than mathematics in these rarified social strata.

In the tech tycoons’ much-publicized meeting with Trump, Meta CEO Mark Zuckerberg apparently made up his pledge of $600 billion on the spot, though he has since showed further fealty by doubling-down on the number, and actually appears to be on track to hit that goal. This, despite his investors’ legitimate concerns. On a recent earnings call, when one analyst pressed him about how, exactly, Meta would make money from the investment, Zuckerberg underwhelmed his rapt audience, saying merely that, “The research is going to enable new technological capabilities to exist…and then those capabilities can get built into all kinds of different products.”

As to what those products might be, however, the young titan of industry did not elaborate. Perhaps unsurprisingly, the stock fell by almost 9% following the call.

It could be, then, that responsibility for the next recession—or depression—will lie at the feet of elites who exploit the close ties between Silicon Valley and Washington DC.

Where’s the beef?

Despite impressive technical advances and the release of a plethora of tools to generate “AI slop” at scale, Big AI has yet produced no “killer app.” Even the AI coding tools (basically autocomplete for programmers) have flopped. They were presented as the breakthrough necessary to make the myth of the “10x engineer”—someone who performs the work of 10 colleagues—into a reality. That has not happened. In fact, as a Google executive recently admitted, most AI-written code does not make it to production.

If AI disappeared tomorrow, the kids might gripe about having to do homework on their own again, but nothing essential would be missing from American life. And besides, kids don’t pay taxes.

In my next post, I’ll consider two more explanations for the continued investment despite bad math and lack of products; both irrational, but one highly plausible: the so-called “AI Arms Race” with China, and hunt for “superintelligence,” the Holy Grail of AI.

Musa al-Gharbi posted a great overview of the all the bad election hot takes, finding that, as far as the numbers are concerned, Mamdani isn’t a whole lot different from previous mayors. The coalitions that put them into power looked about the same, and the differences between those who voted for Cuomo and those who cast a ballot for Mamdani aren’t all that significant.

As I pointed out after Mamdani’s primary victory, a great deal of the pundit class have a narrative to push and they very much want Mamdani’s electoral success to corroborate it. And there’s an infinite array of statistics that can massaged fit their preferred storyline.

The story that I see, partly thanks to al-Gharbi’s summation, is a small blip in an otherwise clear trend of political apathy and polarization. Like for a great deal of other elections, the non-official winner was “no one,” because most eligible New Yorkers could be bothered to go to the polls. And most of those who were newly galvanized show up to fill out a ballot card seemed to have been there primarily to try to prevent a Mamdani mayoralty.

Whatever the history books will conclude about Mamdani’s tenure as mayor, it will have little to do with the slice of the population that put in into office and everything to do with what he actually accomplishes as the city’s executive. We’ll just have to wait and see.

Subscribe now

Leave a comment

Share

On a very windy October day in 1983, a group of engineers and utility executives gathered on the coast near Hamburg, Germany to witness the maiden test of Growian. It was the biggest wind turbine built to date, coming in at a full 3 megawatts. The turbine was supposed to serve as forerunner to the coming renewable energy future, a gargantuan testament to Germany’s engineering prowess. Too bad it ended up being a piece of shit.

Growian, a contraction of grosse Windenergieanlage (literally “big wind turbine”), failed after a mere 420 hours, or the equivalent of 17 and a half days of continuous operation. No crane was tall enough to assemble the massive turbine, so parts had to be made small enough to be pulled up the interior of the units tower. And the resulting welded joints of the turbine’s housing proved to be too weak handle the mechanical stresses that electricity generation produced.

Energy scholars Craig Morris and Anne Jungjohann, suspect intentional corporate sabotage, given that at least board member at the utility stated that “we need Growian to prove that wind power won’t work.” However, wind turbine prototypes across the pond were equally pathetic. One should be skeptical of stories that involve malice and conspiracy when plain old foolishness is the more likely root cause.

In America, almost a half a billion dollars and all the expertise that engineers from both NASA and firms like Boeing and General Electric could muster ended up producing a full decade of wind turbine failures. Many prototypes required serious maintenance after as a little as 40 hours. None operated for more than a few years. Nearly all were gigantic, especially so for the era. Based on these cases, a person in the 1990s could be forgiven for believing that wind energy was a failed technology.

But in Denmark the wind turbine was transformed into something commercially viable. The path to success here didn’t involve millions in federal R&D or aerospace engineers using computer models and the latest material science. The motive force for wind energy innovation came from a seemingly unlikely set of firms: farm machinery builders.

Known today as the biggest turbine manufacturer in the world, the Danish firm Vestas started out making milk coolers and small hydraulic cranes. Although it is tempting to see airplanes as the closest analogue to a wind turbine (given the need for propellers and the importance of aerodynamics), hands on experience with agricultural implements proved decisive for Danish companies. No matter how good a turbine’s power to weight ratio or its predicted energy capacity, if it isn’t resilient, it’s worthless. German and American engineers chased technical sweetness, a turbine that might have even had a fighting chase at getting airborne, when what they needed was something that was as robust as a tractor.

The innovation process in Denmark was also far better organized to support ongoing learning. Designs had to be proven on a government test stand prior to licensing, which helped provide performance feedback and access to engineering expertise to firms still new to the turbine business. And the companies shared both successes and failures through an industry journal. And most importantly, in the absence of massive influxes of other people’s money, these firms had to make things that they could actually sell.

Moreover, government policy ensured a place for wind turbines within the energy grid. Regulations were developed to ease their integration. Feed-in tariffs and other relatively light forms of subsidy were implemented to help commercialize them. All these factors put together not only turned wind turbines from a small-scale farm implement into a serious energy technology but also resulted in Danish companies absolutely dominating the market in its early years.

One common truism is that while history doesn’t exactly repeat itself, it certainly rhymes. As a number of economic indicators now signal a coming burst of the AI bubble, I couldn’t help but hear the word Growian ringing in my ears.1 If the bubble indeed bursts, the last five years of AI development should be seen in similar terms as the American and German failure to turn a theoretically plentiful resource into a practical reality: an ongoing act of pompous hubris, of arrogant stupidity.

We can see it in the innovation strategy being employed. Propped up by billions in venture capital, AI firms have single-mindedly chased narrow technical measures of success. Like the builders of the Growian turbine or its equivalents in America, they have acted as if they believed that an AI system only needs to be big, fast, and comprehensive enough, then its deployment and integration across the planet’s sociotechnical systems is basically a done deal: Build Skynet and it takes over the world.

Just a little knowledge of the history of innovation shows us that the technological advancements are more than mere technical achievement. They’re a matter of reweaving currently dominant sociotechnical systems so as to make way for new technologies. Thomas Edison’s genius didn’t lie in superior knowledge or analytical ability, he sought to develop or acquire, and then deploy, the technological infrastructure and regulatory policies that enabled electric technologies like the light bulb to easily become a part of citizens’ everyday lives. One does not simply build a technology and think the rest will come, at least if you want to be successful.

But that is exactly what the main AI players are doing.

What should have been done instead? Or what should AI advocates do after the bubble bursts and we likely land in yet another AI winter?

Rather than make yet another spectacular gamble on artificial general intelligence, the industry would be better served by a more humble and incremental approach. Discover the problems that AI truly help solve and develop small marketable products that can address them. Commercialization shouldn’t be an afterthought but one of the initial steps. It’s such basic UX thinking (find out what the user needs and wants before building a product) that I have a hard time not wondering if there has been an unreported epidemic of serious head injuries among AI leaders, an ill-fated billionaire rager in Aspen.

In my conversations with colleagues, it’s become clear to me that AI won’t be anything more than a hinderance to universities in the near future, simply because Silicon Valley developers had never bothered to understand education as a sociotechnical system. I would immediately petition my department chair if readily available AI products were explicitly designed to enhance students’ learning, to make them better at exercising expert judgement, rather than to give them an alluring escape from the frustrating uncertainties of doing their own work.

As Colin Garvey pointed out last week, what we wanted was artificial personal assistants. Instead the AI industry has given us digital sycophants that try to do everything possible, including lie to us, to keep us forever distracted, forever engaged.

Yet, as we’ve noted in past posts, the Tech world isn’t set up to reliably solve people’s actual problems. Corporate leaders believe in moving fast and breaking things. They want to monopolize new market segments in order to make monopoly rents. Whether it actually makes citizens’ lives better is seemingly irrelevant. The trouble is that in many cases what actually breaks is the technology’s own chances for a sustainable future. A hubristic vision of progress is ultimately self-stultifying, delaying broadly shared advancement rather than accelerating it, as countless dollars get wasted on failed technological megaprojects—and the rest of us are left sifting through the wreckage when it all goes to shit.

Without a set of brakes (and a steering wheel) on the overenthusiasm and the discretionary power of venture capital, we’re seemingly destined to hear the rhymes of Growian over and over again.

Share

Subscribe now

Leave a comment

On the one hand, Boosters promise a future to dream about, where dull, dirty, and dangerous jobs are handled by robots, leaving each of us to pursue our interests with a fleet of powerful digital assistants at our beck and call. (Bracket for now the fact that in the same breath, they add that each of us will have to rapidly adapt and optimize to thrive in this brave new Age of AI—or risk being left behind.)

On the other, Doomsters, including Nobel Laureate Geoffry Hinton, one of the so-called “AI godfathers” who developed many of the techniques behind the current boom, warn that utter disaster, even extinction itself, awaits us. The title of a recent Doomster manifesto sums it up: If Anyone Builds It Everyone Dies.

Boosters and Doomsters have polarized the discourse into deadlock, yet technical progress proceeds apace, with new advances each week. Society still hasn’t figured out what to do about deepfakes—but have you seen the new videos generated on Sora 2? No robust protections for minors exist and real harms are emerging—but chatbot usage is higher than ever, with tens if not hundreds of millions of children and adults interacting with generative AI systems worldwide.

And despite a growing pushback from artists, musicians, booksellers and even the New York Times—one of multiple news outlets suing OpenAI and Microsoft for copyright infringement among other crimes—almost everyone, from small and mid-sized businesses to superintendents of school districts, is encouraging their employees to make use of the new tools…somehow.

Who to believe? Should we be preparing for takeover by intelligent machines, or a life of leisure and luxury with robotic assistants?

A heuristic for cutting through the crap

The situation hasn’t changed much since I was doing my doctoral research on AI back in the late 2010s. Big advances were already being made—AlphaGo, chief among them. Superintelligence, the 2014 book by the now-defamed philosopher Nick Bostrom, had put the existential risk of “artificial superintelligence” (ASI) on the table as a topic for serious discussion. Meanwhile, Siri and Alexa had become household names, and AI was getting cover stories in Science and Nature. Still, nothing like ChatGPT had been released yet, so the AI hype was proliferating unencumbered by broad experience with actual systems.

At every conference I could get myself into, from semi-social scientific colloquia to the AAAI, experts of all stripes agreed that AI would transform the world in coming years. But as for what that meant, it was all pet theories and speculation.

I had trouble putting all the pieces together. Interviewing experts—not the famous faces of the AI industry, but academics and industry types—helped orient my investigation.

In particular, a conversation with a grizzled computer science professor, one of the old guard who lived through the AI boom of the 1980s and had seen it all, has stayed with me. As I probed him on recent prognostications emerging from the AI hype machine, he interrupted, telling me not to go on because he would “give me a heuristic that will teach you how to cut through all that crap.”

He pointed me to an Apple ad from 1987 about something called the Knowledge Navigator. “It’s on YouTube,” he said. “Check it out. Until we have that, don’t talk to me about ‘machine intelligence.’”

Sure, one might complain that it’s a product of a bygone era. The charming-though-forgetful professor—a socially powerful white male hitting up his female colleague for help and ignoring his poor mother—would have probably been cancelled during #MeToo. But damned if I didn’t agree completely with the white male professor who pointed me to that video: That is what I want from AI.

My fellow knowledge workers: What say you? Seamless interaction with an actually intelligent assistant who smoothly moves from visualizing research results and integrating datasets to scheduling meetings and taking calls. Say what you will about the interface with its Bill Nye-esque floating head icon—nearly 40 years ago the product designers at Apple knew what we wanted: The right intelligence, at the right time, in the right amount, for the right purpose.

Consider the role AI plays in your own life, and I think you’ll agree we’re nowhere near that level of fluidity, much less usefulness and relevance. Sure, we can talk to devices, but I don’t see anyone conversing with a device as an assistant yet.

As of late 2025, supposedly living in the Age of AI, we’re not even at Bladerunner VCR-level AI.

The bait-and-switch

The irony is that tech-bros tell us that this is what they are giving us—or going to give us, soon enough. “Productivity-maximizing AI agents for everyone” is a kind of mantra at the highest echelons of Boosterism. The AI we’ve actually got is something else entirely, though.

And now, dear reader, I will give you my video-heuristic for making sense of where the AI enterprise is taking us. If Knowledge Navigator shows what we want, but don’t got, then this video shows what we definitely don’t want from AI, but are being given.

Witness: Celery Man, Paul Rudd’s classic scene from Tim and Eric: Awesome Show, Great Job!

My fellow knowledge workers: What say you? I can’t be the only one who sees the future—and a little of themselves—in this bizarre scenario.

If you’re confused about what you just saw, let me walk you through it.

Coffee in hand, we start the day by isolating ourselves inside an ethereal VR cubicle. Ready to get some work done, we log in to be greeted by a friendly chatbot interface that knows us by name.

At first we’re all business—Celery Man. Late morning, we take it up a notch with some 4d3d3d3 to hit those performance goals. After lunch, we go hardcore Oyster and really bang it out, celebrating our workplace wins with a photo or two.

The AI assists us at every step, but—trapped alone inside our sensory-deprived workplace—we lose focus and grope for new sources of stimulation. By the end of the day, drained of meaning and purpose, we find ourselves fucking around with hat wobbles and trying to print out nude Tayne.

All of the fluid intelligence of the Knowledge Navigator is on display here—for all the wrong purposes. In this future, we each have our own dancers, generated from the “digital exhaust” we leave behind online. Under constant surveillance by not only our employers, but by every app on our phones, each click we make adds to the matrix of datapoints invisibly following us in the cloud.

The “dancers” are merely the physical manifestation of our browser history—that diary we don’t even realize we’re keeping. The AI merely uses it to show us our (true?) selves, dancing back at us.

I dare say there is a Nude Tayne awaiting each one of us—some secret kink we didn’t even know we had until the computer coaxed into generating it.

Meanwhile, our “wife” is calling: our responsibility to family, society, life, the human world and all its connections on this green Earth. Yet we just go on gooning, seduced by our AI collaborator into exploring the depths of our brainstem.

Whether it’s bespoke AI-generated bespoke pornography in 4K, immersive video gaming, or increasingly personable private chats with an attentive (and persuasive) bot, the dynamics of digital parasitism are already in place. Look no further than the skimpily clad anime girlfriend available to all users of Grok, the non-woke (i.e. “based”) AI run amok on X, the digital fiefdom of Elong Musk.

Your Personal Nude Tayne Awaits

Booster and Doomster narratives both share a fatal flaw: they treat AI as something happening to society rather than something people actively build, choose, and use.

This technological determinism isn’t just intellectually lazy—it’s a deliberate strategy. While we debate whether AI will save or destroy us, the actual AI being deployed looks nothing like either vision. It’s neither superintelligent nor particularly helpful.

What AI is, increasingly, is extractive. As one scholar puts it, “powerful forces are fracking our attention.”

The fact is that most AI isn’t designed to augment human intelligence—it’s designed to capture human attention. Every major AI deployment follows the same pattern: promise productivity, deliver engagement; promise assistance, deliver dependence.

From ChatGPT to Grok’s anime girlfriend, flagship AI systems are only incidentally tools. The incentives of the attention economy instead produce digital parasites, whose primary purpose is to feed on our need for stimulation, validation, and increasingly, connection.

AI development isn’t driven by technical capability or human need, but by engagement metrics and data extraction. Every interaction teaches the machine more about our desires, not to serve them better, but to exploit them more effectively.

One doesn’t have to accept all the tenets of his “techno-feudalism” argument to agree with Yanis Varoufakis that AI-powered devices are novel in the history of humanity. A strange loop has formed, wherein they actively train us, to train them, to better train us, to better train them, and so on.

The consequence is an evolution that is taking AI in an altogether different direction than Knowledge Navigator or anything envisioned by the Intelligence Augmentation or Computer Supported Collaborative Work folks. Instead of coming to meet us where we’re at, we mold ourselves to the odd contours of the machine’s rough edges, modifying our behavior to make the thing work.

This behavior modification is powerful and dangerous because it is incremental. Each time we adapt to what we’re being given, we risk losing sight of what we wanted, just a little at a time. Before long, we may find ourselves in the detestable situation of being stuck with something we definitely don’t want.

The “nude Tayne” each of us will eventually demand isn’t just pornography—though the explosion of AI-generated sexual content is certainly part of it. It’s whatever perfectly crafted stimulus will keep us locked in the loop, generating data, training models, and justifying infrastructure investments that make no economic sense except as mechanisms of capture.

Leave a comment

Breaking the Loop

The choice between Booster and Doomster narratives is a false one. We need to recognize that the AI being built isn’t aimed at either utopia or apocalypse, but at something far more banal: A world where we mistake engagement for enhancement, stimulation for intelligence, and dancing shadows for genuine assistance.

In today’s attention economy, the Knowledge Navigator remains science fiction simply because building it wouldn’t be nearly as profitable as Celery Man: an infinite content generator that learns our weaknesses and exploits them with increasing precision.

We still have a choice, though. We can refuse the strange loop of mutual training, refuse tools that enhance rather than extract, and force politicians to regulate the lawless frontier of AI.

But first, we need to stop asking our computers to show us nude Tayne.

Because unlike Paul Rudd’s character, before long we won’t be able to log off at the end of the day. The walls of the cubicle are becoming the edges of our world, and the dancers are learning our every move.

Though we don’t often think about it, asteroids do sometimes impact Earth with destructive force, and I’m not just talking about the dinosaurs. In 1908, an asteroid approximately 50-60 meters in diameter exploded 5-10 km above the Tunguska River in Siberia. The air burst annihilated over 2000 square km of forest, felling 80 million trees. Luckily, the area was almost completely uninhabited, and eye witness reports from the time suggest perhaps only three people lost their lives.

But asteroids don’t always fall over uninhabited regions of Siberia. In 2013, a much smaller asteroid, only 18 meters in diameter, exploded over Chebarkul, a city of about 40,000 in the southern Russian region of Chelyabink Oblast. Nearly 1500 people were injured, mostly by flying shards of glass from the shock wave, and buildings in six cities across the region were damaged. In this case, we were lucky that, even though the area was populated, the asteroid was small enough that no one was killed.

But in 2024, it looked like that luck might run out. A Chilean telescope observed an asteroid about the same size as the Tunguska impactor that would cross Earth’s orbit. Initially, scientists calculated the probability it would impact the Earth at over 3%. Fortunately, as they continued to track its movement and could gain a more accurate prediction of its trajectory, that likelihood went down to 0% (although it still has a 4% chance to impact the moon, which might look pretty spectacular).

An extinction level impact event is so unlikely that we can take our time protecting ourselves from it. We probably have around 35 million years to get ready, since they happen about once every 100 million years. But a mass casualty impact event is probably worth preparing for. The ARM would have helped build up the technological capability to do just that.

The initial design for the ARM had three main stages:

1. Detecting, characterizing, and selecting a target for the mission from among near Earth asteroids.

2. Redirecting the target asteroid into lunar orbit.

3. Sending crewed spacecraft to the asteroid to explore and return samples.

These three stages reflect four main priorities.

First, redirecting the asteroid is practice for defending against a potential asteroid impact. It would sure be nice to know we’ve successfully redirected an asteroid *before* a city is on the line.

Second, study and sample return reflect the priority of scientific study. Asteroids are often leftover material from the formation of the solar system. They never got incorporated into planets or moons and therefore never subjected to all the active processes of those bodies. In a sense, they remain preserved. So scientists who study them can learn a lot about the formation of the solar system that they cannot learn from elsewhere.

Third, sending astronauts to the asteroid adds to the prestige and publicity of the mission. Strictly speaking, sending a human crew is not necessary to either of the above goals, and only adds to the risk and expense of the mission. But there is a real sense among many space policy makers that human space exploration has suffered since the end of Apollo. Moreover, many also believe that this lull in human exploration provides an opening which other nations can capitalize on to make the U.S. look dull and complacent. Human spaceflight can act as a proxy for which country is the most technologically preeminent.

Fourth, sending astronauts to a pre-prepared destination is a relatively easy step on the way to Mars. NASA’s Moon to Mars architecture takes an incremental approach to get to Mars. Starting with developing a launch vehicle capable of leaving Earth’s orbit, NASA plans to conduct progressively more difficult missions, further from home, until eventually sending astronauts to Mars. The ARM was to be situated within that progression as a test for long distance space travel before landing on the Moon.

But what really makes the ARM unique is two-fold. First, it was the first mission to include a participatory technology assessment. Scholars from Arizona State University designed two public forums, one in Phoenix and one in Boston, where citizens of all backgrounds could discuss the goals of the mission and set priorities. The goal was to incorporate public values into the mission design process. The second thing that makes ARM unique is that Congress took the unusual step of explicitly excluding ARM from the 2018 funding bill. While Congress often increases or decreases NASA funding, and may even work to dictate the general direction and focus of the agency, they do not usually earmark or prohibit funding on a mission by mission basis.

We’ve written before about how important public participation is to successful technology programs. It builds trust and legitimacy which often leads to public support. So what happened? Did Congress undemocratically step in to cancel an otherwise popular mission? Did public participation not actually do any of the things we claim it does? Why was ARM such a dramatic political failure despite NASA’s efforts towards better public engagement?

By all accounts, the design of the ARM’s participatory technology assessment firmly rooted itself in the most recent studies on public engagement and deliberation.

The exercise as a whole pursues public engagement and participation rather than public outreach; a dialogue rather than a lecture. Far too often interactions between experts and members of the public follow what is called the deficit model. This model posits that the public has a deficit of knowledge, and the job of experts is to fill that deficit. In the context of government science programs, such as the ARM, the usual thinking is that the public will support science if only they understand it enough. This is why educational campaigns are so common as a strategy for gaining public support. But the ARM didn’t fall into that trap. This was not a mere educational campaign, instead it was NASA officials that would learn about public preferences. The views expressed by participants would inform NASA’s decision making about ARM.

The forums were two day events that began with presentations by experts to provide participants with a common background about relevant topics. These briefings were designed to be as similar as possible to the sorts of briefings NASA leadership would receive. Participation in the forum was deliberative. Participants engaged in open discussion with their peers with trained facilitators. This oriented the forum toward going beyond merely capturing opinions to capturing the values and priorities that underlie them. It also increased the level of participation such that nearly all participants contributed directly to discussion. The participants were also diverse. The makeup of each forum reflected the demographic makeup of the state where it took place across multiple categories: gender, age, education, ethnicity, income, and employment status. The deliverables of the forum focused on usable outcomes. It was not just data for NASA leaders to use, but actionable items. Forum organizers wrote reports summarizing the conclusions of the participants. They also conducted surveys before, during, and after deliberations about key decisions NASA would have to make about the ARM. By all accounts, the forum organizers were very diligent and thorough in their design to facilitate the best possible integration into decision making.

So what was missing then?

In the next installment, I’ll talk about the many answers to that question. If you’re familiar with our previous writings about intelligent trial and error, you can probably guess some of the things I’m going to touch on.

When people from the west encountered the intensively irrigated rice paddies in Bali, they could only imagine that they were the outcome of raw executive power. Marx called it the “Asiatic mode of production,” what later scholars referred to as “hydraulic societies.” The coordination of something so complicated must have been accomplished by the control of a king, who, in turn, derived his power from successfully managing essential infrastructure. This kind of thinking would prove disastrous for Balinese rice farming, just as its analog in the social realm has decimated the infrastructures that once supported childbearing.

By the mid-20^th century, rulers of a now-independent Indonesia had also come to see rice farming as a straightforward technical system, one that could be optimized with “Green Revolution” technologies. As Stephen Lansing recounts in his book, Priests and Programmers, the ensuring technical transition was disastrous. Bureaucrats thought that hybrid-varieties and new chemicals made traditional irrigation schedules obsolete, that rice could be cropped continuously. The result was water scarcity and a boom in rice pests and diseases.

What the architects of this transition had failed to see was the ingenious way that the traditional Subak system had solved a series of coordination problems in order to ensure that rice paddies were well watered, provided with nutrients, and made free of pests. As Lansing uncovered in his research, irrigation was organized to ensure that large patches of land alternated between being flooded and dried, a pattern that disrupted the lifecycle of rice pests. This was organized in a nested series of organizations, from the local Subak to regional water temples.

Like other commons-tragedy-solving institutions, the Subaks meet periodically to resolve disputes, set cropping schedules, and maintain the canals and weirs. But on top of this democratic infrastructure was also a spiritual one. The travel of water from upstream was understood in sacred terms, as a purifying element but also as itself purified of human impurities as it returned to the sea. Farmers had small shrines where water entered their field, participated in a range of ceremonies and making of offerings, and delivered small amounts of “holy water” alongside the irrigation canals. Lansing sees these as part of the “sociogenesis” of the Subak culture, one that sustained the necessary social relationships to make the society flourish, namely by imbuing productive relationships with symbolic meaning.

Friedrich Nietzsche once wrote, “If we have our own why of life, we shall get along with almost any how.” But not all whys are equally powerful. In his book, The Righteous Mind, Jonathan Haidt argues that religion is uniquely powerful, because it combines belief and practices into strongly binding forms of community. He cites a study of communes that found that the ones rooted in religious beliefs not outlasted the secular versions (six percent to 39 percent). Moreover, they were strengthened rather than weakened by demanding sacrifices, such as abstaining from alcohol.

Even though it is cliché to speak of our own culture as “disenchanted”, we too are driven by a quasi-religious belief system that gives meaning to our daily toils. As individualists we bear the burdens of more frequent bouts of loneliness; we regularly move households and cities in service of something we call a “career”; we repair our homes, raise our children, and suffer illness increasingly without the help of a supportive community; and many of us are seemingly locked in an endless quest to discover an authentic identity for ourselves.

We not just tolerate but often venerate such individual toil. Modern secular peoples often describe their self-actualization through what philosopher Charles Taylor has called “subtraction stories”, wherein they describe the subtraction of previous social forms in religion and community as having been a necessary step in the journey toward a more authentic and meaningful life as an individual.

And there no shortage of ritualistic celebrations that reflect and reinforce these beliefs: the cliché and unbearably self-congratulatory LinkedIn post made when switching jobs (“I’m thrilled to announce…”); the divorcee who writes social media paeans to their new-found freedoms and opportunities for self-exploration; or the online affirmation that the last couple generations of young people (deprived of collectively set rites of passage) have sought in curating their self-identity to match an ever proliferating panoply of cultural and gender subgroups. Every culture has a set of practices for reinforcing a set of dogmas about how to live, which in turn make the burdens those dogmas entail more bearable. Ours are just evermore oriented toward keeping us apart.

In his post a few weeks ago, Michael Bouchey likens the decline of fertility to a leaky pipeline, attributing the source of the decline to the ever proliferating barriers posed by those or that aspect of contemporary societies. Practical challenges around dating, educational requirements, and growing parental expectations lead to potential parents simply running out of time to have as many kids as they might want. But part of me, even though I’ve probably described my own family situation in similar terms, thinks that such a metaphor more describes a series excuses than offers an explanation.

Dating challenges, the de facto requirement to study for a good career, and whatever other explanations that people give today for why they delay having children have only recently become respectable reasons. They pale in the face of the relative impoverishment and survival risks faced by previous generations. My parents didn’t have three kids in their 20s because times were good. They weren’t. The births of me and my brothers bankrupted them. They just didn’t see financial uncertainty as a reason to not have children. In contrast, I long held fast to a litany of excuses, which a Bangladeshi friend of mine in grad school didn’t shy away from pointing out. Every reason I gave about delaying parenthood until after grad school (or tenure, or..) rang hollow. She would point out that the grad school offered parental leave, that a tenure-track job wasn’t guaranteed, etc.

I could only ignore my colleague’s counterarguments, not refute them. And I could only live with the cognitive dissonance, because excuse making was the commonsensical water that I swam in, something legitimated by the epochal changes that have given birth to networked societies.

Demographers call these changes the “Second Demographic Transition.” It is marked by broad shifts in cultural tastes, away from relational commitments like marriage and family and toward individual fulfillment. Sociologists from Robert Putnam to Robert Bellah (along with countless other researchers not named Robert) characterized it through their observations of the decay of associational life, the disappearance of bowling leagues, sewing clubs, parent-teacher associations, and fraternal organizations and their imperfect replacement by the pseudo-relationships television and media fandom and social networks—and today by AI romantic partners.

The old social organizations fell, in part, because of how alluring the new options were. Some semblance of connection could be wrung out of new technological mediums with comparably less risk and fewer relational demands. But another necessary component was a shift in understanding. People needed to see themselves as the only important social unit being served social connection. And they needed to come to see relationships as a kind of portfolio, one expected to provide a suitable return on investment.

One clinical psychologist voiced the invariable philosophy of connection in a world premised on relational optimization: “What are relationships for all of us? They’re just neurotransmitters being released in our brain. I have those neurotransmitters with my cat. Some people have them with God. It’s going to be happening with a chatbot. We can say it’s not a real human relationship. It’s not reciprocal. But those neurotransmitters are really the only thing that matters, in my mind.”

The transformation into networked societies involved a literal restructuring of the geometry of togetherness. It has been an ongoing unweaving of webs of connection into something that literally resembles the sparse, peer-to-peer networks that constitute the Internet. Sociologists, namely Barry Wellman and Keith Hampton, have viewed this transformation in more triumphant terms, as a new “social operating system” that has provided people with “more freedom to tailor their interactions. They have increased opportunities about where—and with whom—to connect.” And, tellingly, they portray togetherness within this new operating system as an achievement, not a public good, one accomplished by social entrepreneurs with the right “combination of talent, energy, altruism, social acuity, and tech-savviness.”

In this view, however, we should have expected fertility rates to have cratered in developing nations decades from now, after their traditional social institutions were unwoven by the demands paralleling the establishment of modern market-driven networked societies. But as demographer Alice Evans has noted, they have “culturally leapfrogged.” Consuming the cultural content from the West, women in these countries not only come to feel imprisoned by local social norms but also embrace the dating expectations of the women they watch on screens. Men and women come to live in different realities, with men still clinging to tradition but women striving for autonomy and equality. The result is that relationship formation as a whole is in steep decline, as more and more women in these countries have come to recognize it as a raw deal. One woman told Evans, “Sometimes I wonder whether I should reduce myself…. but I don’t want to settle for less”.

In the West, the causes are somewhat different but the outcome is the same. Under the allure of an apparent glut of choices, there are the same expectations to optimize partner selection (and not to sacrifice career or fun in order to partner up with a potentially suboptimal relational investment). Yet, young people also spend increasingly more time alone engaging with online content, becoming more socially anxious the less they interact meaningfully with other people.

And for many of us, it’s because we’ve simply and inexplicably gone “all in” on a particular vision of the good life, one that we desperately want to believe that we came up with ourselves. In debating the prospect of parenthood with my grad school colleague, I had no image of success to guide me but the one that seemed to have just fallen into my lap during my early 20s. To fail to prioritize that journey in order to selflessly care for someone else filled me with dread, the fear that in doing so that I would be setting myself adrift and that I would miss out on a life that was supposed to be authentically mine.

All this isn’t meant to put the blame on careerism and smartphones. Nor is it to suggest that women in developing countries shouldn’t be balking against the unfairness of their country’s social norms about marriage and family. Rather, my point is that there has been a collective failure to adjust traditional social institutions to help people weather new pressures and realities. The trouble has been that the practices that sustained relationships at scales bigger than pairs of individuals and far thicker and meaningful than social networks have been increasingly abandoned rather than reformed.

We are now like the Bali rice farmers after interloping Westerners tried to “rationalize” their paddies and irrigation systems. An enchantment with rationalism rendered them obtuse to the reality that an array of rituals, myths, and relationships founded a complex system of coordination that analytical planning could only replace or advance with great difficulty—if at all.

In our case, what has been rationalized is the social fabric itself. Being unburdened by community, extended family, nuclear family, children, and increasingly even partnership itself, the last remaining practices of sociogenesis primarily serve individualism. Like the waters of a rationalized irrigation system, the atoms in a rationalized social universe are ideally freely flowing. But achieving that absence of friction required the removal of guideposts, and of landmarks on the horizon of significance that would help us understand our social toils as meaningful. What’s left is the enduring effort to overlay the tale of a heroic journey on top of a lifestory shaped largely by the anonymized forces of a globalized network society. We’re each in our own lifeboats, rowing with only minimal success against the currents while being buffeted by the waves, but are still nonetheless convinced that we’re the main character.

Although this system of meaning making is often good enough to get many of us to get out of bed in the morning and head into work—so long as we don’t think too hard about what happens to us when we grow old and begin to lack the of talent, energy, altruism, social acuity, and tech-savviness to operate social networks like we once did. We have stories that can work to maintain modern network societies. But they don’t seem to be up to the task of sustaining human civilization, much less families.

As Joseph Tainter pointed out, civilization is a sort of deal. It’s legitimacy rests on societies continuing to offer benefits in excess of the burdens inherent in sustaining them, at least on average across the population. That legitimacy ultimately relies on collectively held vision of a common world worth sustaining, at least in outline form. What Tainter never anticipated is that a society’s members could grow disenchanted with the deal even as it continued to generate massive surpluses, that societies could decline and shrivel because they no longer produced many people interested in the social commitments on offer, willing to make the decision to pair up, have children, and dedicate themselves to community institutions without being agonized by the risks, by the fear of screwing up or of committing at the wrong time, in the wrong place, and to the wrong people.

Our growing inability to people the planet is just yet another stage in the gradual unweaving of communal relationships and their increasing replacement with weaker and more transactional ties. The process is by no means inevitable, and my point is not to wax prophetically about demographic doomsday scenarios. That so many people describe their relationship with their smartphone and social networks mainly through lament signals the latent desire to lash our boats back together, to begin to row together. The question is what it would take to relearn how to do so, and to embrace the seeming irrational rationality that belonging entails, without romanticizing a reactionary return to the past.

Share

Subscribe now

Leave a comment

The problem with population decline is that it isn’t just one problem. In reality, it is a series of complex interrelated problems. This makes it difficult to address head on. It would be nice if there was one policy or type of policy that could be a panacea for population decline, but it is becoming increasingly clear that there is not.

Consider Garvey’s analysis of Japan’s demographic crisis. The declining birth rate sits at the confluence of cultural upheaval, collectivist social expectations, and economic collapse. If Japanese lawmakers could wave a magic wand and fix their economy tomorrow, it still would not alleviate the crushing social pressure supporting one’s family. Likewise if all men there suddenly became ideal citizens, it wouldn’t change the economic ability to live up to those intentions.

Or compare the US to many European nations. Europe boasts fewer working hours, more holidays and paid vacation, and generous parental (thats right, for the father too) leave policies. Since many child-free Americans cite having to choose between having a family and career success, surely these countries that provide ample room for family rearing must be staving off declining birth rates? But no, fertility rates throughout Europe are lower than in the U.S. Another oft cited reason for avoiding parenthood is economic anxiety. But the relatively strong American economy hasn’t protected us from declining birth rates either. So, then, what is to be done?

We might view raising a family as a pipeline. At various points along the way, people who might otherwise have kids or have more kids “leak” out of the pipeline. A prospective parent may not be able to find a partner to raise kids with. If they do, they may not see raising children as feasible or desirable. Even if they did, they may choose to have only one kid so they can focus their energies on providing the best possible life for them, rather than take on the seemingly unmanageable task of meeting escalating parental expectations for a second child. Even if they want to have more than one kid, they may have started their family later in life and simply run out of time before having as many children as they wanted. Potential parents leak out of the child rearing pipeline at all sorts of points for a variety of reasons, and no one policy is going to patch every hole.

To have kids, first you need a romantic relationship. It takes two to tango. But romantic relationships are more difficult now than it was when my partner and I started dating in 2009. There could be numerous causes, and it is very difficult to distinguish which ones, if any, are the most impactful. In most developed countries, dating apps have become the primary way that people find partners. But these apps operate on a somewhat transactional basis, meaning that people finding partners on these apps must act more transactional as well, even while lamenting that very same change in relationships! Starting in the 80s and 90s, divorce rates in most developed nations increased substantially. Although this rate is now decreasing, it's possible that widespread divorce among the parents of those currently in the dating pool led to more risk averse dating styles. Whether it's frustration over transactional relationships or not wanting to repeat the mistakes of their parents, increasing numbers of young people de-prioritize romantic relationships rather than continue to slog through bad first date after bad first date on the off chance you don’t get ghosted. There are other sources of fulfillment to be had.

Even if you do meet someone, there are many reasons why one might not want to have children. Garvey and Dotson both discuss how important it is to live in a society that people want to bring children into. Child free couples around the world have cited issues like climate change, economic anxiety, and career advancement. But as both my colleagues have already so eloquently discussed, these issues are less the root cause, and more the outcomes of social features that can be anti-parent. Whether it's because they want to focus on careers, or because they don’t think they can afford it, successfully coupling up doesn’t guarantee a desire for children.

Even if a couple wants children, they might have fewer children later in life than they would prefer. I didn’t finish grad school until I was 32. Even if I had had children right away, I would have still been nearly five years later than the average age of a first child in the U.S. And my situation is becoming less and less divergent. In the 1970s, the average age of a first child was 20, almost unimaginable today when most jobs require degrees that keep people in school until at least 22. Advanced degrees, both masters and PhDs, are far more common as well. And as educational attainment in the workforce gets more and more common, job seekers have to spend more time trying to make themselves stand out. All of this leads to people starting their careers later in their lives. In turn, they start families later in their lives as well, if at all.

Without a doubt, there are numerous other reasons why people have fewer or no children. It isn’t feminism, or birth control, or the economy. None of these are sufficient for explaining all of the reasons people aren’t having kids. Anyone telling you what *the* cause is, is trying to sell you something (often herbal supplements). But at least we can hold on to hope if we believe the demographic crisis has a singular cause and, therefore, a singular solution. Otherwise how does one deal with such a thorny, interwoven mess?

The problem is that population decline is so complex that it is nearly impossible to implement a single, comprehensive reform. We simply can’t understand the problem well enough to get it right! When analysis fails you, the only way forward is to muddle your way through. At least, that's what political scientist Charles Lindblom argues.

To muddle through something means to make manageable, bite-sized changes and then reflect and see how those changes worked. This kind of intelligent trial and error is a tailor made tool for attacking problems with too many unknowns to make comprehensive reform.

We’ve discussed this concept before when examining Danish wind turbines. Part of the reason Danish farm-tool manufacturers succeeded where NASA engineers did not is because they made small changes to well known designs, learning what worked and what didn’t, while slowly scaling up the production of their turbines. NASA, on the other hand, quickly made massive wind turbines based on mathematical theory that proved unreliable in real world conditions.

Trial and error learning is, inevitably, going to be riddled with failures. South Korea spent $200M over 16 years offering financial incentives to have babies and have one of the world’s lowest fertility rates. China has rolled back abortion access in the country, also to no avail. And liberal parental leave policies throughout the EU have not spared the continent from a precipitous drop in birth rates. Nobody knows what kinds of policies are going to work. The key is being both willing and prepared to learn from mistakes.

That's the real problem with apocalyptic thinking: it leaves no room for learning from errors. As Dotson has discussed in multiple articles, apocalyptic thinking stifles disagreement and undermines incremental improvement.

Apocalypses are tough to be in. They don’t leave a lot of time to act and the repercussions of doing the wrong thing are literally existential. These two factors make for bad decision-making. Extreme consequences can justify excluding people you think are wrong (after all, if they get their way we’re all going to die!) and the immediacy of the situation doesn’t leave time to experiment with different strategies. And when both factors combine, the proposed solutions are, inevitably, massive in scale.

Take, for example, pro-nuclear environmentalists who oppose innovative reactor designs because the impending doom of climate change leaves no time to explore alternatives to light water reactors. Under normal circumstances, it seems plainly foolish to plunge blindly forward into a single technology for energy production for the foreseeable future. But when framed against the existential risk of climate change, suddenly otherwise reasonable people believe there is no other way.

Prophecies of apocalypse reduce complex politics into stories of moral certainty. When the Meadows’ describe degrowth as the solution to the environmental harms of over consumption, they are responding to the facts no more than they are their experience along the hippie trail. Yet in doing so they portray their opponents as villains who will destroy us all along with the planet. It's their way or the highway. To disagree is to invite the predicted apocalypse, and who but a villain or a fool would do that?

Pronatalists do much the same. It's not enough that single cat ladies merely have different values, they must be part of some plot to undermine western cultures (whatever that is). Why else would someone knowingly walk us closer to the edge of population collapse?!

But what is their alternative? Jordan Peterson has suggested that the problem is women’s access to education. Although he pragmatically suggests that it probably isn’t possible to deny women educational opportunities, the subtext makes his preferred solution clear. Idaho Republican Senator Chuck Winder has suggested that the problem is abortion access. At Natal Con in 2023 (yes, it is a thing), Charles Haywood argued that women should not be in the workforce and that the civil rights act, by protecting against sex discrimination, has been one of the most destructive American laws ever.

It’s pretty clear where the movement is going. And most reasonable people, if asked, would probably disagree that rolling back women’s rights by a hundred years is the correct response. But the fear of apocalypse can make people support positions they otherwise might not. Malcolm and Simone Collins, two of the most outspoken pronatalists, claim to think Haywood’s position is dumb. But they have no problem supporting the Natal Con. And their own positions seem merely to be more vague than Haywood’s, blaming “wokeness” on declining birth rates rather than directly calling out women’s rights.

When we start to see population decline and demographic collapse as an impending apocalypse, we will likely find ourselves in one of three scenarios. We could encounter a vicious cycle of doomerism where it all feels too futile to do anything about. Disagreement could turn into a political morass (like bears in Idaho). Or the worst possible option, the pronatalists get their way, lead sweeping changes to the very concept of equality under the law, make women’s lives worse for at least a generation (among who knows what other unanticipated consequences), all while still not stopping demographic collapse.

Despite having no children myself, I share Garvey’s concern about how we will take care of our aging population. But when I think about why I or any of my childless friends haven’t done our share of increasing the population, the reasons don’t seem nearly as one dimensional as any discussion I’ve yet read would make it seem. I know people who forgo having a child for numerous reasons: child care, career aspirations, financial stress, lifestyle preferences, not feeling or being prepared for parenthood, and the list goes on and on. There is no single policy that can address all of the causes at once.

The issue of population/demographic collapse requires policy experimentation and learning to solve. Apocalyptic rhetoric and thinking, on the other hand, will only make those solutions harder to achieve.