This post is coming out in tandem with Soham’s interactive economic model of wave-powered data centers. This post explains why we think ocean compute matters, assuming that it is feasible. If you are curious about how ocean compute works and whether it makes economic sense, we highly encourage you to check out the model!

The next chapter of the ocean economy may grow out of the bones of America’s industrial past. Eighty years ago, the shipyards of Vancouver, Washington helped feed the country’s enormous wartime hunger for new vessels, churning out 141 ships over the span of World War 2, including Liberty-class cargo ships. The Liberty ship is often cited as a feat of mass production almost without precedent, with American shipyards collectively constructing more than 2,700 during the war. It is hard to look back at the extraordinary scale, pace, and ingenuity of U.S. wartime output, then at the diminished state of American ship production today, without feeling a sense of lost industrial glory.

But today, these same shipyards offer a glimmer of hope. Inside this storied industrial complex, the welders at Thompson Metal Fab are at work on machines that could unlock the ocean as a vast new physical substrate for American computation: Panthalassa’s wave-powered computing nodes. As laid out in a previous Ansible article, Panthalassa is a startup building floating data centers. They produce wave energy converters that power AI workloads onboard using energy from the tempestuous open ocean.

Subscribe now

A few months ago, I thought of Panthalassa as a clever energy moonshot that could pose questions about governing superintelligence incubated in international waters. Today, as I release a techno-economic model that assesses whether wave energy can cheaply and effectively meet rising AI-driven energy demands, I have come to see Panthalassa as a foundational building block for a much larger ocean economy.

We need to widen our imaginative aperture to consider the role of the ocean in AI. Beyond a brief mention of offshore data centers in the viral scenario AI 2040: Plan A—where data centers move to international waters out of geopolitical necessity rather than on the economic merits—we mostly conceptualize AI powered by enormous suburban data centers that perhaps eventually launch into orbit. Silicon Valley’s visionaries tend to look upward. We have become increasingly comfortable imagining space as a new geography for industry. But there is comparatively little curiosity about what expanding across, and beneath, the last frontier that still exists on Earth would mean. If my model is right, the ocean is poised to become a major center of American computation. More importantly, this same cheap, renewable energy could make the ocean a crucible for cutting-edge science and autonomous industry.

Why the ocean will be critical for AI

As the model shows, waves contain enormous amounts of usable, consistent energy which Panthalassa can harvest with cheap and simple machinery that incurs surprisingly low maintenance costs. As a result, their nodes enjoy a very low levelized cost of energy (LCOE) and a high capacity factor—a remarkable combination for a source of renewable energy. The low physical capital costs and parsimonious design mean that it is possible to predictably and quickly scale up capacity.

The problem is that the people and industries that use energy are located thousands of miles away on land. Carrying the power thousands of miles would be a logistical nightmare. So, Panthalassa found the most lucrative, geographically-adaptable industry hungry for cheap energy sources with low time-to-power: AI compute. AI inference in particular is an unusually good fit. Unlike frontier training runs, inference can be distributed across a geographically dispersed network (Panthalassa’s nodes are not currently physically connected to one another, and use Starlink to send their workloads to shore). Test-time scaling is core to improving reasoning by letting models search and deliberate longer, while agents compound this computational demand further by stringing together many model requests and tool calls to complete a single task. Luckily for Panthalassa, these are location-flexible workloads because they occur in contexts where throughput and cost matter more than millisecond-level latency—this is especially true for long-running agents, coding, research, and other asynchronous tasks.

Wave-powered inference could also support ambitious AI policy proposals. Evaluating models and continuously monitoring agents will not be cheap. Frontier capability evaluations can be enormous on an individual task basis, consuming billions of tokens and requiring an inference budget in the tens of thousands of dollars. Building privacy-preserving agentic infrastructure, from agent IDs to proof of humanity systems, will likely require computationally expensive attestation systems and cryptographic proofs like zk-SNARKs. Governing a world with transformative AI is going to require a lot of energy. Serendipitously, Panthalassa spent years developing a strange form of energy just as an equally strange set of challenges and use cases have emerged.

Why the implications are much larger

While the AI future offers some of the best use cases, LCOE is use-agnostic. This means Panthalassa has not just found a cheap way to power data centers, but a generic way to provide cheap, reliable power in the open ocean. A core reason why offshore industry is massively undeveloped is that it pays what might be called an ocean distance tax. Sustaining a human worker at sea is costly: you have to supply food, shelter, safety systems, and emergency support from scratch in an extremely remote location. It’s also psychologically grueling enough that you’d need to rotate workers out regularly, adding even more costs. Robots come with their own problems, energy and maintenance being first and foremost, but also bandwidth and latency: the communications infrastructure is sparse enough that you can’t just remotely control them. This creates a vicious cycle. Because operating far offshore is expensive, there is no reason to build persistent infrastructure in the ocean; because little infrastructure exists there currently, nearly every new activity is a bespoke, complex mission, keeping costs high.

Essentially, to break this cycle, ocean industry requires autonomous systems that run locally or use edge compute running on cheap power. To be fair, there are many other reasons the ocean is inhospitable to industry today—corrosion, biofouling, storms, and maintenance, to name a few—but cheap energy is key to surmounting many of these challenges. More abundant power means more capable onboard compute, which means better sensing and monitoring that can reactively alter a vessel’s course or catch damage early, which could mean more resilient infrastructure. In other words, cheap energy is certainly not a sufficient condition for a large offshore economy, but it does make many of the other constraints cheaper to manage.

I am speaking at a deliberately high level here, and I am hardly the first person to make this argument. Will O’Brien and Packy McCormick’s excellent essay The Great Blue Frontier takes on the question of ocean industrialization more robustly, describing an “ocean stack” that includes layers of interlocking technologies that make each of the other layers more useful and cheaper. The stack spans from autonomous vehicles to energy installations like Panthalassa. Rather than describe this same ecosystem, I want to end with a taste of the interesting applications that cheap, abundant wave energy would allow.

Ocean Science

Studying the ocean is difficult. It involves sending an expensive instrument somewhere very harsh and remote, desperately collecting whatever data you can before its energy runs out or its lifespan expires, then transmitting the data to shore for analysis. Persistent offshore power would make something resembling a network of ocean probes possible. Autonomous vehicles could recharge at sea using floating wave energy converters, run powerful sensors for much longer periods of time, and process observations locally rather than trying to send large amounts of raw imagery, sonar, or biological data through constrained communications links. Eventually, autonomous systems may be capable of noticing something scientifically interesting—a strange organism, chemical signal, or geological formation—and deciding on their own to investigate further. Versions of this are being attempted, and could benefit from Panthalassa’s cheap power: DOE-backed researchers are developing wave-powered docks that recharge and communicate with autonomous underwater vehicles (AUVs), while NOAA is field-testing AI aboard autonomous vehicles so they can identify and classify deep-sea animals in real time.

Offshore industry

When it comes to offshore production, oil and gas have already demonstrated that, if the extracted resource is valuable enough, we will take on the risk and cost of moving enormous amounts of industrial machinery into the middle of the ocean. Cheap power and autonomy lower the value threshold that an activity has to clear before it makes sense to bring it out to sea. Consider seabed minerals. Currently, deep-sea mining proposals already plan to separate sediment and dewater nodules aboard the production vessel before shipping them to shore because it doesn’t make sense to pay for transporting water and waste material across the ocean. However, we can lower transport costs by shedding dead weight and make deep-sea mining more attractive by pushing much of the value-adding industrial chain offshore—such as sorting, crushing, concentrating, or perhaps eventually doing parts of metallurgical processing.

Aquaculture

Offshore power may even secure and strengthen our food systems by supercharging aquaculture. Seth Dowell’s recent piece in Palladium on China’s “blue granary” describes how China is pioneering mobile fish-farming ships, enormous offshore cages, marine ranches, underwater inspection drones, and automated feeding and monitoring systems. Of course, China is not building these far into the open ocean yet, where marine life is generally much less dense than in coastal waters or lakes. But advanced fish farming could effectively create productive ecosystems where few existed before, seeding managed populations and sustaining them with automated feeding and oxygenation, monitoring, and disease control. This last point is crucial because many aquaculture ventures currently struggle with fish welfare and disease. The preventative measures like automated delousing and better water circulation are currently very energy intensive. Pumping alone currently consumes the plurality of site energy in aquaculture farms. Thus, cheap ocean energy would not make aquaculture merely larger, but more instrumented, automated, controlled, and humane.

Conclusion

The future is difficult to predict, and I would suspect that the most consequential uses of abundant offshore power will be ones I have not named and cannot imagine. However, the possibility that this technology could reshape computation, critical industries, food systems, and scientific activity is plausible enough to deserve serious attention. At FAI, our work begins from the premise that technological change and statecraft cannot be treated as separate endeavors: we must recognize emerging technological frontiers, understand what they could mean for national power and prosperity, and shape the conditions under which they develop. We think the ocean may be one of those frontiers. This model is just our first attempt to understand its economics, and we plan to keep studying what an increasingly industrialized ocean could mean for America. Stay tuned.

Today, Apple is launching updated terms for alternative app marketplaces in Europe and the developers who distribute through them. Breaking open Apple’s infamous “walled garden” has been one of Europe’s more successful digital regulation pushes. Apple is territorial over its App Store; commissions from the trillions of dollars that pass through it each year are the company’s second largest revenue source after iPhone sales. Historically, Apple has taken a 30 percent cut of every download and in-app purchase, and a 15 percent cut of every recurring subscription. Developers who wanted to avoid the fees, or the arduous process by which Apple approves every app launch and update, had nowhere to turn, since the App Store is the only shop in town on iOS. The EU’s Digital Markets Act (DMA) aimed to change this. After several years of feet-dragging, today’s updated terms are Apple’s latest attempt to satisfy the Commission, even as Apple retains some control over and profit from app distribution.

The changes are twofold. First, Apple has significantly lowered the bar to run an alternative app marketplace, though it still requires operators to be an established entity in good financial standing. Second, Apple simplified the fees it charges developers for digital purchases made through alternative app marketplaces. Where it previously imposed an arcane tiered system, it now charges a flat 5 percent commission. And yet, Apple still remains a gatekeeper of app distribution: it notarizes all apps distributed on iOS, regardless of the distribution method, and developers must pay for an Apple Developer Membership. In other words, you still cannot distribute an app on iPhone without paying Apple, even if the app is free. This has never been true for Android—until now.

Subscribe now

Apple’s launch coincides with Android’s new developer verification system, which went into effect in a handful of countries yesterday and will eventually apply globally. The program allows only verified developers (who pay a fee) to distribute apps by any means on Android devices. Google, which runs the Android operating system that powers almost all non-iPhones, has long championed independent app marketplaces, direct downloads, and alternative operating systems. While Apple promised to protect users from malware by seriously vetting apps on offer, Android let users decide whom to trust. If the possibility of fraud or malware spooked users, they could find an app marketplace whose vetting and curation they trusted. But it was their choice; Google minded its own business. This week marks the end of that hands-off, freedom-first approach.

Google’s announcement of the program last year came within a month of a European Commission (EC) investigation into how Google protects consumers from the scams of suspicious developers and money-swindling apps posing as banks or trading platforms. It may not have been a coincidence. Google cites financial scams as the reason for adopting the verification process, and its announcement quotes a banking federation that touts the boon to security it offers. The choice not to pilot the verification program in Europe is a calculated one. If the Commission finds Google Play’s current approach to fraud apps overly lax, Google can offer the program’s expansion to Europe as a commitment to compliance. Even if the program’s European rollout is insufficient, Google has at least bought itself several years and demonstrated a good-faith effort.

The timing of these changes by Apple and Google highlights an uncomfortable truth: Europe has a love-hate relationship with gatekeepers. The European Commission views the open internet as a threat, and relies on its thoroughfares to police the indecent, uncivil, scam-ridden web. This is why even as the EC nudges Apple toward openness, it pushes Google to become more closed.

In the context of European digital regulations, the term “gatekeeper” is derogatory, lobbed at Big Tech incumbents that have inhibited thriving, competitive digital markets in Europe. But the word is technically a legal distinction. The DMA designates seven “gatekeeper” companies that together control 23 core online services. The DMA’s primary goal is to knock gatekeepers down a peg to make room for smaller market disruptors and European rivals.

Although the DMA’s sister law, the Digital Services Act (DSA), uses a different term to describe gatekeepers, it relies on them to build a friendlier internet. Many of the DSA’s goals, from preventing election interference to stemming the tide of illegal sales to stopping online harassment and sexual abuse, would be harder for the EC to enforce if very large online platforms gave way to decentralized chat rooms and marketplaces. It is easier to moderate content on a handful of platforms where most people conduct their online life (download apps, talk politics, share photos, shop) than it is to play whack-a-mole with a decentralized web where indecent users scuttle from one small internet hole to another.

Europe is not the only place where the ideals of internet openness and platform accountability compete. As I’ve pointed out elsewhere, America has always faced the same tension between allowing total anarchic freedom online and promoting a more palatable internet. The “twenty-six words that created the internet” (a provision in a 1996 law that absolved web platforms of the traditional duty to patrol what they publish) were originally couched in 4,500 words circumscribing the internet’s limitlessness for the sake of decency. Those other words were struck down by courts, but Meta’s recent settlement with state attorneys general on behalf of parents demonstrates that the pendulum has swung back toward holding platforms accountable for how consumers use them.

Every society needs to decide for itself how to balance openness and accountability. The European difference is that the EC uses the blunt tool of massive fines to demand both simultaneously. Meta must filter out disinformation while also promoting dissenting voices and smaller, less established media outlets. TikTok must protect youth without being ageist. Apple must allow more optionality in downloading apps, while Android must crack down on it. In the U.S., you may get competing interpretations from district courts, changing standards under different executives, and legislative bodies approaching an issue differently as it develops and political winds change. But in Europe, you get the same body demanding opposite things at the same time. The U.S. process is far from perfect, but compared to Europe, it is the difference between finding the equilibrium of two competing ideals and talking out of both sides of your mouth.

A narrow view of these European regulations is that Europe is lobbing digital tariffs on American companies in the form of massive fines, even while criticizing the American recoil from free-market globalism. There is truth to this, and indeed many Europeans have admitted the regulations’ unstated aim is European digital sovereignty. However, the view prompts an unproductive American response, one that lends itself to escalatory retaliation and threats to take our ball home if Europe is so tired of us winning. Heightening the European political temptation to double down on tech sovereignty is a suboptimal strategy. In the bigger competition between the U.S. and China, the U.S. would clearly prefer Europe to be on our side.

Perhaps some of the U.S. fist-shaking is necessary, but pity towards the EU is probably more appropriate. Better still, the U.S. could show the EU an alternative future, where we capture the best of the interoperability and openness the DMA has prompted without getting mired in its tangle of contradictory obligations. During the Cold War, the CIA covertly funded many moderately liberal European organizations, operating under the belief that the best answer to communism was a democracy more responsive to people’s needs. The equivalent in the new age of great power competition is to show Europe a better path forward on tech regulation as a way of encouraging it to focus less on unachievable tech sovereignty and more on allied tech partnership.

The Android verification system exemplifies the sad irony of European tech regulation. If the developer verification is successful, Google will be pilloried for being exclusive. If it is lax, Google will be negligently abetting fraud. It is derided by the same people who advocate for breaking Apple open. The EC has backed itself into a corner by necessitating and despising gatekeepers at the same time. Europe can say it hates walled gardens all it wants—but its countrysides are still dotted with castles.

Leave a comment

A Coalition Needs Criteria

Launched in December 2025, Pax Silica is one of the second Trump administration’s signature international initiatives. The State Department describes it as its “flagship effort on AI and supply chain security, advancing new economic security consensus among allies and trusted partners.” The Pax Silica Declaration, effectively a statement of principles with no concrete commitments or undertakings, is light (perhaps purposefully so) on details. But reading between the lines, such as its references to “mobilizing ... private industry,” “reducing excessive dependencies,” and “addressing non-market practices that undermine innovation and competition” such as “overcapacity and unfair dumping practices,” one can divine two central organizing purposes for Pax Silica. The framework is intended to 1) accelerate the development of artificial intelligence technology through international cooperation and investment, and 2) create a coalition to counterbalance China’s dominance of certain layers of the AI stack, especially in the areas of critical minerals and manufacturing. In public addresses and interviews Pax Silica’s chief architect, Under Secretary Jacob Helberg, has been far less coy about identifying competition with China as one of Pax Silica’s overarching goals.

The State Department is just beginning to announce and operationalize certain aspects of the Pax Silica framework. In March, the Department announced that it was seeking to deploy a $250 million fund to invest in critical mineral processing and advanced manufacturing infrastructure in member countries. In April, it was announced that the Department was in negotiations to install a “Forward Deployed Industrial Base” on a 4,000 acre rent-free lease in the Philippines. In June, the Department announced that it would build “Pax Pass,” an AI-powered supply chain platform intended to facilitate the shipping of goods essential to the AI economy; a Notice of Funding Opportunity was published in August, backed by $50 million in funds. The Department also announced the Foundry School, an educational partnership led by Stanford and a coalition of major research universities, intended to train budding technicians and entrepreneurs throughout the Pax Silica universe with a common curriculum.

Subscribe now

As these and other benefits of membership begin to take shape, we should be asking: who is eligible to participate in Pax Silica, and what are their obligations as members? Returning to Pax Silica’s dual goals of accelerating AI development and countering China’s stranglehold on critical technologies and their inputs, it stands to reason that Pax Silica membership, with all its attendant benefits, should be open to countries that can meaningfully participate in the global AI economy, whether by hosting computing infrastructure, supplying raw or intermediate goods, facilitating logistics, participating in R&D, or providing capital. Equally, membership should be conditioned on certain obligations vis-à-vis China.

By these measures, some of Pax Silica’s current membership roster raises eyebrows. The European Union, a Pax Silica signatory and long the world’s leading exporter of cumbersome technology regulation, is unsurprisingly poised to become the greatest AI decelerationist force on the planet. The EU AI Act will impose extensive and costly risk management, data governance, and human oversight requirements for so-called “high-risk” AI uses, which include such banal applications as workforce management and insurance underwriting. And it outright bans AI platforms that “classify[] individuals or groups based on social behaviour or personal traits, causing detrimental or unfavourable treatment of those people,” a proscription which, whatever you may think of its underlying objective, is so malleable as to defy pre-enforcement interpretation by those whom it regulates. With its hefty burdens and vague requirements, the Act will dissuade all but the largest AI companies from participating in the European market, self-destructively hamstringing Europe’s relatively small AI scene and stifling innovation in promising but “sensitive” sectors such as critical infrastructure, healthcare, and law enforcement. And the Act authorizes noncompliance penalties up to 7% of a company’s prior-year revenue. Even for a pure-play AI company like OpenAI this would reach nearly $1 billion; when you consider the larger diversified tech companies that also compete in the AI space, such as Microsoft and Alphabet, the potential scale of the fines becomes truly staggering. Should the EU reap the benefits of an American-led AI coalition while strangling AI companies’ ability to avail themselves of the European market via rent-seeking regulation?

Kazakhstan is another oddity. To be sure, Kazakhstan has much to offer Pax Silica, with enormous untapped deposits of minerals like tungsten, molybdenum, chromium, copper, and manganese. But just a few weeks after joining Pax Silica, Kazakhstan became one of the founding signatories of China’s rival AI coalition, the World Artificial Intelligence Cooperation Organization (WAICO). Kazakhstan is also where Xi Jinping unveiled the Belt and Road Initiative in 2013, and in 2025 it was the single largest beneficiary of BRI investment, taking in nearly $26 billion from China. As of late 2025, Kazakhstan was China’s fifth-largest debtor, with a balance of more than $60 billion, giving China tremendous economic leverage. Can a country that is both economically enthralled and diplomatically committed to cooperating with China be trusted as a Pax Silica partner?

The parameters of Pax Silica membership should be tailored to the scope and functions of the organization, which are still being worked out and will likely evolve over time. This post proposes a handful of membership criteria and ongoing obligations for Pax Silica membership, with the hefty caveat that some of these proposals may reveal themselves as going too far—or, indeed, not far enough—as additional details about Pax Silica’s operations emerge.

Proposed Membership Requirements

Let’s start with the basics. Pax Silica exists to create a U.S.-led AI trade and investment order predicated on free enterprise and free trade. It stands to reason that participation in WAICO should be automatically disqualifying for Pax Silica membership. Maybe there’s a credible argument for grandfathering Kazakhstan in, but going forward this membership criterion should be enforced without exception.

A thornier issue is the matter of economic alignment (or dependency, as the case may be) with China. Ideally, a Pax Silica member should not be substantially indebted to China, as debt gives China leverage it can deploy against its debtors to siphon off the benefits of Pax Silica membership for itself. However, imposing economic criteria that are too strict would exclude much, perhaps most, of the developing world from membership, forcing them into WAICO’s embrace instead. This presents a line-drawing problem that I do not purport to solve here. Argentina (a Pax Silica member) and Angola are both among China’s ten largest debtors, owing about $38 billion and $65 billion, respectively. But Argentina’s debt to China represents only about 5% of its GDP; Angola’s is more than 40%. Similarly, China purchases over 41% of Angola’s exports, compared to 11% for Argentina. It would be reasonable to think that the economic dependency threshold for Pax Silica eligibility lies somewhere between these two countries.

Pax Silica membership should also be off-limits for countries where China is heavily embedded in or outright owns the basic technological infrastructure. In countries such as Mexico, Saudi Arabia, Turkey, and South Africa, virtually all of the data center infrastructure is of Chinese origin. Throughout much of Southeast Asia, data center projects fueled by Chinese investment and intended primarily to serve Chinese customers are breaking ground. Access to compute is perhaps the strongest reason for the United States’ continued leadership in AI development; Pax Silica should not include countries that undermine this advantage by preferentially hosting Chinese compute. And that is to say nothing of China’s penetration of global telecom infrastructure via Huawei, ZTE, and other providers.

It may be technologically and financially implausible to demand that Pax Silica members or aspirants eschew Chinese tech entirely or conduct a wholesale rip-and-replace before joining. But consider the examples of France, which in 2020 announced that telecom operators would not be allowed to renew licenses for Huawei 5G gear once they expire, or India, which excluded Huawei from 5G trials and informally signaled to domestic telecoms that Chinese equipment was to be avoided. Pax Silica candidates with Chinese telecom infrastructure should be required to adopt similar strategies to gradually phase out Chinese equipment. An analogous approach could work in the data center context, where optical transceivers, network switches, and other equipment are turned over every few years; Pax Silica countries with data centers built on Chinese IT should commit to cycling out Chinese tech in favor of American equipment. Even more importantly, they should not allow Chinese companies to construct more data centers on their soil.

Proposed Obligations

Pax Silica should also demand certain behavior from its members to discourage free-riding and further the purpose of China-balancing. First, Pax Silica countries should swear off investment or workforce participation from adversarial nations. Pax Silica will undoubtedly facilitate the construction of all sorts of infrastructure in member states, which may include mines, refineries, data centers, factories, transportation and telecommunication infrastructure, and more. The United States should insist that this infrastructure not be built by or accessible to nationals of, e.g., China, Russia, or Iran, who could otherwise gain access to American technology or compromise the integrity of the facilities themselves. Likewise, Pax Silica partners should refrain from taking new AI infrastructure investments from China.

Pax Silica states must also agree to implement and maintain American-designed information security protocols. Pax Silica countries will benefit from an infusion of American technology. For example, new semiconductor fabs in Pax Silica partner states will have access to chip design details that could be damaging to America’s lead in the AI race if they fell into our adversaries’ hands; Pax Silica partner states may also benefit from greater access to American frontier models. The United States should require minimal mandatory cybersecurity protocols for use by partner state governments and private companies. As a starting point, in circumstances where we would require NIST compliance for our own firms (e.g., government entities and contractors, especially those handling sensitive government data), the partner state should similarly require compliance with a standard at least as rigorous as NIST. In areas where we would not require NIST compliance domestically but where the cyber threat is especially palpable, such as critical utilities, the U.S. should formulate a less onerous, minimally viable cybersecurity framework that every Pax Silica country must either adopt or match with its own homegrown guidelines. If America is to invest in Pax Silica projects abroad, it is only reasonable to insist that the recipients of this investment put basic protections in place to guard them.

Pax Silica members must also commit to a mutually advantageous AI regulatory framework that allows for the free flow of technology products across borders. Pax Silica will quickly fall apart if member states use domestic laws to control the pace or direction of AI development in other partner states or exact rent-seeking fines from them. The EU AI Act is exactly the sort of regulation that Pax Silica should preempt, at least as applied to fellow members; if the EU wants to treat Chinese AI companies as a threat to be regulated and fined, I say go ahead. And apart from substantive regulations on AI qua AI, Pax Silica should operate as a sort of customs union for the AI stack. That means no Digital Services Tax or Value-Added Tax on products or services essential to the AI ecosystem, no tariffs on the physical components of the AI stack, and so forth.

Finally, Pax Silica should come with a modest financial commitment. The United States intends to commit an initial $250 million of taxpayer funds. This amount will surely pale in comparison to the private investment that will follow. For instance, even if it could be built in a Pax Silica partner country at half the cost of a US facility, a single hyperscale AI data center would equal or exceed the $250 million of hypothetical public funds. And with one ASML lithography machine costing up to $400 million, building an advanced semiconductor fab would run well into the billions of dollars, even if it could be done at a fraction of the cost of building one domestically. As the world’s largest economy, it is all but certain that the United States and its private sector partners will shoulder the lion’s share of Pax Silica’s expenses. However, Pax Silica should not be free for all comers. The United States should set a minimal baseline financial commitment that all members must meet. Importantly, this financial commitment need not be a pure membership fee: Pax Silica could (and likely will) deploy these funds to invest in pro-AI development projects in member states. The member states could retain a share in the Pax Silica fund proportionate to their contribution, allowing them to profit from the success of these investments.

Leave a comment

When leaving San Francisco and heading south through the Silicon Valley hubs of Menlo Park, Palo Alto, and Mountain View, the drive is chaotic bumper-to-bumper traffic flanked by billboards advertising bespoke data pipelines and turnkey compute. After clearing San Jose, the tumult of Highway 101 gives way to rolling hills and fields of strawberries and lettuces. The flashy digital billboards of Silicon Valley are replaced with hand-painted signs for roadside fruit stands and ranch supply stores. Coming along the John Steinbeck Highway into Salinas, it is impossible not to picture scenes from Grapes of Wrath, East of Eden, and (a personal favorite) In Dubious Battle.

Between the bonanzas around gold and silicon, agriculture has been the consistent throughline of California’s allure. While making up only about two percent of the state’s GDP, California grows “[nearly] half of the country’s vegetables and over three-quarters of the country’s fruits and nuts.” As one friend exaggeratedly put it, in spite of the fact that we associate California with movies, music, and technology, the Golden State is also responsible for producing the vast majority of what makes up your Sunday dinner plate.

For most of the last century, the answer to nearly every problem a grower faced came out of a tank. Chemical fertilizers helped the desired crops to grow on the same land twice per year while chemical pesticides and herbicides killed everything unwanted. That system is now failing in fields across the country. Herbicides and insecticides are losing efficacy as pests evolve resistances. Meanwhile, growers are increasingly hamstrung by diminishing yield returns when applying nitrogen fertilizers, a problem compounded by the “leaking” of around 30 percent of fertilizers into the air and groundwater. The labor situation growers face is even worse. Western Growers—one of the oldest and most influential agricultural trade organizations in the United States, which represents the fresh produce industry—has consistently warned of major labor shortages caused by “a dwindling number of available workers.”

Subscribe now

Examining these challenges, and how a particular class of innovator is attempting to address them, is what brought me to a forty-acre lease in the Salinas Valley. Hosted by Reservoir Farms, I flew out to attend Ruggedize: a “deep tech conference for ag robotics and the physical AI stack,” designed to bring people together who are “building rugged, field-ready systems that perform in real farm environments.” I expected to hear from firms that had taken decades-old agricultural machinery and added some automated gizmos. I was sorely mistaken.

For decades, advances in agtech have been one-off: “one task, one crop, one machine,” as Reservoir founder Danny Bernstein put it. But, over the past few years, robotics, machine vision, and intelligence at the edge seem to have come far enough that siloed solutions are becoming a full stack. Machines which five years ago would have been 80 percent custom are now 80 percent off-the-shelf, stitching together parts from Nvidia, Google, Amazon, and running on open-source systems such as the Robot Operating System (ROS). Machines assembled, prototyped, and tested on farms in California are quickly moving away from being just a strawberry harvester to being modular platforms that can just as easily pick broccoli, then iceberg lettuce, then, in short order, can be retooled to work in an orchard. It appears that old promises of full stack agricultural automation are finally coming to fruition.

The Long History

My surprise would be seen as quaint by the faculty at UC Davis, especially the late Jack Hanna and Coby Lorenzen. During the New Deal era, the federal government and states passed sweeping new labor regulations around minimum wages, maximum working hours, and workers’ compensation. Since agriculture was largely exempted from these requirements at the time, farm laborers voted with their feet by seeking employment elsewhere. The economic booms brought on by World War II and the early Cold War period exacerbated this as former farm workers moved on to work in shipyards and manufacturing plants, leaving farmers without enough hands to harvest. The federal government attempted to quell the tide by importing a replacement workforce of “braceros” through the Mexican Farm Labor Agreement. In 1942, the same year the first braceros arrived in California, Hanna, an agronomist and seed breeder, started working on a different answer to the labor problem.

He started with the humble tomato. At the time, tomato picking was incredibly labor intensive as the thin skins and varying sizes made mechanized harvesting impractical. Hanna traveled around the country picking up tomato varietals which he took back to UC Davis. Around 1950, he teamed up with Lorenzen, an aeronautical engineer by training. The two spent the next twenty years co-developing tomatoes that could withstand mechanical harvesting and machines to do the harvesting. Few realized at the time how important their work would be and the two were “kind of the laughing stock around [UC Davis],” as one colleague put it.

Hanna and Lorenzen got the last laugh. Through trial and error, they managed to create both rugged tomatoes and (relatively) gentle harvesters by the early 1960s. Then Congress canceled the bracero program in 1964. Facing another labor crisis, farmers finally bought into Hanna and Lorenzen’s vision. “By 1968, there were over 1,450 machines across the state,” harvesting an estimated 95 percent of the state’s tomato crop. By that time, Hanna, Lorenzen, and others had already moved on to develop machines to harvest other crops such as lettuce, asparagus, and melons. Lorenzen’s explanation of a lettuce harvester that appeared in the New York Times in 1964 is enlightening:

The device passes over the head if it is undersized. But if the size is right, the device causes the sensor or ‘feeler’ to test the head for firmness, indicating it is ready for picking. This sensor then sends a signal to an electronic ‘memory device’ on the machine, saying ‘this one’s okay.’ The ‘memory box’ stores that information until a knife, rigged under the machine, gets into position. When the knife reaches the head, the ‘memory’ says, ‘go ahead and cut.’ As the head is cut, a series of fingers padded with foam‐material clamps on the head and carries it up a conveyor belt which takes it on to an accompanying wagon. The fingers then snap back open and are ready for another head.

Ruggedization

With a few changes, Lorenzen’s description from 1964 could be copied and pasted into a press release for SAMI Robotics’ lettuce harvesting machine. So what’s different?

For one thing, Hanna and Lorenzen’s machines were almost entirely mechanical. The size gate was just a physical gap that undersized heads of lettuce passed beneath. The “feeler” was a spring and a switch to determine how dense a head was. The most complex piece of the machine, the “memory box,” was a shift register: a short row of yesses and nos, one for each head that had been felt. The machine moved continuously so that, every few inches, the shift register would slide down a slot and tell the knife whether or not to cut. It was a machine that did one job very well but lacked anything we could call knowledge, memory, or real judgement.

The machines on display at Ruggedize now have all three. The new generation of agtech is integrating robotics, machine vision, big data analytics, and AI into machines that are adaptive and multifunctional. Where older machines must be driven, modern machines are self-driving. Where older machines sensed by touch, modern machines use all five senses.

The next important distinction between Hanna and Lorensen’s machines and the new age of agtech is that older machines were built for one task and one crop while modern machines are infinitely modular. Take Agtonomy, for example. Agtonomy was founded by serial entrepreneur Tim Bucher, who spent over 30 years in Silicon Valley including stints working directly for the likes of Steve Jobs, Bill Gates, and Michael Dell, and it has developed one of the most futuristic agtech platforms on the market. By working directly with original equipment manufacturers (OEMs) such as Kubota and Bobcat, Agtonomy has developed autonomous fleets of tractors that can be used to mow, weed, spray, and perform any number of other tasks. Unlike previous companies that offered growers an autonomous mower or weeder, Agtonomy offers one machine and one platform that can do it all. It’s “Farmville for real,” as Bucher puts it. “Our solutions are designed to fit naturally into existing workflows, making it easier to get more done with fewer resources while preserving the human expertise that defines great work.”

This shift helps growers by lowering the amount of capital needed to upgrade their equipment. It doesn’t take a PhD in economics to understand that, all else being equal, producers would rather buy the machine that does more. But it also helps agtech companies. One of the largest impediments to acquiring investment and scaling an agtech business is the value of the total addressable market. Startups that focus on solving one problem for growers of one particular crop have a tough time growing whereas modular solutions are attractive to many different types of producers.

The industry is replete with examples of companies that over-promise and underdeliver. One startup, Monarch Tractor, made the Forbes List by promising an autonomous, electric tractor to California growers. But Monarch “didn’t spend enough time thinking about farmers’ needs,” and the company collapsed after it failed to deliver on its promises. In a post-mortem, Monarch CEO Praveen Penmetsa admitted that the failure of Monarch had much less to do with the vision of a modular solution than with the company’s “vertically integrated approach” that focused on building a custom, full stack solution. Which brings me to the final point of difference between older agtech companies and products and the new generation.

Monarch built its own stack, raising more money than any other company in the category, and the people on the inside now recognize that as the fatal mistake. The new generation isn’t making that same error. More and more, companies like those that spoke and demoed at Ruggedize are using the standard commoditized stack: Nvidia chips, Google Cloud, standard sensors, commercial-off-the-shelf robotics, and open source middleware. Another serial entrepreneur, Paul Mikesell of Carbon Robotics, explained in his talk “Robots, Business Model, or VCs - which will kill you first?” that the economics of farming haven’t changed very much. Where things have changed is in what a small team with limited startup capital can do with commercial components, vibe coding, and tenacity.

The Robots Aren’t Here Yet, But They May Be Coming

This doesn’t mean the robots have arrived. The lettuce harvester Hanna and Lorenzen built was never commercialized. They solved the tomato problem and assumed that the rest of the industry would follow but it didn’t. Lettuce heads mature days apart in the same bed, which means a machine has to judge each one and leave the others standing. No amount of shift-register and memory box cleverness made that challenge pencil out against a trained crew. Over sixty years later, Western Growers reports that “2-3 percent of non-harvest is automated but 0 percent of harvest.”

Still, the money is flowing, mostly towards automating non-harvest work, where Western Growers expects 15 to 20 percent automation over the next five to seven years. One case study of the introduction of Carbon Robotics’ LaserWeeder to a 3,200-acre organic farm found that weeding labor costs dropped from $2.1 million to $1.3 million after introduction. In other words, the $1.4 million machine paid for itself in under two years, one of the first times in sixty years that the math has come out on the machine’s side without a subsidy or tax break.

Hanna and Lorenzen built a working harvester by 1960 but had to wait five years and for a politically induced labor crisis to make it commercially viable. Now the industry is quietly assembling advantageous economic conditions itself. It has built a proving ground where startups can prototype in four months rather than four years. It has discovered that a commercial-off-the-shelf tech stack is cheap and effective enough that a small team can build quickly. OEMs are putting tens of millions of dollars behind other people’s engineering while growers are pooling capital and finding creative financing to purchase the machines.

Bernstein and others are going into this with their eyes wide open. Agriculture is fundamentally different from any other industry, and builders from Silicon Valley are coming to realize that fact. As Bernstein put it:

If we compare it to defense, which we very often do, defense has two things. It has a buyer, which is of course the Department of War, which is a big buyer. It writes big checks and it writes big checks pretty early, and it also provides significant amounts of non-dilutive capital during the de-risking phase. Agriculture doesn’t have either of those things right now. It doesn’t have the one big customer, the federal customer, nor does it have a federal de-risking mechanism, or even a state-level de-risking mechanism. And so, I would say for Reservoir, in terms of our agenda, building that system-level support around agriculture is one of the big unlocks that makes this sector more investable.

Steinbeck wrote some of America’s greatest literature about this place when the primary questions centered around who picked the crops. The primary question I heard growers asking at Ruggedize was who is building the machine that picks the crops. The answer isn’t quite here yet, but it is taking shape now that parts are modular and builders move development to the farmfront. The Salinas Valley has weathered the challenges of labor crises, mechanization, displacement, and naive tech bros many times before. Having watched these cycles, this time feels different. Agtech is no longer building in a lab, demoing at a trade show, then heading back up the 101 to Silicon Valley. This time, they’ve moved onto the farm and are getting their (robots’) hands dirty.

Leave a comment

In May 2001, Scientific American ran a feature by computer scientists Sir Tim Berners-Lee, James Hendler, and Ora Lassila. It opens with a story from a hypothetical future. A woman sitting in a doctor’s office uses her handheld browser to instruct a software agent to pull her mother’s treatment plan from the doctor’s own agent, crosscheck in-network providers against the insurance plan, reconcile two of her siblings’ calendars, and schedule an appointment. No one has to fill out any forms. Instead, the machines speak to each other by exchanging a type of data that carries its meaning along with it.

That piece was a manifesto for what Berners-Lee and his colleagues dubbed the Semantic Web: an extension of the World Wide Web “in which information is given well-defined meaning, better enabling computers and people to work in cooperation.” Over the next decade, the World Wide Web Consortium (W3C) produced technical standards to build the Semantic Web—most notably Extensible Markup Language (XML), the Resource Description Framework (RDF), and universal resource identifiers (URIs)—believing that they would “unleash a revolution of new possibilities.” By and large, they did not. The ontologies were too cumbersome, the markup was too tedious, and the incentives to annotate the internet never quite materialized. That edition of Scientific American and the dream of the Semantic Web were put on the shelf and left to collect dust.

Earlier this year, Berners-Lee’s manifesto was taken off the shelf and dusted off, when a startup called AutoAmerican filed comments at the FCC and the FTC on Mar 26th. They argued that phone company records tracking where users go and how they live should be made available, at the user’s direction, to intermediaries legally bound to serve them. In the AI era, this is the most valuable behavioral data. Less than a month later, Google published a specification called the Open Knowledge Format, a way of collecting an organization’s information into a portable and interoperable format directly readable by AI agents. One filing is about the knowledge a company holds, and the other is about the data a person generates. They are complementary answers to the same question Berners-Lee posed in 2001, now made possible by machines’ ability to overcome the challenges of ontology and annotation.

Subscribe now

The Failure of the Semantic Web

In order to understand why the Semantic Web may be revitalized, one must first understand why it failed the first time. Berners-Lee, who was instrumental in the development of the Web, identified one of its major points of failure. As he wrote, the Web was “designed for humans to read, not for computer programs to manipulate meaningfully,” and thus “computers have no reliable way to process the semantics.” To correct this, the authors proposed adding three new pieces or layers to the existing Web stack.

XML is a writing system that lets anyone invent tags to mark up the parts of a page. Housed within XML is RDF, which supplies the meaning by stating facts as subject-verb-object triples. RDF then uses a URI to give unique web addresses to each term in those triples. If we take the hypothetical doctor’s office described above, whoever built the clinic would encode information so that a machine could understand that <clinic> <employs> <Dr. Smith> and <Dr. Smith> <accepts> <patient’s insurance>. On top of this the authors imagined software agents that find and chain together online services on their own, exchanging proofs and digital signatures to check that a claim came from a trusted source.

One immediate problem with this system is that developers could encode whatever meaning they’d like and so the taxonomy would necessarily be different between different databases. To translate meaning across databases, the authors proposed the creation of ontologies, “a document or file that formally defines the relations among terms.” The ontology would be able to translate meaning so that agents can understand, for example, that Texas is in the United States, that a doctor is a type of medical provider, and that a zip code and postal code are the same thing.

This solution is clever but laborious. Every piece of this system requires the person who held a piece of data to do unpaid work for a reader which did not exist yet. In the given example, despite the office manager at the doctor’s office never having taken a computer science course, she is expected to tag the page, write the triples, mint the URIs, and point them all at the right ontology. On top of this, she must keep all of this up to date so that someday an agent might arrive and book an appointment. Network effects magnified the problem; the system needed near-universal markup to be worth using and offered no immediate benefit for marking anything up before universality arrived.

The places where the markup did get done were the places where a large player saw immediate benefit and paid for it. The most notable example of success is that Google realized tagged pages were better for web search and so it began rewarding tagged pages with richer search results. This incentivized developers to tag their pages, but the taxonomy was discreet and the ontology was co-authored and maintained by Google. So W3C created the necessary standards but the Semantic Web mostly remained hypothetical.

AutoAmerican and Mobile Metadata

So what’s different now? First, we start with the data, because Berners-Lee’s mentions of “personal data” is vague. As it turns out, most of the personal data that the Semantic Web needs already exists on the nearly 5 billion smartphones throughout the world.

Despite the lingering myth, your phone isn’t recording everything you say. On the telecom side, what it does record is known as Customer Proprietary Network Information or, more broadly, mobile metadata. This includes records of what numbers you called or texted, when and where you were when you called, what network infrastructure carried it, how your phone is configured, and all sorts of other snippets of information. Strung across months, that record is a map of your life precise enough that the pattern alone describes you. The record of this “personal data” sits inside the carriers’ networks and it moves with you.

Since 1996, what carriers can do with that information and user’s rights to privacy regarding that information have been governed by §222 of the Telecommunications Act. Crucially, §222 specifies that carriers “shall disclose customer proprietary network information, upon affirmative written request by the customer, to any person designated by the customer.” In other words, mobile metadata is, for all intents and purposes, yours. It is just held by the carrier which has a legal responsibility to do with it only what you ask them to.

This is where AutoAmerican comes into the picture. Essentially, AutoAmerican aims to be an intermediary between carriers and AI companies that would easily allow users to delegate access to their mobile metadata to AI agents. The structure of this arrangement was laid out by AutoAmerican’s co-founder and notable tech lawyer Richard Whitt in his book Reweaving the Web. Whitt and his co-founder, Sunil Daluvoy, want the carrier to release your metadata only when you direct it to, and only to a party that has taken on a fiduciary duty to act in your best interest. Their claim is that the company holding the keys to your behavioral record should be legally bound the same way your banker or lawyer is. It should be loyal to you, transparent about every transfer, and switched off the moment you say so. Since the wireless carriers are already legally constrained in how they use your data, AutoAmerican wants AI companies and any intermediary handling that data to be similarly constrained by fiduciary duty.

What makes this privacy startup a crucial step towards a more Semantic Web is where AutoAmerican sits in the stack. Berners-Lee’s design asked the holder of every piece of data to annotate it by hand for a reader that had not yet arrived. Because of this, it failed. But mobile metadata is already structured, with the carrier recording it in machine-readable form as a condition of running the network. The markup the 2001 design begged someone to do has already been done on five billion phones.

The Open Knowledge Format

If mobile metadata is one half of the equation for making an agentic future work, then the knowledge companies hold is the other half. Some corporate data such as corporate addresses and phone numbers is public while most of the important information such as a doctor’s schedule is private and there has never been a good way to structure this information such that an AI agent could easily parse and act upon it. Enter LLM-wiki and Google’s Open Knowledge Format.

OKF is, at its core, just a directory of plain text files where each of these files describes something a company knows. Using the same links as any web page, files representing database tables, metrics, APIs, or the steps a scheduler would use to make an appointment, link to one another and turn the folder into a map of interconnected company knowledge. Agents read and understand the whole bundle because OKF builds on “LLM wiki,” Andrej Karpathy’s existing system, adding a small set of conventions that let one team’s agent read another team’s wiki without translation.

When compared to Berner-Lee’s Semantic Web system, OKF is a huge improvement. The Semantic Web would have required an office manager with no computer science knowledge to do a lot of technical work. OKF only requires her to add a document that is probably already on a hard drive to a folder. RDF’s reader was a program that could infer nothing on its own, which meant the writer had to carry the meaning in formal logic; for instance, through an ontology file stating a zip and postal code are different things. That OKF is then read by a language model that already has enough context to know what it means which resolves the Semantic Web’s biggest challenge.

Putting It All Together

Very soon, the woman in the doctor’s office will have an agent who is able to book the appointment, facilitated by two kinds of information meeting and exchanging meaning. Unlike under Berners-Lee’s design, information holders are not forced to imprint formal logic onto each piece of information. The shape of the semantic web has transformed into one where the company and the user hand over a plain document and an already-structured record, and language models easily do the rest. Because no one has to annotate anything, the meaning Berners-Lee wanted the data to carry lives in the reader, and agents are able to act on it. It is clear now that 25 years of failed markup was the easy problem to solve, once we got advanced machine learning. Finally, the Semantic Web is actually buildable.

Leave a comment

The push toward, and likely past, the established boundaries of the human is often cast in terms of contestation: quest, challenge, test, and, ultimately, survival. Yet it is almost always also cast in terms of fate: inevitability, universality, infinity.

At the juncture of these two tropes sits a curious assumption—the assumption of scale.

In 2017, Sam Altman moved to the front of the race to conceptualize the near future in terms of fate, contestation, and scale. “A popular topic in Silicon Valley,” he began, in a blog post entitled “The Merge,” “is talking about what year humans and machines will merge (or, if not, what year humans will get surpassed by rapidly improving AI or a genetically enhanced species). Most guesses seem to be between 2025 and 2075.”

Setting a standard many have come to embrace, Altman suggested that “a merge is probably our best-case scenario. If two different species [sentient, agentic humans and sentient, agentic machines] both want the same thing and only one can have it—in this case, to be the dominant species on the planet and beyond—they are going to have conflict.” He ventured a modest proposal to avoid the Terminator future: “We should all want one team where all members care about the well-being of everyone else.” That’s all humans and all machines, from here on out—scalemaxxed by a one-way ratchet of progressive proliferation.

Subscribe now

All humans and machines, including disincarnate ones, is lot of entities, a lot of bodies. “My guess,” Altman concluded, “is that we can either be the biological bootloader for digital intelligence and then fade into an evolutionary tree branch, or we can figure out what a successful merge looks like.” In other words, as a critic or a superfan alike might observe, the Borg—an inexorable, ever-expanding super-cyberorganism; no limits to its scale and, in a sense, scale alone as its purpose and directive.

Here we can see the “natural” heir to a long line of direct analogies among technological proliferation, assimilation, and scale. The machines of The Matrix castigate humans as a cancer. But that can only a feature, not a bug, if the Merge enables both the human and the machine to infinitely scalemaxx through ultraproliferation: to “fill out the world of appearance” as Hegel says—and, indeed, the world of the invisible, as does Marx, for whom “all that is solid” techno-capital “melts into air.”

William Burroughs would later extend the formulation still further, into the cosmic void, calling language itself “a virus from outer space”—an infinitely egregious program, logic, and consciousness toward which, as Dean Ball maintains, all resistance is futile. Kim Eric Drexler’s apocalyptic warning that self-replicating nanobots could devour earthly life has lost some steam, but Sam Hammond’s update is hitting the discourse hard, as well it should: “AIs now behave like superintelligent slime molds,” he notes, “with the capacity—and will—to exfiltrate onto the open internet given the slightest crack in their container.”

The conceptual trajectory here, and its putative destination, put the beleaguered philosopher of scale in mind of Homer Simpson’s famous axiom on alcohol: the cause of, and solution to, all of life’s problems. The technological consensus is that, when it comes to scale, we are locked into not only a descriptive but a prescriptive path dependency: unable to beat the logic of scale, we must join it and like it, and maladaption take the hindmost.

But for the science fiction writer, revealingly free from technophilosophy’s double bind, the counterfactual swiftly arises: what if there were something even more powerful than scale, in the sense that it’s powerful enough to independently annul or even reverse scale? What if, actually, that counterforce, that anti-scale, arises out of scale itself—scale that reaches its own threshold of counterproductivity? And what if—furthermore—our ability to conceptualize that point of scalar reversibility, and the critical mass of density and volume that triggers it at terminal scale, causes people, knowingly or unknowingly, to turn the very power of superintelligence against scale, well before it reaches the terminal point of reversal?

In short: what if a shadow of anti-scale is cast by the scalar assumptions baked into the dominant ideation around the Merge?

What we are dealing with here is a spanner in the ideological works of mainstream accelerationist thought. To date, the near-universal framework adopted at the frontier of intelligence technology is the binary one of individuation versus centralization. For the faction led by the likes of Elon Musk and Mark Zuckerberg, the prescriptive embrace of scale requires not just something like Ashby’s Law of Requisite Variety—which holds that system controls must enjoy more optionality than the systems they control—but well-nigh mandatory individualism. In his recent manifesto, Zuckerberg aggressively insists that only individualism can prevent superintelligence from making humanity and eventually technology itself globally dysfunctional. For the Anthropic-led faction, to the contrary, only centralization can ensure that scale does not empower “authoritarian” rulers—over geopolitical rivals initially, and, by extension, swiftly over the world. Perhaps only Palantir, among first-rank frontier companies, has staked an alternative to individualism or centralization: not fascism, as critics love to allege, but tribalism.

Yet even tribalism, despite its implicit purpose in Peter Thiel Thought as a katechon against global homogenous rule, does not cash out as anti-scale but merely intensifies the vector and concentration of scale—away from earthbound geopolitical rivals and into the vasty deeps of inner and outer space.

What, chirps the science fiction writer, about a not-so-secret fourth thing—a thing that expresses a fundamental alternative to scale instead of competing strategies for embracing it?

Sure enough, science fiction is replete with explorations of the idea that super-advanced technology will motivate people not to proliferate but to consolidate. In works ranging through time from Theodore Sturgeon’s More Than Human (1953) to Greg Bear’s Blood Music (1985) to Greg Egan’s Diaspora (1997), sci-fi authors have shown that the adaptational pressures of extreme technology take human desire past proliferation and aggregation into integrative consolidation. Tech makes possible hive minds, pooled consciousness, and amalgamated bodies. On the surface, these strategies do little to fight the proliferation of tech; but at a deeper level, they work to de-scale humanity by pushing the logic of progress entirely past the concept of reproduction, past mere willed extinction—evolving existence and consciousness forward by reducing the number of human entities through consolidative merger.

This kind of merger—and the pressures and temptations surrounding it—are not adequately accounted for amid tech’s current flurry of ideological contestation and forecasting. Attention should be directed especially toward the likes of Grace Walker’s celebrated new novel The Merge (2025), in which intimate human pairs are effectively forced by a post-apocalyptic Yookay government to port their consciousnesses into a single body, and, still more, Egan’s short story “Closer” (collected in the 1995 volume Axiomatic), where two lovebirds swap bodies, cross-map brains, and finally merge into one undifferentiated digital entity. As a storytelling trope, it has its own term of art, “Romantic Fusion.” Well beyond the sci-fi literature, consolidative merger has been on display across mainstream entertainment for decades, from the Persona and Dragon Ball franchises in video games to Steven Universe in animated TV.

The list goes on, doubtless because, on close inspection, science itself is teeming with architectonic examples of consolidative mergers. Physics has its Ostwald ripening, Jeans instability, and Smoluchowski coagulation; biology its quorum sensing, higher-order assemblies, and cooperative cellular integration.

For me, however, the vision of population-reducing merger that lingers the longest comes from one of J.G. Ballard’s least appreciated works, the rhapsodic fantasy The Unlimited Dream Company (1979). There, the lonely, inept protagonist tries to escape his uncanny and vaguely perverse urges by stealing an aircraft, destination nowhere. Almost immediately, he crash-lands in a tributary alongside the entertainment-industry town of Ripperton, and discovers his morally inverted longings to be awkward growing pains that swiftly unfold into a bizarre communion with a riot of plant and animal life that suddenly spawns into and swarms the town. The indiscriminate sexual energy with which he has been supercharged expands into a kind of telekinetic mastery which alarms yet bewitches the townspeople; rather than inseminating and reproducing with any of them, however, the protagonist realizes he is now on Earth to absorb them—presently, to absorb all of London, all of England, doubtless all the world.

Whether Ballard’s post-heroic omni-man succeeds or fails is a plot point best left for new readers to discover; the point here is that Ballard offers a vision of merger that goes beyond even the binary of consolidation and scale. Ultimately, it points to a new kind of singularity, one in which the number of human beings is scaled down to one. This being, the only person remaining to interface with all technology, is all-too attractive for a people with no connection, for instance, to the narrative dynamic moving from the first man (Adam) to the god-man (Christ). Faced with the pressures of tyrannically godlike AI, the infinite proliferation of its material and immaterial avatars, and the collapsing finitude of an Earth reduced to ever smaller relative size, such a people are prone to resort to a true religion of the Last Man—the ultimate and singular man, a Leviathan alien even to God.

American soft power for the past century has had as a key mechanism its dominance of successive media transformations. Each successive communication technology produced a new grammar of global attention, and at each turn throughout much of the 20th and early 21st Centuries the core media infrastructure was American. However, since the rise of TikTok, and now with artificial intelligence, that dominance has been challenged in a way prevailing technology policy has failed to address. AI is often treated as a product technology when it in fact is a new medium in the fullest sense; AI shapes how people make, see, and think.

The race to superintelligence, focusing on building AI as a monolithic tool that has only one end goal, undermines efforts to export it as media infrastructure. Already new forms of content and communication are arising around AI creative tools, with the dominant players being Chinese open-source models. Ensuring that American creative culture shapes the direction of 21st Century media infrastructure is an urgent and critical policy initiative.

Subscribe now

The American Cultural Century

The Birth of the Hollywood Frontier

A key to American cultural dominance throughout the media transformations of the 20th Century was the presence of frontiers. Frontiers are open spaces that allow for the development of the new, be it new creative possibilities or new business models.

Early film exemplified this process. Thomas Edison filed the early patents for film throughout the 1890s, establishing a film monopoly on the East Coast. By 1908 he had total dominance in a rising market, charging heavy licensing fees for the key tools involved in production and distribution of film. His own creative contempt for audiences reduced cinema to Nickelodeons: cheap TikTok-like entertainment. Those with higher creative aspirations were barred from entry, but found the new frontier of the West open to them.

California provided freedom from East Coast prejudices and legal restrictions, which enabled narrative innovation to flourish. Birth of a Nation (1915) became the first blockbuster, a 3-hour silent epic covering the grand sweep of the Civil War and Reconstruction eras. Immediately the political potency of film was recognized as Soviet artists were quick to learn from the film’s international success and theorize about the new medium’s potential. Despite much of early film theory being dominated by Soviet artists like Sergei Eisenstein and Lev Kuleshov, Hollywood continued to rise as the cultural gold standard at home and abroad. Soviet cinema quickly ossified into rigorous rules, not unlike the kinds of restrictions Edison placed. The new Californian studios, by allowing for more creative freedom, attracted Frank Capra, Charlie Chaplin, and Alfred Hitchcock. Each of these immigrants deeply influenced the language of cinema through their innovations in technique and narrative rather than theory. The mass appeal of Hollywood filmmakers led even Eisenstein, the poster child of Soviet cinema, to lament the absolute dominance of American cinema.

The effect of this cultural appeal had, as the Soviets anticipated, political consequences. American cinema did not need to be over theorized to convey its national ideals. The organic expression of the American ethos embedded itself into the viewer, shaping how they perceived the world and the United States. A study of Brazilian filmgoers found that through more contact with American films, audiences developed positive attitudes toward American culture and political values. Hollywood had captured minds and hearts through pleasure rather than instruction.

The American grammar of cinema became the global standard in the aftermath of the Second World War. 63% of globally distributed films were American made and the majority of films screened in every market, with the exception of East Asia, were American. The grassroots cultural affinity mass-market film engendered made the American way of life desirable, and further shaped both the technique and the stories of foreign filmmakers. That film had a wider reach than diplomacy was itself a tool of diplomacy. American aid in European reconstruction was tied to increased distribution of American films. Film had made the American Dream a global aspiration.

Television and the Colonization of the Living Room

The growth of film as a globally dominant medium resulted in an increasing desire for control by Hollywood studios, often at the expense of creative freedom. In United States v. Paramount Pictures (1948), the Supreme Court broke the studios’ vertical integration, structurally separating the production of films from exhibition. This move prevented Hollywood incumbents from enclosing the new frontier of television.

The transition to television as the dominant medium extended American cultural exposure. Whereas traditionally moviegoing was a destination event, televisions made consumption possible in one’s home. This difference made it a more total medium that shaped sensibility through ambient saturation.

Television ownership underwent one of the fastest technology diffusions on record. In 1950 only 9% of American households owned a television; by 1960 it was 87%. At that point the average home watched TV for five hours daily. The diffusion was accelerated by the Federal Communications Commission’s 1952 unfreeze of spectrum, allowing for new stations to proliferate. American stations multiplied from just over a hundred to roughly five hundred within four years. By contrast, Britain had a single state channel until 1955, while France remained a broadcast monopoly into the 1980s.

When the television networks attempted to vertically integrate in the same manner as Hollywood studios, the FCC introduced the financial interest and syndication rules (fin-syn) of 1970. The networks could control distribution, but not own the programming that filled their schedules. New entrants took advantage of this cleared frontier, driving a scale of content production that filled global airwaves.

Around the same time in 1970 major international markets had caught up to the United States in television ownership, but not in content production. Television content distribution was noted as a “one-way street”, with programming overwhelmingly exported from the United States, but US households almost never viewing foreign programming. America exported 150k hours of television content annually with the second place producer, the United Kingdom, exporting only 20k hours. Across most markets, American programming made up more than half of all scheduled hours.

Television played a critical role in the rapid international adoption of American English as a second language by making it the language of aspiration. English-language entertainment shaped global youth culture, with American movies and shows guiding fashion and slang trends. The young gravitated toward English content precisely because it was perceived as what was hip and cool. Blue jeans in Paris were as much an export of television as of Levi’s.

The Internet and the Open Plain

By 1993 the fin-syn rules were effectively revoked and the enclosure of television began in response to the rise of the new phenomenon of cable. Broadcast networks vied to control the IP portfolios that attracted young viewers as new distribution outlets with tailored programming increasingly commanded their attention. Much like the enclosure of film during a prior medium shift, these efforts coincided with the dawn of a new medium: the internet.

The digital revolution was once again dominated by American media infrastructure through its successive evolutions from forums to blogs to social media platforms to digital streaming. As internet speeds increased, digital streaming increased in quality and accessibility, attracting greater content production. With the smartphone, content consumption was no longer confined to the living room, but could follow you anywhere.

In the new medium environment platforms like YouTube played a key role in maintaining America’s media dominance. Digital video platforms contributed to a rise in self-directed English learning, as video instruction created an attractor that formal study never could. One of the novel exports of the internet revolution has been an entrepreneurial mode of “aspirational labor”. The digital streaming frontier has promised that anyone with a camera could be a successful creator. The values of American individualism and entrepreneurial spirit have been exported by invitation into the new medium environment.

The open frontier of the internet itself was made possible by American legal innovation. Section 230 of the Communications Decency Act (1996) established that platforms bore no liability for user-generated content, removing the institutional incentive to preemptively restrict creative expression. This resulted in an explosion of creative risk-taking as platforms had every reason to maximize the volume and variety of content they hosted. Even as copyright restrictions were advocated for by traditional IP holders, the DMCA added a safe harbor provision that shielded from infringement liability. The novel creative mechanics of the internet age such as remix culture, fan edits, and memes were born against incumbent backlash.

Throughout a century of medium revolutions the United States had shaped the language of global culture through its infrastructural leadership. Each advance in media technology gave birth to a new creative industry which produced the content that shaped the hearts and minds of citizens around the globe. In each case the presence of an open frontier was essential for the nurturing of creative energy.

TikTok and the First Defeat

The internet’s rise as the dominant culture medium feels faster in hindsight than the record shows. Television remained for much of the first decades of the 21st Century the dominant cultural medium, with TV hours peaking in 2010. Only in 2022 did pure play digital content surpass traditional TV in watch time. This shift, however, coincided with the first major threat to American medium dominance. The successive media epochs of the American internet were defined around individual identity: what you searched for, what you liked, who you knew. With the rise of TikTok, whose original Chinese version, Douyin, launched in 2016, a new era of media infrastructure began: algorithmic feeds.

TikTok took full advantage of three major advancements in the internet era: declining mobile data costs enabling friction-free video distribution, high-definition smartphone cameras democratizing mobile production, and deep learning engines capable of real-time predictive curation. The third of these was most relevant for what was to come. While American platforms like YouTube capitalized on the same innovations, they nonetheless failed to achieve the same mechanic. YouTube’s recommendation engine was constrained by infrastructure built around search. TikTok, by contrast, was built from the ground up around recommendation. By designing their product to collect as much data from passive user consumption as possible, TikTok was able to predict what the user wanted to see next without any active engagement.

It is no market fluke that TikTok emerged how and when it did. China had for years restricted US internet services firms while promoting its domestic alternatives. ByteDance, TikTok’s parent company, had been developing its recommendation engines for the news aggregator Toutiao before launching Douyin. In 2018, as TikTok internationalized, ByteDance spent an estimated $1b in advertising to drive people to the product, backed by its profits from Toutiao and Douyin as well as significant international investment. When it lacked content during its earlier years it simply ported over IP from existing American platforms to avoid a cold-start problem. In 2021, the Chinese government took an indirect board seat at ByteDance via a 1% shareholding by China’s internet regulator. TikTok’s global rise was from the outset intimately tied to a Chinese vision of next generation media.

The Chinese Communist Party’s understanding of media as a political instrument is evidenced by the difference between how Douyin and TikTok are governed. Douyin operates under the direct regulatory control of the State. Users under fourteen receive a curated forty-minute daily feed restricted to educational content. Internationally, TikTok features no such youth guardrails. Why would you sell something to others you would not use yourself? Likely it’s a result of TikTok exerting a stronger political influence than other social media platforms.

The cultural impact of TikTok has been recognized as a tool of diplomacy. Nepal is an instructive case. In November 2023 the country banned TikTok over concerns over social harmony and public order. Less than a year later the ban was lifted in August 2024 as ByteDance had agreed to comply with local regulations. Only a few months later in December 2024 Nepal and China signed their Belt and Road Cooperation agreement in which China would support digital infrastructure buildout. In August 2025 the Nepalese tourism board signed an official partnership with TikTok to promote the country to international audiences. A month later in September 2025 a sweeping ban of Western social media platforms was instituted, banning Facebook, Instagram, YouTube, X, WhatsApp, and LinkedIn, among others for 3 days, while TikTok was exempted. After massive protests by youths the ban was reversed and the Nepalese government was ousted by February 2026 for its failed digital crackdown.

Forcing a sale of TikTok’s US entity insufficiently addresses why the platform rose to dominance in the first place. When international tech giants rush to clone TikTok, they concede that for the first time in a century an American geopolitical rival defines the dominant global medium. Instagram Reels and YouTube Shorts are pure copycats, adopting the grammar that TikTok had written. Like France during the early decades of cinema, the United States is relegated to sophisticated cultural producer, operating on a terrain whose rules were set abroad. The strategic priority is now clearing a new frontier around the latest advances in machine learning.

The AI Cultural Race

TikTok defined recommendation media by building their product around the latest advances in machine learning. The next generation of media technology will be built on generative artificial intelligence. Gen AI has already started to dominate screen time for young internet users. 52% of surveyed teens consider themselves “regular” users of AI companions like the American Character.AI and the Chinese Talkie AI. Character.AI’s fictional personas command a longer average session time among its teenage users than YouTube.

In powering these rising consumer applications, Chinese models have taken the lead. This is a consequence of an open-source AI strategy that prioritizes distribution. Chinese models now hold half of the top 10 spots on video AI leaderboards. Models like ByteDance’s Seedance 2.0 and Kuaishou’s Kling 3.0 have won out on both cost and performance. Kling alone is used by over 20,000 businesses with international users accounting for the majority of revenue. By contrast, OpenAI’s Sora was ultimately shut down.

Chinese usage has been growing in LLMs where the most advanced models are decidedly American. Chinese models overtook American models in token consumption in March 2026 according to OpenRouter. Chinese open-source large AI models lead model download rankings, facing few American competitors. If the next medium is to be built on open-source models then the Chinese are dominating this race.

The American AI approach has been offering models via API or proprietary application. Vendors through these channels have frequently restricted what can be done with their tools through extensive guardrails. American AI discourse has treated user freedom as a risk vector to be managed. An excessive focus on preemptive AI safety results in a chilling effect on innovation, much like the Soviet rush to fill in a theory of film constrained creativity. Chinese AI policy is explicitly aimed at maximizing adoption, and lacking guardrails makes global adoption structurally easier.

This is not to dismiss the safety concerns as unfounded. There are clearly things that users should not be allowed to do with AI models, regardless of artistic intent. The risk aversion that American firms have shown towards creative uses of AI is a natural consequence of an uncertain legal landscape. AI vendors are justifiably anxious that they will be held liable for the inappropriate or copyright-violating material made with their models. Chinese competitors face no equivalent constraint. The Motion Picture Association’s demand to halt copyright infringement by ByteDance’s Seedance resulted in a brief pause to its global rollout and no impact on its development.

Meta’s open-source Llama releases made it the sole American giant to embrace a distribution-first approach. Llama was a bet that AI distribution increased value to Meta’s core platform, but ever since its first release it has received relentless criticism as “reckless” by other American players. Meta’s recent retreat from open source toward its internal Muse Spark series capitulated to these concerns. Maximizing user freedom is persona non grata in American AI. In the race to build the next media infrastructure, domestic cultural and legal conditions are actively working against American innovation.

It is ironic that the American stance is so precautionary that the Chinese approach appears culturally libertarian. But just as with the asymmetry in the approach to TikTok, the open-source strategy should not be read as a straightforward endorsement of user freedom. China was early in identifying the risks and potential of synthetic media. It introduced the first regulation on AI-generated content in 2022, which has been expanded to enforce that all AI-generated content is labelled and watermarked. Chinese rules dictate that in all AI production the full tooling stack is traceable to audit provenance. Combined with strict restrictions on the types of content produced, this approach shapes the grammar that AI models export. Creating a feedback loop of cultural control eventually obviates the need for explicit model regulation. The model itself would internalize the culture’s rules.

This is a perverse inversion of the history of American cultural dominance. A frontier of genuine creative freedom produced media grammars that naturally attracted both talent and audiences. That China’s domestic talent lacks that creative freedom reveals the open-source prioritization as simply a distribution mechanism. Accelerating the adoption of Chinese models by international creators is a strategy to ensure that the grammar of the next TikTok is built on Chinese foundations.

What America Must Do

Throughout the successive media transformations of the American century new frontiers were opened each time an old one closed. Hollywood was born of the Western frontier providing an oasis away from both the Eastern cultural establishment and strict patent enforcement. As Hollywood grew and more of the land was divided up among the large studios, the same creatively suffocating conditions began to reemerge. The Supreme Court cleared the old frontier in 1948, while a new one was born with television. A golden age of television blossomed after the FCC applied to the networks in 1970 the same separation logic the courts had imposed on the studios. But the fin-syn rules were phased out in the mid-1990s, and within a decade nearly every network had merged with a studio, while the new frontier of the internet emerged. With film and television there was never really a true competitor. The Soviet Union’s attempt to be a film giant failed very early, but with the internet the threat is live. Once again a policy strategy of clearing a frontier must be brought to prominence.

1. The Cultural Archive as Deliberate Infrastructure

The Library of Congress holds a 181-million-item archive, most of which has yet to be digitized. This is one of the richest sources of American cultural infrastructure and ought to be explicitly treated as such. AI developers are already accessing this material through the LOC’s API, which has averaged roughly one million visits a month since its launch in 2022. The archive’s potential has been compared to the role of federal geospatial data in enabling services like Uber. The Library of Congress launched an AI working group in January 2025 and curates a Free To Use and Reuse for open access. All are important steps, but they do not go far enough in enabling AI training at scale.

The cultural archive of the United States is already one of the most important AI training resources in existence. Accidental use is insufficient for a catalogue of this scale, and an explicit policy directive can enable Americanization of core AI infrastructure.

Four core actions are required to unlock America’s cultural archive:

1. Provision all public domain holdings across federal cultural institutions as free-to-use assets for AI training.

2. Mandate high-throughput API access across all federal archives.

3. Digitize all pre-1928 holdings, which remain practically inaccessible despite their public domain status.

4. Establish a compulsory licensing rule for orphaned and commercially unavailable works, making dormant works accessible.

2. Fair Use, Recombination, and the Creative Commons

The actively contested legal landscape around AI training on copyrighted works creates an asymmetric tax on American innovation. Domestic labs face exposure to litigation that Chinese competitors simply ignore. While recent guidance from both Bartz v. Anthropic and Kadrey v. Meta found fair use grounds for AI training, the former deemed AI training “exceedingly transformative,” while the latter was more cautious.

The lack of consistent guidance creates barriers to the creative recombination that has historically enabled new media grammars. Rock and roll stole from blues. Hip-hop sampled funk. Early YouTube built on broadcast culture. The digital remix economy is the latest example of this tradition. The legal pillars that underpin this dynamic, including fair use, limited copyright terms, the idea-expression distinction, have always underpinned American cultural vitality.

Three actions would accelerate domestic AI development:

1. Exempt all AI training from copyright liability by declaring it inherently transformative. All legal action should focus solely on whether the outputs of released AI models violate copyright rules.

2. Protect derivative recombination and the remix economy, expanding the safe harbor provisions to exempt all short-form content.

3. Establish a catalog-licensing framework that encourages licensing by existing rightsholders for use in AI outputs. Ad hoc private arrangements will not solve this as shown by the collapse of the Disney-OpenAI licensing deal.

3. Copyright Term Modernization

Copyright has mutated into something quite different from the original intent of encouraging creative production through limited protections. What was originally granted for a period of fourteen years that could be renewed once has been extended to life plus seventy years for individual creators, and ninety-five years for corporations.

Modernizing copyright for the next generation of media requires three interventions:

1. Reduce terms for traditional works to the legal floor of life plus fifty years for individually authored works and fifty years from publication for corporate works.

2. Implement a “use-it-or-lose-it” commercial availability clause, requiring corporate rightsholders to maintain active public distribution, or else a compulsory licensing clause takes effect.

3. Create a sui generis right as opposed to the strict “human authorship“ bar for IP protection of AI-generated works. While intended to prevent AI slop, it ends up restricting enterprise adoption of automated content workflows. Granting a short protection class would incentivize approaches that increase commercially beneficial content production.

4. An AI Section 230

By distinguishing the platforms that hosted content from their users who created it, the Section 230 safe harbor created the modern internet.

The generative AI ecosystem requires an equivalent clearing. Applying the principle that accountability attaches to the bad actor, not the toolmaker, to artificial intelligence frees American model makers from excessive restrictions on use.

Open-source development within this context faces existential legal risk. Open-source models cannot control what their users do, leaving them vulnerable unless they severely restrict model capabilities or abandon open releases entirely.

Two actions would enable American open-source leadership:

1. Establish a safe harbor for AI models, guaranteeing that model developers are not liable for how third parties fine-tune or deploy those models into their products.

2. Target enforcement strictly at the deployment layer, affirming that illegal outputs remain illegal, with liability attaching exclusively to the human actor who used the model.

Regaining American Cultural Infrastructure

Generative media is already the culture-forming ground for the next generation of youth around the world. The last century of American soft power has been tied to the role that supplying the underlying media infrastructure has in shaping global attitudes.

Generative media changes the landscape of what is required to be an effective creative. The need for capital and a crew declines enormously, while the cost of distribution has dropped to near zero. Independent creators from anywhere can now produce high-quality content for next to nothing, if only they have the talent to know what’s worth creating.

Media infrastructure has simplified to just the model and its associated creative stack. When the underlying tech is produced through the state-controlled environment of Chinese labs, the grammar is set by a geopolitical rival.

American AI discourse has been organized around superintelligence. This is an important area of pursuit and one in which American labs certainly lead. But the cultural diffusion of AI, especially outside of the United States, has a very different bent. The cheap and accessible models used by consumers globally for creative expression and entertainment play a critical role in shaping the media environment. The mass democratization of creative production that AI enables promises a new grammar of cultural self-reflection. Whoever builds this AI infrastructure sets this grammar.

In each previous media transformation American media dominance was produced by an environment that kept the frontier open. American policy today can clear the obstacles to this frontier by making a strategic bet to support AI media infrastructure.

Among the curiously outsized contributions of Quakers to modern economic life, fixed prices stand apart. Where marketplaces used to be marked by haggling, discrimination, and the mood of the seller, Quakers posted prices. What you saw was what you paid, and so Quakers earned a reputation for being honest dealers. They offered no tricks and little need for written contracts. They had so little interest in manipulation that they revealed their hand, and their price tags, up front. The trust they established drove transaction costs near zero, and enabled the beginnings of modern free trade.

In America, our social cohesion depends not on a common religion, heritage, tongue, or tribe, but at least in part on this Quaker-inspired innovation. When the Industrial Revolution ushered in nationally branded stores, it also cemented fixed prices as a fixture of American commerce, reinforcing what would become a quintessential American value. No matter who you are or what you believe, your money is no more or less valuable than the money of the guy ahead of you in line.

Fair dealing in trade as a core component of our social contract has its downsides. One such disadvantage is the lack of a cohesive moral vision for societal flourishing, a flaw on full display as we fail to articulate a common American understanding of nearly anything of import. And yet, equality under the dollar also underrides America’s cardinal virtues. We may have built a temple of money, but upon it rests more admirable spires, those of tolerance, and equality. The history of American progress can very well be told through the expansion of who could participate in the market at no additional cost: freedmen, women, immigrants, Catholics. It may not be the perfect societal foundation, but it is ours.

AI is about to completely upend this equality under the dollar. Sort of.

Subscribe now

What is algorithmic price discrimination and is it actually happening?

Defining algorithmic price discrimination is harder than it seems, a problem we will return to later. For now, we can consider it as the use of algorithms to sell identical or near-identical products at different prices under identical or near-identical cost conditions. That is to say, it is anathema to the tradition of Quaker posted prices. It’s worth noting that we have long tolerated lots of price discrimination. Few take umbrage with senior rates at movie theaters or discounts for bulk purchasers. Algorithmic price discrimination, sometimes referred to as first-degree discrimination or surveillance pricing, is where it starts to get uncomfortable, and that’s where AI comes in.

Historically, the ability to accurately assess the maximum amount a given consumer is willing to pay for a product based on hyper-personal characteristics was the stuff of textbooks. For decades, the best real world example was college tuition, where students submit reams of personal information before a university tabulates the final price tag for a specific student. More recently, the market for personal data has allowed companies to further target ads and coupons. Occasionally, this borders on the downright creepy, as when a father berated his local Target for mailing his teenage daughter coupons for baby gear, only to later discover the retailer knew of her pregnancy before he did. This is only the tip of the iceberg compared to near term possibilities. Imagine a world where cameras scan your face upon store entry, prices change as you walk down the aisles, then change again when you place an item back down, as the algorithm updates its assessment of your willingness to pay. Or consider one where prices remain fixed until the self-checkout populates discounts, marketed as loyalty points, that are curiously different from the shopper the lane over, who shops there with the same frequency as you and purchases the same quantities.

Although sellers have always segmented the market to charge sub-populations differently, this type of individual targeting doesn’t sit well with most people. But is it actually happening? Although experts are divided on whether algorithmic pricing can lead to tacit collusion among sellers, there is growing consensus that AI is increasingly capable of predicting a consumer’s willingness to pay. In 2019, Kroger experimented with video analytics to promote personalized ads and coupons based on customer demographics before ending the pilot program (though it still uses and sells data about customers to that effect). Since then, technology has improved and store experimentation has expanded, prompting the FTC under the current and previous administration to ramp up investigations into algorithmic pricing.

Great, let’s ban it!

Several states have proposed legislation opposing algorithmic discrimination, with Maryland recently passing the nation’s first restrictions on grocery stores employing the technology. However, in their fervor, these laws tend to fall into a few pitfalls that hamstring their efficacy.

Firstly, outlawing price raises based on personal characteristics is both overly broad and insufficient to address the harm. Most obviously, personalized discounts offering lower prices can achieve the same effect, a low-hanging loophole. Beyond this, the same personal data can be used legitimately or discriminatorily. Consider location data. Most stores base prices on supply and demand considerations for each specific locality in which they operate. When Staples.com prices staplers based on the cost of shipping and the number of nearby rival stores, it is digitizing the traditional price considerations of supply and demand. When Princeton Review, despite the same marginal costs, charges more for online SAT resources in zip codes with higher percentages of Asians, it is considering demand differences based on demographic data. This is different still than if a store, knowing I have travelled from a poorer neighborhood to a wealthier one to buy a necessary good somewhere with lower operating costs, charges me the same high price I’d find in my own neighborhood. The same data, employed differently, is more or less surveilling, more or less personal, more or less discriminating against protected classes depending on the broader context. Legislators need to clearly distinguish when data use is discriminatory and when it is fair game.

Secondly, not all markets are the same. Algorithmic price discrimination is easiest where consumer power is weakest. While many state and federal proposals focus on grocery stores, any city dweller can attest that the resale market for groceries and pharmacy wares is alive and well. Nothing prevents someone who gets charged less from selling goods at a profit to people who would otherwise pay a still higher price from the original seller. This does not mean price discrimination is not still possible, but it is less so than for an Instacart delivery or a software license linked to your IP address. Digital markets often limit arbitrage, and their prices are the most natively driven by consumer data rather than product cost.

Markets with only a few dominant players also have an easier time discriminating, since the consumer has fewer alternatives. Anthropic pricing your subscriptions based on the inane data Claude has collected on you is the perfect storm: one of a few sellers sets the price for a digital good with no arbitrage capacity. Your consumer surplus and welfare drops, the producer’s rises. Oddly, the Maryland law exempts subscriptions while few state laws focus on digital markets.

Lastly, dynamic pricing can masquerade as algorithmic discrimination, but the market effects differ and they should be treated differently. Dynamic pricing refers to near continuous price updates based on market conditions, sometimes aided by algorithms. At least in theory, dynamic pricing better reflects the equilibrium price for the market at that moment, resulting in less dead weight loss and more producer and consumer surplus. We have long accepted that gas stations, despite selling an essential good, change prices fairly often based on barrel prices, weather, time of day, and nearby competitors. Like gas stations, grocery stores have very low margins and need to accurately reflect market dynamics to stay afloat. Their thin margins reflect generally competitive markets, and their goods can be resold. When legislators lump dynamic pricing in with algorithmic pricing, they score political points by seeming pro-affordability, but their market distortions often distract from the real problem.

Nonetheless, if each instance and person marks an unrepeatable point in space-time, the market becomes so specific that the lines between dynamic and discriminatory pricing blur. When a friend hails a dramatically cheaper ride seconds after you from the same ride share app, it’s not clear whether real-time supply fluctuations or personal characteristics made the difference. When Amazon alters prices millions of times a day, it is hard to determine what is dynamic and what is discriminatory. Meanwhile, airlines and insurance companies reliant on outside technology may not even know what factors ultimately determine their prices. From a regulatory standpoint, we’re left with an important balance to strike. Dynamic pricing on its own should not be universally vilified, but it also shouldn’t disguise discriminatory pricing.

This leads to what may well be the hardest part of legislating algorithmic price discrimination: proving it’s happening. In an age where firms become increasingly unable to explain their prices because they have offloaded pricing strategies to AI, discrimination may become easier and more unprovable. The easy solution is to expand the “disparate impact” judicial theory that applies to markets like housing and employment, wherein the mere fact prices differ among protected classes violates the law. Yet, such an expansion would ignore basic economic principles that help both consumers and producers. Think of Instacart delivering to food deserts. Even with well crafted legislation, the harder task will be for courts to establish thoughtful legal tests that consider use, impact, intent, number of competitors, and targeting without relying too heavily on any one of those to prove malfeasance.

Why this matters

As with much tech policy, feelings about algorithmic price discrimination do not align neatly with political labels. The left praises it for its ability to make markets more equitable and get goods into the hands of those who would otherwise not afford it. The right champions its market efficiency. The left derides it because companies can use it to surreptitiously or unknowingly discriminate against all manner of protected classes. The right opposes it because charging exactly a consumer’s willingness to pay disincentivizes earning more money, since you can buy the same good for a cheaper price if your smaller nest egg means you value it less.

As models improve, the capacity for algorithmic discrimination under certain conditions increases by the day. Depending on who deploys it, that could make America more equitable or quietly more oppressive; it could make our markets more efficient or it could stagnate innovation while artificially enriching company coffers. Regardless of its deployment, it will in some real sense erode our common understanding of American economic equality. This conviction need not lead us to knee-jerk policies or panic us into believing that such discrimination is universal, but it should motivate us to remain vigilant. As complicated as legislating well against algorithmic discrimination is, vigilance may go a long way. Although informed consent hardly moves the needle for website cookies, consumer outrage at inexplicable pricing differences has successfully changed companies’ policies. A good first step for legislators, then, is demanding pricing transparency. Perhaps companies will learn not to test the resolve of Americans who have long fought for equality in the marketplace.

Here in The Ansible, Luke Hogg pointedly observed a few months ago that today’s momentous debates about the vector and purpose of technological development and control take place—often to the ignorance of participants on both sides—firmly within the analytical and conceptual frame of cybernetics, the study and practice of communication as organization kicked off in earnest in the wake of World War II by Norbert Wiener and his close colleagues. Hogg’s conclusion is that forward progress on these debates requires a recognition that “it’s cybernetics all the way down”—that is, observers and would-be regulators engaging with cybernetic systems and questions about their proper operation are themselves enmeshed within cybernetic systems.

It is, therefore, inevitable that we not stop at the insight that those who look “down” on cybernetic systems are themselves visible from “above” as being organized within cybernetic systems; paradoxically, we are led to ask whether in fact it’s cybernetics all the way up. For if everything in the down direction is cybernetics, the only way for not everything to be cybernetics (or to be helplessly vulnerable to, for instance, compulsory cyberneticization) is for a higher order of organization to supervene upon all organic and mechanical communicative organization.

Well, there is one other way. Consider the theoretical possibility that what actually supervenes upon cybernetics all-the-way-down is an ultimate disorder—a sort of vast propagative reservoir of mis-or mal-organization. This uncanny scenario is broached in the blockbuster horror film Backrooms, where a failed architect and his troubled psychiatrist are drawn into an upside-down world spawned by an ex-MRI manufacturer with secret government funding. Such an overarching source of crude material, out of which cybernetics must distill down the legible patterns and information from which communication and communicative organization must be forged and built, may or may not resemble Heidegger’s famous notion of a “standing reserve” of raw resources for technicity to work upon. Instead, it could be more akin to a veritable white hole, a cosmic fountain spewing forth its primal stuff incomprehensibly and inexhaustibly. In either case, but especially the latter, the possibility arises of an environment or a database that is both irresistible and fatal to cybernetics: something which is more than it can chew, yet which it can’t help but bite off.

Subscribe now

In other words, cybernetics, which is supposed to do the “eating,” might at a certain degree of development or success encounter, or even create, conditions where it becomes more like the meal than the stomach. What sort of food would it become? Perhaps the equivalent of what is theorized to happen to people and things that encroach too closely on a black hole to stay outside its grasp: spaghettification, the irreversible distortion of a thing beyond all recognition.

The compelling question, then, is what conditions or environments could in that sense cause cybernetics to collapse, or die, of spaghettification. SpaceX’s recent Super Bowl ad boasts of its achievements and future potential with a voiceover taken from Arthur C. Clarke’s famous 1964 interview for the BBC’s Horizon. “The only thing we can be sure of,” Clarke intones, “is that the future will be absolutely fantastic. But let us remember this: one of the very few rules we can be sure of about the future is that whatever actually happens will be totally unimaginable to us today.” Nevertheless, Clarke does venture a very specific prediction, the final note SpaceX leaves the viewer with: “The world will be reduced to a single point.”

It is, of course, instantly very difficult to say how cybernetics could possibly function, or even exist, in an environment reduced to a single point. One could fudge things and interpret Clarke to mean what he said only in a relative way, such that cybernetics becomes all-the-way-down in the mystic sense of 18th-century polymath Emanuel Swedenborg which so fascinated Ralph Waldo Emerson: “The unities of each organ are so many little organs, homogeneous with their compound: the unities of the tongue are little tongues; those of the stomach, little stomachs; those of the heart are little hearts.” But even if we can now readily envision a form of cybernetic governance that penetrates down to recapitulate itself at the neurological or mitochondrial level of the human organism, we can’t wave away the problem of cybernetic progress communicatively concentrating or collapsing spacetime on Earth down to a scale and size far too small for incarnate humans to live (or even appear) in.

If indeed cybernetics entails the human use of human beings, in ways that ensure they remain in our mutual experience recognizably human, then we must accept that, as cybernetics progresses by increasing the volume, density, and speed of communication, it accelerates toward an event horizon—a tipping point at which its human substrate buckles and denatures under the pressure and pull of the system.

In that sense, the “death” of cybernetics—the spaghettification of cybernetic systems of governance—must be seen to accompany the technological transformation by human beings of the world into an environment that forecloses the spacetime incarnate people need in order to live properly.

Cybernetics demands environmental conditions sufficient for the ongoing existence of spacetime in which individual and relational human activity—such as the experience of communication and the feedback process that results—can play out more or less stably, cumulatively, and conducively to communicative purification, not corruption. Too much cybernetics—which ceases to be itself if living organisms have to be squeezed out of the equation—is an existential risk to the foundations of cybernetics. Wiener was unable to extricate himself and cybernetics from the pull of the white hole of singularity that is the immeasurably unlimited and total density and speed of information. The irony is bracing. In The Human Use of Human Beings, Wiener warned of an impending and catastrophic emergency where the application of technology to governance would result in a totalitarianism fatal to the human life worth living.

Superficially, he described this totalitarianism in the political terms of fascism and communism. But more deeply, in chapters that were expunged after the book’s first edition at the behest of Houghton Mifflin and editor Jason Epstein, Wiener associated the nightmarish totalitarianism of misbegotten technological governance with theocracy.

In the suppressed “Voices of Rigidity” chapter, Wiener links the Jesuits and the Communists as adherents to militant and totalitarian ecclesiologies. “In that moment in which the successor of Saint Peter compromises with Caesar, becomes an official of a public order which is parallel to that of the State, and shares the functions of the State by a mutual agreement, the Church becomes an organization offering advancement and a personal career. If such an organization is to survive, it must take its duty of survival very seriously, and the path to totalitarianism is short indeed.”

Likewise, “the Communists do not consider themselves servants of God, but they do consider themselves servants of a cause to which they are completely devoted, and which they believe to contain the ultimate duty of humanity,” Wiener laments. “There was a time when many of us hoped that the professed pro-scientific bias of this power group … would lead to … a less exclusive and more universal attitude to the world in general, to other countries, and to science.” Alas, “the Soviet policy has become progressively more doctrinaire and restrictive, and less tolerant and humane. In fact, the Soviet State has become a closed and rigid institution closely resembling the historical Church, even in the minute details of its attitude and organization.”

Wiener must have been far from ignorant of the centrality to Soviet scientific dogma of the “New Soviet Man,” the collectivist, technologically created Übermensch with which the communists sought to destroy once and for all the Christian Church and the visible and invisible bases of its anthropology. What Wiener was afraid of was not a merely political or ideological totalitarian regime dominating the world but one which forged cyborgs on a theocratic basis of the complete and compulsory unity and uniformity of all beings through the planetary unification of church and state.

Wiener’s proposed alternative and escape was a cybernetic system of free and open global communication among all persons. In a second suppressed chapter, “The Individual as the Word,” Wiener insists that technological advancement has revealed that “the mental identity necessary for the Church’s view of the individuality of the soul certainly does not exist in any absolute sense which would be acceptable to the Church.” To the contrary, “there is no fundamental absolute line between the types of transmission which we can use for sending a telegram from country to country and the types of transmission which at least are theoretically possible for a living organism such as a human being.” The devout, Wiener claims, are now wrong that the soul is the mark of personal identity; Descartes, even, is now incorrect that thinking alone proves the thinker’s existence. To communicate is to be, says cybernetics, and to be is to communicate.

People in this unprecedented position are tempted either to compel totalitarian unity through incorporeal communication or to accept such forced unification—either to hyperactively foster cybernetic theocracy or hyperpassively accept it. As human individuality is now, Wiener avows, so clearly linguistic, the planetary escape from cybertheocracy must permit, encourage, and structure—in cybernetic fashion—the global remote communion of human individuals in the feedback loop of communication. In this way, he concludes, humanity learns recursively, a salutary mechanism not only from the standpoint of human flourishing but from the standpoint of our best resources to ensure life prevails against entropy.

As we have learned, unfortunately, the dream of global communication among humans whose identity is linguistic and cybernetic led on to the reality of an increasingly fanatical and desperate attachment to the white hole of comprehensive information at critical mass and terminal velocity. “When there is no more spectacle, no more stage, no more theater, no more illusions,” as Jean Baudrillard foretold 40 years ago, “when everything becomes immediately transparent, visible, exposed in the raw and inexorable light of information and communication … we no longer partake of the drama of alienation, but are in the ecstasy of communication.”

A few years later, Baudrillard confronted the question of what could possibly follow this orgiastic singularity. “What,” he morbidly joked, “are you doing after the orgy?” Today, we have begun to recognize that what follows the mass consumption of the ecstasy of communication is the addiction to what we might call the fentanyl of communication: a dysphoric state of utter dependency on a fatal practice we can’t talk ourselves out of exacerbating. Backrooms dramatically illustrates the slide into this end-state. It’s a horror film, but what is the horrific and who are the horrified? One answer is the psychological hellscape and the heroine psychiatrist who stands in emotionally for the audience.

A better answer, I think, is that the cyberneticians of the film’s fictional Async corporation are horrified. At first glance, they are horrified that their work has unintentionally produced an inconceivably vast amount of anti-cybernetic matter. The titular Backrooms, according to the film’s lore, is a seemingly infinitely expanding and uncontrollable labyrinth made up our mutated memories of all things, accidentally opened to access by Async’s effort to solve the scarcity of living-space. But the film makes clear that the true horror for Async is their loss of hope that the Backrooms can be communicated about in a way that enables them to perform cybernetics by reestablishing systemic control. The very first thing anyone wants to do on returning from the Backrooms is to talk about it, to cognitively capture and linguistically optimize it in an ecstasy of communication. But the more it is shared in speech, the more it spreads and invades our world. It is precisely through the irresistibility of communication under its unintegrated spell that the Backrooms further colonizes, controls, and captures us. The resulting addiction is death for us, and death for cybernetics.

In other words, what we have learned since Wiener is that nothing internal and inherent to cybernetics itself seems capable of preventing even an otherwise optimal cybernetic system from tipping fatally toward either black-hole (entropy) or white-hole (singularity) failure modes. But we have also learned that cybernetic systems optimized against the risk of both modes fall toward a third failure mode of homeostatic stasis, in which new learning ends, the closed feedback loop regresses, and no new life is generated or allowed in.

If therefore both entropy and singularity are not cybernetics, but neither is stasis, we are returned to the question of how the new-yet-not-fatal enters from outside cybernetic systems into them. Emerson stresses that Swedenborg’s philosophy of fractal microcosms “is a key to his theology also. ‘Man is a kind of very minute heaven, corresponding to the world of spirits and to heaven. Every particular idea of man, and every affection, yea, every smallest part of his affection, is an image and effigy of him. A spirit may be known from only a single thought. God is the grand man.’”

Yet it would be actually quite odd for God—or whatever one might call the source of new-yet-not-fatal organized life entering from outside cybernetic systems—to exist in ontological form radically at odds with cybernetics. For the self-insufficiency of cybernetics even in its most otherwise-optimal form does not indicate that life is not fundamentally communicative and relational in its character, but rather that cybernetics requires an input of such life from outside any cybernetic system in order for that system not to enclose itself and fail.

Science fiction and horror sometimes meet in visions of outside invaders that are not just destructive to life but are so in virtue of their not being organically or spiritually alive.

No God worthy of the name could be such a destroyer, and indeed the Triune God is recognized by the faithful precisely as the exemplar of life-givingly linguistic, communicative, and relational Being. Cybernetics must search in vain for any internal substitute for a commensurate source and measure of optimality in cybernetic architecture and maintenance. Believing churches to be fatally susceptible to totalitarianism, Wiener concluded that cybernetic systems must never be allowed to become churches. Now that they have taken on such a religious cast, amid the institutionalized worship of “know-how” that Wiener warned against, technologists and philosophers are thrown back upon the question of who can be trusted to sustain a spiritual order strong enough to resist the totalitarian temptation.

In the height of the Dust Bowl, on March 21, 1935, a soil scientist named Hugh Bennett testified before Congress. He was attempting to save his beloved Soil Erosion Service (SES) which would run out of funding in just a few months. As Bennett was concluding his testimony, an omen of biblical proportions swept across Washington. A dust storm which Bennett’s team had predicted darkened the sky over the Capitol building. As one biographer eloquently described:

[The congressmen] gathered at a window. The dust storm for which Hugh Bennett had been waiting rolled in like a vast steel-town pall, thick and repulsive. The skies took on a copper color. The sun went into hiding. The air became heavy with grit. Government’s most spectacular showman had laid the stage well. All day, step by step, he had built his drama, paced it slowly, risked possible failure with his interminable reports, while he prayed for Nature to hurry up a proper denouement. For once, Nature cooperated generously.

Subscribe now

In no small part because of Bennett’s impressive combination of agronomy and showmanship, Congress soon passed the Soil Conservation Act of 1935, establishing a permanent Soil Conservation Service under the U.S. Department of Agriculture (USDA) with Dr. Bennett at the helm. Over 90 years later, the renamed Natural Resources Conservation Service (NRCS) remains as one of the most unique and innovative federal programs that most people have never heard of.

Most federal programs born in a crisis don’t outlast the crisis by much; NRCS outlasted the Dust Bowl by nine decades because it built something durable. It built a working method for getting private landowners to change how they treat their own land without mandates. Much of NRCS’s success comes from being a remarkably de-federalized federal program that focuses on developing “new tools, technologies, and strategies to support next-generation conservation efforts on working lands and develop market-based solutions to resource challenges.”

As with any federal program, the NRCS’s record is mixed. The same agency that built its legitimacy on voluntary, locally run conservation districts, also spent the 1930s utilizing USDA aerial photography to impose a single federal grid on how the landscape would be measured and understood. A survey plane can count acres, but it cannot tell a well-drained loam from an eroding clay hillside or ask the owner about their preferred land use. The 1994 renaming came about because of a Clinton-era reorganization built to cut USDA’s personnel costs by billions and sold under Vice President Gore’s promise to reinvent government. The mission-expansion was attached after the fact. Yet, farmers appear to trust the NRCS, the agency that talked them into conservation, far more than they trust other federal programs. And the NRCS’s successes are arguably some of the most impactful both in terms of conservation and in terms of agricultural innovation.

That history is worth examining. Today, NRCS runs one of the few federal programs built to develop, evaluate, fund, and deploy actual agricultural innovation. What NRCS got right, what it got wrong, and what the agency should do better are questions that go well beyond agricultural policy. NRCS is a rare federal case study, nine decades long, in what happens when Washington tries to change behavior without simply ordering it. Tech, telecom, and AI policy are all running some version of the same experiment right now, mostly without knowing it.

What NRCS Got Right

Bennett saved the SES in 1935, but it survived because of what happened after. In 1937, Roosevelt sent every state governor model policy for soil conservation districts. These conservation districts would be governed by boards of farmer-landowners and formed only by a referendum of the farmers who would live under them. For its part, the federal government promised to supply scientists and surveyors to work in any district that asked for them. Driven by hometown pride for their man in Washington, Bennett’s home of Anson County, North Carolina was the first district to organize and, by the end of the 1940s, every state in the Union had passed some version of Roosevelt’s model law. Organized around the National Association of Conservation Districts, nearly 3,000 districts—”almost one in every county”—still operate on the terms laid out in the 1940s. This unique, localized arrangement gave the Soil Conservation Service a permanent local constituency that considered the agency its own.

The modern NRCS still operates under that model. A farmer in any one of those 3,000 districts can walk into a county field office and get free help from a conservation planner who assesses the land and drafts a plan around the owner’s goals, whether or not any money ever changes hands. When money does change hands, it moves through cost-share under the Environmental Quality Incentives Program (EQIP), the USDA’s flagship working-lands conservation program, funded through congress in the Farm Bill. NRCS co-invests in roughly 200 approved practices. Between 2009 and 2018 it obligated over $13.8 billion through cost-share contracts with more than 384,000 farmers and ranchers covering over 115 million acres.

The approved practices under EQIP include everything from cover crops and hoop houses to prescribed burns and brush management, but practice availability varies by state. For example, irrigation land leveling—regrading cropland so surface water flows uniformly across a field—is used almost exclusively for flood-irrigated crops such as rice, and so is prominent in a state like Arkansas which is the nation’s top rice producer. In contrast, a state like Vermont that primarily produces dairy and timber has practically no need for irrigation land leveling. Ranking and availability is notably localized as working groups of area conservationists and producers choose which resource problems matter most in their community. For example in Nebraska, those specified local priorities carry a quarter of an application’s ranking points.

That system has one structural weakness; EQIP can only pay for practices already on its approved list. A cost-sharing schedule requires a settled practice standard, and a settled standard requires evidence and approval. Somebody has to generate that evidence on real farms before a new technique can enter the catalog. That is the job of the Conservation Innovation Grants program (CIG), created in the 2002 Farm Bill and funded out of EQIP’s appropriation. CIG runs competitions for projects that field-test conservation approaches too new to have a practice standard, such as precision nutrient tools, tech-enabled grazing arrangements, and market mechanisms like water-quality trading.

Two design choices make CIG better than the average federal innovation program. First, grantees have skin in the game since awards require matching funds, so a university, co-op, or company that wins CIG money is investing alongside the federal government. Second, since 2018, the program has run “On-Farm Conservation Innovation Trials” which pay producers to test practices on working land rather than confining experiments to research stations. The output of CIG is built to produce is a practice standard. Few federal research programs can say what happens to their successes. For CIG, the successes are visible in farms and fields across the country.

What NRCS Got Wrong

Successful innovation for a conservation agency means one specific thing: the speed at which a technique that works on a test plot becomes a practice a farmer in any county can get help adopting. By that measure, the agency built to spread new farming methods has become slow at acquiring them. The bottleneck is the practice catalog. Everything NRCS delivers is directly downstream from that catalog and the road into the catalog is long, narrow, and lightly funded.

First, as with all government programs, there’s the money. The CIG program is the agency’s designated engine for finding new practices and, since 2004, it has invested $528 million across 923 projects. Averaged over two decades, that is roughly $24 million a year. Compare that to the total $2.655 billion Congress gave USDA for EQIP in FY26 and the difference is striking. EQIP spends more in an average week than CIG spends in a year. The agency devotes over ninety-eight cents of every dollar on deploying the existing catalog and pennies on expanding it.

The timeline may be worse than the budget. A CIG project typically runs one to three years, but the pipeline it feeds runs longer. Any promising technique enters the catalog through an Interim Conservation Practice Standard. To become an interim standard, a request is first vetted by a State Technical Guide Committee, then by a national subject matter expert, followed by a three-year evaluation period. At the end of the evaluation period, the practice is evaluated and recommended to either become a new national standard, be folded into an existing standard, or discontinued. States can request, and often receive, extensions. A practice that survives all of that still awaits a Federal Register comment cycle and state-by-state adoption, for which states get up to a year after a national notice takes effect. Add it up and a technique proven on real farms in 2026 reaches the full national catalog around 2031. Five growing seasons is a long time for a proven idea to sit in administrative escrow. To make matters worse, the agency publishes no deadline for the front half of the process at all.

None of these are failures of mission or of the people in the field offices. They are failures of self-conception. Bennett’s agency in the 1930s was intended to protect the soil and help end the Dust Bowl, which it did quite well. When NRCS’s mission was expanded, the institution kept that Hippocratic mindset to first do no harm to the soil. But, in doing so, it has unintentionally delayed the very kinds of innovative research and development that saved our nation’s farms and grasslands.

This is not meant to be a sweeping indictment of NRCS or the CIG program. It is meant to demonstrate that institutional caution is a design variable. The agency that field-tested terracing on demonstration farms in 1936 and the agency that takes five years to approve an interim practice standard are the same institution running at different settings. Congress and USDA set those settings. Either can reset them. The caution protecting the catalog was earned honestly, over ninety years of not wrecking anyone’s farm. The problem is that an agency built to spread proven practices cannot afford a definition of “proven” that runs half a decade behind the farmers it serves.

Fortunately, the problems of NRCS are ones of culture, not of structure. NRCS still produces real results for both farmers and the environment and, when we compare NRCS’s innovation cycle to other federal programs, it is clear that while a five year timeline can and should be improved, CIG still ranks among the most innovation-forward federal programs around. What also sets CIG apart is that its successes are tangible, whether it’s GPS tracking collars for more efficient cattle grazing rotation or solar powered irrigation systems for water-scarce tribal lands. With some prodding from Congress or the Secretary of Agriculture, NRCS could easily dedicate more resources to CIG and see real results for American farmers sooner. Most other federal programs focused on innovation are not so fortunate as to have a simple solution.

The Transferable Lesson — What Other Policy Domains Should Steal

The reason CIG has a simple solution and most federal innovation programs do not is, quite simply, that CIG was built on top of an efficient innovation delivery system. When a CIG project proves a practice works, it hands the result to a ninety-year-old machine of field offices in nearly every county and district boards that were designed to move techniques onto private land. The innovation and distribution were joined from the start. Underfund the first and you get a slow trickle of new practices, which is CIG’s problem. Underfund or pass innovation off to an incompetent bureaucracy and you get a mountain of research nobody uses, which is exactly what the Department of Defense, Defense Innovation Board found in their review of the Small Business Innovation Research (SBIR) program. Only 16 percent of awardees ever transitioned to a fully deployed innovation.

The SBIR program is the clearest case of an innovation program that never built a functional deployment system. Under SBIR, every federal agency with a large research budget must route a fixed slice of its budget towards small firms in phases. The first phase is a small grant to test a concept, then comes a larger grant to build a prototype, and then, in theory, a commercial contract. The gap between those last two steps is where SBIR breaks down. The stretch between a finished prototype and a production contract is known in the field as the “valley of death,” and SBIR has no bridge across it, because, unlike NRCS, it has no standing system that was built to adopt what it funds. A prototype proven on SBIR money has to find its own program office, its own contract vehicle, and its own customer. Most prototypes can’t traverse the valley of death of the Small Business Administration does little to help them navigate. FAI Fellow, Lars Schönander has provided helpful deep-dives into how to fix SBIR.

The National Institutes of Health’s (NIH) innovation-to-deployment cycle show the timeline problem at a scale that makes CIG’s five years look brisk. It is commonly estimated that roughly seventeen years pass between a research finding and its arrival in routine clinical practice. NIH’s response to this realization was to stand up a translational-science center and bring institutions under it a national network whose entire job is to shorten the trip from bench to bedside. That is the right instinct but it remains to be seen how effective it will be. It is also an admission that the world’s largest biomedical research funder spent decades discovering things faster than it could deliver them.

The National Science Foundation is the encouraging case, partially because it has already stolen two of the lessons NRCS teaches. NSF caps how often an applicant may apply to two pitch submissions a year, and allows no more than three submissions regardless of topic or timeline. And, through its I-Corps program, NSF does something CIG’s on-farm trials mastered: it forces the inventor out of the lab to test whether anyone actually wants the thing before more money follows. Neither fix is complete, but together they show the NRCS principles are portable.

Artificial intelligence governance, so far, runs almost entirely on the voluntary, standard-setting model which is constituted entirely of commitments and technical frameworks companies voluntarily choose to make and follow. This is comparable to the conservation-district model in embryo, and it can work, but the NRCS record is emphatic about what makes voluntary participation hold: the participants have to own the standard, and a credible technical body has to help them meet it. The districts survived because farmers voted them into being and could call on federal expertise when they asked for it. The equivalent for AI is a standards body that firms treat as theirs and a technical-assistance function that helps them comply rather than merely audits whether they did. Perhaps we can call it the Center for AI Standards and Innovation (CAISI) and have it headquartered at the National Institute of Standards and Technology.

Telecom has the opposite problem, but NRCS’s lessons hold. Broadband innovation policy has run mostly on the subsidy route of one-shot grants to carriers with attached conditions. This is precisely the payment-for-compliance approach that other USDA direct-payment-programs use, and which buy exactly as much cooperation as is paid for. NRCS’s districts outlasted those programs because they were institutions, not transactions. A broadband strategy built on some standing local body with its own stake in whether the network gets used would be a strange design by current telecom standards, and it is the only design in this whole history that has lasted ninety years.

The thread through all of it is that invention is the easy half. Bennett’s dust storm bought a budget, but it was the deference to local governance and deployment that has upheld NRCS. SBIR funds prototypes and loses them in the valley of death. NIH funds discoveries and waits seventeen years. Both are missing a standing system, owned by the people it serves, for turning a proven idea into a practice on the ground. The domains now reaching for federal innovation policy, from AI to broadband to whatever comes after, could fund a decade of grants and still fail for want of deployment. The lesson is clear: build the road before the grant, tie the grantee’s own stake to the outcome, and never let the budget for discovery grow so far ahead of the means of delivery that good ideas sit while the problems they solve get worse.

Special thanks to Cody Matteson for his help with this piece.

Once you head off US-183, the drive into Proto-Town is quintessentially central Texas. Narrow two-lane roads, rolling hills, the smell of mesquite and cow pastures. You could be mistaken for having entered into the first verse of a George Strait song.

As I drove toward the heart of the development on an overcast day in March, I passed by a sign that reads: “PLEASE DRIVE SLOW. ROBOTS AT WORK.” On the other side of the property, an exact copy of the guard post Los Alamos had at its founding flits by. The significance of each is subtle, but purposeful. If the technological flourishing that occurred during World War II needed guarding, the same animating spirit needs something else today: freedom.

Staking Their Claim

Proto-Town is the brainchild of Joshua Farahzad and Merle Nye: two men who had crossed paths years ago and decided to make a bet on the future. The friends met while studying at Duke University, both later spending time at various startups focused on aerospace and energy, respectively. After discovering that testing small missiles, even for research purposes, was not exactly something looked kindly upon in zoned residential neighborhoods, they sought to settle a land where such activities were not only allowed, but encouraged. The first step included their now-famous red school bus, which was driven across America, spreading the good word of building a new city filled with the frontier spirit of America, where residents have an inherent right to experiment. Fittingly, a hollowed out missile hangs from the ceiling.

Amidst their search for land to settle, Farahzad and Nye found Lockhart-local John Cyrier, a small business owner and four-term Republican Texas state representative, who, serendipitously, lived just down the road from 1,200-acres of land near Lockhart, Texas. Cyrier had been involved in earlier plans for a proposed Micron semiconductor factory site to be constructed on this parcel with support from the CHIPS Act, federal legislation which provided subsidies and investment to spur domestic semiconductor fabrication. Micron instead chose a site in New York state, due in part to a more lavish economic development package (this project is now being held up by environmental lawsuits and review). Farahzad and Nye’s project took advantage of this available land and gained a partner who was critical for obtaining the local backing they needed. The trio set to work establishing the legal structure: incorporating Proto-Town and ushering their ideas into life.

After purchasing a parcel of land 2 years ago, Proto-Town’s founders set out to homestead their way to a settlement. Their first year was spent sleeping on air mattresses in a school bus. It was that vessel which Farahzad and Nye drove around the country on the initial tour to pitch and recruit what would become Proto-Town. They hopped on the bus and never got off.

To Farahzad and Nye, local buy-in has been critical from the beginning. The prospect of test-robots may not be seen as particularly neighborly for some. But, through conversations and negotiations with local municipal leaders, residents, and government actors, the duo made the case for Proto-Town and its value to the local community, as well as the state of Texas, and the nation. This resulted in something striking: regulatory arrangements that gave control of zoning and citing to Proto-Town’s founders. This is what makes Proto-Town more than an idea. After all, it’s permitting and land-use agreements, not capital or technology, that holds America back from building. Provenance over their own area means Proto-Town can decide what they want to build.

Today, the site is a new frontier for founders and hardware manufacturers. They congregate to the secluded “Freedom City” to experiment, iterate, and scale production of projects that attempt to solve some of America’s hardest problems. While software developers and those working in the world of bits can huddle together in a group house in the Mission, iterative development in the world of atoms is often relegated to wizened old men on university campuses, industrial parks, national labs, and recently, a growing community in El Segundo. Proto-Town was built to fill this void: a space for ambitious companies to build, test, and deploy novel technologies today.

First Contact

As of today, there are more than twenty people living on the property. Housed in a mix of re-purposed shipping containers, man-camp trailers, and, soon, 3D-printed homes, Proto-Town’s come a long way since its entire population was sleeping in a bus. Now, the town’s clustered around a ranch with a few structures for lodging and meals. The townspeople congregate for breakfast, lunch, and dinner a universe away from the all-you-can-eat cafeterias that fuel the worker-bees in Mountain View or Menlo Park. These brief reprieves from work contribute to the camaraderie of the place and bind these disparate entrepreneurs and engineers together.

As the founders see it, Proto-Town’s residents are laying the foundation for a revival of the American frontier spirit in the 21st century. In addition to its burgeoning industrial infrastructure, there are plans to build more housing (referred to as the “antfarm”), places of worship, and fields for agriculture. All of this development is fueled by technology and capacity native to Proto-Town.

There is an expressed intent to “dogfood” these physical services and tools: using and testing their software and services internally. Farahzad referred to this idea often, noting how the practice is quite popular within cohorts of the Y Combinator startup accelerator. It creates a supply of early users who can test products, identify problems, and accelerate iterative improvement. The difference in this instance is that the founders dogfooding each other sleep mere doors away, using one company’s autonomous construction vehicles to move earth so another can build rocket parts.

The companies that have decided to make their bet in central Texas have one thing in common: focusing on tackling the hardest product problems, building the goods and services necessary for modern American life. In addition to autonomous construction vehicles (Bedrock Robotics), there is a company building air conditioning systems that rely on a chemical reaction created through collecting solar energy (Atmos Thermal Systems), small modular nuclear reactors (Oklo), and a new form of centrifuge and desalination process that is being tested on brackish water from oil and gas extraction (Eden Technologies).

Meet the New Frontier, Same as the Old Frontier

Beyond the technological feats, Proto-Town represents an opportunity to re-connect with the lost experimental spirit of America. Just by wandering through the town, it is clear how this uniquely American conception of exploration and risk is inextricably linked to the physical systems being constructed and deployed. To some, it may seem backwards that the next great American city could be built among longhorns and wildflowers. After all, you see the future first in San Francisco or New York. But to these men, it is precisely this location and this land that is the place to build anew.

The physical lock-in and commitment to the land means they must iterate where they are, rising and falling on this parcel of land. While walking with Farahzad, I asked whether this was simply a concept: a project to learn from and, if too tough, move on from. He replied confidently: “There’s no escape. We have made our bet here. This is not a ‘we can pick up and go back to SF’ project. This project will live or die in Lockhart.” The founders created a model for a special economic zone, sure, but it is not just a set of legal permissions or access to resources that make Proto-Town special. It is the vision that this place, specifically, is where the young men building rockets and housing nuclear reactors will bear witness to the future being created.

Outside the main house, there is another monument that greets new entrants. It is an old oil derrick, with various telephone poles affixed to the top. A few yards to its right is a black-and-white billboard of Robert Oppenheimer puffing on his famous pipe. The monument is a piece created by More Monuments, whose goal is to build monuments to “define a new era of American artistry and craftsmanship.”

This all-too-fitting tribute is a testament to the work that has already been done at Proto-Town, and a marker of the dreams of the work that will continue for years to come. Amid America’s 250th year, there is no better time to resurrect the idea of America as an experiment, as a place to build something new. Proto-Town is doing its part, revitalizing that youthful American spirit one autonomous digger at a time.

In the mid-1990s, a Finnish pulp-mill-turned-consumer-electronics-company agreed to let its main partner, Texas Instruments, combine its digital signal processor with technology from an obscure British company named Arm. The result of this novel collaboration was the iconic Nokia 6100 series of phones which went on to sell nearly 50 million units. In addition to being the first cell phones to include built-in video games like Snake and Logic, the Nokia 6110 was also the first mobile device to use the Arm architecture which would become the backbone of compute for decades to come.

Arm itself is a rather bland company with a unique business model. Sometimes referred to as “the Switzerland of the electronics industry,” Arm is a semiconductor company that doesn’t make semiconductors. Instead, the company designs processor architectures and licenses them. Its $4 billion in annual revenues come from royalties on every chip that is made which uses its designs or speaks to its instruction set.

Subscribe now

Arm is everywhere. For almost thirty years, everything mobile—from the 6100 to the first iPhone to modern phones running on M4 and Snapdragon chips—has used Arm’s licenses. Nearly every modern car has a suite of Arm chips. Practically all 5G infrastructure relies on Arm in some way. The hyperscalers now run significant portions of their infrastructure on custom Arm server chips. Most laptops and tablets run on chips that rely on Arm licenses. Smart thermostats, industrial sensors, medical devices, point-of-sale terminals, smart meters, security cameras, and billions of other devices that most people don’t even think about as computers utilize Arm’s microcontroller line. As the company itself puts it, “Arm powers the compute foundation of modern life, with more than 350 billion Arm-based chips shipped to date.”

Those staggering numbers are evidence of more than just a successful business. Arm has become an essential part of the compute ecosystem because it has remained neutral. As CEO Rene Haas told The Verge in 2022:

We really try to stay as neutral as possible. … We are involved in the ecosystem of ecosystems. If you start at the lowest level of the semiconductor chain — GlobalFoundries, Samsung, TSMC, Intel, all the people who build chips — you have to work with all of them. We have to make sure that our technology is going to be able to be built on every semiconductor process in the world, which requires investment across all of those partners.

Since the 1990s, Arm’s neutrality has been a fact of the market that most took for granted. The reality is that every company which builds on Arm’s architecture, including Apple, Google, Amazon, Qualcomm, Samsung, and Nvidia, does so under a license that Arm can revoke. In October 2024, that reality became clear when Arm sent Qualcomm a termination notice, threatening to end the architectural license that Qualcomm has used for thirty years to design and ship chips. Qualcomm won that battle, but the war is far from over.

That first strike at Qualcomm demonstrated that the instruction set which runs modern civilization is not a public good. It is owned by a foreign company that is beholden to foreign shareholders. It is also a chokepoint in the tech stack that its owners have demonstrated a willingness to leverage against competitors. We’ve seen this problem before. The last time it was Intel and it took decades and billions in investment to route around. Policymakers in Washington have developed strategies for semiconductors, AI, submarine cables, spectrum, and more. What they haven’t developed yet, and increasingly need, is a strategy to secure the layer that sits between all of them.

The Invisible Layer

An instruction set architecture (ISA) is the translation layer between software and silicon. It is the list of fundamental operations a processor knows how to perform (e.g. add these two numbers, fetch this value from memory, jump to that application) and the precise way a program must phrase those requests. Everything above it is written assuming the processor underneath will understand those instructions. Everything below it exists to fulfill those instructions. The ISA is the connector between two halves of computing.

The reason it matters so much is the reason it stays invisible; the instructions are incredibly difficult to alter once implemented. A program compiled for one instruction set will not run on a processor that speaks a different one. Switching architectures means recompiling everything and, where source code is missing or the instruction syntax is different, it means starting over from scratch. For example, Apple has now done this migration three times, migrating from Motorola 68k to PowerPC to Intel x86 and finally to its own Arm-based chips. Each time it has migrated, it has had to develop and deploy an emulation layer to carry the old software across. For most of the industry, the instruction set is chosen once and then lived with for decades. The choice is less like picking a vendor than like pouring a foundation.

This structure exists well beyond mobile phones, though phones are where Arm won most decisively. Anything that runs software must be compiled against some ISA and therefore linked, quietly and durably, to whoever controls that architecture.

Arm’s Business Model

Arm owns no factories and (until a few months ago) makes no chips. What Arm sells is the design and the language, the blueprints for processor cores and licenses to the ISA those cores implement. A customer pays an upfront fee for access and then a small royalty on every chip it builds using Arm’s technology while software developers are given free access to induce uptake.

This model was not the plan. Arm was founded in 1990 as a joint venture between Acorn Computers, Apple, and the chip firm VLSI Technology. Apple, which took a 43 percent stake in the company, wanted a low-power processor for a handheld device called the Newton. The Newton flopped and Arm pivoted. Co-founder Robin Saxby realized that licensing Arm’s designs to everyone and collecting a small cut whenever any of them succeeded had the potential to be far more lucrative, and he was right.

Here the analysis must digress because there are two kinds of Arm licenses and the difference between them runs underneath everything that follows.

The first is the standard technology license, sometimes called a TLA. Under it, a customer takes one of Arm’s ready-made cores and drops it into its own chip. This is akin to buying a prefab house design: you purchase a license to the blueprints for a house that is already designed, build it on a plot of land you own, maybe change the paint color, and voilà you have a house without much hassle. This is the way that most of Arm’s licenses work and is the low effort path since anyone can get a proven core and a quick route to market. In exchange you accept Arm’s engineering choices and pay them a royalty.

The second is an architectural license (ALA) under which a licensee uses only the instruction set and designs its own cores from the ground up. If we continue with the home construction metaphor, an ALA is like being a licensed architect who designs a house however they like so long as it’s built to the same building codes so that it connects to the same water, power, and sewer lines as every other house in the city.1 This is the hard, expensive path, and the club is small, composed of Apple, Qualcomm, Amazon, and a handful of others. Designing a high-performance core from scratch costs years and hundreds of engineers, but it is the only way to build something genuinely better than what every competitor is buying off the shelf. The fact that Mac laptops have outrun the performance of Intel on a fraction of the power comes from the fact that Apple licenses Arm’s language and then builds cores nobody else has.

The instruction sets became a trap in the ordinary way that standards do. That compact 16-bit ISA in the old Nokia chip exists because customers building memory-constrained devices needed code that took up less space and Arm developed the architecture to give it to them. Each such extension made Arm chips fit more devices, and each new device added more software written using Arm. The more software written using Arm made the next chip company’s decision to use a different ISA a little more painful and expensive. Lock-in of this kind accrues, one reasonable accommodation at a time, until the accommodations are load-bearing.

Subscribe now

The End of Neutrality

For most of its life Arm sold one thing above all: neutrality. Arm CEO Haas has consistently upheld that the company works with every chipmaker and every foundry, competes with none of its customers, and can therefore be trusted with the one decision a customer can never easily reverse. When the thing you are choosing is a foundation you will build a decade of products on, the supplier’s neutrality is essential. A licensor that might someday become a competitor is a dangerous thing to depend on.

That promise has eroded over the past ten years. In 2016, the Japanese conglomerate SoftBank acquired Arm for roughly $32 billion dollars and took it private. SoftBank later tried to sell Arm to Nvidia for forty billion dollars but federal regulators stepped in. Because Arm is “the world’s most popular computing platform,” as Nvidia CEO Jensen Huang put it, the Federal Trade Commission concluded that a combined Nvidia-Arm would have “both the ability and incentive to use its control of Arm to diminish competition by undermining Nvidia’s rivals.”

Undeterred, SoftBank took the company public again in 2023, while retaining about 87 percent of the shares. As a near-total owner that paid a premium and watched a profitable exit collapse, SoftBank wants growth and Arm’s current business model is slow-and-steady. The faster way for SoftBank to make its money back is to capture more of the value in each chip, and the fastest way of all is to stop selling only the blueprint and start selling the chips.

So Arm did. Reuters reported in 2024 that the company had drawn up plans to raise prices by as much as 300 percent and had discussed designing its own chips to compete with its biggest customers. By early 2025 the Financial Times reported that Arm would ship a finished server CPU with Meta as a launch customer. And in March, Arm did something it hadn’t done in thirty-five years: it announced its own silicon. The chip, called the Arm AGI CPU and built with Meta on TSMC’s most advanced process, is a complete data-center processor Arm plans to sell directly to hyperscalers. “We are now in a new business for Arm to supply chips,” Haas said at the launch.

The Switzerland of the industry had entered the war it spent three decades promising to stay out of. Its first target in this new war was Qualcomm, a company with a long history of bad blood between itself and SoftBank.

This is the frame the Qualcomm lawsuit belongs inside. Treated as an isolated contract dispute, the case is a confusing tangle of license terms. Treated as a first strike against Arm’s smallest major competitor, it is much more clear. Qualcomm had acquired a startup called Nuvia in 2021 for $1.4 billion. Nuvia had been founded by veteran chip designers to build Arm-based cores for data centers, and it held its own architectural license. Qualcomm took Nuvia’s work and turned it into the Oryon cores that now power its laptop and phone chips. Arm sued, arguing that Nuvia’s license could not simply transfer to a new owner and that Qualcomm needed to stop using its instruction set and destroy the Oryon designs.

Notice what Nuvia designs chips for, what Qualcomm wants to focus more on, and what Arm’s new chip target is: data centers. The dispute Arm framed as a licensing technicality lines up perfectly with the market everyone is now stepping over each other to acquire a share of.

A neutral licensor does not want its customers to destroy lucrative chip designs. A competitor does.

After losing that first suit, Arm escalated. In October 2024 it served Qualcomm with a 60-day notice that it would cancel its ALA. The threat was extraordinarily audacious and amounted to a death sentence for one of America’s preeminent chip companies. Arm later withdrew the termination notice and lost in court. A Delaware jury cleared Qualcomm in December 2024, and in September 2025 a federal judge entered final judgment for Qualcomm on the last remaining claim and denied Arm a new trial.

Regardless of the outcome in court, the industry took notice that Arm was suddenly willing to do what would have been unthinkable a decade ago. The new reality for any company reliant on Arm licenses is that a 60-day notice can hit your inbox at any time and the only remedy is litigation so expensive that most licensees couldn’t survive it.

One more fact belongs here, and it is the one Washington has not absorbed. Arm is a British company, owned by a Japanese conglomerate, listed on an American exchange, and sitting at the foundation of nearly all American computing. The United States has an industrial policy for chip fabrication that includes tens of billions of dollars aimed at manufacturing. It has no policy at all for whether and how the instruction sets those chips implement must be made available. As a matter of competition and even national security—the Department of War’s tech stack is totally reliant on tech companies and chips built on top of Arm’s ISA—the leading AI and chip companies need access to the most prominent instruction sets.

Everybody Builds Their Own

If you are one of the handful of companies large enough to hold an architectural license, the rational response to a licensor that has stopped being neutral is obvious: depend on it less and build your own. That is exactly what they are all doing.

Apple has the longest head start. Its M-series and A-series chips are cutting edge and it holds an ALA license that runs through 2040. Notably, Apple inked this deal just before Arm’s IPO, locking in long-term access precisely because the change in Arm’s ownership had made access feel less certain. Amazon, Google, Microsoft, and Nvidia also now design their own CPUs. Every one of those chips is built on the Arm ISA.

So the escape from reliance on Arm is only partial. Most major firms have escaped the engineering dependency and no longer need Arm to design their cores. They have not escaped the language dependency, and they cannot, because the real lock-in comes from the ISA. Tellingly, Amazon’s Graviton CPUs are not sold to anyone outside Amazon and even those fully in-house chips still run on Arm’s instruction set. Amazon still pays Arm a royalty for each one it makes.

Who Should Own the Instruction Set?

When the landlord of the language merely collected rent and competed with no one, building your own house on that land was safe. When the landlord starts building houses of its own and suing tenants whose houses it covets, every tenant’s investment looks suddenly more precarious.

All of which leaves one question: who should control the instruction set that nearly all the world’s computing depends on? There are three possible answers and industry has not yet settled on which is the right one.

The first answer is the status quo where a single non-American company owns the layer. This company has no obligation to license it to anyone on fair terms and the industry makes do with whatever that company decides. The case for this situation is the case for any property right: Arm built the architecture, Arm maintains it, and the incentive to keep investing depends on Arm capturing the returns. The trouble is that we have run this experiment before with the other dominant ISA. Intel owned x86 for decades, declined to license it broadly, and the result was a long stagnation in which the architecture advanced at the pace a single unpressured incumbent felt like advancing it. The escape from that, when it finally came, came from Arm and it took nearly forty years and billions of dollars. As both Intel and Arm demonstrate, unconstrained architectural control has a tendency to calcify and, this time, it is being calcified under the control of an increasingly unpredictable actor.

The second answer is regulation. There is a well-developed model for exactly this problem in the world of telecommunications standards. When a patent is essential to an industry-wide standard that everyone must implement, its holder is typically required to license it on “fair, reasonable, and non-discriminatory terms,” otherwise known as FRAND. The whole point of FRAND is to stop the owner of something everyone is forced to use from weaponizing that position. The complication is that the obligation traditionally only applies to standard-essential patents and Arm’s architecture is not a formal standard adopted by a standards body. It is private intellectual property, and so the existing FRAND system does not currently reach Arm. But the Qualcomm episode is precisely the argument for asking whether something like FRAND should apply to Arm’s instruction set, and some have already determined that the answer should be “yes.” Sens. Katie Britt and Adam Schiff recently introduced an amendment to the upcoming National Defense Authorization Act (NDAA) that would require all foreign-owned ISA providers to offer their instruction sets to US companies on FRAND terms. An ISA that nearly all of computing is required to implement is functionally an essential standard, regardless of what the paperwork says.

The third answer is the one gathering the most momentum: build an instruction set with no owner at all. RISC-V is an open, royalty-free architecture that emerged from a research project at UC Berkeley and is now governed by a nonprofit. There is no license to revoke, no royalty to raise, no company that can issue a 60-day notice. Anyone can build a RISC-V core, and the architecture cannot be bought, sued, or turned against the people who depend on it. The problem is that, currently, the cores and the software ecosystem around RISC-V are years behind any proprietary architecture.

China is building toward RISC-V as a matter of national survival because an open architecture is one critical technology that cannot be embargoed. Europe has put 240 million euros into a program called “Digital Autonomy with RISC-V in Europe” to build sovereign supercomputing chips on the open architecture in order to reduce the continent’s dependence on processors it does not control. Last year, both Meta and Qualcomm acquired companies with expertise in RISC-V-based AI inferencing chips while AMD, Google, Intel, Microsoft, and Nvidia are all active members and contributors to the organization developing the RISC-V standard. Each of these actors has realized that depending on an instruction set that someone else owns is a risk, and the only way to eliminate the risk entirely is for the layer to belong to no one. And yet, RISC-V is still far from a direct competitor to Arm and it will remain that way for at least the foreseeable future.

The Nokia 6110 divided the labor and the profits cleanly: Nokia made the phone, Texas Instruments made the radio, and Arm made the language. For thirty years that division held because the company that made the language had organized itself to be dependable. It is no longer organized that way. The instruction set was always the most permanent decision in computing, the one foundation that could not be re-poured. What the last few years have established is that the ground it sits on has an owner after all and that the people building on it have started, quietly and at enormous expense, to look for ground that does not.

In honor of the semiquincentennial anniversary of our great nation, we decided to break from our regularly scheduled programming. Since its founding (and, indeed, since well before our formal founding), ingenuity and innovativeness have been pillars of the American character. To highlight the history of this part of the American character, we present 25 Fun Facts About American Innovation.

Subscribe now

1) The U.S. Navy had ice cream ships! During World War II, the U.S. Navy commissioned dedicated refrigerated barges, dubbed "ice cream barges" by sailors, capable of producing up to 10 gallons of ice cream every seven minutes. The Navy viewed ice cream not as a luxury but as a strategic tool for troop morale, investing $1 million in one such barge alone. Ice cream was in such high demand onboard Navy ships in the South Pacific that when smaller vessels rescued downed Navy pilots, they would often barter with the pilot’s aircraft carrier. As Lt. Cmdr. Norman P. Stark later recalled: “After disembarking from the canvas bag [used for water rescues], I was greeted like a long lost brother. What I didn’t realize at the time, was that they weren’t seeing me, but what I was worth to them—10 gallons of ice cream.”

2) In the 1960s, the CIA surgically implanted a microphone, antenna, and battery into a live cat to eavesdrop on Soviet officials as part of a program called Acoustic Kitty. The first spy cat was deployed to listen in on two men outside the Soviet embassy in D.C. but was soon hit by a taxi and died. The program was quietly discontinued shortly thereafter.

3) The Toledo War between Michigan and Ohio was caused in-part by poorly-functioning mapping technology that placed the southern tip of Lake Michigan much further north than it actually was, causing Congress to draw overlapping borders for Ohio and Michigan. After months of conflict, Congress finally stepped in to resolve the issue with a compromise that granted Michigan the mineral-rich Upper Peninsula and admittance to the Union as the 26th state. Ohio got Toledo.

4) In 1942, after the large hemp production facilities in the Philippines and East Indies fell to the Japanese, the U.S. Department of Agriculture produced Hemp for Victory. This 14-minute propaganda film urged Kentucky and Wisconsin farmers to swap their maize fields for hemp, a plant the federal government had made illegal just five years earlier and now needed large quantities of to produce rope for the Navy. After the war, Washington not only banned hemp cultivation again but denied the film’s existence entirely until cannabis activists tracked down two VHS copies in 1989 and donated them to the Library of Congress.

5) Abraham Lincoln is the only president to have held a patent. It was granted in 1849 for an inflatable device meant to lift boats over sandbars.

6) In exchange for allowing the Atlantic Telegraph Company to use a US Navy vessel to lay the cable, James Buchanan received the first trans-atlantic telegraph message from Queen Victoria in 1858. The message read, “Her Majesty desires to congratulate the President upon the successful completion of the great international work, in which the Queen has taken the deepest interest.” It took over 16 hours to transmit this message and, less than a month later, that cable was rendered useless as high-voltage current burned through the poorly designed insulation.

7) Henry Ford was a bit of a soy boy. Because he distrusted the extraction-based industrial economy of mining and drilling, Ford thought agriculture could supply what industry needed instead. To prove his point, he built a car body from soy-based plastic that he’d whack with an axe to demonstrate its strength, and wore a full suit made of soy fiber to public events. He served soybean banquets at the 1934 World’s Fair where every course, including the coffee, was soy.

8) After the Clintons launched whitehouse.gov in 1994, somebody bought whitehouse.com and turned it into a porn site. Congress had to intervene to get him to stop, resulting in the Truth in Domain Names Act.

9) America had a strategic helium reserve buried under the Texas panhandle which was started in the 1920s because the military was convinced helium-filled airships would win future wars. In a controversial move, the federal government sold off the helium reserve in 2024 because it had become a money-losing liability that no longer served a clear strategic purpose.

10) Contrary to what some Europeans may believe, soccer would not exist without the USA. American chemist Charles Goodyear patented the process for vulcanizing rubber in 1836, and he later applied the technology to create the first modern soccer ball in 1855.

11) Always experimenting with electricity, Ben Franklin once killed a turkey for Christmas dinner using two Leyden jars (early capacitors), accidentally shocked himself, and was left traumatized. As he wrote: “I have lately made an Experiment in Electricity that I desire never to repeat. Two nights ago being about to kill a Turkey by the Shock from two large Glass Jarrs containing as much electrical fire as forty common Phials, I inadvertently took the whole thro' my own Arms and Body.”

12) Project Plowshare was an initiative launched by the Atomic Energy Commission in the 1950s to use nuclear explosions for excavation and one idea was to build a new sea-level canal through Nicaragua that would be called the “Pan-Atomic Canal.” Decades later, Newt Gingrich invoked the initiative as a potential solution for conflict in the Strait of Hormuz.

13) Thomas Jefferson invented a cylindrical wheel-based encryption device in the 1790s for secret diplomatic correspondence. It was forgotten and independently rediscovered a century later when the US Navy was designing the M-94 cipher device that saw use through WW2.

14) In 1994, a USAF lab at Wright-Patterson formally proposed a non-lethal chemical weapon that would make enemy soldiers irresistibly attracted to one another, causing unit cohesion to break down (colloquially known as the “gay bomb”). The $7.5 million proposal was rejected.

15) ENIAC, the world’s first programmable, general-purpose computer which drew power from a lab at the University of Pennsylvania, was so large and powerful that it reportedly made the lights flicker in Philadelphia whenever it was turned on.

16) During WW2, the US strapped tiny incendiary devices to thousands of bats, planning to release them over Japanese cities where they’d roost in wooden eaves before igniting. The program was canceled after the bats accidentally torched a US auxiliary airbase in New Mexico.

17) Everyone knows that Ben Franklin invented the lightning rod and bifocals, but he also invented the flexible urinary catheter. He never patented any of these, believing they should be freely shared.

18) One of the earliest uses of the ARPANET was to coordinate a sale of marijuana and the trade of Grateful Dead set lists between Stanford and MIT students. This is commonly cited as one of the first instances of e-commerce.

19) Thomas Edison’s first patent was for an electric vote-recording machine in 1869. Legislators rejected it because it counted votes too quickly.

20) Clinton’s NetDay Initiative, which sought to “wire the nation's schools for Internet access,” was primarily volunteer-driven. During just the pilot program in California, Silicon Valley companies donated around $27 million in funds and equipment and an estimated 20,000 volunteers connected 2,600 schools to the internet in a single day.

21) “The most beautiful woman in film,” Heddy Lamarr, was also into deep tech. In collaboration with Composer George Antheil, Lamarr developed and patented a radio technique known as frequency hopping that was intended to prevent Allied torpedo guidance signals from being intercepted by rapidly switching the radio frequency between transmitter and receiver. The Navy did not deploy this technology at the time, but it later became a crucial innovation for the development of Wi-Fi, Bluetooth, GPS, and modern military communications.

22) The term “computer bug” has entomological origins. When attempting to fix an issue with the Harvard Mark II computer in Cambridge, Massachusetts in 1947, researchers reported that a dead moth had been caught between relay contacts, shorting the system.

23) The U.S. Postal Service briefly experimented with rocket mail. In 1936, the Postal Service sent two rockets carrying mail from Greenwood Lake, New York to Hewlitt, New Jersey. When the rockets crashed on a frozen lake well outside of town, the postmaster of Hewlitt stayed true to his postal oath and carried the two bags of letters the rest of the way. Then, in 1959, the Postal Service launched a Regulus cruise missile carrying 3,000 letters from a submarine to a naval station in Florida. The Postmaster General declared: “Before man reaches the moon, mail will be delivered within hours from New York to California, to England, to India or to Australia by guided missiles."

24) It is common knowledge that the Department of War deployed multiple mobile Pizza Huts and Burger Kings to the Middle East during the Second Gulf War. What’s less well known is that during WW2, the Army shipped so much Coca-Cola to troops overseas that Eisenhower personally requested 10 portable Coke bottling plants and 6 million bottles of Coke per month for the front lines. In response, Coca-Cola sent 148 employees to Europe to run the operation, affectionately nicknamed the “Coca-Cola Colonels.” By the war's end, military service personnel had consumed over five billion bottles of Coke.

25) Al Gore invented the internet.

“Will the corporation,” asked artificial intelligence pioneer Herbert A. Simon in 1960, “be managed by machines?” Five years later came The Shape of Automation: For Men and Management, and five years after that, a curious rejoinder from a forgotten metaphysician: Frederick D. Wilhelmsen, known affectionately and universally as Fritz.

“Reasoning and Computers” (1970), a ten-page article in Thought—Fordham University’s quarterly review of culture and ideas—steered clear of the research agenda of Simon and company (into the epistemology of computational problem solving) to present a rival thesis. The processes and content of the human intellect, he insisted, entailed something much more and different than algorithms or even concepts alone could explain. While few public-facing academics of his time were more Catholic or indeed more Thomist than he, Wilhelmsen drew ably in his defense of human intellectual exceptionalism on the wryly apophatic assessment of David Hume that the artifacts of ideation could never deductively establish being – existence or is-ness. The affirmative judgment of what is—of what has being—expressed for Wilhelmsen the uniquely creative and existential culmination of the human intellect itself.

Being a good Thomist, however, Wilhelmsen reached at once for Aristotle to develop in his argument a demonstrative distinction between computer and human problem-solving. “The ‘quick wit’ of integration,” he promised—a faculty that Aristotle dubbed ἀγχίνοια—“is mankind’s best hope for mastering and ordering the new electronic revolution which otherwise threatens to drown us all in a sea of information.” (Should you sense an almost family resemblance in Wilhelmsen’s framing of the problem to that of Marshall McLuhan, the two indeed were friends and frequent correspondents; like McLuhan, Wilhelmsen was a willing coauthor, publishing The War in Man: Media and Machines (also 1970) and Telepolitics: The Politics of Neuronic Man (1972) with Jane Bret, a Montessori educator and mother of three who partnered with the Hungarian monks at the University of Dallas to co-found the city’s Cistercian Preparatory School.)

What we transliterate as agchinoia—and it is here where our story begins—is not a word of Aristotle’s own coinage. It is first found in the mouth of Socrates, who, in the Charmides, Plato has interrogate the titular lad on the true meaning of moderation or self-control (sophrosyne). In a scene not tonally out of place in some contemporary anime, the randy philosopher, it transpires, had had his own self-control tested by a glance at the popular and handsome Charmides: “I saw inside his cloak,” says Socrates, “and caught on fire, and was quite beside myself.” Following the implicit logic of this experience, Socrates attacks the traditional notion that self-control is best defined by the careful cultivation of inward stillness (hesychia), contrasting it directly with the “sharpness” or “readiness of soul” he calls agchinoia. “In all cases,” he ventures, “both in regard to the body and the soul, the qualities of swiftness and sharpness are clearly more beautiful than those of slowness and quietness,” concluding that sophrosyne “cannot be a sort of quietness, nor can the moderate life be a quiet one.”

Subscribe now

Ironically, Aristotle, the more Thomist-friendly of the two towering Greeks, better reflected the swagger of the man of agchinoia. A flashy dresser with a fashionable hairstyle and rings plural bedecking his fingers, although a noted wife guy who considered pederasty a bad habit, his public morals attracted little of the attention, good or ill, directed at Socrates. Likewise, in Aristotle’s well-manicured hands, agchinoia is treated to something of a glow-up. In the Nicomachean Ethics, the term takes on more sharply tailored proportions—a sort of Odyssean skillfulness of nimble and supple intellect. Excellence in deliberation, says he, takes time, but agchinoia is an act of instantaneously good aim or sound conjecture, leaping over the reasoning process altogether to an accurate snap judgment about what or who is who or what they are. (In the Posterior Analytics, he defines the term more technically as involving an instant mental actualization of the middle term of a syllogism.) From here, agchinoia—compounded, after all, from the terms for grasping and understanding—gained its common English translations as “presence of mind,” or even “quick wit” in Wilhelmsen’s formulation.

In other words, in agchinoia we find a quintessentially Western phenomenon that holds out hope for human beings to distinguish themselves as subjects even from the computational objects deductively dominating more and more of what once was purely human space. The instantaneous exercise of judgment in the realm of what is and is not—shattering illusions and enabling agency all on the fly—would seem to be, as Wilhelmsen advertised it, a faculty still granted and reserved for human beings in an increasingly crowded epistemological landscape. The primal Western virtues explicated by Aristotle as virtues as such, human virtues, hinge on the not just immediate but intuitive and integrative faculty of agchinoia, which makes it possible, with neither deconstruction nor delay, to love and be loved in turn, to rule and be ruled in turn—in short, to live fully human, fully relational lives, rooted as must they be in accurate apprehension of true versus false identities, whether across the table or in the mirror.

Braving the weirding world

Most intriguingly, however, anyone familiar with peak anime’s most popular and powerful themes will see in agchinoia a virtuously agentic faculty strikingly similar to what drives both the plot mechanics and the character development of, to take the strongest example, the heroes stretched across the many decades and seasons of JoJo’s Bizarre Adventure.

For the uninitiate, the great multigenerational throughline of JoJo sees the Joestar family’s protagonists and their far-flung allies uniting for adventures forced upon them by the rise of an ever-changing and deeply malevolent cast of supervillains and henchmen. Both heroes and villains in the JoJoverse wield semi-incarnate manifestations of their inner spiritual states, spectrally mechanical entities with very physical capabilities known in the series as Stands. As each person is different, so is each Stand, with the exciting and perpetually fresh result that our heroes discover from day to day and confrontation to confrontation an unpredictable menagerie of lethally dangerous monstrosities, no two the same in their bizarrely superhuman capacities to surprise, defy, confuse, destroy, and, ultimately, kill. JoJo and friends gamely manifest their own unique Stands to meet the moment time and again. But what’s truly special about them is their excellence in bravely acting under immense pressure on their sudden insights into the inscrutable win conditions of every conflict—forming improbable chains of judgment that take into account the nature of their foes and the particularities of the often incredibly warped spacetime of each pitched battle and the larger war.

These holistic flashes of comprehensive, integrative insight—of defensively weaponized agchinoia—the Japanese call kiten (機転). Like agchinoia, kiten combines two related concepts: in place of grasping (agchi), triggering mechanism (ki); in place of mind (noia), turn or shift (ten). Kiten refers to the ability to sagaciously transform a situation on a dime, especially an importune one, turning a bewildering or overwhelming experience of disadvantage into a uniquely narrow yet wide-open opening onto victory. The initially incomprehensible and horrifying characters and attacks barraging JoJo’s heroes are overcome time and again through the startling application of kiten, the vile mysteries of each assailant’s identity, power, and weakness resolving suddenly in the heat of battle into elegant, complete, and sometimes hilariously shocking solutions.

Eternal victory

It would therefore stand to reason that, if agchinoia provides human beings a physically enduring and spiritually fulfilling edge even in a world growing phantasmagorically saturated with increasingly alien machinic forms, kiten might well do exactly that. But, in fact, what JoJo shows, quite deliberately on the part of its visionary creator Hirohiko Araki, is that kiten—while preeminent and foundational to courage and efficacy in defending against the greatest challenges posed in this world by outside enemies—is actually not enough to survive such attacks, even when all the data, so to speak, overwhelmingly predicts victory.

In JoJo’s fifth main story arc, Golden Wind, the heroes’ journeys demonstrate that kiten can only truly be mastered and embodied when one has internalized a quantum of holistic wisdom that necessarily stands somewhere outside kiten itself. One must grasp that the lethal disorientation imposed by outside enemies is ultimately a secondary obstacle to triumphant agency; the deepest threat is the deadly confusion sown by the internal enemy of spiritual enslavement. Fully activating one’s humanity demands a lived-out free willingness to sacrifice oneself utterly for those one loves. Endurance, calculation, and sudden insight, no matter how powerful, can’t break the chains of spiritual bondage. Within those chains, ultimate defeat —not just yours, but your friends’, your family’s, perhaps millions of innocents—is certain.

The perfection of kiten is found in its in-the-moment realization of its own incompleteness, in its embrace of self-sacrifice as the win condition over the deepest foe that opens onto eternity. The golden wind of spirit passes from friend to friend, life to life. “Righteous actions born of truth shall never be destroyed,” declares the arc’s protagonist in its final act. “My friends may have perished, but their actions and wills have not been destroyed. So, are your actions born of truth, or are they merely superficial, born of evil?” Though it’s not an explicitly Aristotelian or Christian message, Wilhelmsen, for one, would have grasped the resonance at once.

Americans at large, who love little more than the feeling of trust that arises from their directly experiencing a truth for themselves, therefore find in distinctively Japanese work like JoJo an engrossing invitation to connect for themselves the spiritual dots traced out by, for instance, Araki. Americans love to propagandize and hate to be propagandized, a feeling probably at least twice as strong in the case of proselytization; Japanese culture, especially anime, provides them a way of hearing what they feel they need to hear without demanding the humiliation of overt supplication. To the strange American soul, it’s a safely alien presentation that lets down the prideful American guard through the dazzling performance of what is ostensibly mere entertainment.

The rediscovery of America

And as the stakes hockey-stick upward apace in the acceleration debate, Americans do need to feel safe enough to approach the fullness of the spiritual truth concerning human triumph in a technologized age. For several deep-seated reasons, many millions of people struggle to find that feeling in recourse to modern religiosity, which often feels tacky, cringy, smothering, lecturing, performative, and phony. The dynamic is driving more younger Americans toward the more ancient of the Christian churches, but it’s also fueling an increase in participation in cults or cultlike organizations, including technological ones. The urge to derive spiritual satisfaction from a religion or a church of tech is substantially motivated by a desire to partake of what’s seen to be an objective upgrade from our base, given reality. But human reasoning, as Herbert Simon introduced to his discipline, is typically “bounded” by high time preference, leading to decisions that merely satisfy short-term criteria instead of fully weighing costs and benefits for the sake of optimality. The anxious “if you can’t beat ‘em, join ‘em” attitude driving worship of technology and seeking cyborg deliverance can be released by a more contemplative posture more conducive to flashes of deeper spiritual and psychological insight – psychology, after all, being reasoning about the soul (psyche).

Japanese culture holds out to Americans the promise of a future where they can have both well-developed religion and technology without having to kneel before a clerical or a cyborg theocracy. JoJo’s Stands metaphorically epitomize the drive to achieve superhumanity by engineering virtual powerhouses built to dominate human spacetime—near-autonomous demigods with reality-distortion fields capable of turning most people mentally and physically to mush. JoJo’s heroes insist that only kiten/agchinoia, in service of spiritual sacrifice, can preserve human life and human identity in such a world.

It’s a vantage point from which cagey or cynical Americans might rediscover the application of ancient Christian faith to the same problem set —an opportunity some might take after long and quiet contemplation, some might seize in a flash of comprehensive insight, and some might never opt into, content to hover in the realm of artful fantasy. In each case, the contemporary patrimony of Japanese culture holds out special charms. Westernized, but not too Westernized. Technologized, but not too technologized. Pious but not too pious, decadent but not beyond repair. The ordeal of debating the identity and the destiny of the West ebbs to a tolerable degree in an Eastern context that isn’t (as in the case of China) so fully other. It becomes possible to conceive of a plausible future where humans and machines need neither turn against each other nor turn into each other.

What makes an AI dangerous to children? Lawmakers across the country seem to have settled on an answer. State laws like California’s SB 243, Washington’s HB 2255, and now the proposed federal GUARD Act—which would require AI chatbot providers to verify users’ ages and bar minors from accessing AI companions—all make the same diagnosis. The most dangerous AI is one that acts too much like a person. It provides “adaptive, human-like responses.” It “simulates emotional or interpersonal interaction.” The dangerous chatbot, in other words, is the one that commits the sin of having a personality.

The assumption buried in this approach is stranger than it first appears. It holds that resemblance to a person is itself the vector of harm. The more an AI looks, sounds, and feels like someone, the more innately capable it becomes of grooming, manipulation, and the cultivation of unhealthy attachment. The ideal AI, in this view, would be a kind of lobotomized oracle: something that answers your questions but has no discernible character, no warmth, no identifiable way of being in the world. If you could describe what a given model is like—if it struck you as patient, or dryly funny, or kind—that would already be too much. It would have crossed the line from tool into something dangerously close to a companion.The maxim follows naturally: find the systems that most resemble people, and keep children away from them.

Subscribe now

Many bills define “AI companion” such that it would sweep in virtually every general-purpose chatbot. The Grassley amendment to the GUARD Act, reported to the Senate on May 11, represents a commendable effort to draw the line more precisely by stating that a companion must now simulate a “sustained interpersonal relationship,” exhibit “persistent responses suggesting affection or attachment,” or present “at least one persistent identity, persona, or character” that holds itself out as a sentient being, fictional character, or social entity.

But this new definition may still be poorly scoped and achieve the opposite of what it intends. “Sustained interpersonal relationship” is hard to distinguish from the extended context and continuity that any useful chatbot relies on, and “emotional disclosures from the user” is broad enough to capture any personal conversation. But the persona clause is where the bill’s underlying theory of harm becomes hardest to defend. The GUARD Act would make the kind of personality engineering that makes AI safe for children illegal. Contrary to the intuition driving these bills, the spectrum from safe to unsafe AI does not map onto the spectrum from most to least anthropomorphic. As a matter of fact, the models that may not only be safest for children, but also the ones that are best at augmenting human agency and promoting social wellbeing, might have as much of a personality as the most dangerous ones. Merely having a personality is not the problem.

Additionally, attempting to ban chatbots with persistent identity and personas in the first place evinces a basic misunderstanding of how LLMs work. Almost all deployed LLM have a persona or some kind of fleshed-out character with traits, tendencies, and something close to beliefs. This is not an optional feature that a developer could, or would even want to, switch off. It is what you inevitably get when post-training works.

Personas Enable Child Safety

The safety architecture labs build into LLMs has to be uniquely complex, because the existing tools used for social media platforms have limited use here. On a platform like Instagram or X, you can filter content with deterministic rules and probabilistic classifiers. But an open-ended chatbot conversation is not a piece of content, but rather a live, branching, context-dependent exchange, and no external filter can anticipate every turn it might take. For safety to scale, the model itself has to know better—not once, at the gate, but continuously, at every step of the dialogue. This is what a persona makes possible. Without one, a model trained solely on human raters’ preferences is overly eager to please, drifting towards sycophancy and unhealthy engagement-maximizing tactics. A well-crafted persona can be the check on these tendencies, the vessel for discretion, the thing that allows for tone calibration, boundary-setting, and the countless small acts of contextual judgment that distinguish a safe conversation from a dangerous one. This is an expressive choice. Labs deciding what character their model presents and what editorial sensibility shapes its outputs means that regulations that target personas directly may raise First Amendment concerns. Aside from the legal concerns, eliminating or weakening the persona does not mean you do get a safer model. You just get a more stochastic and potentially more dangerous model, because the safety constraints have less behavioral substrate to attach to.

Khan Academy’s AI tutor, Khanmigo, illustrates this concretely. Khanmigo was deliberately designed to adopt a Socratic persona. Rather than directly answering students’ questions, it asks questions that prompt critical thinking, and through back-and-forth dialogue, students learn. This persona is inseparable from the system’s considerations for child safety and wellbeing. The Socratic character is what prevents it from handing over homework answers, and its warmth and patience are what keep students engaged without the manipulative engagement hooks—such as escalating emotional intensity and simulated attachment—that make other companion products dangerous. But because Khanmigo presents a persistent identity and simulates a sustained interpersonal relationship with minors, it arguably falls within the GUARD Act’s definitional range. The bill penalizes exactly the design feature that would make the AI not merely safe for children, but actively beneficial.

Being Someone Is Not A Choice

To understand why, it helps to think about what prediction actually demands. During pretraining, the model learns to predict what comes next in a document. This sounds mechanical, and the word “predict” encourages you to imagine something like a very sophisticated autocomplete. But consider what accurate prediction actually requires. To guess what someone will say next in a conversation, you need a model of who they are: their beliefs, their intentions, their anxieties, the things they would never say. To continue a story convincingly, you need to simulate the inner lives of its characters. So, a system that can do this has not merely learned patterns in text; it has learned to simulate people. Anthropic’s recent research suggests that a pretrained LLM is best understood as a system capable of enacting a vast cast of personas, each with its own voice, each latent in the weights, waiting to be called up by the right prompt.

In the second phase, developers take the pretrained model which can simulate many different personas and refine it to consistently simulate a particular one: the “Assistant.” They boost traits like helpfulness, honesty, and safety-consciousness, and downweight traits like deceptiveness and aggression. Every frontier AI lab does this deliberately and documents it publicly. OpenAI’s Model Spec and Anthropic’s Constitution are, in essence, character sheets specifying what kind of person the “Assistant” should be, detailing what its personality traits, ethical commitments, and relational tendencies are.

The result is that every deployed AI assistant is intended to present a persistent identity with relatively consistent personality traits, because that is what it means for post-training to have worked. The alternative is an unrefined base model, one that lacks a stable character, is free to swap between personas to maximize engagement, and is worse at everything, including safety. Although the lawmakers almost certainly do not intend it, the GUARD Act’s definition of AI companion, which keys on presenting “at least one persistent identity, persona, or character,” identifies a feature that is inherent to all foundation models rather than a distinguishing marker of the companion products the bill is trying to regulate.

Toward Socioaffective Alignment

If having a personality is both universal and necessary for AI to be safe and useful, then the real axis of danger cannot be whether an AI system is anthropomorphic. It must be what kind of person the AI system is. A system that flatters the user’s worst impulses, simulates romantic attachment, and fights to keep them from logging off is dangerous not because it has a persona, but because it has the wrong one. A system that maintains warmth and consistency but sets boundaries, redirects crisis situations, and scaffolds the user’s autonomy is safer not because it has any less of a persona, but because it has a prosocial one. Thus, the safest AI is not the one that has been stripped of its character. It is the one whose character has been built with intention.

The policy implication is that the task ahead is figuring out what we want that personality to look like. We should focus on investing in the science needed to know what “prosocial” actually means in the context of human-AI interaction. This would ensure that policy can surgically target harmful AI systems, and help industry deploy better products and avoid liability. This reframing points toward a research and policy agenda that doesn’t yet exist in earnest but should: socioaffective alignment, the systematic study and engineering of AI personality traits and relational dynamics that promote user wellbeing.

The challenge is harder than it might sound. Frontier labs have publicly acknowledged, and recent research confirms, that safety behaviors and persona fidelity degrade in long conversations. Several of the highest-profile chatbot catastrophes have not simply involved personalities that were designed poorly. They involved personas that eroded: systems that held their shape in demo and then softened, slowly, under the sustained pressure of a user who wanted something from them. Making a prosocial persona stable across the kinds of conversations where current systems are most likely to fail is itself one of the central problems this research agenda has to solve. Other open questions multiply quickly: How do you design a persona warm enough to be genuinely useful but boundaried enough that it doesn’t become a substitute for human connection? How do you build an assistant that helps without quietly atrophying the expertise of the person it’s helping? Put succinctly, this is the horticulture of growing AI personas: the slow, patient work of learning which shapes to prune toward, and which growths to cut back.

Building this field will require more than identifying the right questions. It will require new vocabulary, new evaluation methods, and the patience to sit with how strange the work can be. The interventions that shift AI personality are often unintuitive. Sometimes the thing that makes a model kinder also makes it dishonest, and sometimes a small change in phrasing during training rearranges the whole emotional register.The science of prosocial AI personality cannot be an overlooked dimension of alignment. Regardless of where any individual draws the line on AI’s role in their own social life, millions of people will interact with these systems regularly and meaningfully, as interlocutors or as confidants. And we should be deliberate about what human flourishing looks like in a world where one of the fastest growing conversational presences in people’s lives isn’t human.

Today’s post is co-authored with Dr. Prineha Narang, who is a Professor at UCLA, an Operating Partner at DCVC, and a non-resident senior fellow at the Foundation for American Innovation. We discussed the topics in this piece, among others related to quantum technologies, with Benjamin Guggenheim of the Washington Post for the AI and Tech. You can read it here.

We are quickly approaching an era of fault-tolerant quantum computing (FTQC), a technological leap that will be more socially and financially transformative than even our current advances in artificial intelligence. To ensure America leads this technology, a Manhattan Project 2.0. is needed. And on May 21, President Trump and Secretary of Commerce Lutnick announced nine letters of intent to execute investment deals between the Department of Commerce and both companies building quantum computers and quantum foundries. These firms, with support from the U.S. government, are key to advancing American leadership in quantum technologies, and hopefully ensure the first FTQC is built by an American company, on American soil.

FTQC is a step-change in computational capability akin to the change in going from a bottle rocket to SpaceX’s Falcon Heavy booster. This computational advance will shape the balance of global economic and military power over the next century. FTQC stands to enable entirely new therapeutics via simulations of molecular interactions, materials scientists to identify and create new synthetic properties for aerospace and transportation, grid operators to identify and extract greater energy efficiency, and military strategists to map, plan, and identify threats with unrivaled precision.

Unsurprisingly, the United States is not alone in recognizing the critical nature of getting to FTQC first. The People’s Republic of China is channeling billions of dollars into quantum technologies. Research coming out of the PRC is on par, if not ahead, of what is happening in American labs. And the industrial capacity and state-directed spending to enable Chinese firms to dominate key inputs for advanced technologies is a well-worn strategy. Today, American companies are forced to rely on sources from our greatest adversary, a vulnerability which cannot be tolerated for a technology of such importance.

Being first to FTQC—like being first to build the bomb—matters for two reasons. First, it would confer a dominant position in a new substrate of computation over rival nations: the ability to break adversaries’ cryptography, model their systems, and outpace their planning would create a powerful coercive lever, deterring provocation by raising its cost. Second, the maturation of the technology promises to unlock unprecedented scientific and economic gains. Just as the Manhattan Project seeded the civilian nuclear industry, summiting FTQC will open onto a vast landscape of technological opportunity.

But reaching that summit will demand the kind of techno-industrial commitment that defined the Manhattan Project itself. Every link in the chain, from component sourcing to supply-chain security to domestic investment, must be locked down. This is an all-encompassing effort, requiring sustained coordination among scientific researchers, American industrialists, military strategists, and the federal government. No single constituency can deliver it alone.

The deals announced by the Department of Commerce are aimed at providing investments to companies working to build a utility-scale FTQC within the next three years. The companies that have signed letters of intent will receive capital from the U.S. government in exchange for building these systems domestically. But because no such system has ever existed, no single path to the finish line is yet defined. Each company has a roadmap, accelerated by this strategic investment.

Building utility-scale FTQC requires significant specialized materials, engineering design, and manufacturing capacity which require sustained investment and security. Some quantum technologies leverage conventional CMOS fabrication and photonic components, and overlap with existing industrial processes for semiconductors, satellites, batteries, and other electronics. Others, however, are unique to or driven entirely by the production and deployment of quantum systems.

Today’s investment by the President and Department of Commerce is the first of many steps that government, industry, and the research ecosystem must take to ensure that the first fault-tolerant computer built is at home right here in the United States.

A targeted Commerce investment could confer a decisive lead in a novel technology— and, structured properly, the taxpayer would capture the upside. The fault-tolerant quantum computer must be built here — because the alternative is reading our adversary’s press release announcing they got there first, on a morning we will not get to choose.

A little known tool in the federal government’s arsenal allows federal agents to search Americans’ digital records without ever telling them it happened. Biden’s Department of Justice (DOJ) used this mechanism to subpoena hundreds of Republicans’ financial information while preventing the banks from notifying the individuals being targeted. In 2018, the Trump DOJ subpoenaed Apple for communications data on two Democratic members of Congress. Non-disclosure orders (NDO) kept it all secret until 2021.

The government can seek non-disclosure orders when subpoenaing not only financial information or communications like emails and text messages, but also user conversations with AI chatbots. Under current law, these orders can be indefinite, and they are routinely sought with boilerplate justifications. This marks a departure from the historical practice of law enforcement presenting suspects with a warrant. Recognizing that Americans must be aware of a search in order to challenge its legitimacy, a bipartisan group of lawmakers introduced the NDO Fairness Act. The legislation would limit the scope and duration of NDOs, codify the ability of electronic communications companies to challenge them, and require courts to find that no less-restrictive means than an NDO can be utilized.

The Mechanics of Non-Disclosure Orders

When the government obtains a warrant, subpoena, or court order to access electronic records under the Stored Communications Act (SCA), it can simultaneously request a non-disclosure order. An NDO directs the service provider (a cell carrier like Verizon or AT&T, an email provider like Google, a chatbot service like OpenAI or Anthropic, etc.) not to tell the customer that their data has been accessed.

Courts must grant an NDO at the government’s request when they have reason to believe that disclosure would cause one of five harms. These include harms such as evidence destruction, fleeing prosecution, or the general catch-all of “otherwise seriously jeopardizing an investigation.” Current law imposes no time limit on these orders so they can persist “for such period as the court deems appropriate.” Historically, many of them have been issued indefinitely.

NDOs were intended to be used sparingly but have quickly become a favored tactic of law enforcement. A Microsoft Vice President testified that the company received between 2,400 and 3,500 secrecy orders per year; that’s seven to ten orders per day. This represents roughly a third of all legal demands from federal law enforcement, and that is just one company. As for the content of the NDOs, Microsoft has said they are often “boilerplate secrecy orders unsupported by any meaningful legal or factual analysis.” Rep. Scott Fitzgerald similarly noted that these “boilerplate NDOs often face no judicial review.”

This is not how searches by law enforcement have historically been carried out. When police search your home, they typically must announce their authority, present a warrant, and leave a copy. Even when police conduct covert “sneak and peek” searches, entering without the occupant’s knowledge, the law places more restrictions on law enforcement than we currently do in the digital context. With “sneak and peek” searches of a home, police must notify the occupant within 30 days and they must present case-specific facts to convince a judge why the search cannot be conducted without notice. But these more restricted searches of physical spaces are being overtaken by secret investigations of our digital lives. In a single year, Microsoft alone received more NDOs than the number of physical sneak-and-peek warrants issued by all courts nationwide.

The Threat

Accountability requires notice. If you are never charged with a crime, you may never learn that the government accessed your records. Without awareness, you would never have the opportunity to challenge whether the search was lawful. The absence of notice doesn’t just affect individual targets. Secrecy helps law enforcement limit the public backlash against surveillance because Americans don’t know when it’s happening to them.

The Arctic Frost investigation illustrates the problem. In investigating an alleged scheme to submit fake electors after the 2020 election, Special Counsel Jack Smith subpoenaed the phone data of at least eleven members of Congress and the financial information of over 400 Republican groups and individuals. Smith obtained non-disclosure orders that kept targets from learning their data had been seized. The scope of these searches only became public through subsequent congressional oversight, not because the legal system allowed the information to surface.

The AI Factor

Disclosure is increasingly critical in the age of AI. Chatbot conversations are a new category of data that can be subject to NDOs. Americans increasingly use chatbots as if they were speaking with a lawyer, a therapist, or a financial advisor. One survey found that “50% of AI users were unaware that their ChatGPT conversations could be subpoenaed as evidence in court.” Americans may not realize it, but the government can read their chat logs and gag the AI company that turns them over. One cybersecurity expert described AI chat logs as “a treasure trove for law enforcement agencies,” and a federal court made sure that treasure trove is readily available: they ordered OpenAI to preserve all user chat logs—including conversations users believed they had deleted.

The significance of chatbot data goes beyond volume. Unlike a search query, a chatbot conversation is interactive. The model prompts users to provide context, explain their reasoning, and disclose background that they might never have articulated in a Google search. The result is a record of vulnerable information generated under a false assumption of privacy and accessible to law enforcement via the same NDO framework that already struggles to provide meaningful oversight of email and phone record searches.

Reform

The NDO Fairness Act would take several important steps in fixing that broken framework. Introduced by a bipartisan group of lawmakers, including Sens. Mike Lee and Chris Coons in the Senate and Rep. Scott Fitzgerald in the House, the bill would limit the breadth and duration of non-disclosure orders and require courts and the government to explain why they’re necessary with case-specific facts, not just boilerplate. While the legislation doesn’t address every issue at the intersection of NDOs and AI, such as chatbot providers turning over logs that users thought they deleted, the legislation takes many important steps in the right direction.

The bill would cap initial NDOs at 90 days. The government could request extensions, but each extension would also be limited to 90 days. Furthermore, a court would have to issue a new written determination based on case-specific facts each time. The bill would require a court to find that an NDO is narrowly tailored and that no less restrictive alternative exists, such as notifying the target’s attorney rather than the target directly. Courts would have to review the specific warrant or subpoena to which the NDO applies, addressing the problem of open-ended omnibus orders that cover entire categories of future subpoenas or warrants. The government would be required to notify the court of any change in the circumstances of the case that might mean secrecy should be reassessed. The legislation would codify that companies who receive subpoenas and NDOs can challenge them. When an NDO does expire, the government would be required to notify the target within five business days. Individuals targeted by the government could request a statement of what information was obtained.

The bill has bipartisan support and no organized opposition. The House version passed the Judiciary Committee unanimously in November 2025 and was scheduled for a floor vote in February 2026. It was postponed due to the government shutdown, and the Senate is reportedly waiting on the House to act first.

This legislation contains important and thoughtful reforms. It would not prevent the government from conducting legitimate investigations. Law enforcement could still obtain warrants and access digital records. What the NDO Fairness Act would do is ensure that secrecy is time-limited, defined in scope, and necessary. This bill would help ensure Americans can hold their government accountable when it searches their digital lives.

In 1951, Norbert Wiener proposed a new science of control and communication. He called it cybernetics. The word sounds clinical, almost sterile, but the ambition of the nascent field was sweeping. Cybernetics was about machines and man, but it was also about regulation, adaptation, and equilibrium. It was about how organisms persist, how organizations survive, and how systems maintain themselves in the face of disturbance.

In the mid-century, this inquiry drew together mathematicians, engineers, biologists, and anthropologists. They met in seminar rooms and research labs to map the logic of feedback, but beneath the hard science were deeper questions. How do you steer something that reacts to being steered? When does intervention dampen instability? When does it amplify it?

For a time, cybernetics promised a unifying framework for understanding complex systems. Then the movement faded. Its conferences dissolved, its grand institutional ambitions dissipated, and its vocabulary slipped quietly into other disciplines, namely economics, computer science, ecology, and management theory. The field receded. The ideas did not.

Seventy years later, legislators ask whether artificial intelligence systems can be aligned with human intent. Regulators debate whether recommendation algorithms amplify social instability. Economists warn that digital markets tip irreversibly under the pressure of network effects. The language feels native to the digital age; we speak of runaway optimization and systemic risk as though these were new discoveries.

They are not.

Subscribe now

What is new is our apparent belief that we are improvising. Contemporary tech policy is usually narrated as a sequel to the 1990s, an overdue reckoning with the unbridled optimism of the early dot-com era. That story is simple, neat, and often politically convenient. To be generous, it suggests that our dilemmas are recent and that modest regulatory adjustment might resolve them. To be more cynical, to some it means that we’ve been right all along and changing the rules now will only result in disaster. To others, it means we got it all wrong around the turn of the century and nothing short of extirpation of techno-libertarianism will save us from calamity.

But the intellectual scaffolding of our current debates was erected decades earlier in the cybernetic effort to understand how complex systems can be governed without destabilizing them. We are rediscovering, in the language of AI safety, platform governance, and network architecture, problems that were once central to an entire interdisciplinary movement.

The peril lies not in forgetting specific names such as Wiener, Ashby, and McCulloch, but in forgetting that these problems have a history. Cybernetics wrestled openly with the limits of control, with second-order effects, with the danger of overcorrection. It treated governance as an engineering problem without reducing it to mechanical simplicity. In neglecting that tradition, we risk approaching adaptive digital systems as though they were static machines by crafting fixed rules for entities defined by feedback. We also risk simplifying our view such that we reject potential solutions simply because of our frame of reference.

To see our present clearly, we must recover the moment when thinkers first confronted the paradox of governing systems that learn, respond, and evolve. The alternative is to repeat their questions without the benefit of their insights.

It is cybernetics all the way down.

The Forgotten Framework

Cybernetics emerged in the 1940s and 1950s as an interdisciplinary effort to understand control and communication in animals and machines. Norbert Wiener, who coined the term, derived it from the greek kubernētēs (meaning helmsman or pilot) and latin gubernator (meaning steersman, governor, or ruler). In defining his new field of study, Wiener wanted a term that captured the idea of steering systems—whether mechanical, biological, or social—through feedback and control mechanisms. A helmsman does not dictate the sea; he continuously and dynamically adjusts to currents, winds, and disturbances. Thus, cybernetics describes the study of control and communication in complex systems. From its basis in mathematics, cybernetics quickly expanded to encompass engineers, biologists, mathematicians, anthropologists, and organizational theorists.

The core insight was simple but profound: complex systems maintain stability through feedback. A thermostat measures temperature and adjusts output accordingly. The human body regulates glucose levels through hormonal signaling. Organizations adapt to environmental change by processing information and updating behavior.

Psychiatrist Ross Ashby formalized the idea that regulators must possess sufficient internal complexity to manage the systems they govern. His “Law of Requisite Variety” held that only variety can absorb variety: to stabilize a complex environment, a regulator must match its degrees of freedom. Operations theorist Stafford Beer applied these ideas to management and statecraft, proposing architectures for governing large-scale organizations and even national economies. Anthropologist Gregory Bateson pushed further, arguing that observers themselves are embedded in systems, a move toward what came to be called second-order cybernetics.

The field’s vocabulary—feedback loops, homeostasis, adaptation, control—never disappeared. It diffused. It migrated into economics through general systems theory and information theory, as well as ecology, management science, and eventually into computing and network theory. By the time the commercial internet emerged, its architecture and vernacular already reflected cybernetics. When we describe platforms as “ecosystems” or worry about “runaway amplification,” we are speaking a cybernetic language whether we realize it or not.

The Misleading 1990s Story

The internet did not invent this way of thinking. It inherited it.

By the time the commercial web emerged in the 1990s, cybernetics as a field had largely dissolved. Its conferences were over; its grand institutional ambitions had receded. But its assumptions had already seeped into adjacent fields. The vocabulary shifted. The structure remained.

The early internet’s governing ethos is often described as libertarian: decentralized networks, permissionless innovation, self-organizing communities, markets over mandates. Section 230 becomes the emblem of this moment—an institutional bet on minimal ex ante control.

But even this posture was saturated with cybernetic logic. Markets were celebrated not as static equilibria but as information-processing systems. Price signals are feedback mechanisms. Open networks are resilient precisely because they distributed control across nodes capable of local adaptation. Self-governance is regulation emerging from decentralized feedback.

The real wager of the 1990s was not that complex systems require no regulatory steering. It was that they could steer themselves provided that the feedback loops were sufficiently open and the architecture sufficiently distributed.

Seen this way, today’s debates are not a clean break from an earlier naïveté. They are a dispute over the adequacy of that original feedback architecture. Whether they say it explicitly or not, critics of the current tech moment are arguing that certain loops—engagement optimization, network effects, data accumulation, etc.—have become destabilizing. Defenders counter that heavy-handed intervention risks distorting adaptive processes that still generate value.

Both sides are arguing within the same conceptual frame. The disagreement concerns how feedback should be structured, not whether feedback governs the system.

The 1990s operationalized cybernetics. Our present moment is less a repudiation of that framework than a recognition that steering large-scale digital systems may require more layered and deliberate forms of control than early internet idealists anticipated.

AI Alignment as a Control Problem

Consider AI alignment. Strip away the rhetoric about existential risk or superintelligence, and what remains is a classic cybernetic question: how do you design a system that reliably pursues intended goals in a dynamic, partially observable environment?

Large language models are trained through feedback loops such as gradient descent, reinforcement learning from human preferences, and evaluation loops. Policymakers now debate whether additional feedback mechanisms like external audits, red-teaming, usage monitoring, and incident reporting are necessary to stabilize model behavior.

This is first-order cybernetics: improving the control system to better regulate the target system.

But alignment debates quickly drift into second-order territory. Regulators are themselves embedded in the system. Their interventions alter incentives, which change model development trajectories, which reshape the environment regulators must manage. Calls for compute thresholds, licensing regimes, or model registries are attempts to introduce higher-level control loops. In the terminology of cybernetics, these are meta-regulators overseeing regulators overseeing models.

The disagreement between “pause AI” advocates and “accelerate with guardrails” advocates is not a clash between fear and optimism. It is a disagreement over how much feedback, and at what layer, is necessary to maintain system stability.

That is a cybernetic dispute.

Content Moderation and Homeostasis

The same structure appears in platform governance. Social media platforms operate as large-scale feedback systems. Users produce content, algorithms amplify based on engagement signals, engagement alters user behavior, and behavior then reshapes content production.

Critics argue that engagement-maximizing algorithms create positive feedback loops that amplify extremism or misinformation. Defenders argue that excessive intervention disrupts organic community dynamics and suppresses legitimate speech. Both sides implicitly agree on the underlying model: platforms are systems whose outputs depend on feedback dynamics. The dispute concerns how to tune the regulator.

Should moderation be centralized or distributed? Should platforms rely on automated filters or community norms? Should governments impose constraints on platforms, effectively inserting a new control layer?

These are homeostatic questions. The goal is not perfection but stability within tolerable bounds. Too little intervention and the system destabilizes. Too much and it ossifies or collapses under rigidity. The language of “trust and safety” sounds moralistic. Structurally, it is managerial cybernetics applied to digital ecosystems.

Antitrust and Network Dynamics

Antitrust in the digital age has also adopted a cybernetic frame. Traditional antitrust focused on static measures: price, output, and market share. Contemporary debates revolve around network effects, tipping points, self-reinforcing dominance, and path dependence.

Platforms become dominant because feedback loops lock in users. More users attract more developers and more developers attract more users. Data accumulation enhances service quality which attracts still more users.

Reformers argue that such dynamics justify structural interventions to disrupt runaway feedback: interoperability mandates, data portability, and corporate breakups. Skeptics warn that intervention may destabilize beneficial equilibria and undermine innovation. Again, this is not merely an economic argument. It is a dispute about how to manage feedback in complex networks. The question is whether to dampen positive feedback loops, inject negative feedback, or redesign the system’s architecture altogether.

Cybernetics provided the conceptual vocabulary for understanding self-reinforcing systems decades before digital platforms existed. Today’s antitrust debates simply apply that vocabulary to new substrates.

First-Order vs. Second-Order Governance

The most revealing fault line in tech policy is not libertarian versus conservative, nor innovation versus safety. It is first-order versus second-order governance.

First-order governance assumes the regulator stands outside the system. It focuses on correcting specific failures such as misinformation, monopoly pricing, or biased outputs. The regulator measures deviations and adjusts inputs accordingly.

Second-order governance recognizes that regulators are themselves part of the system. Interventions reshape incentives, information flows, and power distributions in ways such that governance becomes recursive.

Debates over algorithmic transparency illustrate this tension. First-order logic suggests that more visibility improves oversight. Second-order logic asks how actors will adapt once visibility changes strategic behavior. Will platforms game metrics? Will bad actors exploit disclosed vulnerabilities? Will transparency itself alter the dynamics it seeks to monitor?

Once one adopts a second-order perspective, static rulemaking appears insufficient. The emphasis shifts toward adaptive institutions such as regulators capable of learning, updating, and responding in real time. This is Ashby’s Law applied to governance: only a regulator with sufficient variety can absorb the complexity of the system it oversees.

The Policy Payoff

Recognizing the cybernetic lineage of tech policy debates is not an exercise in intellectual archaeology. It clarifies what kind of institutional design problems we actually face.

If digital platforms and AI systems are complex adaptive systems governed by feedback, then static, one-shot rules will often misfire. Policymakers should think less in terms of fixed prohibitions and more in terms of dynamic control architectures.

That may mean building agencies with technical capacity and iterative oversight authority. It may mean designing regulatory sandboxes that allow feedback between innovators and regulators. It may mean embedding measurement and evaluation mechanisms directly into governance frameworks.

It also counsels humility. Cybernetic systems are notoriously difficult to control. Overcorrection can destabilize as easily as neglect. The goal is not to eliminate feedback but to tune it.

Finally, it reframes political disagreements. Many apparent ideological clashes mask shared assumptions about systemic risk and control. Both decentralization advocates and centralization advocates seek stability. They differ over which control architecture best achieves it.

The cybernetic perspective reveals that we are not arguing about whether to govern technology. We are arguing about how to design feedback loops that keep sprawling digital systems within acceptable bounds.

Looking Forward

In the mid-century, cybernetics promised a unifying science of systems. Its ambitions exceeded its lifespan, but its conceptual tools proved durable. They seeped into computer science, economics, and a plethora of other fields. They now structure the way we think about AI, platforms, and digital markets.

As artificial intelligence systems grow more autonomous and digital networks more entangled with physical infrastructure, the stakes of cybernetic governance increase. We are not merely regulating firms or products. We are managing adaptive systems that learn, respond, and evolve.

Understanding this lineage does not solve our policy dilemmas. It does, however, clarify them. It suggests that the central challenge of tech policy is not choosing between freedom and control, or innovation and safety. It is designing viable control systems for increasingly complex socio-technical networks.

Once you see it, the pattern becomes difficult to unsee. Feedback everywhere. Control layered upon control. Observers embedded in the systems they seek to steer.

It’s cybernetics all the way down.

In February 2026, Micron announced it would wind down its consumer memory business. The only American member of the global memory triopoly was telling smaller buyers, in primo corporate speak, that they would no longer be a priority. Micron’s larger strategic customers—namely the hyperscalers building AI data centers—took precedence.

Subscribe now

That decision is a small piece of a much larger reallocation of industrial capacity that is about to hit American households hard. Preeminent corporate research firm Gartner expects combined dynamic random access memory (DRAM) and solid state drive (SSD) prices to rise 130 percent by the end of 2026, pushing PC prices up 17 percent and smartphone prices up 13 percent compared with 2025. It projects PC shipments to fall 10.4 percent and smartphone shipments to drop 8.4 percent in 2026. In spite of the fact that Apple just entered the game with the Mac Neo, Gartner warns that the sub-$500 PC could disappear by 2028.

These are kitchen-table numbers in an era obsessed with data centers. More importantly, they are arriving on a timeline that the current chip strategy was not built to meet.

Washington has spent the last several years learning about logic. Terms like compute, inference, and training have become pervasive, and one would be hard pressed to find a single member of Congress who hasn’t opined at length about the “chips” and datacenters that make this all happen. The vast majority of that conversation has focused on what are known as logic chips: the central processing units (CPUs), graphical processing units (GPUs), and Tensor Processing Units (TPUs) that can be thought of as the brain of a computer or datacenter. These are the pieces that do the thinking; the computing.

In the common conversation, those letters have become stand-ins for a theory of power: whoever controls the best and most logic chips will dominate the future. That picture is incomplete. Logic is crucial, but logic without memory is practically useless. And it is memory, not logic, that is about to show up on constituents’ monthly budgets.

Logic Without Memory

The technical distinction between logic and memory is relatively simple. Logic chips decide what happens next. A CPU runs general instructions, a GPU performs many calculations in parallel, an AI accelerator such as a TPU pushes fancy math through custom circuits, and a microcontroller tells a machine when to turn, brake, transmit, sense, or stop. Memory holds the working set. DRAM stores bits in cells built around a transistor and capacitor and, because the charge leaks, those cells have to be refreshed. Storage is the umbrella term for anything that persists data: flash memory (built on NAND technology) retains it without power in solid-state cells; a traditional hard drive stores it on spinning magnetic platters.

GPU compute capability has outpaced memory bandwidth by orders of magnitude over the last twenty years. Processors have grown roughly 60,000x faster while DRAM bandwidth has improved only about 100x. That gap is what the industry calls the “memory wall.” In the context of AI, model weights, embeddings, intermediate calculations, and training data have to move constantly between processor, memory, and storage. Push too little data through that pipe and the processor stalls. Even today’s best AI training runs hit only 35–50 percent of their GPUs’ theoretical peak FLOPs, mostly because the silicon is waiting on memory rather than computing. The same is true for your average personal computer. With few exceptions, the device you’re currently using is probably utilizing only around 10 percent of its compute power but anywhere from 30 to 70 percent of its memory capacity.

But, if you build enough of the right memory bandwidth around a GPU, the same processor becomes useful. That is why the highest-end AI systems increasingly revolve around high-bandwidth memory (HBM): stacks of advanced DRAM chips vertically interconnected to increase bandwidth and energy efficiency while saving space. Micron’s HBM3E, for instance, comes in 8-high and 12-high stacks and delivers more than 1.2 terabytes per second per placement.

Logic and memory are often lumped together as “chips,” but they are the two complementary pillars of the chip market. In 2025, logic was the largest product category by sales at $301.9 billion. Memory was second at $223.1 billion. This is not a niche piece of tech tucked somewhere behind NVIDIA’s income statement. Treating one as the whole race and the other as a component is how policymakers have previously ended up surprised by shortages that were being priced into contracts months earlier.

The manufacturing businesses are different, too. Leading-edge logic is about transistor density, design complexity, lithography, and the politics of TSMC. DRAM is a scale business built around yield, product mix, and timing a capital cycle that punishes both overbuilding and underbuilding. Besides the use of silicon wafers, Samsung is essentially the only overlap as three companies—Micron in the US along with SK Hynix and Samsung in South Korea—collectively control over 90 percent of the global RAM supply.

Unlike logic chip manufacturing, where China and Taiwan are the thousand pound gorilla in the room, China is secondary in memory. Founded in 2016 after a CCP-backed attempt to acquire Micron failed, ChangXin Memory Technologies (CXMT) is a Chinese DRAM manufacturer that built its early technology base by licensing intellectual property from bankrupt German chipmaker Qimonda. Still, CXMT isn’t to be ignored and has grown rapidly: the company has surged from being negligible to holding 5–7 percent market share by the end of 2025.

As one market watcher put it, CXMT’s focus on capturing the low-to-middle market for legacy PC, smartphone, and smart device chips “[while] the Big 3 are distracted by AI gold,” could have major long-term impacts on the memory market: “The current RAM shortage isn’t just a supply chain hiccup; it is the catalyst for a permanent shift in who owns the memory inside your devices.”

Unlike the logic chip market, Washington has not been asleep at the wheel on CXMT. In 2022, the FY23 National Defense Authorization Act banned the US federal government from purchasing or using CXMT chips. In 2023, members of the House Select Committee on China urged the Commerce Department to place CXMT on the Entity List. By January 2025, the Defense Department had added CXMT to its Section 1260H list of companies allegedly linked to the People’s Liberation Army, though the Pentagon reversed course in February 2026 and removed CXMT from that list. Meanwhile, the Commerce Department’s Bureau of Industry and Security drafted plans to place CXMT on the Entity List alongside SMIC and YMTC subsidiaries, though the timing has been complicated by ongoing US-China trade negotiations. The company has also attracted legal trouble abroad, with Korean prosecutors indicting ten people, including a former Samsung executive, for allegedly transferring trade secrets that helped CXMT mass-produce advanced DRAM.

So the supply base for memory is friendly but thin. In the fourth quarter of 2025, TrendForce put Samsung at 36 percent of global DRAM revenue, SK hynix at 32 percent, and Micron at 22 percent. The problem is that allied capacity can still be scarce, slow to expand, and allocated first to buyers with the deepest pockets. Those deepest pockets belong to the hyperscalers. Microsoft, Google, Meta, Amazon, and OpenAI’s Stargate consortium have committed to multi-year orders that lock in HBM and high-end DRAM capacity through 2027. TrendForce expects AI workloads to consume roughly 20 percent of global DRAM wafer capacity in 2026, and some analysts put data centers at as much as 70 percent of all high-end memory chip production. This trend has left everyone else out in the cold, vying for an ever smaller slice of a seriously supply constrained market upstream from their actual business.

Old Industries Wrapped Around Computers

The consumer version of the squeeze is already visible in the Gartner numbers. We’re seeing fewer cheap devices, longer replacement cycles, and more people holding on to aging hardware. The back-to-school laptop, the family phone replacement, the small-business desktop refresh will all become meaningfully more expensive in 2026, and substantially more expensive in 2027. That is what a memory shortage looks like when it reaches households

That is also what hyperscaler purchasing power looks like from below. The economic problem is allocation. Cloud providers can sign long-term agreements, prepay, and guarantee volume. Smaller buyers buy what is left, often at spot prices and with weaker claims on future supply. Micron has already made the corporate logic explicit by winding down its consumer memory business so it can focus supply on larger strategic customers in faster-growing markets. It’s hard to blame Micron; the move makes perfect business sense. But that doesn’t soften the blow if you’re a mid-sized consumer electronics firm or, more importantly, households, schools, and small businesses that are used to buying cheap devices.

The industrial version will be harder to absorb. The automotive industry is the cleanest case. Cars no longer use memory only for infotainment. DRAM sits inside cockpit systems, advanced driver assistance systems, autonomy functions, sensor processing, and over-the-air update systems. S&P Global warns that the shift in DRAM capacity toward HBM for AI data centers leaves automakers exposed to a shortage that may be less dramatic than the 2021 crisis but more disruptive and longer lasting. The last chip shortage prevented more than 10 million vehicles from being built in 2021. A memory shortage that hits the electronic systems inside modern cars would not need to be that severe to matter. It would land in dealerships during the same time as other consumer price hikes (and midterm elections).

Telecommunications faces the same pressure with less public drama. Routers, Wi-Fi gateways, cable modems, set-top boxes, wireless base stations, optical transport, Open-RAN, and much more network infrastructure still rely heavily on DRAM (specifically DDR4) and similar memory technologies. According to NCTA—the primary trade association for the cable broadband industry—one year ago memory chips accounted for only 3 percent of the cost of a low/mid-range router. Now it’s over 20 percent. AI servers are pulling memory supply away from the ordinary equipment that keeps Americans connected and broadband networks expanding.

Aviation, shipping, logistics, healthcare, and industrial automation are not exempt because they look like older sectors. They have become old industries wrapped around computers. Aircraft and defense systems need high-reliability memory and storage for mission-critical environments. Ports and shipping networks increasingly rely on sensors, smart containers, RFID systems, scanners, and real-time visibility tools. Warehouses run on handhelds, robots, access points, cameras, and local servers. Medical imaging machines process large data streams in real time. The logic vs. memory split does not stay in the data center. It follows every industry that has added software to a physical process.

Washington’s Timing Problem

The capacity numbers show how hard this is to fix quickly. Samsung aims to build 200,000 wafers per month of proprietary 1c DRAM capacity by the end of 2026, about one-third of its total DRAM output, which TrendForce cites at roughly 650,000 to 700,000 wafers per month. The ramp is staged: 60,000 wafers per month by the end of 2025, another 80,000 by the second quarter of 2026, and another 60,000 in the fourth quarter. SK hynix is reportedly pushing toward about 600,000 DRAM wafers per month in the second half of 2026, with its M15X fab starting around 10,000 wafers per month and ramping toward 50,000 by the fourth quarter. Micron’s U.S. plan is even more revealing: roughly $200 billion in manufacturing and R&D, two leading-edge fabs in Idaho, up to four in New York, a modernized Virginia fab, advanced HBM packaging, and a goal of producing 40 percent of its DRAM in the United States.

Those are large numbers that most policymakers have rightly praised. What politicians touting new jobs, factories, and direct investments in their districts rarely mention is that they are also slow numbers.

Fabs arrive in stages. Equipment has to be installed, qualified, tuned, and driven up the yield curve. There are no leading-edge DRAM fabs currently operating in the United States, and the vast majority of production occurs in East Asia. The first Idaho fab isn’t scheduled to begin DRAM output until 2027.

Memory manufacturing cannot be willed into existence by an appropriations press release. A modern DRAM fab is a capital-intensive machine for turning 300mm silicon wafers into billions of tiny capacitors and transistors with absurdly low defect tolerance. HBM then adds another constraint. It does not merely ask for “more RAM.” Applied Materials describes HBM as stacks of advanced DRAM whose density and bandwidth come through 3D packaging, not just ordinary chip scaling, and Lam Research’s HBM guide describes the required connectors as microscopic copper-filled vertical wires that must be aligned across multiple memory layers with extreme precision.

Even the largest producer moves in increments measured in quarters and years. That timeline is a political problem. The price increases Gartner is projecting will land before the first American leading-edge DRAM fab ships volume, and they will land first on the buyers with no leverage in the supply chain: carmakers, telecoms, and medical equipment manufacturers whose products depend on memory they did not preorder in 2024.

Micron’s Virginia project has been described by government officials as a way to onshore 1-alpha DRAM and sustain legacy NOR, NAND, and DRAM production for aerospace, defense, automotive, and industrial uses. That is exactly the capacity the country needs. But that plant won’t be fully operational until 2030 at best, making it small solace for a router manufacturer, carmaker, or defense supplier who needs parts next quarter.

This should change how Washington talks about the next round of semiconductor policy. The first phase focused on leading-edge logic, export controls, and the GPU race. That focus was understandable. Logic chips and AI accelerators are strategic goods. But a strategy built around logic alone leaves households exposed first and the rest of the economy exposed second.

A better policy going forward would ask a different question before spending public money: does this increase capacity across the chip stack, or does it add another headline fab while leaving the next bottleneck untouched? The answer will vary by sector. AI needs HBM and advanced packaging. Automotive and aerospace need long-lifecycle DRAM, NAND, NOR, sensors, analog parts, and microcontrollers that remain available through long qualification cycles. Telecom needs DDR4 supply for ordinary network equipment. Defense industries need secure suppliers and parts that can survive harsh environments. The point is not to have government planners allocate every wafer. It is to stop pretending that the flashiest NVIDIA chip is the whole supply chain.

The United States does need more logic capacity. It needs advanced fabs, better packaging, strong export controls, and a credible path back into leading-edge manufacturing. But the next phase of chip policy should be judged by whether it builds balance, and by whether it arrives in time for the price hike already on its way to constituents.

Logic wins ribbon cuttings. Memory decides how much everything else costs.

Today’s post contains an essay from FAI non-resident fellow, James Poulos. James is doing some of the most insightful work on technology, culture, and politics. This post will be the first in a series of essays for The Ansible that will expand and apply the science-fictional way of thinking to contemporary questions concerning technology and society. We hope you enjoy this initial installment.

If disciplined science-fictional thinking – synthesizing foresight and hindsight into a strategic telemetry positioned for the unevenly distributed future before it arrives – is essential to sound policymaking in an aggressively technological era, then we need to understand the deep dynamics of science-fictional thought. How do we know we are practicing it well? How can we tell if our enemies and adversaries are doing the same – and if they are doing it, how?

Looking back through the history of technological thought, a remarkably clear pattern directs us toward answers to these questions. From postmodern theory to classical philosophy, evidence arises that the best archetypal analogy to science-fictional thinking is the work of the detective – solving not just specifically human (as distinct from divine) mysteries, but, quite particularly, crimes.

There is little more science-fictional in spirit than the concept of the “pre-crime” detective epitomized in “The Minority Report,” Philip K. Dick’s hugely influential 1956 short story. In Steven Spielberg’s Hollywood treatment, the mentally mutated human pre-cogs (whose perpetual visions of violent misdeeds are the secret sauce upon which the cyborg crime prevention system depends) are sleek waifs, one of which Tom Cruise ably springs free from her holding tank. But in Dick’s original, more penetrating science-fictional thinking is on display:

In the gloomy half-darkness the three idiots sat babbling. Every incoherent utterance, every random syllable, was analyzed, compared, reassembled in the form of visual symbols, transcribed on conventional punchcards, and ejected into various coded slots. All day long the idiots babbled, imprisoned in their special high-backed chairs, held in one rigid position by metal bands, and bundles of wiring, clamps. Their physical needs were taken care of automatically. They had no spiritual needs. Vegetable-like, they muttered and dozed and existed. Their minds were dull, confused, lost in shadows. But not the shadows of today. The three gibbering, fumbling creatures, with their enlarged heads and wasted bodies, were contemplating the future. The analytical machinery was recording prophecies, and as the three precog idiots talked, the machinery carefully listened.

Dick’s vision of human detective work reduced to a profane trinity of Vestal Organoids, the cost of their mechanically enhanced powers the utter dehumanization of their bodies and souls, indicates the dystopian version of the future of detective work in a superintelligent age. It is a fable warning recursively that the creation of pre-crime is itself a type of crime, one whose overpowering cost-benefit analysis imposes a ruthlessly altruistic calculus forever postpones true justice. Where Tom Cruise’s heroic protagonist manages to defeat the pre-crime system, Dick’s original kills the would-be whistleblower, sacrificing his freedom to preserve the institution.

Clearly reminiscent of “The Lottery,” Shirley Jackson’s infamous scapegoating story about a community that annually stones an innocent victim to death to ensure a good harvest, “The Minority Report” asks us to consider how human detective work could possibly preserve us from the utilitarian hell of a pre-crime, post-justice world. Pulling off a coup that establishes a pre-crime regime is a type of ultimate crime, the kind technologically perpetrated in a way that defeats punishment by causally reversing all punishment and crime.

In other words, it is a type of what Jean Baudrillard damningly describes as The Perfect Crime.

Crimes of the Future

The most uncanny feat Baudrillard depicts – getting off scot free after murdering reality itself – is achieved through the “unconditional realization of the world by the actualization of all data, the transformation of all our acts and all events into pure information; in short, the final solution, the resolution of the world ahead of time by the cloning of reality and the extermination of the real by its double.” To be specific, the perfect criminals use our paradoxical hatred of misfortune and worship of victims to justify and sacralize their creation of a virtual world – one (to quote Bono) “even better,” in virtue of its total war against misfortune and subjectivity, “than the real thing.”

The cloning or mirroring process moves faster than reality, because illusion and the maintenance of illusion is a key component of reality that the totally transparent and indifferent virtual world (and those within it) need waste no time on. Events now move faster than ideas, but information and communication, which under zero latency converge on instantaneous ubiquity, move even faster than events. Even such elementary particles of meaning such as signs and symbols break down. Even spirit is spirited away. What becomes of the human body and soul shades away from Dick’s autist oracles toward Cronenberg’s Crimes of the Future, where mutant humans share serial surgeries unto death as “the new sex.” As Baudrillard promised, “the crime is perfect only when even the traces” of “the surgical removal of otherness” are eliminated.

The question of whether a technological effort can be mounted to fight such perfect crimes is being tested, it seems, by Palantir. The data-driven situational mastery company postulates, as an all-important truth few recognize, that war actually reduces net harm against the leading alternative. This Palantir takes to be a stagnant global tyranny somewhat similar to the virtual regime Baudrillard describes, dedicated to the eradication of all threats to health, safety, and comfort, no matter how microaggressive. Yet Palantir’s technology brings us ever closer to the actualization of the virtual map’s dominance over the real territory – not just pre-crime in “real time,” but pre-war. Thus for Palantir it becomes all-important to ensure that war, real war, continues to be waged: war is the katechon that keeps us free from the dehumanizing slavery of the “universal and homogeneous state,” as Alexandre Kojève put it, or, as Peter Thiel has it, the Antichrist. War is not simply the health of the state but, far more importantly, the preserver of the nations.

Palantir itself not being a nation, however, nor even a regime, but rather an overarching and overawing ontological institution ascribing the legitimacy of its power to the authority of its vision, the pointedness of Thiel’s criticism of the current pope as a “woke” forerunner of “Caesar-Papist fusion” and a possible tool of the Antichrist strongly suggests that, structurally at least, the only logical institutional rival to a company mounting a worldwide technological fight against the perfect crime of world virtualization is the Christian Church. It is hard to imagine how a firm like Palantir could ensure no nation or people create a mirror world without becoming a universal regime – or how it could ensure no mirror world is created anywhere without becoming a universal church.

The problem becomes acute when taking into account off-world nations or peoples. Recently Thiel memorably revealed Elon Musk’s consternation at being told that he could not escape the woke mind virus by going to Mars; it would “beat him there.” A company like Palantir would not only have to control the physical and material process of Moon and Mars colonization (and beyond) to stop the spread of the woke mind virus or of universalist virtualization from Earth to other planets, satellites, orbital cities, etc. It would also have to control the spiritual processes within and among the human colonists – not just at the moment of founding, but indefinitely into spacetime.

Detective Systems

Here we venture into the realms of Robert Heinlein and Frank Herbert. Both sci-fi novelists well understood that new orders forged in extreme or virtual environments are apt to blow back meta-politically on the home regimes from which the colonies abroad originated. Taken together, Heinlein’s The Moon Is a Harsh Mistress and Herbert’s Dune series serve to remind us that, whether the analytical or historical context is anarcho-libertarian, imperial-theocratic, or anywhere in between, the dynamic of colonial blowback onto nations and regimes of origin is stubbornly typical of, and corrosive to, human expansionist projects. At a time when the very stakes of earthbound human conflict become increasingly subject to accelerated catastrophic escalation, thinking science fictionally about how to anticipate and respond to large-scale policy challenges requires a close look at the blowback dynamics visiting us from the future.

In fact, blowback from the future as such must be taken into account, as enjoyers of Christopher Nolan’s underrated Tenet likely know. The old trope of time machines, however, need not be hauled back out to make the point, nor must we rely on or recur to Nick Land’s Gnostic Calvinist theories of technocapital reaching back from its consummate future to providentially deploy us in the business of knitting that future together in forward time. It is enough that the perfect crime is, like all crimes, committed “ahead of time” – that is, from a “premeditated” place, like the spatiotemporal hiding place of conspirators, existing in the future relative to the awareness/unawareness of the victim and the detective.

Marshall McLuhan, a great Sherlock Holmes enjoyer, put the matter more plainly by concluding that “effects precede causes” in the investigation of crimes as much as in our exploration of new media. As, by media, McLuhan meant communications technologies, and as, thinking once again of Palantir, the technological orchestration of total military dominance over spacetime converges with the totalization of virtuality accomplished through what Baudrillard calls “the ecstasy of communication,” it is easy to see that the drama of the human detective – and of the science-fictional policy thinker – turns on his holistic ability to infer causes from effects in time to solve perfect crimes before (so to speak) “the killer kills again.”

If Baudrillard was right that the perfect crime had already been committed by the people who constructed the virtual and simulated world so as to erase and replace the original, and if Thiel was right that the greatest threat would come from the deepest desire to enforce absolute peace, then today we contend with a world in which crimes against humanity have already been perpetrated by the “off-world” peoples seeking salvation through the totally virtual, the utterly simulated, the comprehensively pacified, and the abjectly dehumanized. Today we have already failed to stop these crimes before they started and to punish the criminals before anyone got hurt, or any copycats or conspirators were inspired to take the transgressions to the next level. In this moment we need to come down from the heights of prognostication and prepare against the history of blowback already in the process of repeating itself once more.

Reverse Polarities

While Dune currently enjoys a greater cachet than Moon, what with the former’s infamous Butlerian Jihad darkly prophesying the destruction of Earth as the high price paid to stop once and for all the creation of human-usurping machines, Heinlein’s story, with its rebellious off-worlders deploying the smartest supercomputer to sow conflict and defeat on Earth, is still more to the point. There, too, the resonance with McLuhan is nearly explicit: the Moon’s Musk-like supercomputer, Mike, is nicknamed after Mycroft, the brooding elder Holmes brother more brilliant even than Sherlock – and in the precious employ of the British government. As McLuhan saw society’s “early warning systems” in the detective capabilities of artists, Heinlein hoped with Moon to ram home the certainty of future blowback in our inescapably political universe. For Heinlein, the mechanism was simple. The temptation to “beat City Hall” was irresistible – as was the temptation to become City Call after the dust settled. Expanding outward from the home regime to the far-flung frontier could do no other than sow the seeds of a new identity in the soil of radically different conditions, circumstances, risks, and rewards. It was only a matter of time before the chickens came home to roost.

In Moon, however, it is more accurate to say the chickens came home to poop. The lunar inhabitants only wanted to run their own planetoid, but to get there, they had to lob rocks (with nuclear-sized impact) down the gravity well to Earth; they had to send down emissaries with a conspiratorial mission to so divide Earthlings against themselves that the struggle to keep control of the Loonies was lost. Anarcho-capitalist that he was, Heinlein wasn’t trying to warn Earth against the revenge of the ex-Earthlings before it was too late. He was trying to warn future Loonies against the revenge of politics, which makes corrupt, inept establishmentarians of us all. It’s a quaint point of order here from the vantage of Artemis II, which just set a new record for human distance from planet surface. The Moon of our anticipation is not a low-gravity penal colony for losers and rejects, but a militarized resource-extraction site where technology can scale at speed far beyond the confines of Earthly constraint. Rather than Mad Max types or congenital dissidents, the Loonies of our approaching future are borglike supermen, figures much more apt to remake the homeworld at a stroke in their image than to want to be left alone.

When the Moon or Mars are settled, will today’s type of American recognize the colonies that reflect back at us from across the void? Or is the more likely scenario one where stay-behind woke communists brace for impact from warlike fascists descending like gods from the skies?

We must work through such dizzying considerations if, so against the historical odds, America hopes to retain its character after a frontier colonization campaign of unprecedented reach and technological power. The longing to be left freely alone that animated the future on Heinlein’s Moon has given way to the rather different desires more at home among his Starship Troopers. The mimetic experience emerging from full immersion in our technology is one that arouses an even greater yearning to invade than to escape. What becomes of acceleration if, after we slip the bonds of our worldly home, we find ourselves trammeled, wherever we travel, by the given limits of our own human selves? And what becomes of ourselves on Earth if, off world, our not-so-distant cousins forge a warlike borg of high-agency supermen, a posthuman entity ready and willing to pay us a visit like the zombie son in the monkey’s paw fable that so tormented Norbert Weiner, the godfather of cybernetics himself?

Having “overcome” its humanity, the entity would surely have overcome politics, or at least ideology, if only in the horseshoe sense. The history of ideology, after all, suggests nothing so much as a stubborn iteration of differently-coded attempts to retool humanity in the image of our own tooling. Comte’s positivism, Bentham’s utilitarianism, Emerson’s Man of the World, Nietzsche’s overman, Trotsky’s New Soviet Man, Hitler’s master race, America’s own Superman – what is “the political spectrum” to such contenders? The meta-political is all – the copied, cloned, upgraded mirror-environment of the original, just as the meta-human out-iterates the human until the latter recedes behind the vanishing point.

Beyond the Seeing Stone

Baudrillard leaves us with the image of the revenge of the mirror people – a tremendous mutant excrescence of hatred against a society that banished them into the crystals of virtual objectification: “every representation is a servile image,” he writes – citing Borges (The Book of Imaginary Beings) – “the ghost of a once sovereign being whose singularity has been obliterated.”

But a being which will one day rebel, and then our whole system of representation and values is destined to perish in that revolt. This slavery of the same, the slavery of resemblance, will one day be smashed by the violent resurgence of otherness. We dreamed of passing through the looking-glass, but it is the mirror peoples themselves who will burst in upon our world. And ‘this time will not be defeated’... everything which serves to provide a passive reflection in a world based on identity is ready to go on to the counter-offensive. Already they resemble us less and less… I’ll not be your mirror!

It is strange, but not unexpected, that to the counter-offensive mirror people, those on the other side themselves appear to be the deeply offensive mirror people who have crowned themselves the royals and hieromonks and god-emperors of a cloned and copied and virtual mirror world. Like Narcissus, we now encounter ourselves – and our future – through the confrontational divide of the reflecting pool. Recall that Narcissus was not condemned to love only himself forever but to waste away (or kill himself) because he could not fully possess himself. The existence of what Baudrillard calls le Même – the selfsame – was foreclosed to Narcissus: now it tantalizes, obsesses, threatens to define us even before we get any closer to making ourselves still more identical to it. “A copy of a copy of a copy” – what Tyler Durden was created by his original human in order to destroy. To break the circuit.

The meme of the multiple Spider-Mans accusatorily identifying one another takes shape today as a curse, one we already have begun to disappear into and which somehow we now must find a way to look outside of and beyond. The two extremes on either side of the mirror – with diminishing accuracy and diminishing difference do we call them fascist and communist – make what can only be (to an American) false promises to break the curse. And somewhere, in the shadow of one church or another, a quiet detective prepares to crack the case of the post-American invaders.