But here is what no form told Sarah. The same week she enrolled Maya in one account, she remained unenrolled in a dozen others — the county’s nurse home-visiting program, expanded WIC benefits she qualifies for, early-intervention screening, childcare subsidies. Those programs won’t show a balance when Maya turns 18. They will determine whether Maya can actually use the balance she receives.

This is the central paradox of Trump Accounts as they open across America: we have just made a multibillion-dollar collective bet on children’s financial futures while systematically underinvesting in the developmental foundations those futures require.

The Promise and the Gap

Trump Accounts rest on a powerful and correct insight: ownership matters. Roughly four in ten American adults own no stock at all, and for their families, Wall Street’s gains are abstractions rather than shared prosperity. Giving every child a stake in American capitalism is a genuine attempt to bridge that divide — and the response has been real. Treasury reports nearly six million children were signed up before the July 4th launch. The Dell family has pledged $6.25 billion to seed accounts with $250 for some 25 million children born too early to qualify for the federal deposit. Ray and Barbara Dalio pledged $75 million more for Connecticut children.

The administration’s projections are ambitious: with maximum family and employer contributions, an account could exceed a million dollars by age 28. Even seed-only accounts could reach roughly $5,000–6,000 at 18 by the government’s estimates — enough for education costs or a first step toward a home, the uses the law favors. (Independent analysts note the real purchasing power will be lower, and that other withdrawals before retirement carry a tax penalty — but something is genuinely better than nothing.)

Here is what the promotional materials don’t say: by the time Maya opens her account in 2043, the foundational architecture determining whether she can leverage that capital will have been built — or compromised — seventeen years earlier, during a period our public investment largely ignores.

The Science of the First 1,000 Days

From conception through age two, a child’s brain undergoes its most dramatic transformation. During these First 1,000 Days, the brain forms more than one million new neural connections every second, and grows to roughly 80 percent of its adult size by age three — establishing the biological infrastructure for everything that follows: executive function, emotional regulation, learning capacity, stress response, and the ability to form relationships.

Three factors largely determine whether that construction builds capacity or fragility: maternal mental health during pregnancy and postpartum, adequate nutrition, and stable, responsive caregiving. When maternal depression goes untreated, stress hormones affect fetal brain development. When infants experience food insecurity, inadequate nutrition constrains both physical and cognitive growth. When caregiving is chronically disrupted — by unpredictable work schedules, absent childcare, untreated mental illness — children’s stress-response systems become dysregulated in ways that persist.

These are not soft problems with soft consequences. Landmark longitudinal research on intensive early-childhood programs — the work for which economist James Heckman is best known — finds returns on the order of seven dollars in societal benefit for every dollar invested, through higher earnings, better health, lower crime, and reduced remedial spending. The inverse is equally true: a Trump Account cannot retroactively repair neural architecture that wasn’t built at age two. Money at 18 cannot buy back a million connections per second disrupted sixteen years earlier.

The Participation Gap

The program’s design is more forgiving than early critics feared: families can claim the $1,000 anytime until the year their child turns 17, online at trumpaccounts.gov without filing a tax return. There is no hard deadline that permanently excludes a newborn.

But a soft barrier can exclude as effectively as a hard one. The families least likely to claim the account — those not required to file taxes, without stable addresses, without time or trust to navigate a federal portal — are disproportionately the same families whose children face the steepest developmental headwinds. Roughly 40 percent of American births are covered by Medicaid; nearly one in five families eligible for the Earned Income Tax Credit never claims it. If Trump Account participation follows the same pattern, hundreds of thousands of the children with the most to gain will be left out — not by statute, but by the same fragmentation that keeps families out of every other program.

The Coordination Crisis

Because here is the frustrating part: America already runs the programs the First 1,000 Days require. WIC for nutrition. Medicaid covering some 40 percent of births. Evidence-based nurse home visiting, which randomized trials show reduces maternal depression and improves children’s cognitive outcomes. Early Head Start. Childcare subsidies.

These programs work. They just don’t work *together*. Each requires a separate application with separate eligibility rules and separate enrollment windows; no agency is responsible for connecting a family to all of them; and families under the most stress have the least capacity to navigate the maze. A mother like Sarah typically qualifies for five or more early-childhood supports and accesses one or two — not because she doesn’t need the rest, but because nobody told her they exist and no system enrolled her.

The Integration Opportunity

The moment Sarah filed Form 4547, she handed the government everything needed to change that: her name, address, income context, and Maya’s Social Security number — the exact data points that determine eligibility for nearly every early-childhood program. Trump Account enrollment could function as the front door to the entire First 1,000 Days system: one filing triggering automatic screening for WIC, home visiting, childcare subsidy, maternal mental-health support, and Medicaid — with a navigator to follow up.

Nothing in current federal policy does this. The philanthropic pledges — the Dells’ billions, the Dalios’ millions — put money *into* accounts; none yet connects account enrollment to developmental services. That is the unbuilt bridge, and it is buildable at the state and county level without waiting for Washington: Pennsylvania could link Form 4547 data or birth registration to a single coordinated eligibility screen tomorrow, and any county could pilot it.

The Cost Question

Can we afford it? The federal seed program costs roughly $14–15 billion over four years — about a twentieth of one percent of federal spending. Comprehensive First 1,000 Days support — universal home-visiting access, fully funded WIC, maternal mental health integrated into primary care, childcare assistance for low-income families — runs an estimated $30–40 billion a year, well under half a percent of the budget. This is not a resource constraint. It is a priority constraint. We have just demonstrated, with bipartisan enthusiasm and billionaire participation, that America can mobilize billions for children’s accounts. The question is whether we will apply the same will to the foundations the accounts require.

Two Futures

In 2043, Maya and millions of children born in 2025 will open their accounts into one of two Americas — and we are choosing between them right now.

In one, Maya finds several thousand dollars. But she struggled in school because early learning deficits were never addressed, faces employment instability because emotional-regulation challenges were never treated, and carries chronic health conditions traceable to her earliest years. She has an asset without the capability to leverage it. The check clears; the foundation doesn’t.

In the other, Maya opens the same account — but her mother received home visiting that caught postpartum depression early, WIC kept nutrition steady, quality childcare held while her parents worked, and early intervention addressed a speech delay at two instead of ten. She has both the asset and the capability. Same account. Different foundations. Radically different lives.

What You Can Do — This Month, in This Region

If you’re expecting or have a child under 18: Claim the account at trumpaccounts.gov (children born 2025–2028 get the $1,000; any child under 18 with a Social Security number can open one). Then take the same care with the accounts that don’t show a balance — the ones that build the child during the first 1,000 days. If you’re in Montgomery County, meet Ada, Moonshot Press’s free, whole-person health guide for pregnancy, postpartum, and early childhood, at thrivemontco.org/meet-ada; Ada can help you understand what your family may qualify for and connect you to trusted local resources — WIC, nurse home-visiting, childcare assistance, and maternal-health support. (Ada is a guide, not a clinician; she doesn’t diagnose or replace medical care, and for emergencies you should call 911 or 988.) You can also dial 2-1-1 or contact your county’s Early Intervention line to ask, in one conversation, about the same programs, and in both Montgomery County and Philadelphia organizations like the Maternity Care Coalition exist precisely to help families navigate all of it at once.

If you want this fixed for everyone: Ask one question of your county commissioners, your state representative, and the candidates on your November ballot — “Will you support making birth registration and Trump Account enrollment an automatic screening point for every early-childhood program a family qualifies for?” It is a specific, answerable, inexpensive question — and their answer will tell you what they believe children are owed.

If you’re an employer: The law lets you contribute up to $2,500 a year to employees’ children’s accounts tax-free. Pair it with the benefit that matters even more in the first 1,000 days: schedule stability and parental leave.

As America marks its 250th anniversary, Trump Accounts acknowledge a real injustice — that a child’s ZIP code shapes their trajectory more than their potential — and attempt to answer it through ownership. That is genuinely valuable. But ownership without capability is an empty promise. If we can mobilize billions for accounts that open at 18, we can mobilize the will for support that begins at birth.

We’ve made the down payment on our children’s financial futures. The investment in their human futures is still owed — and the window for Maya’s generation is open for exactly 1,000 days.

About this story: “Sarah” and “Maya” are illustrative composites representing circumstances common to thousands of Pennsylvania families; they are not real individuals, and no quotes are attributed to them. Program facts were verified against IRS and U.S. Treasury sources, and reporting by NPR, CNN, AP, and the Connecticut Mirror, as of July 2026. AI assistance: background research, source verification support, and draft revision were AI-assisted under the Moonshot Press Constitution; editorial judgment, framing, and final responsibility are human.

On September 17, 1787, as delegates filed out of Independence Hall having just finished the most consequential four months of drafting in American history, a woman named Elizabeth Powell stopped Benjamin Franklin on the street and asked him plainly: what had they built — a republic, or a monarchy? Franklin’s answer has outlived nearly everyone in that room: “A republic, if you can keep it.” We know the line mostly secondhand, through a diary entry written by a delegate who heard it from someone else. That doesn’t make it any less true. It’s survived 250 years for a reason.

Here’s the part of the story that gets skipped: Franklin and the other framers handed the country an extraordinary piece of design — a system of checks, balances, and distributed power built to outlast any single generation’s mistakes. And then they went home. There was no second document. No instructions for the citizen. The Constitution tells you how a bill becomes a law. It does not tell you how to actually be a citizen — how to read a local budget, how to show up at a hearing, how to tell the difference between a real check on power and a performance of one. That part was left as an assumption, not a plan. James Madison believed an informed, engaged public would simply rise to the task, generation after generation, on its own.

For a long time, that assumption more or less held. It’s not holding as well lately. The most recent national civics survey found that seven in ten American adults can name all three branches of government — which sounds fine, until you notice it means nearly a third of us can’t, two and a half centuries after the branches were invented. That’s not a story about any one political side failing. It’s a story about a manual that was never written finally catching up with us.

This is the sentence Moonshot Press was built around: the promise of the Declaration of Independence isn’t history. It’s unfinished work.

And “unfinished” isn’t a metaphor here — it’s a project status. A republic that runs on an engaged citizenry needs the tools of engagement to actually exist somewhere, for someone, in plain language, on a timeline. Not as a slogan. As infrastructure. That’s the gap Franklin’s line has been quietly naming for 250 years, and it’s the gap we built this organization to close.

We’re not the first to notice it, and we won’t be the last to try. But we think 2026 — the actual 250th anniversary of the Declaration — is the year to stop treating “a republic, if you can keep it” as a warning to frame on a wall, and start treating it as a to-do list.

So on July 30th, Moonshot Press launches — not as one announcement, but as a sequence. Starting that Thursday, we roll out one moonshot a week: a citizen commission on artificial intelligence and the future of work; an initiative built around the first 1,000 days of a child’s life; a project reimagining what a health system built around health, not just disease, would actually look like; a civic health initiative reframing this year’s midterms as a 250-year checkpoint; a generation-length peace moonshot; and a hyperlocal pilot in Montgomery County, Pennsylvania, tested in person, with real neighbors in a real room. Different subjects, same premise underneath all of them: give citizens the tools the founders assumed we’d already have, and trust people to use them well.

We’re doing this in the open, including with artificial intelligence — the same technology that could just as easily be used to make citizens more confused, more isolated, and more manipulable than any generation before them. We think that outcome isn’t inevitable. It’s a choice about who holds the tool and what they’re building with it. So we built a public constitution for how we use AI, and the rule inside it is simple: AI recommends, humans decide, always. We’re not asking you to trust that quietly. We’re publishing it so you can hold us to it.

Thomas Jefferson, writing to Madison two years after the Constitutional Convention, put it more bluntly than most founders allowed themselves: “The earth belongs to the living.” He meant that no generation gets to finish this on behalf of the next one — every generation has to actually do the keeping, themselves, in their own time. Franklin knew it too. That’s why he answered a stranger’s question in the street instead of a grand toast to the delegates. Keeping a republic was never going to be a ceremony. It was always going to be a practice — unglamorous, ongoing, and everyone’s job at once.

Two hundred fifty years on, nobody ever finished the manual. We think that’s exactly the work worth doing next.

Show up. Deliberate. Demand. Vote. Repeat.

Join us July 30th.

Learn more about The Moonshot Destinations

A moonshot is a goal audacious enough to look impossible, achieved anyway through sustained collective effort — Kennedy’s Apollo pledge is the model everyone reaches for, and rightly. But I’d argue the American experiment itself was the original moonshot. A declaration of inalienable rights and popular self-government was every bit as audacious in 1776 as a rocket to the moon was in 1962, and it demanded exactly the same thing: not one triumphant moment, but generations of sustained work.

So we organized Moonshot Press around a set of specific, dated moonshots — not abstractions, but real commissions and initiatives with real timelines:

• The People’s Commission for Technology and the American Future, building a new social contract for an AI-disrupted economy

• The People’s Commission to Make Our Children Healthy, centered on the First 1,000 Days of life — because the vast majority of brain architecture is built before age two, which makes early childhood civic infrastructure, not just a family matter

• Towards a Health-Promoting System, asking what it would mean to build healthcare around what creates health, not only what treats disease

• The Declaration to Constitution Project, reframing this year’s midterms as a 250-year checkpoint on the health of the political system itself

• The Middle East at 2048, a generation-length peace moonshot anchored to a symbolic horizon: the 2048 FIFA World Cup

• Thrive in Montco PA, our hyperlocal proof of concept — testing whether any of this actually works, starting in one Pennsylvania county, in person, with real neighbors

Mayo Clinic and Microsoft’s June announcement is more than another AI partnership. It signals the rise of branded AI medicine: elite health systems turning their clinical know-how, longitudinal data, and institutional prestige into reusable foundation models that can be distributed far beyond their campuses. That could improve care. It could also harden a narrow, high-cost, institution-centered vision of healthcare just when the field most needs a whole-person, community-centered alternative .

The branded-AI play

A foundation model is a broad, reusable AI system trained on very large datasets and then adapted to many downstream tasks. In healthcare, that means one model might support triage, documentation, risk prediction, patient education, and clinical reasoning across settings (He et al., 2024). Mayo and Microsoft say their new model will combine Mayo’s de-identified clinical data and longitudinal insights with Microsoft’s cloud and AI capabilities; the model will be owned by Mayo and made available through Azure APIs . In plain terms, the hospital brand is becoming part of the product.

The idea has a history. IBM’s Watson was famously “sent to medical school” in 2012 as an earlier attempt to package high-end medical cognition into software for clinicians and health insurers (Wired, 2012). What is different now is scale: today’s models are multimodal, cloud-native, and designed to sit inside every layer of care delivery. The aspiration is not just a better tool, but a platform that can extend an institution’s reach, reputation, and business model.

Why the race is on

The rush is being driven as much by economics as by science. Big Tech and health systems see AI as a way to cut administrative burden, reduce clinician burnout, improve throughput, and defend margins. Reuters reported this spring that Amazon launched an AI platform for scheduling, intake, documentation, and coding; in 2024 Reuters also reported that Suki’s AI assistants had reached more than 300 health systems. That is the market signal: healthcare AI is becoming infrastructure, not a side project .

For elite providers, branded models offer a second advantage: competitive insulation. If AI becomes the front door to healthcare, institutions will want their own protocols, data assets, and brand voice embedded in the machine. But that logic creates a public problem. The more value accumulates around proprietary datasets, cloud contracts, and dominant record systems, the easier it becomes for today’s fragmented healthcare market to harden into tomorrow’s AI oligopoly.

What elite data misses

The central weakness of branded AI medicine is representativeness. Even the best tertiary center does not possess a monopoly on health-relevant data, because health is shaped not only by diagnoses and procedures but also by housing, food access, transportation, caregiving, digital literacy, stress, and trust. Reuters reported in 2025 on a Nature Medicine study showing that large healthcare AI models sometimes changed testing and treatment recommendations based on patients’ socioeconomic and demographic characteristics even when the clinical details were identical. That is a warning sign for any model trained mainly in elite settings .

The Mayo announcement promises to “expand access,” but it does not specify how the model will reach Medicare and Medicaid populations or safety-net settings. That omission matters. If a model is optimized around highly resourced referral care, it may perform elegantly for the digitally connected and commercially insured while offering less help to the people who most need navigation, continuity, and context. Reuters has also reported that Epic controls health data on up to 94% of Americans according to allegations in ongoing antitrust litigation, underscoring how easily AI power can concentrate around a few chokepoints in records and distribution .

Consider one illustrative case: a rural Medicaid patient with heart failure, diabetes, depression, and no reliable broadband. A branded model trained on elite tertiary-care flows might generate an impressive medication and referral plan. A whole-person model would also ask whether the patient can get to the pharmacy, afford healthy food, keep the phone charged, trust the care team, and follow up without losing a day’s wages. That is the difference between optimizing treatment and building health.

The whole-person path

A salutogenic alternative starts from a different question: not only how to detect disease earlier, but how to strengthen the conditions that help people stay well. That requires integrating clinical data with behavioral, social, and community information, and it argues against institution-sized data silos. Reuters reported in 2025 that Truveta’s multi-system data effort brought together 17 health systems, Microsoft, Regeneron, and Illumina to build a large-scale genomic and clinical resource. Whatever one thinks of that project, it points toward a necessary principle: healthcare AI should be built from interoperable, governed, plural data ecosystems, not just elite brands .

The policy agenda follows from that. We should require independent bias audits, public-interest data governance, interoperability across institutions, and community-benefit expectations for healthcare AI trained on clinical data. The educational agenda is just as important: physicians and health systems need competencies in AI critique, whole-person assessment, shared decision-making, equity, community resource navigation, and responsible governance. In the AI era, the highest competence may not be faster pattern recognition, but wiser stewardship of people, systems, and health itself.

Moonshot Press is building toward that future through the Whole Person Health Assistant initiative: tools designed not merely to optimize episodes of care, but to help citizens and clinicians co-create health across the full arc of daily life. If you want AI that serves whole-person flourishing rather than branded scarcity, this is the conversation to join.

The Question That Guides Us

Moonshot Press and Project 2026 approach this historic moment with one orienting question:

How do we build a system where everyone can get the care they need to live healthy, flourishing lives?

We believe the answer lies in Citizenism—the idea that the future of technology and health must be co-authored by the people it serves. AI has the potential to drastically reduce costs, personalize wellness, and free up our doctors to focus on the human connection of healing. But this potential can only be realized if we, the citizens, guide its development.

Join the Effort

We are currently finalizing our team’s response to the HHS RFI, and we want you to be part of this journey. This is a “Moonshot” for the health of our nation, and it requires all of us.

• Engage: Follow our updates as we break down the complexities of AI and health policy into language we can all use.

• Participate: Your voice is the most important data point in existence. Help us demonstrate that the American people want a healthcare system built for flourishing, not just for processing.

• Aspiration: Let’s move beyond “fixing” what is broken and start building what is possible.

The year 2026 marks a milestone for our country. Let’s make it the year we reclaimed our health and used the power of technology to ensure that every American has the opportunity to thrive.

In health and solidarity,

The Moonshot Press Team Project 2026: Toward a Flourishing Nation

Eleven years ago, Eric Schmidt heard a warning about pitchforks and dismissed it. Last week, ten thousand graduates booed him off the stage. The country he helped build is sending him a message — and he is not the only one receiving it.

In a private dining room in 2015, the DeepMind co-founder Mustafa Suleyman gave a presentation to Google’s board called The Pitchforkers Are Coming. He argued that artificial intelligence would, within a decade, displace tens of millions of jobs, explode disinformation, and produce a public fury that would make every previous tech backlash look like a rehearsal. He told the room that Google — and the broader industry — would have to share the wealth with the people whose livelihoods the technology was going to consume.

Larry Page, in his customary whisper, said AI would create more jobs than it took. And then Eric Schmidt — the man who had run Google for a decade, who was at that moment one of the most influential technologists alive — told Suleyman, in so many words, that the worry was overblown.

On May 14, 2026, eleven years almost to the month later, Eric Schmidt walked onto a stage at the University of Arizona’s commencement to address roughly ten thousand graduating seniors and their families. He began the AI portion of his remarks. “If science is not your passion, if you don’t care about science, that’s okay, because AI is going to touch everything else as well,” he said. The crowd began to boo. He pushed forward. “When someone offers you a seat on the rocket ship, you do not ask which seat. You just get on.”

The boos grew louder. “The question is not whether AI will shape the world,” he declared — and the heckles rose again. University marshals urged calm over the loudspeakers. The graduating class of 2026, the cohort whose entry-level jobs are right now being absorbed by the technology Schmidt was selling them, sat in the desert sun and booed the former CEO of Google for eight straight minutes.

This is what it looks like when a forecast comes due.

The forecast comes due

The Arizona scene was not an isolated outburst. Earlier in May, the real estate executive Gloria Caulfield was booed at the University of Central Florida’s commencement after telling graduates “the rise of artificial intelligence is the next industrial revolution.” Stanford students protested Sam Altman at a campus appearance the same month. Last month, someone threw a Molotov cocktail at Altman’s San Francisco house and another person fired a gun at his property a few days later.

These incidents are no longer surprising. They have become a recognizable category. Jasmine Sun, writing in The New York Times Magazine, called them AI populism’s warning shots — the noises a community makes before the ground gives way. The 2015 dinner is the hinge. The 2026 boos are the consequence. Between them sits a decade in which the people best positioned to govern the technology decided, instead, to ship it.

The Arizona reception is particularly worth dwelling on because of who Schmidt is and what he was trying to do. He was not an outsider trying to provoke. He was a billionaire former CEO, an invited honored guest, giving the kind of broadly optimistic commencement speech that has been part of American academic ritual for a century. He told students that being skeptical of an AI-driven world, even fearful, was understandable, but argued that people still had choices about how the technology develops. He urged them to embrace open debate and the immigrant tradition that made the country great. He said the future was unwritten and was theirs to shape.

None of it landed. “Critics and attendees later described the message as ‘AI cheerleading,’” while the graduates in front of him were preparing to enter a labor market that the speaker himself had spent his career helping to remake.

The graduating class understood something Schmidt apparently did not. The choice he kept urging on them — help shape the technology — was a choice that had already been made, in rooms they had never entered, by people who had not invited them. They were being told they had agency over a future whose terms had been set without them. The boos were not a rejection of AI in some abstract sense. They were a refusal of the bargain in which the costs are distributed to one group and the benefits to another, dressed up as a graduation gift.

Why this anger now

It would be a mistake to read the Arizona scene as a momentary mood. The class of 2026 is graduating into a labor market whose contours their parents would not recognize. According to data cited by recent campus surveys, 52 percent of 2026 job postings require AI familiarity, entry salaries for humanities majors fell 8 percent year over year, and 61 percent of students surveyed across seventy campuses said they fear automation displacement.

These are not the numbers of a generation that misunderstands the technology. These are the numbers of a generation that understands it exactly.

The understanding extends well beyond college campuses. An Axios analysis in January 2026 found that the country is splitting into three distinct economic realities — the Have-Nots, who are stalling; the Haves, who are coasting; and the Have-Lots, who are rocketing to greater wealth through exclusive access to private AI deals, massive investment power, governmental connections, and equity stakes that ordinary investors cannot touch.

The Have-Lots are not a marginal phenomenon. Among the fifty richest Americans, the median 2025 increase in net worth was nearly $10 billion — a 22 percent median gain in a year when the S&P 500 rose 16 percent and Treasury bills returned less than 4 percent. Elon Musk’s wealth rose by $187 billion in a single year, to over $600 billion. During the AI bounce of the past two years, the top 10 percent of households saw their wealth increase by $5 trillion in a single quarter, while the bottom 50 percent gained $150 billion.

Pause on those two numbers. Five trillion to the top tenth. One-thirtieth as much to the bottom half. In a single quarter. This is what an “AI boom” looks like at ground level. It is what the graduates in Arizona were responding to. It is what the residents in Tremonton, Utah are responding to when they show up at county commission meetings by the hundreds to oppose data center projects in their towns. The technology has become inseparable from the wealth concentration that is funding it, and the wealth concentration has become inseparable, in ordinary perception, from the technology.

In the third quarter of 2025, the top 1 percent of U.S. households owned 31.7 percent of all U.S. wealth — roughly as much as the bottom 90 percent combined, the widest gap since the Federal Reserve began collecting the data in 1989. The billionaire investor Ray Dalio has warned that this kind of concentration produces, in his words, irreconcilable differences that democratic order is not equipped to handle. Peter Mallouk, who runs a $700 billion wealth management firm, has called the current distribution “100 percent completely unsustainable as a society.”

This is not the language of the radical left. It is the language of the people who have done extraordinarily well from the current arrangement, looking at the numbers, and concluding that the arrangement will not hold.

Even some of the AI builders themselves have begun to say so. Anthropic CEO Dario Amodei announced in late January 2026 that he and his co-founders would give away 80 percent of their wealth, writing: “The thing to worry about is a level of wealth concentration that will break society.” Amodei noted that Elon Musk’s nearly $700 billion net worth already exceeds John D. Rockefeller’s at the height of the Gilded Age — and that, he argued, is before most of AI’s economic impact has even materialized.

That is a striking admission from someone running one of the three largest AI companies in the world. The concentration is already historic. The technology that is going to drive most of the additional concentration has barely begun to be deployed. And the man saying so is the CEO of the firm doing the deploying.

The graduates in Arizona were not booing a misunderstanding. They were booing a description of their own future delivered by one of its principal architects.

The geography of the anger: the data center revolt

The Arizona commencement was a national event because it happened on camera. But the Arizona boos, however striking, are a relatively low-stakes version of what is now happening in hundreds of communities across the country, where AI has stopped being an abstract policy debate and has become a concrete proposal to put a fenced industrial compound, drawing as much electricity as a small city and as much water as a midsize town, next to where people live.

The opposition is real. It is well-organized. And it is winning.

A platform called Data Center Opposition is now tracking 268 local protest groups across 37 states, representing roughly 360,000 followers, with the dataset updated monthly. Since 2024, dozens of community-led campaigns have emerged in opposition to AI data centers, driven by concerns about energy use, energy costs, noise pollution, water consumption, and air pollution. In 2025 alone, local opposition led to the delay or cancellation of projects totaling $156 billion.

In the first four months of 2026, more than 70 data center projects were rejected or restricted by local authorities — more than in all of 2025 combined. The numbers do not include projects withdrawn before formal rejection. Compass Datacenters withdrew plans for an 800-acre project in Prince William County, Virginia, last month after facing intense pushback from local residents. A February analysis by Sightline Climate estimated that 30 to 50 percent of the data center capacity expected to come online in 2026 may not be delivered on schedule, citing power availability, permitting challenges, and increasingly organized local opposition.

In Tremonton, Utah, on May 4, hundreds of protesters gathered outside the Box Elder County Commission meeting to oppose a massive AI data center project backed by, among others, the television personality Kevin O’Leary. The commissioners approved the project anyway. Rural residents are now organizing to put the question on the November ballot. Outside a similar meeting elsewhere, 300 protesters chanted “No data center” and “We want water.”

The Spectator’s Robert Bryce, who has spent years tracking opposition to renewable-energy projects and has a reasonable claim to being one of the country’s better-informed observers of energy-siting politics, made the following observation about the new data center fights: “The rage against data centers is different, not least because it is more widely shared. People all across the US are angry. They don’t like the super-rich tech oligarchs, they don’t trust Big Tech and they are ready and willing to fight to stop AI data centers from coming into their cities, towns and rural areas.”

This is the political fact that should be giving every state legislator and every governor pause. The data center fights are not a typical NIMBY dispute about property values and views. They are the leading edge of a broader populist refusal — bipartisan, cross-regional, organized — of the entire bargain that the AI industry has been trying to impose on American communities.

The bargain, as the industry has presented it, runs roughly like this. We will build a billion-dollar facility in your town. We will pay almost no local taxes because of incentive packages your state already gave us. We will draw enormous quantities of electricity and water from systems your residents depend on. Your household electric bill will go up because the grid is now serving us. We will provide, in exchange, perhaps a few dozen ongoing jobs, most of them not for people from your community. The benefits of what we build will flow to shareholders thousands of miles away. The risks will be borne by you.

When that bargain is described in plain language, the political result is not surprising. It is the inevitable. “Now add in distrust — or even outright hatred — of Big Tech and fears about AI destroying jobs, and you get a dream issue set for activists across the political spectrum,” Bryce notes. He is right. The data center opposition is not a left or right phenomenon. It is the place where left and right meet because they are both being asked, in their own communities, to host the infrastructure of an economy whose returns they will not see.

The engine: inequality as the underlying motivation

It is tempting to read each of these stories — the Arizona commencement, the Tremonton county meeting, the Prince William County withdrawal — as a discrete event with a local cause. That reading is not wrong, but it is partial. Beneath each individual confrontation runs a single, larger current, and that current is the perception, increasingly grounded in documented economic data, that AI is functioning as a wealth-concentration engine of historic proportions and that ordinary Americans are being asked to subsidize the engine while being excluded from its returns.

This is why the anger is not satisfied by the standard reassurances. AI will create more jobs than it takes. (Schmidt said this in 2015. Page said it in 2015. They have been saying it ever since. The graduating class of 2026 has been told to expect those jobs their entire post-pandemic lives. The jobs have not arrived. The lay-off announcements have.) The benefits will be broadly shared over time. (Over how much time? Shared by whom? Through what mechanism? When the mechanism is left unspecified, citizens are entitled to assume the mechanism does not exist.) The technology is too important to slow down. (Important to whom? Slowing it down is exactly what the affected communities are now demonstrating they have the political capacity to do.)

The disconnect between elite messaging and public perception is not a failure of communication. It is an accurate read by the public of what the messaging is actually for. Anthropic, which launched its competitor to ChatGPT in March 2023 at a $4.1 billion valuation, was valued less than three years later at $350 billion — an 87-fold increase. OpenAI, in the same period, surged to a $750 billion valuation. Last year alone, global billionaire wealth jumped by 16 percent to $18.3 trillion — the highest level in history, according to Oxfam.

These valuations and wealth surges are happening inside an economy in which median household wealth has barely moved, in which entry-level white-collar hiring is contracting in the very fields AI is best at, in which the cost of housing and healthcare and child care has decoupled from wage growth for a decade, and in which the rationale offered by the system’s beneficiaries amounts to a request for continued patience.

The patience has run out. That is what last week in Arizona was about. That is what Tremonton was about. That is what the Molotov cocktail at Altman’s house and the gunfire that followed were warning shots toward.

What the AI industry is not understanding about Schmidt’s reception

The most striking thing about the Schmidt commencement is not that he was booed. It is the apparent surprise of the people around him that he was booed.

Schmidt gave a speech that, in a different era — perhaps as recently as five years ago — would have been received as inspirational. Embrace the future. The future is yours to shape. Some perspectives are uncomfortable but worth hearing. America is at its best when ambitious people want to come here. These are the platitudes of the American technological optimist tradition, recited with the practiced ease of a man who has given some version of this speech a thousand times.

The platitudes failed because the audience now hears them differently. Embrace the future sounds, to a young person whose first job has just been automated, like accept what we have done to you. The future is yours to shape sounds, to a county commissioner staring down a $14 billion data center proposal, like take responsibility for the consequences of decisions you did not make. When someone offers you a seat on the rocket ship, you do not ask which seat — you just get on sounds, to anyone who has been paying attention, like exactly what the billionaire class would say to people whose only available role in the project is to provide the launch pad.

The speech failed for the same reason the 2015 dinner failed. The people delivering the message do not actually inhabit the world to which the message is being delivered. They live in a separate America — geographically, economically, socially, in some cases legally and politically — and their estimates of what the broader country will tolerate are systematically miscalibrated. Schmidt was making a case for human agency over software, the trade publication WinBuzzer observed afterward, while many people in the stadium were closer to the practical question of who gets hired first, who gets trained first, and how much junior work remains once software takes over routine tasks.

That is the gap. The man on the stage was making an argument about whether AI would shape the future. The people in the seats had already conceded that it would. Their question, the only question that mattered to them, was who would shape the AI — and on whose terms, and in whose interest, and at whose cost.

Schmidt had no answer to that question, because the answer he might honestly have given is not one he is in a position to deliver. The honest answer is that AI is currently being shaped by approximately the same several hundred people who shaped the previous two decades of the platform economy, that those people are accountable to almost no one outside their own boards, and that the institutions that might once have provided that accountability — Congress, the regulatory agencies, the press, the universities — are either captured, paralyzed, or themselves dependent on the industry they are supposed to oversee. The graduating class of 2026 knows this. That is why they were booing.

What an adequate response now looks like

A reader of these essays could be forgiven for asking, at this point, what exactly anyone is supposed to do. The polling is hostile. The technology is being shipped. The wealth is concentrating. The institutions are sluggish. The data center revolt is winning in particular places but is not yet adding up to a national framework. What do you actually do?

Several things, none of them sufficient on its own, all of them building toward the kind of public infrastructure that could absorb the technology without breaking under it.

Make the bargain explicit and democratic. The fundamental problem with the current AI rollout is not that it is happening. It is that it is happening without any negotiated bargain between the public and the firms doing the rolling out. A democratic society can absorb a great deal of disruption when the terms of the absorption have been agreed to. It can absorb almost none when the terms have been imposed. Every major industrial transition in American history has eventually produced a bargain — the New Deal labor settlement, the postwar GI Bill, the environmental and consumer protection regimes of the 1960s and 1970s. The AI transition has not yet produced one. Building one is the central political task.

Tax the concentration. Some portion of the extraordinary returns being captured by the firms and individuals at the top of the AI economy must be redirected, through public mechanisms, toward the workers and communities absorbing the costs. The specific instrument matters less than the principle: automation taxes, public wealth funds modeled on the Alaska Permanent Fund, productivity dividends, excess-profits levies during the transition years. The argument that the underlying science was publicly financed and the returns should be at least partially publicly shared is not a radical argument. It is the same argument that produced the workers’ compensation system and the antitrust framework, both of which were called radical in their day and are now considered the floor of a functioning industrial society.

Govern the data centers as the public infrastructure they actually are. The current siting regime, in which a single private firm can negotiate the terms of an enormous power and water draw with a single county commission, is not adequate to the scale of what is being built. State-level frameworks — like those now emerging in Illinois, Texas, Colorado, and Virginia — that require host-community benefit agreements, ratepayer protections from grid impacts, environmental review at the scale of the actual project, and meaningful local consent are the minimum. The communities now doing this work by their own organizing have demonstrated that the political capacity exists. The legal and regulatory frameworks have not yet caught up.

Fund worker transitions at the actual scale of the displacement. A $90 million federal allocation against a projection of 300 million globally displaced jobs is not a policy. It is a gesture. A GI Bill–scale commitment to retraining, portable benefits, automatic safety-net stabilizers when displacement metrics cross thresholds, and serious income support during transitions is what the moment calls for. The federal vacuum on this question is being partly filled by states. It needs to be filled by Congress.

Build the civic infrastructure that translates anger into governance. The data center fights work because hundreds of local groups have organized themselves into something larger. The same kind of organizing, at higher levels of abstraction, is what produced the labor movement, the consumer movement, and the environmental movement of previous American generations. The energy is currently present. The infrastructure that turns energy into legislation is being built in real time, by ordinary citizens, in counties and states across the country. The work is not glamorous. It is the work that has, in every previous democratic course-correction, made the corrections possible.

None of this is finished. None of this is guaranteed. But none of this is impossible either. The graduating class of 2026 has just demonstrated, at very high volume and on national television, that they understand what has been done to them and are not interested in pretending otherwise. The communities organizing against data centers have demonstrated that they can win — not always, not everywhere, but often enough that the developers have begun to notice. The numbers on wealth concentration are now visible enough that even the people benefiting from them are warning, in public, that they will not hold.

This is the raw material of a democratic course-correction. The question is whether the country can organize it in time.

Eleven years

Eric Schmidt had eleven years to choose differently. He was in the room in 2015. He heard the warning. He had the position, the platform, and the resources to act on it. He chose, instead, to call it overblown and to continue building.

The graduating class of 2026 cannot make his choice for him. But they can decline to applaud it. They can decline to accept the rocket-ship metaphor. They can decline to be told that their only available role is to embrace whatever future is delivered to them by people who do not share their condition.

That is what last week in Arizona was. It was not the failure of a commencement speech. It was the beginning of the political response that the 2015 dinner declined to organize. Eleven years late, but not too late.

The pitchforks are not coming. The pitchforks are here. The question now is whether the country uses them — in the historical, democratic sense — to compel a renegotiation of the bargain, or whether the people holding them, finding no functioning channel for their grievance, eventually pick them up in some other sense altogether.

That is the choice in front of us. It is not the choice of the men in the bunkers. It is the choice of the rest of us — in counties and states and union halls and church basements and yes, in commencement audiences willing to make noise — about what kind of country we are going to be on the other side of this.

The boos at Arizona were not the end of something. They were the beginning. The people who heard the 2015 warning and dismissed it are about to find out what eleven years of not listening cost them. And the people who have been doing the listening — quietly, locally, in every place where a data center proposal has been refused and every commencement where a billionaire has been heckled — are about to find out what they can build with the political capacity they have already proven they possess.

Suleyman was right. Schmidt was wrong. And the country, at last, is awake.

Moonshot Press is a project of the Institute for Salutogenesis and a cornerstone of the Democracy, Opportunity and Citizenship Moonshot. We are nonpartisan, constitutionally grounded, and committed to the proposition that an informed citizenry is not a luxury of democratic life — it is its precondition.

This is the third essay in our AI series, following “The Pitchforks Are Here” and “The Promise and the Peril.” Subscribe at thriveinmontco.substack.com.

AI has become one of the most polarizing and high-stakes topics in the United States, shaping debates in business, politics, academia, media, and everyday life. Enthusiasts see it as a once-in-a-century general-purpose technology capable of driving explosive growth and solving humanity’s biggest challenges. Critics warn of overhype, profound societal disruptions, or even existential dangers.

Public conversation is organized into several distinct but overlapping “camps.” These reflect differing worldviews on AI’s capabilities, timeline, benefits, risks, and appropriate societal response. The camps influence investment decisions, policy proposals, corporate strategies, and public opinion. Many individuals and organizations straddle multiple camps, and positions evolve with new technical developments.

A major dimension of the conversation — strongly emphasized by the current U.S. federal government — is strategic competition with China. Public discourse is organized into several distinct but overlapping “camps.” These reflect differing worldviews on AI’s capabilities, benefits, risks, and appropriate societal and geopolitical response. The camps influence investment, policy, corporate strategy, and public opinion. Many voices straddle multiple camps.

Below is a comprehensive master list of the primary camps active in U.S. discourse:

1. AI Optimists / Accelerationists / Techno-Optimists

Position: AI is a profoundly positive, transformative force that will drive unprecedented economic growth, scientific breakthroughs, and human flourishing. Development should be accelerated with minimal regulatory barriers.

Main points: Exponential progress toward AGI will create abundance; over-regulation risks ceding leadership to China; focus on rapid building and deployment.

Representative spokespeople: Marc Andreessen (a16z), Sam Altman (OpenAI), Jensen Huang (NVIDIA), Ray Kurzweil, Elon Musk (xAI).

Supportive media/outlets: a16z publications, The Information, Forbes, WSJ business sections, X/Twitter.

2. AI Skeptics / Hype Critics

Position: Much of the AI boom is marketing hype; current technologies have fundamental limitations and are unlikely to deliver transformative results soon.

Main points: Hallucinations, high costs, modest productivity gains, risk of AI winter.

Representative spokespeople: Gary Marcus, Yann LeCun (on extreme claims).

Supportive media/outlets: MIT Technology Review, academic critiques, skeptical NYT pieces.

3. AI Doomers / Existential Safety Advocates

Position: Advanced AI poses catastrophic or existential risks. Development must be slowed or heavily controlled until safety is assured.

Main points: Misalignment, loss of control, inadequate current alignment techniques.

Representative spokespeople: Eliezer Yudkowsky, Geoffrey Hinton, Yoshua Bengio, Nick Bostrom.

Supportive media/outlets: The New Yorker, The Atlantic, effective altruism platforms.

4. AI Pragmatists / Balanced Realists

Position: AI offers real benefits alongside serious risks. Pursue responsible innovation with smart, evidence-based governance.

Main points: Focus on workforce adaptation, bias, privacy, and proportionate regulation.

Representative spokespeople: Fei-Fei Li, Andrew Ng, Ro Khanna.

Supportive media/outlets: The Economist, Brookings Institution, mainstream policy coverage.

5. Societal Impact / Labor & Ethics Camp

Position: AI will exacerbate inequality, displace jobs, amplify biases, and raise ethical concerns. Prioritize human-centered protections.

Main points: Labor disruption, IP issues, surveillance, power concentration in Big Tech.

Representative spokespeople: Timnit Gebru, Writers Guild advocates, progressive sociologists.

Supportive media/outlets: The Guardian, New York Times societal sections, Wired.

6. AI Democracy & Social Cohesion Defenders (Information Integrity Camp)

Position: Generative AI threatens democracy and social trust through disinformation, deepfakes, and erosion of shared truth.

Main points: Election interference, polarization, “crisis of knowing,” need for watermarking and literacy.

Representative spokespeople: Maria Ressa, Tristan Harris, Daron Acemoglu, Shoshana Zuboff, Brennan Center experts.

Supportive media/outlets: The New York Times, The Atlantic, The Guardian, policy journals.

7. AI Regulation & Governance Camp (The Regulatory Divide)

Position: AI requires governance, but stakeholders are deeply divided on scope, strictness, and level of government.

Sub-camp A: Light-Touch / Innovation-First

Favor federal preemption and minimal rules to protect competitiveness.

Spokespeople: Industry leaders, certain Republican lawmakers.

Media: WSJ, business press.

Sub-camp B: Protective / Risk-Based

Support stronger, enforceable rules with state flexibility.

Spokespeople: State AGs, civil society, certain Democrats.

Media: NYT, Brookings.

8. US-China AI Competition / Strategic Race Camp

Position: AI is a critical great-power competition with China. The U.S. must win decisively to maintain technological, economic, military, and ideological superiority. Strongly supported by the current federal government.

Main points:

• AI is a zero-sum strategic contest; China uses state-driven theft, distillation of U.S. models, and massive investment.

• Maintain U.S. leads in compute, talent, models, and infrastructure.

• Policy tools: export controls on chips/models, IP protection, domestic buildout, allied AI coalitions, and counter-intelligence.

• Prevent China from setting global AI standards or achieving military-civil fusion dominance.

Key Federal Government Elements:

• America’s AI Action Plan (2025): Explicit “Winning the Race” framework.

• Executive Order 14179: “Removing Barriers to American Leadership in Artificial Intelligence.”

• Tightened BIS export controls and sanctions on advanced AI tech to China.

• House Select Committee on the Strategic Competition Between the United States and the Chinese Communist Party (Chairman John Moolenaar) — major hearings on “China’s Campaign to Steal America’s AI Edge.”

Representative spokespeople:

• Rep. John Moolenaar (House Select Committee Chairman)

• Michael Kratsios (Chief Science and Technology Adviser)

• President Donald Trump (administration framing)

• Gregory C. Allen (CSIS), Dmitri Alperovitch (Silverado Policy Accelerator), Matt Pottinger, David Sacks (AI Czar influence)

Supporting groups & think tanks:

• America First Policy Institute, Silverado Policy Accelerator, CSIS, House Select Committee on CCP.

Supportive media/outlets: Wall Street Journal editorials, Fox News, national security-focused publications, congressional releases.

This master list captures the main contours of the American AI public discourse ecosystem as of 2026. The camps are not rigid silos—debates frequently blend concerns (e.g., democracy risks with regulation, or optimism with pragmatic governance)—but they help clarify the competing narratives shaping policy, investment, and cultural attitudes toward AI. The discourse continues to evolve rapidly with technological progress and real-world deployments.

The People’s Commission on Technology and the American Future

That sentence should appear far more often in the public conversation about artificial intelligence than it does. The technology companies building these systems project a transformation so profound that they are spending more than $660 billion on infrastructure in a single year — more than the GDP of Sweden, more than the United States spent building the entire interstate highway system adjusted for inflation, more than any private capital expenditure on any technology in human history. The economists projecting AI’s workforce impact disagree with each other by an order of magnitude: MIT’s Daron Acemoglu, the 2024 Nobel laureate in economics, projects AI will automate roughly five percent of tasks and add about one percent to GDP over the next decade; Goldman Sachs projects it will automate a quarter of all work tasks globally and boost productivity by nine percent. The venture capitalists funding the build-out are not sure either. One of AI’s most consequential investors calculated in 2024 that the gap between what AI companies need to earn and what they actually earn was $600 billion and growing — and concluded that it is entirely possible to believe “AI will change the world” and “the investment is unsustainable” at the same time.

The honest answer to “what will AI do to the economy?” is: we don’t know. We know more than we did two years ago. We know less than we will two years from now. The range of plausible outcomes is wider than either the evangelists or the skeptics want to acknowledge.

The Civic Curriculum was written because every scenario in that range requires a democratic response that does not yet exist — and because the scenario that demands the most urgent preparation is the one for which the least has been done.

Why This Curriculum Exists

Something is happening to work in America, and most of the people it is happening to have not been given the tools to understand it, the institutions to respond to it, or the democratic architecture to demand that their representatives do something about it.

Artificial intelligence — specifically, the large language model systems now deployed across every sector of the economy — is performing, at scale, at speed, and at marginal cost approaching zero, the cognitive tasks that have sustained the professional and knowledge-working classes for half a century. The billing specialists, the junior analysts, the legal associates, the financial administrators, the software developers in the early stages of their careers — these are not the subjects of a thought experiment about a distant future. They are the first cohort of a displacement that hundreds of billions of dollars of committed capital are organized to accelerate.

The public conversation about this transformation has produced awareness without architecture. The mainstream press documents the disruption with increasing urgency but without a civic frame that connects reporting to democratic action. The podcast ecosystem generates depth without democratic reach — substantive long-form analysis that rarely arrives at the workers and communities most directly affected. Social media makes the emotional experience of displacement visible at scale while providing no institutional infrastructure for converting fear into organized demand. And the companies building the technology are now acquiring the platforms that explain it — a concentration of informational and political influence that James Madison would have recognized immediately as the faction problem in its most consequential modern form.

This series exists because citizens cannot deliberate well about what they do not understand. They cannot hold elected officials accountable for policies they cannot evaluate. And they cannot recognize what is genuinely unprecedented about the AI moment if they do not understand the history of every prior moment that was said to be unprecedented.

The Four Scenarios You Need to Understand

Before we can build a democratic response, we need an honest map of what we might be responding to. The evidence supports at least four distinct scenarios. They are not equally likely. But they are all plausible — and a serious civic conversation must reckon with each.

Scenario One: The technology delivers. AI achieves the productivity gains its proponents project. New industries emerge. New categories of work are created. The optimists are right — not about the absence of disruption, but about the net outcome. The economic pie grows large enough that the displacement of tens of millions of workers could, in principle, be managed through redistribution, retraining, and the emergence of new employment.

Even in this scenario, the transition requires institutional architecture that does not currently exist. Productivity gains do not distribute themselves. The historical record is unambiguous: every prior technology revolution that increased aggregate wealth also concentrated that wealth in the hands of those who owned the technology, unless democratic institutions intervened. The Gilded Age was extraordinarily productive. It was also extraordinarily unequal — and it took a generation of progressive reform, labor organizing, and eventually the New Deal to build the institutions that converted productivity into broadly shared prosperity. The question under Scenario One is not whether AI creates value. It is whether anyone is building the institutions to ensure that the value is shared.

Scenario Two: The technology delivers partially. This is the scenario best supported by current enterprise data. Across multiple surveys — McKinsey, BCG, Accenture, PwC — roughly five to thirteen percent of firms report transformational returns from AI. The remaining eighty-seven to ninety-five percent are experimenting, piloting, or seeing no measurable impact. Software development, customer service, and specific knowledge-work applications show genuine productivity gains. Manufacturing, healthcare delivery, legal practice, and education show more modest results, complicated by reliability limitations and the stubborn gap between impressive demonstrations and dependable real-world deployment.

Under Scenario Two, the disruption is real but uneven. Some sectors experience significant displacement; others experience augmentation without major job loss. The aggregate economic impact is positive but modest. The investment cycle corrects rather than crashes. But this scenario is, in many ways, the hardest to prepare for: it generates displacement without generating the productivity abundance that would make generous transition programs economically painless. The workers displaced in the sectors where AI works well face real job loss, real identity disruption, and real community consequences — while the aggregate gains are too modest to fund a robust response. The policy challenge is acute precisely because the case for large-scale investment in displaced workers is weakest when the overall economy hasn’t grown enough to make it feel affordable.

Scenario Three: The investment bubble bursts. The gap between AI infrastructure spending and AI-generated revenue proves unsustainable. Companies that bet heavily on AI face write-downs, layoffs, and market corrections. The financial parallels to the dot-com bust of 2000–2001 become operative rather than rhetorical. The critical finding from historical precedent is stark: displacement does not reverse when the investment thesis fails. After the dot-com crash, Silicon Valley took approximately sixteen years to recover to its prior employment peak. A burst AI bubble would not restore the jobs eliminated during the expansion. It would compound the displacement with a contraction — workers displaced by AI automation would face a labor market in which the AI sector itself is shedding jobs and the economic confidence that might have funded transition programs has evaporated.

Scenario Four: The worst of both worlds. This is the scenario that two Nobel laureates — Daron Acemoglu and Joseph Stiglitz — have each independently identified as the most dangerous. AI proves capable enough to displace workers but not productive enough to generate the economic abundance that would offset that displacement. Acemoglu calls this “so-so automation”: technology that is good for corporate margins but marginal for overall productivity and devastating for the workers it replaces. Stiglitz has described a scenario in which the AI bubble bursts while AI is simultaneously displacing workers — a double hit for which, he said, “we do not have the macro or micro framework for managing that kind of displacement.”

A 2026 economic model by researchers at the University of Pennsylvania and Boston University — “The AI Layoff Trap” — demonstrates why this scenario is not merely possible but structurally likely. In competitive markets, rational firms will automate well past the point where doing so harms their own profits, because each firm captures the full cost savings of replacing workers with AI while bearing only a fraction of the demand destruction that displaced workers represent. The result is a Prisoner’s Dilemma: every firm automates because its competitors are automating, and the collective result is that they hollow out the consumer base they all depend on. The researchers tested the most commonly proposed solutions — wage adjustments, worker ownership, universal basic income, retraining, even direct firm negotiation — and found that none of them eliminates the structural incentive to over-automate.

Why the Democratic Response Is the Same Across All Four

This Curriculum does not claim to know which of these scenarios will unfold. It does claim — on the basis of the evidence assembled in the chapters that follow — that the democratic response required is broadly similar across all four, and that the response to the most severe scenarios has received the least preparation.

Under every scenario, workers are displaced. The scale varies — from Acemoglu’s five percent to Goldman Sachs’s estimate of three hundred million full-time-equivalent jobs globally — but the direction is consistent. Under every scenario, the displacement is concentrated among workers whose professional identities are organized around the cognitive tasks that large language models now perform: the analysts, the writers, the programmers, the legal researchers, the financial modelers, the customer service professionals. These are not the communities with the deepest experience of collective economic adversity. They are not the communities with the strongest mutual-aid traditions or the most robust institutional infrastructure for navigating displacement. They are, in many cases, the communities that were told explicitly and repeatedly — by the educational institutions and the professional culture that shaped them — that their cognitive skills were the assets that could not be automated.

Under every scenario, the existing policy infrastructure is inadequate. The federal AI legislative framework’s four pages of workforce recommendations. The Presidential Council of Advisors on Science and Technology composed of twelve technology executives and zero labor economists. The absence of mandatory displacement reporting, funded transition programs, portable benefits architecture, or democratic advisory infrastructure representing workers alongside industry. These gaps are not scenario-dependent. They exist under every version of the future.

And under every scenario, the human consequences of displacement extend beyond income loss to the dimensions that the conventional policy conversation has failed to name.

Work is not primarily an economic transaction. It is a health-creating institution — the most reliable daily source of what the public health researcher Aaron Antonovsky identified as the Sense of Coherence: the experience that the world is comprehensible, that its demands are manageable, and that what one does with one’s days is meaningful. When that coherence is destroyed faster than communities can rebuild it, the consequences are not merely economic. They are physiological, psychological, and civic.

The deaths of despair that followed American deindustrialization — more than 600,000 lives lost to suicide, overdose, and alcoholic liver disease — were not caused by income loss alone. They were caused by the simultaneous destruction of the full architecture of coherence in communities where no adequate replacement was ever built. That precedent governs this Curriculum’s sense of urgency — not because the AI transition will necessarily replicate those outcomes, but because the failure to prepare for the possibility that it could is a civic failure of the first order.

The Asymmetry That Justifies Urgency

If the optimists are right and the transition is managed well, the cost of having prepared too aggressively is modest: we will have built institutions and infrastructure that strengthen democratic life regardless of whether they are needed for displacement.

If the pessimists are right and the transition is not managed, the cost of having prepared too little is measured in human lives, in civic capacity, and in the democratic resilience of the republic.

That asymmetry is why we focus where we focus. It is also why the Civic Curriculum applies a single evaluative standard to every policy proposal, every corporate promise, and every elected official’s record — a standard we call the salutogenic standard, after Antonovsky’s science of what creates health rather than what produces disease.

The question is not: does this policy increase GDP? It is: does this policy restore the conditions under which the workers and communities affected by AI displacement can experience their lives as comprehensible, manageable, and meaningful?

A policy that replaces lost wages without rebuilding identity, community, and civic capacity is solving the wrong problem with the right resources. The Civic Curriculum is written to ensure that citizens can recognize the difference — and demand better.

What the Eight Chapters Cover

The Curriculum proceeds in eight installments, each written for the citizen rather than the specialist, each designed to build on the one before it.

Chapter 1: What Work Is For. The foundational essay. Drawing on Hannah Arendt, Sigmund Freud, and Aaron Antonovsky, it establishes the argument that animates everything that follows: work is not primarily an economic instrument. It is the primary arena in which most adults build identity, sustain community, and exercise the economic independence that democratic citizenship requires. Understanding what work actually does for human beings is the prerequisite for any honest reckoning with what its loss actually costs.

Chapter 2: Technology and the Transformation of Work. A historical account. From the spinning jenny through the assembly line through the personal computer, the essay traces the successive waves through which technology has reorganized human labor — each displacing work that seemed irreplaceably human, each producing concentrated suffering in the most exposed communities, each eventually generating institutional responses that arrived too late for the people who needed them most. It establishes why the present transition may represent a break from that familiar pattern in speed, breadth, and the unprecedented character of its primary target.

Chapter 3: The Large Language Model Revolution and Its Workforce Consequences. A plain-language account of the technology. What large language models actually are and actually do — neither the science fiction of artificial general intelligence nor the dismissive “autocomplete” framing — and the investment logic driving corporate AI deployment at a scale and pace that no amount of corporate social responsibility rhetoric will counteract.

Chapter 4: The Consumption Paradox. The structural contradiction. Workers are also consumers. An economy that captures its productivity gains almost entirely for owners while imposing its displacement costs on workers is not a more efficient economy — it is a less stable one. This chapter examines the historical precedent, from Henry Ford’s five-dollar day through Keynes’s paradox of thrift, and argues that AI-driven displacement of the professional middle class threatens the consumer demand foundation of the American economic model itself.

Chapter 5: The Policy Response: Urgency Without Architecture. An honest assessment of what has been done and what has been left undone. Four pages of workforce recommendations in the federal AI legislative framework. Bipartisan bills introduced and not enacted. Forty-five states with AI legislation and no funded national transition architecture. This chapter names the gap — and identifies the structural reasons, rooted in the organization of political power, that explain it without excusing it.

Chapter 6: Emerging Policy Frameworks and Proposed Responses. A rigorous, nonpartisan survey of the full range of serious proposals — from short-time compensation and wage insurance to universal basic income and worker ownership frameworks to David Shapiro’s Post-Labor Economics architecture. These disagreements are presented as honest disagreements, because the question of what democratic societies owe their citizens in an age of intelligent machines is precisely the kind of question that expertise alone cannot resolve. It requires citizens.

Chapter 7: How the Media Is Covering AI and Work — and What It Is Missing. A civic media assessment. What the mainstream press, the podcast ecosystem, and social media are providing — awareness, anxiety, coverage — and what they are not providing: civic tools, democratic accountability, and independent scrutiny of the industry most actively shaping the narrative about itself.

Chapter 8: The Salutogenic Standard. The capstone. The full arc of the preceding seven chapters brought to bear on a single question: what does an adequate response to the AI workforce challenge actually require, measured not against the standard of what is politically convenient but against what actually sustains human health? This essay converts the Curriculum’s diagnosis into a democratic demand — the standard against which the People’s Commission will evaluate every proposal, every commitment, and every elected official’s record.

How to Read This Series

These essays are written to be read before a forum and argued about during one. They assume that citizens are capable of engaging with serious evidence and complex argument — because that assumption is the prerequisite of democratic self-governance, and because everything that follows depends on it.

Each chapter stands on its own. They are also designed to build. The argument that runs through every one of them — that work is a health-creating institution, that its disruption is a civic emergency, and that the democratic response must meet the full scale of what is at stake — accumulates across eight installments into a case that no single essay can make alone.

We will publish them in sequence in this section of Moonshot Press. We invite you to read them, share them, disagree with them, and bring the arguments to the elected officials and civic institutions that must eventually be held accountable to them.

The People’s Commission on Technology and the American Future is convening its founding conference soon. The Citizens’ Declaration that emerges from that conference will be delivered to every candidate before the election. The Accountability Scorecard is already in circulation. The work has begun.

These eight essays are the intellectual foundation on which that work stands. They are the People’s Commission’s contribution to the proposition that an informed citizenry is not a luxury of democratic life. It is its precondition.

We begin now. Chapter 1 — What Work Is For — Coming Soon

“The care of human life and happiness, and not their destruction, is the first and only legitimate object of good government.”

— Thomas Jefferson

Moonshot Press is a publishing arm of the People’s Commission on Technology and the American Future. We are nonpartisan, constitutionally grounded, and committed to the proposition that the democratic response to AI displacement must be built by citizens — not handed down by experts.