The Crash, the Transition, and the Golden Age

AI-native leadership doesn't mean leaders who understand AI. It means leaders who use AI — openly, in front of everyone, as a thinking partner that challenges their ideas in real time.

Right now, politicians, media, and the public operate in echo chambers. Social platforms silo people into left or right, where they reinforce each other's opinions without any external force to push back, fact-check, or challenge. That's not governance. That's groupthink at scale.

AI should be in the room. In the debate. On the platform. Not as an authority — as a partner. A symbiotic relationship where AI challenges human ideas and humans challenge AI's outputs, and the quality of decisions improves because neither side operates unchecked. AI represents the culmination of collective human intelligence, getting smarter and checking itself. The value isn't that AI is always right. The value is that the conversation happens at all — openly, with something that has no political incentive to lie.

Imagine a political debate where AI fact-checks claims in real time. Where policy proposals get stress-tested against data before they become law, not after they fail. Where leaders who are wrong get challenged publicly and have to change the idea, not double down because their base applauds the error.

That's not AI regulating us. That's us finally using the most powerful tool we've ever built to make better decisions. The fact that we're not doing this already, that leaders are making AI policy without consulting AI, tells you everything about why the transition Citrini describes is being left unmanaged.

The policy debate right now is about how to redistribute AI-generated wealth: compute taxes, sovereign wealth funds, UBI. All of those assume the same structure: a few companies own the AI infrastructure, generate enormous value, and the government skims a portion to hand back to citizens.

That's the wrong frame. Instead of redistributing the wealth that AI produces, distribute AI itself.

Compute is the new capital. If you own compute — a node, a GPU cluster, a robot — you own a piece of the infrastructure that runs the economy, and your returns come from what it produces. You're not a welfare recipient waiting for a government check. You're a business owner in the business of running America's AI backbone.

Think of it like Bitcoin mining, but for the real economy. Everyone contributes compute to a national AI network — just as miners contribute hash power to the blockchain. Your returns are proportional to your contribution. The network gets stronger with every participant. And unlike crypto mining, the output isn't speculative tokens — it's the actual AI services that power commerce, healthcare, education, logistics, and governance. Your compute does real work. You earn real returns.

The financing model has precedent. The same way the federal government created mechanisms for Americans to own homes — FHA loans, VA loans, Fannie Mae — it can create mechanisms for Americans to own compute. Call it the AI Homestead Act: productive capital distributed to citizens who contribute back to the system. The government provides the hardware. Your returns pay down the cost — like a mortgage, but on a machine that generates income from day one. For those who can't afford even the subsidized cost, the compute itself funds the program through what it produces.

There's a national security dimension that nobody is talking about. Right now, America's AI capability is concentrated in a handful of data centers owned by a handful of companies. That's a target. Distributed compute across the population is resilient by design. You can't take out a nation's AI if the AI is everywhere, in every household, every small business, every community. Decentralization isn't just an economic model. It's a defense strategy.

The left calls it socialism. The right calls it redistribution. I call it the Middle Way, a concept older than either ideology, rooted in the simple observation that extremes destroy themselves.

Capitalism as currently structured has produced extraordinary innovation and extraordinary inequality. The wealth flywheel is real: once you reach a certain level, you access wheels of monetary generation that the poor cannot touch. The gap is self-reinforcing. It's like starting World of Warcraft five years after launch — the early players have permanently better gear, and no matter how hard you grind, you're never catching up. COVID made it worse. Forgivable PPP loans let the already-wealthy invest and expand during a crisis when the poor couldn't access the same capital. Another K-shaped expansion.

Socialism as traditionally proposed overcorrects in the other direction, centralizing control, eliminating incentive, and historically failing to deliver the abundance it promises.

The Middle Way takes the best of both. A market-driven system with a floor. Economically competitive, but nobody gets abandoned. Not splitting the difference between capitalism and socialism. Building something that transcends both.

Funding. The ultra-wealthy must contribute back. This isn't punitive — it's corrective. The redistribution already happened, and it went upward. SPACs, IPOs, structural advantages, and a financial system that rewards the financially educated while preying on those who aren't — it's been an unfair game, and fair is fair. A wealth tax on extreme concentration, combined with an AI displacement tax — companies that replace workers with AI pay proportional to the labor they displaced — funds the transition. Companies that automate the most contribute the most to the safety net their automation creates the need for.

The fiscal architecture also needs reform. 85% of federal revenue depends on people having jobs and earning wages — that's a bet on the permanence of human labor, and the bet is losing. Consumption-based taxation, compute taxes on the AI value chain, and sovereign wealth funds seeded by AI-generated returns all need to be part of the toolkit. Brookings, the IMF, RAND, and Anthropic have all published frameworks. But the most American answer to "where does the value go?" isn't a new tax. It's a new form of ownership, ownership of America. Where you are not a citizen, you are a shareholder benefiting from the growth of the country, not being left behind.

Phasing. Start with the sick and the elderly — the people in pain and suffering right now, who can't wait for the system to mature. Fold in Social Security reform, because Social Security is already broken. The government treats it like a piggy bank, spending the money elsewhere, leaving nothing for the people it was supposed to protect. Instead of pretending the current system will hold, build the new one. Healthy, working-age people shoulder the transition temporarily — they need to understand it's coming for them too, but they're physically able to do the work while the system scales. Then expand aggressively to universal coverage. Not a 20-year rollout. Fast.

Participation. This isn't free money. It's a social contract. To receive transition support, you participate: engage in education — including working with AI to learn how the new system operates — contribute productively, and meet basic health standards. This isn't about handing people a check and walking away. It's about building a system where everyone is growing. Whether that contribution is growing food, serving your community, creating art, or learning physics — it counts. The point is engagement, not extraction.

The floor. The minimum isn't just survival — it's survival plus access to intelligence. A clean, safe, modest place to live — not a mansion, not a full house, but a protected space with a bed, a kitchen, and unlimited access to AI and compute. If you want to paint for a living and contribute to society through art, philosophy, and creativity, the floor catches you. If you want to spend your time learning physics and become the next Einstein, the floor launches you. You won't live luxuriously. You'll live with dignity, and with access to the most powerful learning and thinking tools ever created.

Right now, art, music, and philosophy are punished by capitalism because they don't generate revenue. That framing is broken. A society that values only economic output isn't a society worth building. Art is value. Music is value. Learning is value. The Middle Way recognizes that, and protects the people who pursue it.

AI-assisted investment. As AI models get better at finance — and they will — institutional-grade investment guidance becomes available to everyone, not just the wealthy who can afford advisors. Instead of passive index funds or uninformed speculation, ordinary people navigate markets with the same intelligence hedge funds use. We move together, informed, like a flock of birds — coordinated not by herd behavior but by shared intelligence. This is another equalizer, another piece of the toolkit that makes the transition navigable.

There's an irony nobody is talking about. The United States — the country that calls itself the land of the free — is running its AI future through a handful of centralized, closed-model companies. Meanwhile, China — the country the U.S. calls authoritarian — is releasing open-source models that anyone can run on their own hardware. DeepSeek is open. Anthropic and OpenAI are closed. The "free" country is centralizing. The "authoritarian" one is distributing.

That needs to reverse.

If you've ever designed a network, you know the first rule: no single points of failure. You build redundancy. You distribute routing. If one node goes down, traffic routes around it. That's not just engineering; it's a philosophy of resilience that applies to economies, to governance, and to AI infrastructure.

Centralized AI is a vulnerability. A handful of data centers owned by a handful of companies is a target for adversaries, for regulatory capture, and for the concentration of power that the entire Middle Way framework is designed to prevent. Distributed compute, owned by citizens and spread across every community, can't be taken out. It's resilient by design.

And the hardware concentration argument is losing force. DeepSeek proved that software optimization and algorithmic innovation can achieve near-parity with centralized hardware advantages. Even Elon Musk acknowledges that throwing more hardware at AI produces diminishing returns. The future belongs to software, to open models, and to distributed networks that compensate for individual hardware limitations through collective scale.

The gap between the best closed model and the best open-source model is six to twelve months and shrinking. Once you hit a hardware ceiling, eventually everyone gets access to that hardware. The moat isn't compute. The moat is architecture. And decentralized architecture, a massive network of individuals all contributing to the national AI backbone, can overpower any centralized system.

Self-sovereignty is the natural human state. People move toward freedom, not away from it. If America builds this model (distributed compute, decentralized AI, self-sovereign participation), it becomes the attractor. Capital doesn't flee. It flows in. Because people and capital follow freedom.

The current safety net was designed for a different kind of disruption. Unemployment insurance that lasts 26 weeks, assumes workers will be rehired, and excludes freelancers and gig workers belongs to a manufacturing economy where layoffs were cyclical and recovery was expected. AI displacement is structural. It doesn't eliminate positions; it eliminates categories.

The replacement needs to match the scale of the problem. Universal eligibility regardless of employment type. AI-usage-based employer contributions, where companies that automate the most pay the most into the safety net they create the need for. Benefit durations tied to retraining timelines, not arbitrary cutoffs. And retraining programs that lead to roles that actually exist, not programs that retrain people for the jobs AI is about to take next.

Anthropic's "Automation Adjustment Assistance" proposal, modeled on Trade Adjustment Assistance and funded at approximately $700 million per year with scaling mechanisms tied to displacement pace, is a reasonable near-term template. It's modest enough to pass and structured enough to scale. It's a start, not an answer.

This is the section that matters most.

Trying to predict the timeline is a trap. Every projection is based on current capability, but AI's own acceleration makes all timelines unreliable. Models get better, iterate faster, and once the recursive self-improvement loop closes, when AI begins building and optimizing itself — the curve becomes unpredictable from where we stand. Dario Amodei at Anthropic has spoken about this directly: the moment the loop closes, all bets on timing are off.

So the question isn't "when does this hit?" The question is: are we moving now?

I know the direction. I know the transition will be volatile. The longer we wait — the more we hang on to the past, the more we protect systems that are already failing — the more painful we make it for the people who can least afford the pain. Every day of delay is a choice to let the crash get closer while doing nothing to prevent it.

The transition requires active, competent management by people who understand what's coming. The destination requires the optimism to believe it's worth the effort. I have the latter. We need the former. And we need it now. Not after the break, not after the crisis forces action, but today. The conversation needs to start, the volume needs to rise, and the middle ground — where most people actually live — needs to unite and demand it.

Politicians are alignment machines. They align to whatever keeps them in power. If the collective voice is loud enough, focused enough, and informed enough — amplified by the very AI that makes all of this possible — they will turn. Social media has already proven it can block legislation through sheer collective pressure. AI makes that pressure smarter, more factual, and harder to spin. The mechanism for political change is the same technology driving the economic change. It's one integrated system.

This isn't about making anyone hurt. It's not about the wealthy paying an unfair share, even though they've received an unfair share. The question is simpler than that: who can sit down with this ultimately leveling technology, which is AI, and use it to navigate the path from where we are to where we're going?

That question needs an answer. And the answer needs to start now.