The Compact Is the Cage: A Response to Alap Shah’s “Global Intelligence Crisis”
If you read only one of my pieces, let this be the one. Free to read in full — because the argument it’s responding to deserves a wide audience, and so does the rebuttal.
Every so often someone writes a piece so well-reasoned, so honest, and so thoroughly wrong in its conclusion that you have to stop what you’re doing and respond.
Alap Shah has written that piece.
If you haven’t read his three-part series, The Global Intelligence Crisis, you should. Shah is not a crank. He’s a serious man — twenty years investing in public markets, fifteen years building AI companies, sold one of them for over $200 million. He uses frontier AI every day. He has watched, from the inside, as agentic AI began replacing the humans in his own companies. And he has done the intellectually honest thing that almost no one in his position does: he followed the logic all the way to its terrifying conclusion and said it out loud.
His diagnosis is, as far as I can tell, correct. I’ll summarize it, because I want to be fair to him before I tell you why his cure is worse than the disease.
What Shah Gets Right
Shah argues that AI is no longer a tool that complements human labor — it’s a substitute for it. Every previous technological revolution displaced workers from specific tasks while creating higher tiers of work for them to climb into. AI is the first technology that attacks the highest tier itself: non-routine cognitive work. Analysis. Judgment. Management. The refuge that absorbed every previous wave of displaced workers.
There is no higher tier to escape to.
He documents the early tremors: white-collar employment stagnant since 2023, recent-graduate unemployment at its highest level in over a decade, Block cutting 40% of its workforce with the CEO citing AI explicitly, Meta and Oracle reportedly preparing tens of thousands of cuts. He lays out the feedback loop — AI improves, companies need fewer workers, displaced workers spend less, weakened companies invest more in AI to protect margins, AI improves further. No natural brake. He shows how the damage is disproportionate, because the top 20% of earners drive 65% of consumer spending and they’re the most exposed. He traces the contagion into private credit, into the mortgage market, into the government’s own fiscal position.
I have been writing about exactly this for months. In Universal Basic Income: The Great Destroyer of Worlds, I described the same convergence — AI displacement, a fragile economy, a government that responds to crisis the only way it knows how: by spending money it doesn’t have. Shah and I are looking at the same storm. We agree on the clouds. We agree on the wind.
We part ways completely on what to do when it makes landfall.
What Shah Gets Wrong
Shah’s solution is called the American Prosperity Compact. Strip away the careful architecture — the Foundation, the Circuit Breaker, the Backstop, the Accelerator — and here is what it is:
Raise taxes on corporations. Use the money to pay displaced workers. Create a national fund that sends every American a check.
He is too smart to call it UBI. He calls it an “American AI Dividend Fund,” modeled on the Alaska Permanent Fund. But the very first comment on his own essay — from Scott Santens, one of the most prominent UBI advocates alive — cheerfully thanks him for including universal basic income in the plan, “even if you just described it as an AI dividend.”
When the most famous UBI evangelist on the internet reads your proposal and says welcome to the team, you have written a UBI proposal.
I want to be precise about why this fails, because Shah’s version is the most sophisticated one I’ve encountered, and the sophistication is exactly what makes it dangerous. It’s UBI dressed in a tailored suit, carrying a spreadsheet, speaking the language of capital markets. And it will lead to the same place every dole has led for two thousand years.
Here is where it breaks.
One: Taxes Disincentivize the Thing You’re Taxing
This is the oldest law in economics, and no amount of clever structuring repeals it. When you tax something, you get less of it.
Shah’s central mechanism is a corporate “value-added” tax and a “displacement tax” that scales up as companies replace workers with AI. He argues that a “moderately higher rate will not meaningfully deter corporate investment when the opportunity set is so large.”
This is wishful thinking from a man who knows better. Capital is the most mobile substance on Earth. It does not sit still and accept a haircut for the good of the consumer economy. It moves. It restructures. It finds the jurisdiction, the entity type, the accounting treatment that minimizes the bite. Shah even admits this — he concedes the whole scheme requires “stronger global tax coordination” because otherwise “companies cannot trivially reroute profits through jurisdictions that don’t impose one.”
Read that again. His own plan only works if every major economy on Earth agrees to tax in lockstep so that capital has nowhere to flee. That’s not a policy proposal. That’s a fantasy of global fiscal coordination - a hammer so heavy that when it swings its echoes ring of a One World Order - that has never existed and likely never will. And the moment one jurisdiction defects — one Ireland, one Singapore, one UAE, one Próspera — the capital flows there, and the whole edifice collapses.
You cannot build a domestic safety net on the assumption that the entire planet will agree not to compete. Otherwise you undermine the entire foundation of capitalism itself. And I think we can all agree that communism, or even socialism, is not the answer.
Two: It Crushes the Small and the Solo
Here is the cruelest irony of Shah’s plan, and I don’t think he sees it.
He spends an entire section celebrating how AI “levels the playing field for entrepreneurship” — how a one-person firm can now do the work of ten, how the solo operator can finally compete with the giant. He’s right. This is the single most hopeful thing about the entire AI transition. For the first time in a century, the little guy has the tools to fight the conglomerate.
And then he proposes a tax structure that strangles that little guy’s business in its crib.
Large corporations have armies of tax attorneys, offshore structures, lobbyists who write the carve-outs, and the scale to absorb compliance costs. The solo operator and the small shop have none of that. When you raise the corporate tax burden and bolt on a “value-added assessment” and a “displacement tax,” you are not hurting the giants — they’ll restructure around it by lunchtime. You are hurting the nimble two-person company that can’t afford a tax department.
We have already seen this sort of behavior in healthcare - specifically with regard to The Affordable Care Act, and how it has made health insurance more onerous for individuals and small businesses and incentivized employees to move to larger, higher margin businesses that can absorb the increasingly outrageous cost of healthcare.
Every regulation, every tax, every layer of compliance is a moat that protects the incumbent from the upstart. Shah’s Compact, sold as a check on corporate power, would entrench corporate power by raising the cost of competing with it.
Three: It Drives Capital — and the Builders — Out
Shah is an investor. He knows that capital flows to where it’s treated best. He admits his own plan will “increase my own tax burden” and frames this as proof of his sincerity.
I believe his sincerity. I question his arithmetic.
When you raise the cost of building and operating in the United States, you don’t get more building in the United States. You get more building somewhere else. The AI labs, the capital, the talent, the corporate headquarters — they are not bolted to American soil. They are some of the most portable enterprises in human history. A company whose entire workforce is a founder and a swarm of AI agents can incorporate anywhere on Earth with a laptop and an internet connection.
We are not in a closed system. Shah knows this — it’s why his plan requires global coordination. But absent that impossible coordination, a unilateral American tax-and-redistribute scheme is an engraved invitation for the most productive enterprises of the next century to set up shop literally anywhere else. We’d be exporting the golden goose and keeping the bill.
Four: The AI Dividend Fund Is the Monthly SUC
I’ve written about this at length, so I’ll be brief. (For the full argument, see Universal Basic Income: The Great Destroyer of Worlds — it traces the dole from Rome’s free grain through the Great Society to the world of my novel.)
Shah’s “American AI Dividend Fund” — a check sent to every citizen, funded by taxing the productive economy — is UBI. And UBI, however well-intentioned, does to a human being what it did to Rome: it provides survival while stripping away purpose. It keeps the body alive and lets the soul atrophy.
In my novel Complex, the government’s version is called the Standard Universal Credit — the SUC. Val, the protagonist, calls it “the dreaded monthly SUC.” It keeps the Legacy citizens fed and housed and useless. It is the chemical that pacifies a population while the strange loops dim, one direct deposit at a time.
Shah believes his dividend makes citizens “shareholders in the transition.” But a shareholder has a vote, a voice, a stake that can grow. A UBI recipient has a check that arrives at the pleasure of whomever controls the fund. That’s not ownership. That’s dependency with better branding.
And here’s the part Shah doesn’t follow to its conclusion: once you have built the machinery to tax the productive class and distribute the proceeds to the dependent class, you have built the most powerful instrument of control in human history. Whoever administers that fund decides who eats. That power never shrinks. It only grows. It is the foundational architecture of every Complex I have ever written about.
Five: Follow the Power
Here’s what I keep coming back to, and it’s the thing that should make all of us deeply suspicious.
Look at who benefits from Shah’s Compact — not the workers, but the institutions.
The government, facing collapsing payroll-tax revenue as wages disappear, gets a brand-new revenue stream: a corporate value-added tax that doesn’t depend on humans being employed. It solves the state’s fiscal problem beautifully. The government doesn’t shrink in the AI age. It finds a new way to feed.
The largest corporations, facing the terrifying prospect of nimble AI-empowered startups competing with them on equal footing, get a tax-and-compliance regime that only an incumbent can afford to navigate. The moat is restored. The giants are protected from the upstarts.
And the central planners — the people who genuinely believe they should manage the transition, calibrate the triggers, set the thresholds, administer the funds — get exactly what every central planner has ever wanted: control of the flows. Control of who gets what. Control of the money, and therefore control of the people.
I am not accusing Shah of bad faith. I think he’s an honest man trying to solve a real problem. But the solution he’s arrived at — concentrate revenue collection at the corporate level, build a national fund, distribute checks, install automatic fiscal triggers managed by the state — is exactly the solution that a self-interested government and a self-interested corporate aristocracy would also arrive at. When the medicine the patient needs happens to be precisely the medicine that most empowers the people prescribing it, you are right to be suspicious.
The Compact is not a check on centralized power. The Compact is centralized power, wearing the mask of compassion.
So What Would I Do Instead?
It would be cheap of me to tear down Shah’s hard work without offering anything. So here, briefly, is the shape of a different answer.
The problem with the AI transition isn’t that people will lack money. It’s that people will lack purpose — and Shah, to his great credit, actually names this. He admits he feels “the fear of losing a sense of purpose” himself, “as someone without meaningful personal economic worries.” He’s closer to the truth than his own solution allows him to be.
You don’t solve a purpose crisis with a check. You solve it by refusing to build the systems that make humans economically irrelevant in the first place, and by ruthlessly clearing the path for humans to do the things AI can’t.
That means his Accelerator — the deregulation, the occupational-licensing reform, the scope-of-practice reform, the energy buildout — is the genuinely good part of his plan, and it should be the whole plan. Tear out the taxes. Tear out the dividend fund. Tear out the displacement tax. Keep the part that unleashes human enterprise and gets the government’s boot off the neck of the small operator.
Let capital stay. Let the solo founder compete. Let the cost of building collapse so that the displaced accountant can become the displaced accountant who now runs a one-person firm that does things no pre-AI human could dream of. Protect people through transition with the leanest possible support tied to contribution — work, retraining, caregiving, building — never the unconditional check.
And above all, do not build the machine that pays people to disappear. Because once that machine exists, someone owns it. And whoever owns it owns you.
The body is the temple. Purpose is the point. And a dividend that pays you to sit quietly while the strange loop dims is not prosperity.
It’s the Compact. And the Compact is the cage.
Alap Shah’s series is genuinely worth your time. Disagreeing with a serious thinker is a privilege; ignoring him would be a mistake. Follow him at @alapshah1. And if this response resonated, subscribe to Man & Machine — I write about the collision between humanity and AI every week, and I never pull the punch.
If you want to see where the Compact leads when it’s fully built — when the fund becomes the leash and the citizens become the herd — read Complex, Book I of the Silent Beautiful Universe series. The recruitment kiosk is still open.