Pre-public peer review

Every year, AI gets more powerful.
Every year, fewer people control it.

Three to five companies are accumulating the two most critical resources in human history - compute and intelligence. No democratic body governs them. The only check that exists is government itself, and government is slow, external, and only acts after something's already gone wrong.

The math says it doesn't have to work this way.

A research team spent eight years proving something everyone assumed was impossible: a democracy that gets stronger, not weaker, as the systems it governs get more powerful. This is that democracy, already running - a working system with a six-year track record and zero constitutional violations, not a whitepaper promise.

At 126 governors across 30+ countries, this system is measurably harder to capture today than it was at 13. Most networks get weaker as they grow; this one gets stronger - six years of ballots, $29.66M in real decisions, zero violations.

01The idea

A democracy that gets stronger as AI gets smarter

Most systems get more fragile as they scale. This one gets harder to capture with every person who joins.

Networks - what is

Every network you've used gets more centralized as it grows - more users, more power in fewer hands.

What could beA network where the opposite is mathematically true - the more people join, the harder it becomes for anyone to capture it.

AI companies - what is

Right now, a frontier AI lab can be shut off two ways: a courtroom, after the fact, too slow to matter - or a single government letter, with no public standard and no vote, fast enough to take a model offline worldwide within hours. Neither is accountable. Neither leaves a receipt.

What could beA kill switch built into the system from day one, held by 126 people across 30 countries, that the founder himself cannot override - and that produces an auditable record every time it moves.

You - what is

Today, your AI decisions are shaped by shareholder incentives you never voted on.

What could beYour community holds real, mathematically-enforced oversight - not a policy promise, a constitutional ceiling nobody can raise. This already exists. $29.66M governed. Zero violations. Six years running.

02Historical foundations

Historical Foundations

Two hundred and thirty-nine years ago, an orphaned immigrant showed up in America from a small island in the Caribbean with nothing to his name but a reputation for being frighteningly smart. He talked his way into George Washington's inner circle, served at his side through the Revolutionary War, and at Yorktown led a bayonet charge on a fortified redoubt that helped end the war outright. Then he helped defend the country's brand-new constitution, one newspaper essay at a time. Then, as the country's first Treasury Secretary, he built its first monetary system from nothing - the first central bank, the first national currency. His name was Alexander Hamilton.

In one of those essays, number twenty-two, he pointed at something the other framers may not have even noticed they'd done: three separate times, the document capped a minority's power at one-third. Madison gave democracy its separation of powers, in Federalist 51 - "ambition must be made to counteract ambition." Hamilton gave it a ceiling.

Nearly two hundred years later, a computer scientist named Leslie Lamport was solving an entirely different problem: how a network agrees on anything when some of its members might be lying. His answer was a floor - two-thirds honest, or agreement is impossible. He'd never read Hamilton. He found the same number anyway.

Chris McCoy spent a research program connecting them: a governance problem from 1787 and a computer science problem from 1982, landing on the same boundary, independently, twice. STORE calls it the Hamilton-Lamport Convergence - a proposed pattern, headed for public peer review, not yet a settled fact. STORE runs on all three: Madison's equal distribution, Hamilton's ceiling, Lamport's proof. Not three men agreeing. Three problems, two centuries apart, landing on the same number. See the actual Path to Launch →Not yet built

03What you get

What You Get

Layer 0 Operational
You get verifiable cloud infrastructure - every claim checkable, not asserted.
Layer 1 Operational
You get individual AI endpoints already governed, not bolted on after.
Layer 2 Operational
You get democratic oversight of both AI behavior and the compute running it.
Layer 3-4 In development
You'll get the same governance extended across platforms - not just STORE's own AI, any foundation model you already use.
Layer 5 - Breakthrough In development
You get oversight of AI companies structurally incapable of building it themselves. AI companies cannot authentically build democratic oversight of their own systems because it conflicts with shareholder value. STORE was built by people with nothing to protect but the math.
04Enforcement

Swiss Constitutional Enforcement

Your governance rights aren't a company policy STORE could change tomorrow. They're enforced under Swiss law, the same way a constitution is enforced - not by STORE's goodwill, by jurisdiction.

0
Governors, 30+ countries
6 Years
Operational (since 2020)
$29.66M
Governed, zero violations
≤33%
Hamilton's Ratio, always enforced
05Multi-Level Power Distribution

Multi-Level Power Distribution

Most networks get more fragile as they grow. STORE runs the opposite direction: stability rises with every governor who joins, not falls. At 126 governors today, holding well below the constitutional ceiling on their own, the system is measurably more resistant to capture than it was at 13. The framework is called Multi-Level Power Distribution - not "Hamilton's System," which would erase Lamport's own contribution to the same math.

06Real votes

Five real votes, not a hypothetical

Here are five real ballots, with real approval percentages, that actually happened.

BallotTestApprovalStatus
FG.005Reset previous purchases to $0.039/token71.43%Validated
FG.004Reprice to $0.03976.19%Validated
FG.003Allow purchases at $0.03594.74%Validated
FG.002Reset previous purchases to $0.03589.47%Validated
FG.001Reprice to $0.069 (vs $0.049)52.94%Validated
07Economic democracy

Hamilton's Ratio Implemented

Economic democracy that prevents wealth concentration while maintaining incentives.

1%
Security Owned - prevents concentration
1%
Revenue Earned - proportional rewards
1%
Production Costs Shared - fair distribution
1
Vote - equal democratic power, always

Unlike systems where wealth equals power, this ratio holds your vote equal regardless of how much you hold, earn, or contribute - the same math whether you're a single depositor or an institutional allocation.

08Guarantees

Mathematical Guarantees

Human sovereignty mathematically enforced at machine speed.

≤33%
Maximum control - no single entity can dominate
~42
Malicious tolerance at 126 governors
≥67%
Human supermajority required for all decisions
Scale resistant - democracy strengthens with growth
09Timeline

Deployment Timeline

ETA 2026-2027
Preparation - Judicial Branch rights
ETA 2026-2027
Deployment - BFT Democracy launch

First Verifiable Framework for Human Control in the Age of AI

Constitutional mathematics, operational validation, and BFT democracy deployment - all in one system.

Technical paper coming soon

Join the Public Peer Review Process

This research is proposed, not yet proven. That's why it's headed for public peer review rather than being asserted as settled fact - a claim this size should have to earn it.