Qurated: Five People Shouldn't Get to Decide the Future for Eight Billion
Five People Shouldn't Get to Decide the Future for Eight Billion
The Core Problem
Forget the alignment debates for a second. The real risk isn't that AI becomes uncontrollable — it's that it's already controlled, just not by us. A handful of CEOs, a few board seats, two governments. That's the entire electorate for decisions that will reshape civilization. Everyone else gets to watch.
Think of it like nuclear weapons policy decided entirely by whoever happened to be in the room when the button was installed — except the button controls economic structure, information ecosystems, and possibly cognition itself.
The Coin Flip That Should Have Scared Everyone
OpenAI's 2023 board crisis wasn't a corporate drama. It was a stress test that failed silently. One weekend. A handful of people. The trajectory of the most capable AI lab on Earth nearly reversed on a dime — and the mechanism that almost caused it was never designed to survive contact with real pressure.
Mental model: Governance is infrastructure. You don't wait until the bridge is being crossed by millions of cars to test whether it holds weight. If your entire safety architecture depends on five humans not having a bad week, you don't have governance — you have a hope.
This Isn't Hypothetical Anymore
The damage isn't theoretical or future-tense. It's zoning meetings right now:
- Water and land — Data centers are draining aquifers and consuming farmland in communities with zero negotiating leverage. The people affected aren't in the room when the decision gets made.
- Grid strain — Local power infrastructure, built for towns, is being redirected toward compute clusters, with costs socialized and benefits privatized.
- Product philosophy as unregulated policy — When platforms choose permissiveness as a market differentiator, that's not a neutral feature choice. It's a values decision, made unilaterally, deployed globally, with no public input and no recall mechanism.
Every one of these is a policy decision wearing a business decision's clothes.
Why "Just Trust the Market" Doesn't Work Here
Markets work when consumers can exit. You can't "exit" a labor market restructured by AI, a water table drained by a data center, or an information ecosystem shaped by a chatbot's design philosophy. These are externalities at civilizational scale — the classic tragedy-of-the-commons pattern, except the commons is the future itself.
The Framework: Power Should Match Consequence
Here's the actionable principle: the scope of a decision's impact should determine the scope of who gets a say in it.
- Decisions affecting a company → company governance.
- Decisions affecting a community's water, land, grid → community governance.
- Decisions affecting global information ecosystems and labor markets → something we haven't built yet, but urgently need.
Right now every layer of that ladder is missing except the first.
What Actually Moves the Needle
Don't wait for federal AI legislation — it's too slow and too captured. Instead:
- Show up locally. Data center zoning fights are winnable at the county level right now. This is the actual frontline.
- Demand governance stress-tests, not mission statements. Ask labs: what happens when five people disagree? Who has a vote if you don't?
- Treat "permissiveness as differentiation" as a red flag, not a feature. Products that compete by removing guardrails are running an uncontrolled experiment on the public.
- Build alternative infrastructure. Local, private, auditable AI systems reduce dependency on centralized decision-makers — this is a genuine lever, not just a hedge.
The future isn't being stolen in some dramatic villain-movie moment. It's being quietly allocated, water table by water table, board vote by board vote, while everyone watches the wrong debate.