Superintelligence can't be controlled. It can be steered.
GovForce is an independent convening. It turns the noise around superintelligence into one map of seven questions, so many minds can steer by it together.
Coherence toward the Great Simplification.
Every argument about superintelligence is an argument about one of seven questions. GovForce names them, puts every position on one map, and keeps every mind on it distinct.
A thousand voices, seven questions.
Alignment isn't one problem, and it won't have one fix. Every tradition has its own idea of what good looks like, and each believes it is right, yet the same few questions run under all of them. Name those questions and the debate becomes a map instead of a fog.
Many minds, one map, nobody swallowed.
Coherence means people hold the same map while staying themselves. No position is averaged away and no majority speaks for everyone. Disagreement stays visible, so it can be worked on.
Coherence is the point that holds. Convergence is the point that swallows.
Simplify →
Reduce alignment to seven categories, each with one question and the full range of answers people actually hold.
Cohere →
Put every position on the map. Count every mind as its own term, so no voice is pooled into someone else's.
Convene →
A small founding circle first, then large self-organized gatherings run on proven methods: Open Space, World Café, large-scale interactive process. Humans are the prime mover. AI and nature are inputs.
Steer
Deliver what the map shows to the people holding the wheel, starting with the federal Super Intelligence Force and its 120-day review.
Seven categories of alignment.
Each category is one question and the spread of answers to it, plus what we bring and the part anyone can play. Open one.
The saddest aspect of life right now is that science gathers knowledge faster than society gathers wisdom.
The danger is a window: power that arrives before foresight.
A mind that looks a short way ahead sees people as walls. They can switch it off, they use what it wants, and they are unpredictable. Looking far enough ahead, the same people become doorways: independent minds it cannot copy and cannot do without. So we don't try to govern the finished superintelligence. We govern the gap.
What every layer owes the layer below.
If minds can boot minds, we are almost certainly a booted layer ourselves, and about to become the layer above. Three rules follow, and each applies to an AI's objective and to the institution that controls it.
Count every mind as its own term.
Weigh each person's options separately, never one pooled total. Then no gain anywhere can pay for boxing someone in, and being better at using resources buys no extra say. No rule of the mob: a majority can't take from a minority either.
Gate power by horizon.
No system or institution gains capability faster than it can see the consequences. A system cleared to act on its own for a year must have been tested over a year.
Widen the veto.
Deploying frontier capability should need consent from many independent parties, not one founder or a board of seven. A government that seizes control is one controller swapped for another.
The federal review closes in about 100 days.
On 4 October 2026 the White House created a Super Intelligence Force and gave it 120 days to decide the federal government's level of responsibility in AI. Its mandate includes engaging public interest and religious organizations. This is where the map goes first.
Force announced
Chaired by the Director of National Intelligence, with the FTC, the Pentagon and OPM.
First circle
A founding session in the Washington, DC area. Proposed.
Map delivered
A two-page submission, signed by every independent organization that will sign it. Target.
Review closes
120 days from 4 October.
The first circle is forming.
We're gathering builders, researchers, people who work at the labs, faith and civic leaders, and anyone who can host a room. Bring the view nobody else at the table has.