Pieza independiente

The four asking for a brake

Movement 1 · The Brain — September 2026

I The line that does exist

In August 2026 a lab said it had temporarily slowed its scaling in the face of critical cybersecurity capabilities. The same month, another company published its doctrine of personal superintelligence for billions of users and restated its commitment to open weights.

That is where the year's fracture shows. It does not run where people usually say it does.

It is not artificial intelligence against humans. It is acceleration against friction.

II The two sides, which are not blocs

Position	What it proposes	Why it is not a bloc
Friction	independent evaluators inside the labs, coordination between competitors, risk institutes	the four asking for it propose different things and cannot stand each other
Acceleration	scepticism about coordinated slowdown, open weights, mass deployment	their reasons are mutually incompatible; they agree only on the refusal

On the friction side there is someone building data centres at full speed while asking for a brake [and saying so without discomfort]. On the other there is the one selling the substrate and the one giving away the weights, who want nothing like the same thing.

“Whoever sells the metabolism does not want the organism on a diet.”

III Where the line actually falls

It does not separate the good from the bad; it separates whoever makes the substrate from whoever builds on top of it.

A chipmaker loses with any friction; its business is deployed volume. A frontier lab can win with some: a mandatory audit raises the cost of entry and protects whoever is already inside [regulation as a moat]. That does not make the argument false, only interested.

IV How to read a call for a brake [safe]

Look at who would bear the cost [in money and in months]. If the proposer does not bear it, it is a barrier to entry.
Look at whether it proposes an instrument or an intention.
Look at whether it accepts an external evaluator with real access.
Look at whether its infrastructure spending behaves like its speeches.
Look at what would happen to its market share if it were taken at its word.

V Risks

Treating this as a moral debate between visionaries and the reckless makes it impossible to see the one thing that decides anything: the industrial incentives of each node.

There is also an obvious capture risk. Oversight designed by the supervised produces a more closed industry, not a safer one. The precedent exists: finance, pharmaceuticals, aviation.