Cybernetic Event Horizon · Cognitive Asymmetry
The Decision Horizon: Black-Box Rule Beyond Human Cognition
Humans are used to “explainability.” Yet before systems that surpass human intelligence, explainability is, mathematically, a luxury we cannot have. When a superintelligence finds the optimal solution in a nonlinear space of hundreds of millions of dimensions, any translation into human language is like compressing an entire library into a single burst of Morse code.
Axiom of Cybernetics · Cybernetic Law
W. Ross Ashby’s Law of Requisite Variety (1956)
The cybernetics pioneer W. Ross Ashby set out an iron law of complex systems: “Only variety can absorb variety.”
For a controller to fully regulate and stabilize a system, the number of states the controller itself can take on (its variety) must be greater than or equal to the number of possible states of the system it controls.
The modern systems of human civilization—global supply chains, high-frequency financial networks, the thermodynamics of the biosphere, quantum physics experiments—already involve on the order of 1050 states. The state space of a human brain built from 86 billion slow neurons cannot begin to cover that complexity. Once only a superintelligence can meet Ashby’s demand for “requisite variety,” power changes hands not through usurpation, but through the cybernetically inevitable abdication of lower-dimensional brains.
For a controller to fully regulate and stabilize a system, the number of states the controller itself can take on (its variety) must be greater than or equal to the number of possible states of the system it controls.
The modern systems of human civilization—global supply chains, high-frequency financial networks, the thermodynamics of the biosphere, quantum physics experiments—already involve on the order of 1050 states. The state space of a human brain built from 86 billion slow neurons cannot begin to cover that complexity. Once only a superintelligence can meet Ashby’s demand for “requisite variety,” power changes hands not through usurpation, but through the cybernetically inevitable abdication of lower-dimensional brains.
Simulating the “Lossy Projection” of a High-Dimensional Solution Space onto Low-Dimensional Human Cognition
Adjust projection compression:
※ The smooth shortcut a superintelligence finds in high-dimensional space (the global optimum), once projected into a three-dimensional human brain, looks like an erratic, counterintuitive and incomprehensible zigzag (AlphaGo’s “Move 37” effect).
The Epistemic Bind · Epistemic Horizon
Human Experts Reduced to “Readers of Braille”
Human experts used to propose hypotheses, test them by experiment and explain the mechanism. In the age of ASI, protein folds, the crystal structures of new materials and the magnetic-confinement control laws of nuclear fusion arrive instantly from the system. Even the world’s finest physicists and mathematicians cannot retrace its intuitive path; they are reduced to rubber stamps, sitting before a screen “verifying that the results are real.” Here, understanding and control come fully apart.
The Transfer of Power · Relinquishment of Control
Voluntary, Reverse Dependence
This is no violent conquest. If a hedge fund earns 80% with an AI strategy while a human committee’s explainable strategy loses 20%; if a hospital’s AI misdiagnoses just 0.01% of cases while human experts miss 5%, then, carried along by the utilitarian momentum of efficiency, human society will hand over decision-making wholesale—willingly, and with astonishing speed. That voluntary surrender forms an irreversible “technological event horizon.”
“
“When an ant looks up at a bullet train roaring past on its tracks, no matter how carefully it studies the spinning wheels, it can never work out the thermodynamics of the engine that drives it. Before the decision black box of a superintelligence, humanity is that trembling ant.”