Sublime

As in the physics example, if we want to build an AI to work with human behavior, then we need to build the statistical properties of human networks into machine-learning algorithms. When you replace the stupid neurons with ones that capture the basics of human behavior, then you can identify trends with very little data,

The key was control, or self-regulation. To analyze it properly he borrowed an obscure term from electrical engineering: “feed-back,” the return of energy from a circuit’s output back to its input. When feedback is positive, as when the sound from loudspeakers is re-amplified through a microphone, it grows wildly out of control. But when feedback i

Claude Shannon,

This paradox of comprehension was articulated explicitly by a great physicist of an earlier age: “Sir Isaac Newton, when asked what he thought of the infatuations of the people, answered that he could calculate the motions of erratic bodies, but not the madness of a multitude” (quoted from The Church of England Quarterly Review, 1850).

In quantum mechanics, small deviations from a specified initial state tend to cause only small deviations from the predicted final state. Instead, accurate prediction is made difficult by quite a different effect.

In simple cases uncertainty arising from exogenous events can be handled by estimating the probabilities of these events, as insurance companies do—but usually at a cost in computational complexity and information gathering. An alternative is to use feedback to correct for unexpected or incorrectly predicted events. Even if events are imperfectly a

Scientific models that seek to predict the consequences of human actions with some reasonable accuracy—such as game theoretical models of economic behavior—for the most part ignore human individuality in favor of aggregated outcomes.

One of the greatest programmers in history, who did indeed understand every component of the computer from the software down to the chip, Niklaus Wirth, wrote “the belief that complex systems require armies of designers and programmers is wrong. A system that is not understood in its entirety, or at least to a significant degree of detail

Pearl realized that it’s OK to have a complex network of dependencies among random variables, provided each variable depends directly on only a few others.