Thread by @DellAnnaLuca

A mistake that smart people make:

FORGETTING THAT HUMAN BEHAVIOR IS DYNAMIC

aka "The Fence Paradox"

aka "the brain is a homeostatic machine"

aka "human behavior is dynamic and cannot be analyzed statically"

[1/10]

The image 👆, by @DrCirillo, explains all kind of human tragedies:
- student debt (the loans being the fence)
- why the Anti-Braking System actually *increased* the number of car deaths (the ABS being the fence)
- why some financial regulations & bailouts *cause* crashes
[2/10]

First, read the text in the image 👆.

The core principle behind the Fence Paradox is called risk homeostasis.

Humans adjust their behavior to maximize efficiency/effectiveness while keeping a *perceived* risk threshold constant.

For example… [continues below]

[3/10]

…if a new technology makes it safer to drive at 70mph, then humans would rather drive at 80mph to get home faster rather than keep driving at 70mph and be safer.

[4/N]

(Speed limits do not mitigate the example above, because accidents to not happen because someone was going above the speed limit on a straight road, but because someone did not reduce his speed before a turn on a rainy day because he thought it wasn't risky.)

[5/10]

The Fence Paradox predicts that technologies and policies that eliminates high-frequency incidents increase the damage of low-frequency ones (by virtue of reducing risk perception and thus inviting risk-taking, skewing the "frequence-damage" curve).

[6/10]

As @nntaleb described in Antifragile, lack of stressors is bad for antifragile beings (as we humans are).

And as I described in my essay luca-dellanna.com/fragilization, removal of high-frequency stressors (eg. "scares" or "near misses") leads to fragilization, -> catastrophes.

|7/10]

Many politicians and entrepreneurs tend to support the introduction of technology and policies that would work well if the world was static, but are counterproductive in a dynamic world dominated by second-order effects (in particular, how human behavior adapts).

[8/10]

In such case, they invariably appear right for years, because it takes time for behaviors to adapt and for the low-frequency negative events to surface.

[9/10]

I believe that the Fence Paradox model is very useful in evaluating whether a technology or policy will have a positive or negative overall impact on the safety of people, and should be mastered by anyone who is in a position to propose any of them.

[10/N]