Thread by @DellAnnaLuca

ADAPTATION MECHANICS

1/ How do people, species, companies and civilizations adapt?

What are the requirements for antifragility?

How does convexity (the reaction to stressors) change as the entity adapts?

(thread)

2/ For those who didn't read Nassim Nicholas Taleb's book Antifragile, in which the concept is described (you should!):

An entity is antifragile if it increasingly benefits from volatility

An entity is fragile if is is increasingly harmed by volatility

Example in the pic below

3/ An entity is antifragile if and only if:
• It comprises multiple components or sub-entities
• The components can break independently
• New components are generated continuously or at least in response to other components being damaged.

4/ No antifragile entity exclusively reacts to stressors in an antifragile way. Even antifragile entities can exhibit fragile responses.

E.g., humans are antifragile and lifting weights
causes them to become stronger, but attempting to lift a too-heavy weight causes injuries.

5/ Therefore, antifragility is always related to both an entity and to a range of stressors.

The antifragile behavior is expressed only for stressors whose magnitude is between the First Damage Threshold and the Functional Impairment Threshold.

6/ The two thresholds are defined in yesterday's thread on damage mechanics

https://twitter.com/DellAnnaLuca/status/1208420849373589504

7/ Fragile entities react similarly, but do not have a range of stressors producing an antifragile response.

In other words, the two thresholds coincide with no space in between.

8/ An entity is fragile if any of the following are true:
• It is monolithic
• It is brittle (i.e., damage to its components is immediately followed by functional impairment at the entity level)
• It lacks the ability to (re)grow its components

9/ In in antifragile entities, micro-damage causes strengthening and prolonged lack of damage causes weakening (examples in the pic below).

10/ All other things equal (for example, internal structure), the stronger the entity, the more antifragile. The weaker, the more fragile.

11/ Putting together the content of the previous tweets, we can say that micro-damage is followed by antifragilization; prolonged lack of micro-damage is followed by fragilization.

12/ This picture describes fragilization.

Note: it's not absence of damaged that causes fragilization, but *prolonged* absence of damage.

In other words, it's not comfort that makes you weaker, but absence of discomfort.

13/ This picture describes the dependency between convexity at different levels of an antifragile entity, as it undergoes (anti)fragilization.

14/ In general (limitations will be mentioned later),

An antifragile entity tends towards a stable equilibrium with the sample of stressors it recently experienced (assuming no "intelligent proactivity").

Justification in the picture below

15/ Which means that an antifragile entity exposed to a non-representative sample of the environment achieves a partial equilibrium which is fragile and thus temporary.

Justification in the picture below

16/ The points above are expanded in my paper "The Dynamics of Risk-Taking", which contains details, limitations and applications.

luca-dellanna.com/the-dynamics-o…

17/ Of course, antifragility is described much better in @nntaleb's book Antifragile.

The reason I wrote this thread is to link antifragility to the concept of First Damage Threshold & Functional Impairment Threshold (described in yesterday's thread) to prepare for tomorrow's.