However, too big of a variation, and the antifragile breaks.
Similarly, too much small damage too often, and the antifragile breaks too.
The latter is called "fatigue".
Let's see how it works.
1/N
2/ This thread assumes you've read @nntaleb's Antifragile.
Building on it, here is a simplified visualization of how the antifragile reacts to stressors.
3/ And this is a visualization of what happens over time to the antifragile when it is hit by a stressor that causes sparse damage
The antifragile becomes stronger → it becomes both less susceptible to sparse damage and to functional damage (both thresholds move right).
4/ However, and this is key, the previous chart only shows the long-term effect of sparse damage. In the short-term, the antifragile becomes weaker.
Example: two days after a very intense gym session we are stronger, but 10 minutes afterwards, we are weaker.
Healing takes time.
4B/ Similarly, there are events that might seem fragile in the short term (example: a business downturn causing a company to fire a few employees) but that can result in a strengthening over the long-term (the company regrows stronger).
How to visualize them?
5/ If we add a time dimension to the chart, it looks as follows.
Over the next tweets, I explain step-by-step how to read it.
6/ Imagine that a stressor hits with the strength indicated by the THICK black arrow
7/ To know the short-term effect, we vertically project the stressor magnitude towards the short-term axis, and read the damage type.
In the example, point (1) indicates that the short-term reaction is fragile (red). The whole is now less strong than it was before.
However, …
8/ …to know the long-term effect, we vertically project the stressor magnitude towards the long-term axis, and read the damage type.
in the example, point (2) indicates that the long-term reaction is antifragile.
Over time, it recovered and overcompensated, getting stronger.
9/ The slope of the thick black oblique threshold in the chart below represents the rate of healing / growth / regrowth.
10/ If we assume that the distance between the three axis is fixed, we can see what happens in the medium term.
For example, point (X) below would represent the situation a few hours after the hit, but a few days before point (2) is reached.
(Time passes vertically, downward)
11/ Earlier, I said that the slope of the oblique line represents the rate of healing.
As the angle α tends to 90 degrees, healing is instantaneous.
As α is lower, healing is slower.
Let's see why it matters.
12/ This chart depicts a slow-healing entity.
After the first hit, we have the situation at point (1): a more fragile body than we started with.
13/ It immediately begins to heal, but if the second hit comes shortly after, at the point in time described by the X, it hits the body while it's still weak, before it managed to heal.
The second hit causes even more damage.
Damage accumulates → the body becomes more fragile.
14/ In general, we can say that if hits are repeated before time t passes (see chart below), then the body undergoes FATIGUE.
Fatigue is just the mechanical term for accumulated damage.
15/ t depends on the slope, i.e. the healing rate.
The higher the healing rate,
→ the more the oblique bar is vertical,
→ the lower is t,
→ the less time can pass between hits without accumulating fatigue.
16/ Conversely, the slower the healing,
→ the more the oblique bar is oblique (i.e., tends towards horizontal),
→ the larger is t, (i.e., the minimum time to recover between hits)
17/ Note that by "healing", I can intend:
– the healing rate of a muscle,
– the time to regrow population after a natural disaster,
– the time to find a replacement after an employee left the company,
– and so on.
18/ With no healing rate, any hit in the green area will also cause accumulated damage (fatigue).
19/ That said, reality is not as clear as depicted so far.
For example, even in the case of no healing rate, a first small hit that causes sparse damage might, in some cases, cause both fatigue AND strengthening (eg: a downturn causes a company to fire a couple bad employees)
20/ Also, in real life we are subject to many stressors, hence we have dimensions each described by the charts above, and some of them constitute trade-offs
A third consideration: all chart lines are blurred, as individual components (eg muscle fibers) are heterogeneously strong
21 / That said, the charts above should at least clarify the effect of fatigue relative to antifragile entities.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
An example of tragedy of the commons. A city has access to a small forest.
If you are a citizen, self-interest would dictate that you cut a tree to get some wood to heat your house. But if everyone does, the wood disappears.
1/N
2/ The tweet below recently made me notice that this kind of problems can be formulated in terms of #ergodicity.
3/ One of the definitions of ergodicity is "when the outcome of a person performing an action n times coincides to the outcome of n people performing the action once."
Here is how it applies to the tragedy of the commons:
There are three main reasons for irrational behavior.
Thread, 1/11
The first one is an excessively-narrow definition of irrationality.
Example: playing a boardgame, your friend makes a move which is suboptimal to victory. Is he being irrational?
Only if winning at the game is all that matters. But maybe he's optimizing for friendship.
2/11
Another example: working overtime might get you ahead at work but also set you back in personal life.
In general, it's never rational to maximize performance at a given task. It's rational to maximize one's performance across all tasks. And sometimes, the two are at odds.
Too often, consulting engagements end up making the client more fragile.
Why? What can clients AND consultants do differently?
(thread, 1/N)
2/ One possible cause is, of course, psychopathic consultants & self-centered management. But there's a lot that can go wrong even when both parties are well-meaning.
Here are 3 problems that must be addressed to avoid consulting disasters, even assuming competence & good faith.
3/ I often say, "centralization is only efficient to the central observer"
Similarly, consulting is often only effective to the central observer
If it only considers what matters to the central observer, it will only do good relative to what's considered by the central observer
I've been using for a few months an app that sends me daily digests of information *curated by myself*.
It just got better: I can now receive a digests of tweets liked by people I pick. Curation 2.0
(Thread of how it works, 1/N)
2/ I use it to automate repetitive "fetching information" tasks.
It has three advantages:
– I avoid having to manually open multiple websites or apps
– It's fast
– It prevents me from wasting time "mindlessly scrolling": it's an email, and once it finished, I can move on.
3/ Here is the website: mailbrew.com/?aff=DellAnnaL…
(affiliate link, but I've been a paying user for months, and I just love the product for how much time it saves me).
At the beginning I used it "just" to receive a daily email with the top tweets from a few Twitter lists.
2/ First of all, this is a Roam Book (rBook). It is a new format that integrates with your notes, allows for non-linear exploration of its contents, and more. You can read more about it on roam-books.com