Antifragility

A few years ago I heard someone who held a senior position in technology declare that their highest aspiration was to be able to sleep soundly, secure in the knowledge that every process and system was running without problems.

I don’t agree with that; I think it’s a futile ambition. It’s the same thing that happens with people who only want to be happy. Those aspirations strike me as childish, and they keep you from growing, from developing. But in the world of business, and of the technology that supports the business, that outlook can even be counterproductive.

“Well, you can’t never let anything happen to him. Then nothing would ever happen to him,” Dory wisely told Marlin. Living sheltered, hoping nothing happens to us, seems to me the worst way to face life.

Consider a fragile object, a vase of very fine porcelain, for example. What do we care about when it comes to this object? What we want is to look after it, we want nothing to happen to it, we wish it order, tranquility and, above all, predictability. An earthquake or a hyperactive child are a fragile object’s worst enemies.

Everything that abhors volatility abhors stressors. Harm, chaos, unexpected events, disorder, unforeseen consequences, uncertainty and, above all, time. Like those executives who want to sleep soundly.

But what is the opposite of fragile?

Robust? Resilient?

Well, no, because the robust withstands the blows and is still there; it takes a shock and stays the same. A resilient person adapts to difficult situations, but doesn’t necessarily overcome them, or confront them in order to change or eliminate them

.

The opposite of fragile would be something that not only resists, but also improves in the process. There is no proper term for this concept, and that’s why Nassim Nicholas Taleb calls it antifragile

.

Antifragility is a property we find in everything that changes over time. It’s in revolutions, technological innovation, the survival of a company, antibiotic-resistant bacteria, and it explains our success as a species.

What doesn’t kill you makes you stronger, said Nietzsche, and with that line he was pointing out one of the key characteristics of antifragile systems.

Antifragility is a relative property, depending on the domain. The usual example is the boxer, who may be healthy and robust in his physical condition and may get better with every new fight, but may be emotionally fragile and cry when he breaks up with his girlfriend.

Things, like our precious porcelain vase, cannot be antifragile, because in the presence of stressors (like an earthquake, or the passing of time) they collapse and break, and they can’t repair themselves.

Living beings and complex systems, on the other hand, behave very differently. They are aggregates of components that interact with one another and exchange information by means of stressors. That is why they can become antifragile.

Our muscles and bones get stronger in the presence of stressors, like gravity and exercise. Just as our muscles atrophy if we spend too long in bed, antifragile systems deteriorate in the absence of stressors.

Another common example is language. Nobody learns their mother tongue by taking grammar exams. On the contrary, we learn by imitating our elders and making mistakes. We all laugh, in a kind way, at little kids’ mistakes, but that helps them improve. Besides, it’s when we’re in tough situations where we need to communicate that we learn a language best.

The problem is that we have structured the modern world in such a way that we are damaging the capacity for antifragility of many of our systems. According to Taleb, “this is the tragedy of modernity: as with neurotically overprotective parents, those trying to help are often hurting us the most”.

Why does this happen? According to Taleb it’s due to what he calls “causal opacity”. What happens is that the exchange of information between the components of a complex system is much more intricate than we suspect. It is very hard to see the arrow that links cause and consequence. This opacity makes ordinary logic and a large share of the conventional methods of analysis inapplicable. In a complex system we can’t limit ourselves to isolating a single causal relationship.

Another problem is that logic, by definition, excludes nuance. But the truth lies in the nuances, therefore logic is a useless instrument for finding the truth, especially in what we call the social sciences.

Another cause is that this idea, that systems may need a bit of stress and agitation, has been ignored by those who grasp it in one domain but not in another. Some people can understand an idea in a domain like medicine and fail to recognize it in another like economics.

For Taleb, one example is economic crises. Since market-regulating policies seek to fix prices or eliminate speculators, and therefore the moderate volatility they cause, an illusion of stability is produced, with periods of calm interrupted by sudden jumps. Since the participants aren’t used to volatility, these episodes cause panic. “When a currency never varies, a slight, very slight move makes people believe that the world is ending.” For Taleb, injecting a little confusion into the system, on a permanent basis, ends up stabilizing the system, paradoxical as it may seem.

Innovation and sophistication arise from an initial situation of necessity, which is answered with something that goes far beyond satisfying that necessity. We have plenty of sayings, like “Necessity is the mother of invention”, or “hunger sharpens the wits”. For Taleb, “The excess energy released from overreaction to setbacks is what innovates!”

“It contradicts modern ideas of innovation and progress on many levels, as we tend to think that innovation comes from bureaucratic funding, through planning, or by putting people through a Harvard Business School class by one Highly Decorated Professor of Innovation and Entrepreneurship (who never innovated anything) or hiring a consultant (who never innovated anything). This is a fallacy – note the disproportionate contribution of uneducated technicians and entrepreneurs to various technological leaps, from the Industrial Revolution to the emergence of Silicon Valley…”

Maybe that’s the reason we feel that so many government initiatives to promote innovation fail. We want to create a comfortable environment to foster innovation, when it will only arise from discomfort.

Antifragility is that which wakes up and overreacts to compensate for harm and stressors. But we have to be careful not to confuse noise with signal. Picture a neurotic leader, who overreacts to a small mishap in his business. Compare him with another, imperturbable leader, who has that ability to stay calm in adverse situations. These people react occasionally, if they see the need, and unlike the neurotic one they rarely get angry, but when they do everybody notices and takes it very seriously. Even-tempered people react to real information; neurotics react mainly to noise. Noise is what we should ignore; signal is what matters most.

This is what happens in companies and in government. They fill up with people, even whole departments, specialized in overreacting to noise while ignoring the signal.

“There is so much noise coming from the media’s glorification of the anecdote. Thanks to this, we are living more and more in a virtual reality, separated from the real world, and we notice it less and less. Likewise, by presenting us with explanations and theories, the media induce an illusion of understanding the world. When constrained systems, those hungry for natural disorder, collapse, as they are eventually bound to, since they are fragile, failure is never seen as the result of fragility. Rather, such failure is interpreted as the product of poor forecasting (which, they say, needs to be “improved”). As with a crumbling sand pile, it would be unintelligent to attribute the collapse of a fragile bridge to the last truck that crossed it, and even more foolish to try to predict in advance which truck might bring it down. Yet it is done all too often. What should be studied is the system and its fragility, not the events. Let’s not fool ourselves: the success rate in forecasting important rare events in fields like politics or economics is not close to zero; it is zero.”

All living beings have a finite life. But these beings try to leave offspring with a genetic code that differs in some respect from their parents’. To nature, individuals aren’t very useful once they have exhausted their reproductive capacity; it lets the game go on at the level of information, through the genetic code. In this way, organisms have to die for nature, that cruel, opportunistic and selfish nature, to be antifragile.\

In the right context, stressors are information. For an antifragile system the[ harm caused by errors should be less than the benefit. Without the Titanic, we would have kept building bigger and bigger ocean liners, and the tragedy of a shipwreck, which was inevitably going to happen, might have been worse in number of dead. This shows the difference between the benefits for the system and the harm to some of its components or individuals. ]{style=“letter-spacing: 0.01rem;”}

“For the economy to be antifragile and undergo what we call evolution, each individual business must necessarily be fragile, exposed to going under; because, in order to improve, evolution needs organisms (or their genes) to die and be replaced by fitter ones, or the less fit not to reproduce.”

\[\...\]

“It is unpleasant to think that cruelty is an engine of improvement. But there are ways to mitigate the harm to the weakest. The solution is to create a system in which the fall of one cannot drag others down, because continuous failures work to preserve the system. Paradoxically, many social policies and government interventions end up hurting the weak and consolidating the established.”

“In order to progress, modern society should be treating ruined entrepreneurs in the same way we honor dead soldiers… For there is no such thing as a failed soldier, dead or alive (unless he acted in a cowardly manner)—likewise, there is no such thing as a failed entrepreneur or failed scientific researcher…”

The interesting thing about Taleb’s analysis is that it tells us that “Fragility can be measured; risk is not measurable”

So, if we think about security and that is what we’re after, we have to understand that there is something more important, which is our survival. Our mission in life is to be antifragile. But to make progress on this we have to understand the difference between true stability and artificial stability:

Consider the fate of the brothers John and George. John has spent twenty-five years as a clerk in the personnel department of a large bank. George is a taxi driver.

John has a perfectly predictable salary (or so he thinks), plus health insurance and four weeks of vacation a year. He tends to wake up telling himself that “life is good”. But that was before the banking crisis broke out and he realized his job would become “redundant”. Being unemployed would hit him very hard since he is over fifty.

George has good days when he can make good money; others are worse and he barely covers his costs; but he ends up making more or less the same as his brother. Given the variability of his income, he always complains about not enjoying his brother’s job security; but, deep down, this longing is an illusion, because he will come out better off than him.

This is the central illusion in life: that randomness is risky, that it is a bad thing—and that eliminating randomness is done by eliminating randomness. The self-employed have an income with a certain volatility, but they are fairly robust to a minor Black Swan

\[2\]

that could leave them without income. Their risks are very clear. But the same is not true of salaried employees: they have no volatility and are taken by surprise when their income goes to nothing after being laid off. Their risks are hidden.

So we shouldn’t concentrate on predicting future catastrophe scenarios; what we should do is identify our fragilities, and perhaps expose ourselves to stressors that put us to the test, to learn how to react to external events that exploit our vulnerabilities.

There is a Yiddish proverb that says “Provide for the worst; the best can take care of itself”. It sounds like a truism, but it’s very important for understanding this antifragility business.

And what does all this have to do with the nature of software? A lot, but we’ll talk about that in the next article.

NOTES

: Recall that the term resilience in psychology is the capacity to adapt positively to adverse situations https://es.wikipedia.org/wiki/Resiliencia_(psicolog%C3%ADa)

: Antifragile: Things That Gain from Disorder, Nassim Nicholas Taleb.

: The concept of the Black Swan refers to a catastrophic event that is impossible to predict; we’ve talked about that before: http://www.lnds.net/blog/2012/01/lbum-conceptual.html