I don’t know if Nassim Nicholas Taleb considers himself a radical skeptic, but as an already fairly skeptical person myself, his book Antifragile definitely pulled me towards the more radical end of the spectrum. Regardless of his epistemological world view, one thing that Taleb undeniably has plenty of is controversial opinions. From his frequent Twitter arguments with other prominent figures to his bashing rants against psychologists, it’s clear that his ideas are far from being universally accepted. But this was not the reason I started reading Taleb’s works, despite being somewhat of a contrarian myself. In fact, I had no idea that he was such a controversial figure until much after I had already started reading Antifragile.

The book, subtitled Things That Gain From Disorder, is the fourth in a series of books, together called Incerto, dealing with randomness and uncertainty. I had only read The Black Swan, second in the series and arguably Taleb’s most influential work, prior to Antifragile. But as each subsequent book tries to build and expand upon the previous ones, there is a consistent set of philosophies connecting them, like the string running through a set of beads. So while reading these books, I tried to shift aside the beads and get to the core in the middle stringing them together.

Taleb is very fond of coining new terms and adages. In fact, the entire third book in the Incerto series is dedicated to talk about his set of aphorisms. I’m not sure how successful he has been at getting his coinages to catch on, but it is no surprise that he begins the book with a brand new created term, “antifragile”, in the title. “Fragile” is a very well defined term, describing something that is easily harmed by stress, such as a porcelain vase. But the opposite of fragile, Taleb points out, is surprisingly missing from our vocabulary. Most people (including myself) would think of “robust” when asked about the opposite of “fragile”, but this is actually not quite correct. Robustness is merely being resilient to harm from stress, while the true opposite of fragile, i.e. “antifragile”, is being strengthened or improved from stress. In a sense, fragile, robust, and antifragile form the three main types of properties that things can have, denoting their tendency for negative, neutral, and positive changes of state when subject to stress, respectively.

I think the explanation and justification for the new term coinage is fair. In fact, it’s so obvious in hindsight that I was a little surprised that it took this long for the term of antifragile to be coined (Taleb claims that he is the first, after some research) and popularized. Having laid the semantic groundwork, Taleb goes on to assert that most of our modern world, with our fancy technologies and rationalist models, has made us much more fragile while simultaneously giving us the illusion of robustness. Modernity, he defines, began when societal, technological, and engineering progress reached the point where they could eliminate most uncertainty and unpredictability from our lives. Questions like “will I have food for dinner” or “how can I travel across the country to see my family next weekend” do not even enter our minds anymore, because the assumed answers of “the supermarket will have all the food I want” and “an airplane will take me there” are deeply ingrained in us. It’s easy to believe that we are robust, being surrounded by concrete walls with fridges filled with food, but this elimination of uncertainty from our thinking is precisely what makes us fragile, because we have lost the ability to deal with any deviation from the norm. In comparison, an ancient farmer or hunter-gatherer is much more robust than us. Sure their typical quality of life is much lower than us today, but they are much more capable of dealing with random, unpredictable stress, such as the sudden collapse of some infrastructure.

This fragility-disguised-as-robustness extends into many other parts of our society. This is where Taleb begins to pull me in the direction of radical skepticism. He claims that most scholars, politicians, and economists are essentially delusional scammers, who not only fool us into thinking that their nonsense theories are correct, but also actively harm us by applying their theories in bad interventionist strategies.

He illustrates the absurdity of the modern theories and models with an imaginary story of “lecturing birds on how to fly”. A university professor, while giving a lecture, notices that there are birds sitting outside the classroom window. After the lecture ends, the birds fly away, and the professor hilariously concludes that they have learned how to fly from his lecture.

Taleb believes that most models that academics create to explain complex phenomenons fall prey to this type of reversed cause-and-effect error, in manners just as ridiculous, if not more, as the hypothetical story. He points out that the more rigorous fields dealing with simpler systems are much less subjected to this problem, such as most of the natural sciences (e.g. physics). Instead, he focuses quite a bit on pointing out these errors in economics and trading, since he studied business and used to be an options trader.

In the world of trading for example, most academics and bad traders religiously believe in fancy mathematics, and any theory that is backed by complex equations (which nobody can understand anyway) are taken as legitimate, even when they have no track record in practice to back up their credibility. Whenever we encounter new information, it is cherry picked to support the model we have, so additional data only serves to wrongly increase our confidence in the model, without actually improving it (in fact the more data we have available, the easier it is for a “skilled” researcher to manipulate it to fit any model). And whenever the model fails to predict, there are always convenient excuses on why (e.g. randomness, or some factors did not match perfectly with the model’s assumptions; often justified with even more equations), and is soon forgotten.

The smart traders on the other hand, do not care much for abstract mathematical models, and instead make up their own heuristics from experience, which often work in practice but is not accepted by the academic community. They do use math to help them understand the data, but the data comes first, not the model. They reject models invented out of thin air, from some pure rationalist argument in a vacuum (with fancy math). But even when these practical strategies are seen working, most people still will not accept them, until they are retroactively fitted to a set of complex math equations. These equations do nothing to actually explain the strategies, but are made only to appeal to the indoctrinated model-seeking minds of the academia.

Much of this section of the book read like a huge rant to me. I could feel, very viscerally, the frustration and anger that Taleb has for these blindly-model-loving people (whom he calls “Soviet-Harvard academics” in yet another coinage), who not only mislead people with their false theories, but are also honored to the highest degree by society (e.g. the Nobel Prize in Economics). There is perhaps a sense of bitterness, as Taleb realizes that he will probably never win a Nobel Prize, even though he, as the MythBuster, deserves it more than the people peddling the myths. For the most part, I agree with his skepticism towards the rationalist models we make to explain the complicated systems we encounter; most of them are probably wrong and only serve to make us feel more secure. Where I disagree somewhat, is in the usefulness of such models. I think that even wrong models have their utility, and often could turn out to be better than the pure chaos that ensues from abandoning all models altogether. After all, is it so wrong to ask for some sense of security, even when it’s partly false?

So coming back to the title of the book. If our modern world and theories are fragile, what is antifragile? The first obvious answer that the book gives is biology. The human body for example, like all biological organisms, has many antifragile properties. Our bodies are strengthened from stress, as long as the stress isn’t too great to cause instant, irreparable damage (e.g. death). A key component of antifragility turns out to be that repeated small harms do not accumulate. After all, if small harms accumulated in our bodies, we would have died from simply walking, as taking 100,000 steps would probably have exerted the same total force as falling off of a tall building, which is definitely lethal. Exercise is another example showing the antifragility of the human body. We essentially expose our bodies to constant stress by exercising, which in turn causes us to become stronger. We can contrast this to a vase, which is fragile and definitely does not become stronger (or even remain the same) when subject to repeated small stresses (e.g. dropping it from a low height).

A parallel to this that the book brings up is the fact that we tend to prefer smoothness in economic development, which leads us to make up theories, then apply them to intervene in the economy in order to avoid even minor jaggedness (i.e. shocks). These interventions often don’t work out in the way we expect them to of course, but even in principle it’s not good to have an overly smooth economy. Unevenness in the economic development curve is indication of constant, small stresses, analogous to exercises for the body. They help strengthen the economy against future unpredictable shocks. Taleb says that our economy could be much more resilient to large, “Black Swan” style events, if we simply let it run with frequent smaller crises untouched.

Back to biology. Why do biological organisms exhibit antifragility? Because the evolutionary aspect of nature favors the antifragile; things that are fragile simply would not have survived. Although Taleb is not an evolutionary biologist (neither am I), he does seem to know enough about it to avoid the most common pitfalls and misrepresentations about evolutionary theories, at least when judged with my limited knowledge. From here he brings up another key point about the antifragile: the only thing that can truly demonstrate antifragility is time. Since we are very bad at predicting fragility (and in general, after all most of our models are lies), the only sure-fire way to separate the fragile from the robust and antifragile is to let time do its work, i.e. kill off the fragile.

The longer something has been around, the longer its predicted lifetime into the future will be. This applies almost universally to everything that does not have a definite, limited lifespan. The Egyptian pyramids will probably last longer than most skyscrapers; Homer’s Odyssey will probably last longer (in terms of having a relevant sized reader base) than Hamlet, which in turn will probably last longer than most modern books from the past few decades. This is the opposite of how we typically think of longevity. In contrast for example, a person born 60 years ago will probably not last longer than a person born 30 year ago, since there is an inherent limit to the lifespan of humans.

Okay, but what if I’m mortal, and don’t have eternity to wait out and see what is fragile and what is not? Here, Taleb tries to identify some general properties of things that are antifragile, to help us spot them. They all come down to a single, central concept: optionality. Optionality is the general property of asymmetric upside, provided by having the right, but not obligation, to exercise some option. The canonical example used in the book is that of an ancient Greek farmland lessee, who rents/buys the option to use a farmer’s land over several years, for a relatively cheap price per year. If the year’s climate is good, the lessee will exercise the right to use the land, plant and reap the benefits of a great harvest. If the land is suffering from a drought, then the lessee can simply do nothing that year (instead of being forced to plant and lose all crops), and only lose the low annual rent.

Graphically, optionality is represented by convexity and concavity (or anti-convexity) effects. If we plot the potential loss and gain on the y-axis, and some random variable on the x-axis, representing some unknown event X that impacts the outcome, then the antifragile property is represented by a convex curve (i.e. bottom half of a circle) drawn from the bottom left to the top right, and the fragile property is represented by a similar but concave curve (i.e. top half of a circle). These curves illustrates the potential loss and gain between the antifragile and the fragile respectively, contingent on the occurrence of event X. If event X does not happen, the convex curve suffer limited loss, while the anti-convex curve suffer unbounded loss; if event X does happen, the convex curve gain unbounded profits, while the anti-convex curve gain limited profits.

To tie it back with the previous points about stress on the body, one can think of randomness as potential unknown stresses, and in a sense, the role of nature for biology. Then, what Taleb is trying to say is simply: that which benefits asymmetrically from randomness (i.e. unknown stresses) is antifragile; that which is harmed asymmetrically from randomness is fragile. Optionality then, is a key provider of the antifragile property, as it allows for the exploitation of the asymmetric gains.

Taleb says that optionality is everywhere, if you know where to look. Having no debt and some money saved up, living in a large city, and acquiring additional talents not directly related to your current job all increase your optionality. Venture capitalist investing also has antifragile properties by exploiting optionality. In VC funding, the downside is small and capped (i.e. if a startup fails, you lose no more than your initial, relatively small investment), while the potential upside is massive, usually uncapped (i.e. if you manage to hit a Facebook-level success, the return could be many times the investment).

One thing to be careful of when searching for optionalities, Taleb warns, is to make sure that the optionality isn’t already priced in. Certain, usually more mature spaces could have, over time, incorporated the asymmetric gains provided by the optionality into the initial price, in which case there is no more opportunity to exploit. For example, in the ancient Greek farmland leasing story, after a few years, the smart farmer would’ve realized the lessee’s exorbitant financial gains, and when the lease is up for renewal, re-negotiated for a higher annual rent, in order to price in the profits of the lessee’s optionality. It then becomes impossible for the lessee to profit, at least not at the same level. Another, more modern example could be that of the Bay Area. It is a great place to be, offering a tremendous amount of optionality to entrepreneurs and people in technology in general. However, a lot of that benefit has been gradually priced-in to the insane cost of living there, so another up-and-coming city that is developing fast, where the extra optionalities have not been fully accounted for by the market could be a much better place to be.

In the last third or so of the book, Taleb returns to the earlier idea of harmful interventionism, but specifically focused on the version that medical professionals are guilty of, called iatrogenics. People (called “interventionistas”; of course there is a special name for them, it’s Taleb we are talking about here) naively intervene (i.e. using bad models and with no regard for the potential negative consequences) all the time, in areas such as economics and politics. This is bad because naive interventionism has an anti-convex loss/gain curve. Usually an intervention is enacted to fix some known problem (bounded gain, i.e. best case is problem fixed), but has the potential to cause unintended harm that is unbounded (i.e. worst case is everything collapses, since our models are bad at predicting second and higher order effects). In other words, naive interventionism is fragile.

This makes sense. But I find it slightly odd that Taleb, who has little experience in the field of medicine (which he openly admits in the book), chose to spend many pages dedicated to iatrogenics, harm done by the healer. Maybe he’s had some very bad experiences with doctors in his life. Speculations about his motives aside, I do think he has a point, in that this is not a very talked-about topic in society. In fact, I did not even know about the term “iatrogenesis” until this book, although that could also just be due to me having only a high school degree. Apparently it is a very serious problem, and a leading cause of death worldwide, in the same order of magnitude as cancer. I did some quick digging, and the facts presented in the book seem to reflect the reality. In the US, it is estimated that between 225,000 to 284,000 people die annually from adverse effects of medical treatments and drugs. Compare this with the estimated annual deaths of 609,640 from all forms of cancer in the US, it does seem like this is a pretty serious issue.

This is just the tip of the iceberg, Taleb claims, as it is only an estimate on deaths, the worst, most obvious result of iatrogenics. There are many other subtler effects of iatrogenics that cause non-lethal harm, some chronic, some hidden even, that are simply unreported. The real impact of naive medical interventionism could be much worse than this. He brings up older examples such as the harm from the historic use of bloodletting, as well as modern examples like the ubiquity of mammography, which apparently has brought much more harm to women than lives saved.

Taleb pleas to doctors to consider the anti-convexity effects of medical treatments. Obviously in a life and death emergency situation, any treatment should be used, because we are already exposed to the worst harm possible. But under not-so-extreme circumstances, we should be very careful with medical interventions, because we would be in an anti-convex situation, with bounded gains and potentially unbounded harms (i.e. death). Humans, although have antifragile properties under nature’s stress, are in fact extremely fragile to drastic artificial interventions.

Overall this section of the book has the sentiment of “nature knows best, so we should let nature take its course whenever possible”, which I’m not sure that I agree with. But medical over-treatment and iatrogenics definitely do exist, and may actually be a more serious issue than most people realize. In the very least, perhaps we should start to pay more attention to it. It is difficult to even find good sources of data on iatrogenics, partly because it is such a difficult problem to research (medical professionals are probably disincentivized to dig into this themselves). I do personally share the belief that random small stressors are good for the body from a nutritional point of view. Taleb advocates for intermittent fasting in the book, and for a more varied diet, both of which I think are good, or at least better than their alternatives.

Talking about iatrogenics gives a natural transition into the broader sense of ethics in the context of antifragility, which is where Taleb goes to close off the book. Here he introduces the concept of “skin in the game”: you have skin in the game when you personally put something of value on the line. Taleb romanticizes about the warriors and generals of the past who risked their lives in battle, the builders who were liable with their own lives when the house they’ve built collapses and kills the owner (an ancient law from the Code of Hammurabi), then laments on how this kind of heroism is all but gone from our modern world. These people from history had the ultimate “skin in the game” - their lives, and there is something noble and sacred about it that Taleb aspires to.

I don’t think that the civil engineers of today need to put their own lives on the line when building houses and bridges; after all the reputation based system we have now seems to be good enough for adequate architectural safety and quality. But the core issue that Taleb tries to bring to light here is very much true: too many people today are able to exploit the antifragile without risking anything themselves, essentially benefiting from the positive side of optionalities while shifting the negative exposures to others. Forecast analysts for example, often make prediction claims publicly, which other people act on (putting their skins in the game), but they themselves have little on the line. So they benefit when the prediction is correct (lauded as genius prophets), and lose nothing when it isn’t (nobody remembers wrong predictions).

From my understanding, I feel like this “skin in the game” is just a special form of the general concept of externality in economics. With the technologies and social landscape we have today, it is probably easier to exploit that now than in the past, and not just in finance and economics. Truth claims of all kinds are often made on social media without skin in the game, things like accusations of past bad behavior, and grand diagnostic statements about societal problems. It is a balance of course, because forcing people to put their skin in the game all the time will disproportionally silence the more timid. But with all the ridiculous noise we have in the world today, perhaps this is not a bad thing. I think Taleb’s message here is more relevant now than ever.

Overall I would recommend the book. If you are the type of person who often relies on academically-made frameworks to make decisions and think, then I’d say that Antifragile is an especially good book, as a way to balance out your world view. It will seem very radical, and you will find it difficult to accept many of the claims in it, but it is nonetheless a good argument for “the other side”.

Taleb does a very good job identifying antifragility in the book, but I find that there aren’t as many practically applicable, constructive ideas in how to create antifragility, turning something fragile to antifragile. Perhaps antifragility is something that can mostly only be identified then exploited, and not created when there are none to begin with. Or perhaps I have not yet fully internalized and truly digested the core ideas (which is part of the reason that I wanted to write this review, to help me better flesh out the ideas).

Radical skeptics beware: this book (along with Taleb’s other works) will probably just reinforce your already extreme beliefs.