Book Summary: Antifragile by Nassim Taleb

  • Antifragile – a term used to mean the opposite of fragile, and denote things that gain from disorder, chaos, stressors. It is not the same thing as ‘robust’ or ‘resilient’. Fragile things are negatively impacted by uncertainty, stress. Robust things are unimpacted by them. Antifragile things actually grow and flourish in the face of uncertainty and stress. The analogy “wind puts out a candle, but encourages a fire” embodies this.
  • This is an interesting and important concept to think about as it is all around us in nature and evolution – only the antifragile setups could have lasted billions of years in an unpredictable universe. Yet, no known language has a word for the ‘exact opposite of fragile’, leading the author to coin antifragile.
  • Human emotions are a good example to understand ‘antifragile’ behavior – thoughts or emotions such as anger and bitterness grow the more they are suppressed. Countless stories tell the tale of someone whose love/attraction for their chosen ones grows in spite of emotional abuse or mistreatment from the person they desire.

  • Fragilistas – A condescending term the author uses for folks who cause more harm, while trying to do good, by being overprotective and encouraging fragility. This includes overprotective parents who shield their kids from harm, medical experts who prescribe more medication for every small issue rather than helping develop immunity and health.
  • Being antifragile implies thriving on a number of small mistakes. Silicon Valley is cited as an example of antifragile-ism. Research that is narrow and focused is fragile – adding in random tinkering makes it antifragile.
  • Post-traumatic stress disorder is an instance of fragility. Post-traumatic growth (e.g. paraplegics finding themselves gradually more capable and thriving in life after losing their limbs) is an example of antifragility.
  • Chronic stressors are small (in impact), low-level stressors that are repetitive, occur often in our environment and increase fragility. The continuous stress of a boss or bad commute is an example of this. On the other hand, what the author calls ‘acute stressors with recovery‘ are relatively infrequent events, that start with an unexpected challenge (e.g. snake appearing out of nowhere) and end with conquering or containing that challenge. Acute stressors increase antifragility.
  • The economic system is inherently fragile. Unlike, say, a bunch of restaurants in an area (fragile individually, antifragile as a whole – unpopular or bad restaurants will go down in favor of better liked ones), globalized economic systems operate as one entity – errors spread and compound. It is not possible to have small mistakes and learn from them, in a manner like the global aviation community learns from one major plane crash.
  • Common human mistake – absence of evidence (confirming harm) is the same as evidence of absence (of harm itself). This is akin to a group of turkeys fed well months before Thanksgiving, that humans absolutely love them. The butchering days before Thanksgiving comes as a black swan event! ‘Not being a turkey’ starts with figuring out the difference between true and perceived/manufactured stability.
  • Excess intervention (as a promoter of fragility) – it depletes mental and economic resources, it is rarely available when it is needed the most.
  • Time is the mother of all stressors – more time, more events and more randomness.
  • Social, economic and cultural life lie in the Black Swan domain – trying to predict that is being a turkey predicting its fate (with stable but absolutely incorrect info).
  • Move the discourse from “trying to predict risk” to “trying to predict fragility (and develop antifragility”.
  • Thoughts on becoming anti-fragile:
    • Follow stoicism – be immune to misfortune.
    • From the author’s own toolbox of experiences – at a job, start each day expecting the worst case to come true. The rest of the day is then a bonus. Before you start a job, write the resignation letter, put it in a drawer, lock it away and then you’ll find yourself free each day at the job, knowing that you made it easy to quit.
  • “A Stoic sage is someone who transforms fear into prudence, pain into information, mistakes into initiation, and desire into undertaking.” 
  • You want to create an upside-downside asymmetry – if unexpected events create less downside than upside, then you are antifragile. If they don’t create downside (true for stoics), you are at least robust.
  • Implement antifragility like a barbell – with the two ends of the barbells being two approaches – paranoia at one end (for negative Black Swan events) and aggressive risk-taking at the other end (for positive Black Swan events). But these must be implemented as two separate strategies for two different categories (not mixed into one ‘middle of the road’ approach, that leads to nowhere).
  • Stoicism is a domestication, not elimination of emotions. The ‘barbell’ approach (adopt a mix of two extreme approaches) is a domestication, not elimination of uncertainty.
  • Overcompensation and overreaction
    • These are baked into ‘good’ systems.
    • There’s a reason the boss assigns the new pressing task to the busiest person on the team (apart from them being reliable). The more strained you are, the more your intellectual abilities are available for use. (N.Taleb merely touches upon this, Kahneman has a detailed discussion of this in Thinking Fast and Slow).
    • Humans are bad at estimating worst case scenarios, and it is precisely in those low-probability worst case scenarios that the spare capacity or ‘inefficiency’ becomes a boon, or very efficient.
    • Only a fool would think that the tallest mountain in the world is the tallest one that she has visited. However, most humans base their assumptions of the worst case on the worst case that they have seen or heard of in the past.
  • Those from whom we benefit aren’t those who have tried to help us, rather those who have tried, and eventually failed, to harm us.
  • Anything living is inherently antifragile to a certain extent – it is living because it has withstood some unexpectedness, disorder.
  • Fragility (as a property) is more predictable than risk. Prediction models that try to estimate risk are stupid (past is not an indicator of the future) – instead try to predict the fragility (or antifragility) of the system and work on that.
  • Artisanal careers are good lessons in antifragility – when you are a taxi driver, craftsman, dentist etc, small variations in external conditions (market situation, population changes, natural disasters) will make you adapt pretty effectively.
  • Optionality makes you antifragile – the ability to change course lets you benefit from the positive side of uncertainty, without serious harm from the negative side.
  • Option = asymmetry + rationality. It allows for big discoveries, while restricting oneself to small errors.
  • Optionality is Promethean (after the Greek fore-thinker character). Narratives are Epimethean (after-thinking).
  • History is written by the losers (academicians) while the tinkerers move on to newer experiments. Academics rationalize after the fact – like putting a bird in a room, showering lectures at it, and assuming you taught it how to fly! Few academics will admit that we create theories from practice.
  • “Avoidance of boredom is the only worthy mode of action. Life otherwise is not worth living.”
  • “Never let the other person frame the question. In every question, an answer is planted. Never respond straight to a question that makes no sense to you”.
  • Education is an institution that has been growing without external stressors – eventually it will collapse.
  • Procrustean bed – named a Greek inn-keeper who made his guests fit to his bed by stretching or chopping their limbs. You will encounter many logical Procrustean beds in life, where the non-linear is simplified to the linear.
  • Fragility – asymmetry – nonlinearity – concave curves.
  • Antifragility – asymmetry – nonlinearity – convex curves.
  • Financial crisis problems are primarily to do with size – with size, comes fragility.
  • “Figure out if your misforecasts are on balance more harmful than they are helpful – and how accelerating the damage is.” 
  • Property of a convex function – the function of the average of x is greater than the average of the function of x. e.g. for y=x^2, average of the function of a few values is higher than function of the average of these values.
  • When you create optionality for yourself (convex curves), you can do worse than random (be wrong more than 50% of the time) and still making a killing, since the payoffs from being right are high.
  • If you have favorable asymmetries or positive convexity, then in the long run you will do reasonably well outperforming the average in the presence of uncertainty!
  • The more powerful a concept – the more incomplete is our linguistic grasp of it. Antifragility is an example.
  • It is a result of convex curves that we are moving more and more into Extremistan –  while the 80:20 rule held earlier, today we see the 99:1 rule in many ways. For instance, 1% controlling more than 99% of the wealth.
  • Obvious decisions, which are robust to error, require no more than a single reason. By invoking more than one reason, you are convincing yourself to do something – so if you find yourself giving more than one reason for an important decision, look deeper.
  • The future lies mostly in the past – Vibrams shoes of today mimic barefootedness of the past, tablet computers with touchscreens are closer to Babylonian tablets. Both of these are new inventions that resulted after a bunch of ‘unnatural’ iterations.
  • Time brings about disorder – anything time-tested is worthwhile. If you pick a school textbook, most likely any concept you pick in there is still relevant in your life. Meanwhile, papers from a conference of 5 years ago may have been obsoleted.
  • Fragility in the medical field – interventionism by doctors and over-medication leads to harm that is not visible (takes 3/4th of a generation to show up). Humans assume “no evidence of harm” to be “evidence of no harm”, which is dangerous. For instance, synthetic fats like margarine were considered to be a great invention until the heart effects showed up.
  • Interventions almost always have negative convexity effects (mild known benefits, high unknown losses).
  • Antifragility of a system comes from the mortality of its components.
  • Antifragility at the expense of others’ fragility: People can hide risks and hurt others pretty easily without the law ever catching them. 2008 crisis is an example.
  • Basic rules to follow –
    • Don’t get on a plane without a pilot. For any venture, the person advocating it or influencing you must have skin in the game.
    • Have a copilot. Have redundancy, remove asymmetries in your sensitivity to risk (or bias them in your favor).
  • Burn the boats – develop skin in the game.
  • The more complicated the system, the more prone it is to arbitrage by insiders.
  • Look for optionality, and have a non-narrative approach to life.
  • When faced with a variable ‘x’ in life, it is much easier to understand and modify f(x), the impact of it on your, rather than to understand x itself. Predicting how x will vary is useless. Instead try to apply the barbell transformation to f(x) – so you have more benefits on the upside and cap the downside.


2 thoughts on “Book Summary: Antifragile by Nassim Taleb

  1. Antifragile : Wow !! This summary is indeed very useful and you explained it really well with examples. I found Book I, II & III quite hard to understand. In case, you have summary of these books readily available with you please send it to me. Thanks again for your work.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s