# Antifragile: Things That Gain From Disorder

Nassim Taleb's Antifragile had and has a big impact on my life.

I first read Antifragile half a decade ago and came to revisit it, Black Swan and Skin in the Game a few more times over the years.

👇 Here are direct quotes from the book with some of my comments.

Have fun and don't be a turkey.

• All you have to know: suckers exist always. Don’t be a sucker
• We are better at doing than thinking
• Fragility can be measured, risk cannot
• Look for asymmetry
• Discovery <= antifragile tinkering $\cap$ aggressive risk bearing
• Discovery <≠ formal education
• Engineers and tinkerers develop things while history books are written by academics (programmed designs, university research)
• Don’t mistake the unknown for the non-existent. Absence of evidence ≠ evidence of absence.
• Modern culture builds blindness to the mysterious (Dionysian) in life
• Complications lead to multiplicative chains of events — explosive series of branching ‘unforeseen’ responses. Many have to seek sophistication to justify their profession
• To busy being a practitioner to write books
• The Extended Disorder Family:
Uncertainty, variability, imperfect, incomplete, chance, chaos, volatility, disorder, entropy, time, unknown, randomness, turmoil, stressor, error, dispersion of outcomes. Time is functionally similar to volatility, the more goes by the more events and more chance for disorder
• Be the first to be hurt if you’re wrong. Write about what you’ve done — else it’s unethical
• Poets and painters are free. Liberia poetae et pictores
• If you see fraud and do not say fraud, you are fraud
• Tinkering errors are small and rapidly forgotten
• Perceptions and intuitions, as expressed by acting, can be superior to what can be tabulate, discussed in words or classrooms. Apophatic. Being culturally, but not biologically concept blind.
• You need the word ‘blue’ for the construction of a narrative, but not when you engage in actions. (wine-dark sea in odysseee - they had no word for blue)
• Sophistication is born out of hunger. Artificial docuit fames
• Abundance is harder to handle than scarcity. Automation of airplanes is under challenging pilots. Disfluency — moves us into higher gear
• Redundancy seems like a waste if nothing unusual happens. Except that something unusual happens - usually. If humans fight the last war, nature fights the next one.
• Information is antifragile and you can spread it by disguising it as a secret.
• Only when you don’t care about your reputation to tend to have a good one. As in seduction, people lend most to those who don’t need them.
• Learning from mistakes of others. Titanic made future Titanics less likely. If it hadn’t happened it would have occurred even bigger.
• The surviving cohort is stronger than the initial ones (after a trial), but not the individuals themselves! The gulag didn’t strengthen the survivors, they were stronger from the beginning.
• The size of the unit may matter more than the system (eg. comparing countries). If you 10-fold the number of persons in an entity, the properties are not preserved. There is a transformation.
• Techne (craft and know how), not epitome (book knowledge, know what). Apprenticeship.
• (EU) subsidiary: things should be handled by the smallest possible unit
• What makes nations prosper isn’t peace, but freedom.
• There is no restaurant crisis, but banking crisis. Restaurants are small competing units.
• Risk is in the future, not past. Look for potential damage.
• Vienna was then trapped in the nation state of Austria (procrustean bed)
• Maxwell ‘On Governors’: Tight controlling the speed of engines leads to instability (mathematically proven).
• Annealing in metallurgy: Heating and cooling of material. Heat makes atoms unstuck and possible to find better configurations (while moved to higher energy) when cooled
• Modernity: Systematic smoothing of world jaggedness
• Lobbyists, LLCs, MBAs. Dependence on narrative and intellectualization.
• The thinker always needs terminology to act, not the doer. Doer has 1 layer less complexity
• Phenomenology = observation of an empirical regularity without a corresponding visible theory — it is robust and more reliable for decision making.
• Interventionism: give something to an editor and he will propose changes….even if already edited by someone else
• Voluntary omission. ‘Wu-wei’ - passive achievement (Lao Tzu). Procrastination is a natural filter. If you defer doing it all the time it must be eliminated
• Wisdom about decision making is vastly more important than knowledge.
• Possession = asymmetrical. If you’re successful you have more to gain than to lose.
• Mental exercise in morning: assume that the worst possible thing had happened…then you’re inoculated
• The middle is often not the golden middle
• Serious barbell: From doer to thinker. From pure action to pure reflection. Stay away from the middle
• Teleological fallacy: you know exactly were you’re going and assume that you know what your preferences will be tomorrow (error in trusting focus groups and market research).
• Nature is about the exploitation of optionality; it shows how optionality is a substitute for intelligence
• Convex tinkering = getting the upper half of luck
• Option = asymmetry + rationality
• Birds rarely write more than ornithologists
• In intellectual society there is a penalty for simplicity. Not in practice — like wheels on a suitcase
• The significant can only be revealed through practice
• Translational gap: time between discovery and implementation. Currently prolonged by academic interests and excessive noise (and the moving away from simplicity).
• There was a gap (David Wooton) of 200yrs between discovery of germs and the acceptance of germs as a disease
• A delay of 30yrs between germ theory of putrefaction and the development of antisepsis
• and a delay of 60yrs between antisepsis and drug therapy
• Governments and university have done little for innovation. They look for the grandiose, newsworthy, narrated, complicated, lurid.
• Apophatic: things we cannot express clear in language but do rather well. We overestimate the role of the things we can grasp with words
• Random Tinkering (anti fragile) —> Heuristics (technology) —> Practice and Apprenticeship —> random tinkering
• Epiphenomena: When you stand on the front of a ship you may well think that the compass is determining the direction of the ship, instead of just indicate it. Countries with high amounts of research are wealthy and developed…but A doesn’t cause B or vice versa. Many boys have short hair, but the length of hair doesn’t determine your gender
• Granger cause: add a time dimension to your analysis of events and you can see if A always causes B. Debunk fake causations
• Do people read books and then get ideas or the other way round? They look for books that support their ‘mental program’. The master-pupil relationship: are the ideas derived from masters or are the pupils pupils because they are already like minded
• Cherry picking: academia is well equipped to tell us what it did for us, how great its methods. They report what suits the purpose
• The more complex the education system, the less it is linkable to economic growth
• Halo effect is the opposite of domain dependence?
• Does education help families and generations, but disperse for the country. Guy who makes more money keeps it or invests in kids who then get complacent in middle class
• Activity strips things to their simplest possible model
• Theory should be used as a fable to stimulate ideas, indirectly inspire practice, but NOT determine it.
• Epimetheus (after-thinker) and Prometheus (fore-thinker) (2 titan brothers). Optionality = promethium, narratives are epimethean (he opened pandoras box).
• Intellect instills methods that conflict with tinkering (I have to change my modes of mind)
• An idea does not survive because it’s better than the competition, but because the person who holds it has survived (taken from Karl Popper’s ‘evolutionary epistemology')
• Fooled by randomness — overestimating the role of good-sounding ideas
• Practitioners don’t write, they do. History is written by losers with time on their hands
• You cannot reverse engineer everything (the taste of food by looking at the label). Cooking recipes are embedded in cultures. Cooking schools are apprenticeship based
• Physics is an exception, there theories can build on each other (Higgs Boson = entirely expected from theoretical derivations)
• Technologies are more like cooking
• The publish or perish academic model proceeds cosmetic knowledge (like fake watches). Never trust a journalist or blogger if its his only job
• Designed things, designer drugs are theological. The more drugs on the market the more interactions with other drugs to consider.
• Planning makes option blind — locked into a non-opportunistic course of action (the management myth).
• Invest not in ideas but people, someone who can change many times over his career
• History has been written by those who want you to believe that reasoning has a monopoly on producing knowledge.
• The ability of kids to count degrades right after being taught arithmetic. (How many intervals are between 15 poles? 14 (counting) vs 15(algebra)
• Gym rat vs gangster hitman. Some people succeed in schools because they can focus on boring material. As in any system the competition with grades focusses little on performance outside of the environment
• The enterprise has to be totally effortless to be worthwhile —> natural curiosity cures ADHD. The key is to be bored with a specific book rather than with the act of reading
• Problem first, then figure out the math that works for it. Context situation first (language) before grammar
• Most of what other people know is not worth knowing. The treasure lies outside the corpus.
• What I was given to study I forgot. What I decided to read on my own I remember.
• Socrates method to show people how they lack in clarity
• Sterile knowledge (like pharmadrugs…non natural)
• We are killing the things we can know but not express. If you ask somebody to give you the theory behind his bicycle riding he would fall from it. [My counter on that: Theories can help for lateral transfer of experimental knowledge. Making us learn more from our limited sensual experiences]
• Structured learning tends towards the materials and simplifications that are easy to teach (in a classroom or the medium)
• Logic excludes - by definition - nuances, and since truth resides exclusively in the nuances, it is ‘ a useless instrument for finding truth in the moral and political sciences’…what then is the instrument?
• Focus on the payoffs of your actions, instead of studying the structure of the world.
• PAYOFF, EXPOSURE, CONSEQUENCES. Instead of better computation and theories. Modify your exposure
• Fragile: shocks increase harm exponentially with intensity
• Episodic deprivation activates the pathways to better absorb the deprived later
• Economies of scales work in good times…size hurts you at time of stress
• Bottlenecks. The price of wheat tripped from 2004-2007 because of 1% increase in net demand.
• Wealth means more and more is different (non-linear scaling). You are likely to make more errors when you’re wealthy
• If you have optionality you can be wrong more than 50% of the time. If you’re fragile you have to be right about the future most of the time
• The more powerful, the more incomplete our linguistic grasp
• Before thinking to add anti fragility or options. First remove fragility. Michelangelo’s David: 'I took away everything that’s not David
• It’s the negative, subtractive knowledge that’s used by the pros. Become rich by not going bust. Life is about what to avoid. It’s hard to say if a person with skills will be successful, but a person without any skill will eventually fail.
• Subtractive epistemology - we know much more about what is wrong than what is right. Disconfirmation is more rigorous than confirmation. 1 observation can disprove you while even millions can hardly absolutely confirm you
• Focus is saying no. Subtractive productivity. Subtractive purpose: what not to spend time with
• From 80/20 to 99/1. Most things have winner take all effects
• Obvious decisions require no more than a simple reason. If you have 2+ you’re forcing it
• A philosopher should be known for 1 idea. An entrepreneur for 1 ambition. Collection of papers doesn’t do shit. 1 single contribution does
• Think about the future by removing what won’t be there. What we won’t do anymore, rather than adding technologies. We cook with the same tools as the romans; bicycles are getting more popular than cars again
• Why has X survived (water heater, fossett). The most successful technologies are invisible or trying to be (eg. condoms)
• New tech is easier to hype up
• Because all surviving technologies have obvious benefits we mis-assume that then all new technologies with benefits will succeed. Because novels that do well are well written doesn’t mean that well written novels automatically do well
• We notice change not statics (flows over stocks). You will be more pissed if I tell you you lost 10000$as if I tell you your portfolio went from 780.000 to 770.000$. We rely more on water than on cellphones, but water doesn’t change so we undervalue it’s importance.
• Fractal = tamed randomness. Self-affinity
• Scaling causing abstraction. Something humans are horrible at dealing with
• mistaking evidence of no harm for no evidence of harm.
• 50% dose then 150% is better than 100% 2x.
• Be aware of analogies. Rather say ‘I don’t know’ if the situation is not exactly the same. But devise your method from your best analogical guess
• Joint consumption: Fat and carbohydrates were consumed at the same time (meal) and people thought both were bad.
• Getting rid of smoking would outweigh the benefits of curing every type of cancer (p. 360)
• Pursuit of happiness —> avoidance of unhappiness
• Sanatories with curative starvation (fasting)
• Apples and oranges didn’t start out as sweet as they are now. Selective breeding turned them so. Avoid fruits that are not found in ancient eastern mediterranean.
• Comfort has side effects. Shed possession — subtractive well being
• There is a difference between getting the required nutrients together every meal or having them separately and sequentially . Randomize intake
• You can infer that people live longer because of the types of food they eat, but it might just be that the 2nd order effect (variation in intake) is dominant.
• Deprivation causes appetite and pleasure spikes of the meal. Otherwise you achieve the dullness of receptors
• We have a fondness for neomanic complication over archaic simplicity (eg. Hammurabi’s Code, just less violent)
• Redundancy, a margin of safety (and thereby sacrificing optimization).
• No opinion without risk (eg. no talking heads). Talk without embodiment often gives the victory to the most charming, not the most correct!
• In a complex system as Iraq, the invasion was epistemologically irresponsible. The predictability of consequences are very low.
• Post-dictors who had a hint of an idea but didn’t go with it to the logical conclusion will always say they knew it, disregarding the 1000 things they mis-predicted or cherrypicking instances
• An intangible victory has no value (Caesar parading Vercingetorix)
• Predictors should have scars on their body from errors, not distribute them to society
• Suckers try to win arguments, non-suckers try to win
• Wrong ideas that are harmless can survive. Societies with the right heuristics that lead to good outcomes survive. Not the ideas survive (popper on evolutionary epistemology), but people having the ideas that lead to beneficial action survive. Those with wrong heuristics but small harm in events of error also survive. Irrationality is sometimes harmless —> for example mistaking a stone for a bear is less harmful and leads to survival more often than mistaking a bear for a stone.
• If the market goes 50 percent up and then goes back it is more profitable to managers as if it is stable (they can cash stock options)
• Publishers with How To books optimize and put into your hands the thing with the highest time invested…that still can be called a book. Marketing beyond conveying information is insecurity
• Never trust the words of a man who is not free
• Being a lobbyist for tobacco or coke is killing people for profit
• Only he who is free with his time is free with his opinions
• Shame is a good mechanism for ensuring good behavior
• Executives, manager’s who are not the owner, don’t have much downside (they even get bonuses). Owner’s have to report to themselves
• Corporations have the incentive to produce the cheapest piece of rubber that they legally can still call cheese. They study taste buds to fool people into it tasting better than its ingredients (same with books). They are meant to fool you. A corporation doesn't have natural ethics, they worship the balance sheet.
• Corporations don’t have hormonal reactions (visceral) to shame. Agency problem
• The more complicated the regulations (bureaucratic) the more advantage do insiders have (the regulators who get than hired by companies to scam the taxpayer). The insiders don’t want less is more
• Tragedy of Big Data. The more variables, the more correlations can show significance in the hands of a ‘skilled’ researcher. Falsity grows faster than information (convex, nonlinear)
• Many people fit the hypothesis to the experiment. It’s also hard to get funding to replicate or reject existing studies
• Education likes disorder
• Spartan hoplites counter blogger, adventurers contra copy editors, Phoenician traders contra Latin grammarians, and pirates contra tango instructors
• We can control the function of $x = f(x)$ – even if x is vastly beyond what we understand. Exposure to X instead of likelihood of X or event of X occurring. Don’t conflate the event with exposure. x is the strength of the earthquake; f(x) is number of people dying from it
• Narrative or experimental discipline? Narrative disciplines use more statistics and try to fit something into a narrative (bed of procrustes) — good stories...
• Non-narrative action: Doesn’t depend on narrative for the action to be right – narrative is there for motivation, entertainment, or trigger.
• Subtractive prophecy: predicting the future by removing what is fragile, rather than adding