Antifragile

The name for it: the triad

Most people think the opposite of fragile is robust. Taleb says no. Robust only means "not hurt". The true opposite of fragile is something that gains from shocks: antifragile. Your muscles, your immune system, a market of small restaurants, and evolution itself are on this side.

The book's test is simple: does this thing like or dislike volatility, randomness, error and stress? You do not need to predict the next shock. You only need to know which side of the triad you are on.

The name for it: convexity (and Jensen's inequality)

If the curve frowns (concave), a big loss hurts more than a big win helps, so more volatility pulls your average down. That is fragility. If the curve smiles (convex), you have limited loss and open upside, so more volatility pulls your average up. That is antifragility.

In maths this is Jensen's inequality: for a curved payoff, the average of the results is not the result of the average. Taleb's point: you do not need to forecast the shock. You need to know the shape of your exposure. Taleb's own example is traffic: one extra car on a full road can add a lot of delay, but one car fewer saves you almost nothing.

The name for it: the turkey problem

The turkey's confidence is highest when the risk is highest. A smooth past says nothing about a hidden, big shock. Taleb uses this to attack people who say "we have no evidence of risk" when they mean "we have evidence of no risk". These are not the same thing.

In the book, two brothers show the same idea. John has a steady bank salary. George drives a taxi and earns different money each day. John looks safer, but one layoff can take all his income. George gets small shocks all the time, learns from them, and is never one event from zero. Systems that suppress small shocks (smooth economies, rigid jobs, forests where all small fires are stopped) store up a big one.

The name for it: the barbell strategy

Put most of your weight at the two ends and nothing in the middle. Taleb suggests something like 85 to 90 percent in very safe things and 10 to 15 percent in very speculative things. You cannot lose more than the small risky part, so your floor is known. But the risky part has open upside. The "medium risk" middle looks sensible, but its loss in a crash is unknown, and it rarely gives a big win.

The barbell is not only for money. A writer with a dull, safe day job and wild projects at night. A life that is very careful about a few dangers (no smoking, no ruin) and very playful about everything else. The numbers in this lab are a simple model, not investment advice.

The name for it: hormesis and overcompensation

A small dose of a harmful thing can make a living system stronger. Bones get denser under load. The immune system learns from germs. Taleb calls the reaction overcompensation: the system does not only repair, it builds extra capacity for next time.

Two warnings from the book. First, this works only with recovery time and up to a limit; past the limit, stress breaks you. Second, Taleb thinks variety is better than routine: random, acute stress with long rest beats the same chronic dose every week. This lab is a toy model of that idea, not a training plan.

The name for it: optionality

An option is the right, but not the duty, to do something. If a try fails, you walk away and lose a little. If it works, you keep all of it. With options you do not need to be smart or to predict: you need many cheap tries and the sense to keep the good ones. Taleb calls the mix of options and trial and error "convex tinkering".

He says much of history's technology came this way, from tinkerers and practitioners, and that schools then wrote the theory afterwards. He calls the opposite belief, that ideas flow from lectures to practice, the "Soviet-Harvard" illusion. His character Fat Tony does well without theory because he knows where the options are. The lab numbers are a model.

The name for it: iatrogenics and via negativa

Iatrogenics is harm done by the healer. When the problem is small, the visible benefit of acting is small, but the hidden risk of acting stays. So for mild cases, doing nothing often wins. For severe cases, act. Taleb applies this to medicine, economics, politics, and parenting: the urge to intervene ("naive interventionism") is itself a risk.

Via negativa means improve by taking things away. Stop smoking, cut sugar, drop a bad habit, remove a fragile supplier. What you remove is usually clearer and safer than what you add. Taleb: we know much more about what is wrong than about what is right.

The name for it: the Lindy effect

For things that do not age like bodies (books, ideas, technologies, customs), every year that a thing survives adds to its expected future life. A book in print for 100 years will probably stay in print longer than a book from last year. Time is the best filter for fragility, because time breaks fragile things.

So Taleb trusts old wisdom, old foods and old tools over the new and loud. Not because old is good, but because it has passed a test that the new has not yet taken. It is a rule of thumb about the odds, not a law.