Speech: Nassim Nicholas Taleb explains antifragility

[This is largely from a school English writing assignment. Don’t accuse me of plagiarism.]

When Taleb talks, his statements are hard to summarize because his work is so integrated. His points may even sound garbled, because you need to understand the context of his whole body whole work to get the full picture.

So bear with me, I will try to annotate this. I present, the final speech in this series: Nassim Nicholas Taleb’s talk at google.

Pretend you didn’t read the title of this post. What is the opposite of fragile? You might answer something like “sturdy,” “robust, or “resilient.”

What is the opposite of positive? It’s not neutral, it’s negative. What’s the opposite of convex? It’s not flat, it’s concave. In that vein, the opposite of fragile cannot possibly be robust.

When a box is fragile, you write on the box, “handle with care.” What would be the opposite of “handle with care?” “Please mishandle.”

What is fragile is what does not like disorder or randomness. It wants stability and predictability. The fragile is hurt by infrequent events. The antifragile is what benefits from disorder, and random, extreme events. It loves volatility.

How Taleb models fragility:

Screen Shot 2018-12-11 at 2.36.37 AM.png

And the antifragile:

Screen Shot 2018-12-11 at 2.37.19 AM.png

Fragility is a second-order-effect. Harm increases disproportionately with event size. Would you rather be crushed by a 1000 pound stone, or by a 5 pound stone 200 times?

Screen Shot 2018-12-11 at 2.38.49 AM.png

Some options traders (such as Taleb) have a bipolar view of the world. The way they see it, there are things that like volatility, and then things that do not. I found this Wikipedia entry interesting:

Taleb sees his main challenge as mapping his ideas of “robustification” and “antifragility“, that is, how to live and act in a world we do not understand and build robustness to black swan events. Taleb introduced the idea of the “fourth quadrant” in the exposure domain. One of its applications is in his definition of the most effective (that is, least fragile) risk management approach: what he calls the ‘barbell’ strategy which is based on avoiding the middle in favor of linear combination of extremes, across all domains from politics to economics to one’s personal life. These are deemed by Taleb to be more robust to estimation errors. For instance, he suggests that investing money in ‘medium risk’ investments is pointless, because risk is difficult, if not impossible to compute. His preferred strategy is to be both hyper-conservative and hyper-aggressive at the same time. For example, an investor might put 80 to 90% of their money in extremely safe instruments, such as treasury bills, with the remainder going into highly risky and diversified speculative bets. An alternative suggestion is to engage in highly speculative bets with a limited downside.

Such a strategy takes advantage of the fact that a limited number of events have extremely disproportionate impact (“black swans”), (see my last post on recursion).

On a bad year, the gains from your safe investments make up for the losses in the risky investments. On a good year, your risky investments allow you to win big. There is sometimes huge upside but never downside, so you win long run. This is better than a “medium risk” investment strategy, where there is the potential for upside, but also the potential to loose money (and the more money you loose, the harder it is to earn it back, like a zip tie). Continued:

Taleb asserts that by adopting these strategies a portfolio can be “robust”, that is, gain a positive exposure to black swan events while limiting losses suffered by such random events. Together with Donald Geman and Hélyette Geman, he modeled the “maximum entropy barbell” which consists in “to constrain only what can be constrained (in a robust manner) and to maximize entropy elsewhere”, based on an insight by E.T. Jaynes that economic life increases in entropy under regulatory and other constraints, Taleb also applies a similar barbell-style approach to health and exercise. Instead of doing steady and moderate exercise daily, he suggests that it is better to do a low-effort exercise such as walking slowly most of the time, while occasionally expending extreme effort. He claims that the human body evolved to live in a random environment, with various unexpected but intense efforts and much rest.

Now, with respect to that last statement: in defense of runners, I think a marathon qualifies as an intense effort (and, indeed, a type of effort humans specifically are evolved to do). And I’ve done my share of running. But the type intensity Taleb is talking about is strength exerted quickly. For instance, literal barbell training, which I have employed to tremendous effect.

Muscles are an example of something antifragile. That is to say, they benefit from stressors.

Rumors are another example; the more they are suppressed, the more they seem to spread.

We recognize this reality with respect to muscles and rumors, but we often miss the importance of antifragility in other areas of our life.

Taleb believes, like I do, that stability is overrated. A government will shoot for perfect stability, missing the fact that what is organic requires a certain amount of instability.

Ironically, but using interventionism to make a system more stable, you actually make it less resilient, because you reduce its ability to learn how to handle unusual events. So when the bad events do occur, they end up being far worse.

Public policy discourse completely misses the concept of antifragility; anything short of stable is considered bad, with fluctuation to be avoided. After all, the unpredictable is what we can’t control, a problem for authoritarians.

We need instability to kill the weak. Stressors kill the weak items in a system, which, by leaving the strong, gives the appearance of making the whole system stronger.

This, killing the weak in a population, is the selective mechanism of evolution.

The process of evolution involves uncertainty/randomness, so it was considered undesirable and never modeled as an option. A lot of actions later dressed up as acts of design were in fact a result of evolution.

Silicon Valley entrepreneurship knows antifragility. The landscape is not stable. Failure is common.

Stoicism, a popular philosophy among some of the most successful figures in the ancient world, is antifragile because it was designed as a way to inoculate people against harmful events. Always set up a situation so that the best case scenario is more impactful than the worst case scenario. Stoics are not boring and unmovable, only unmovable by bad events.

More notes on fragility

Two more things, I wasn’t sure where to put this.

A “fragile” system entails small to moderate upside in most cases, but large downside in rare cases.

Sometimes in society, the person who bears the upsides is not the same person who bears the cost. For example, with the large banks, there is the old adage of “privatized gains, socialized losses.”

In many ways, we have modeled that life has gotten easier with time. Crime and violence have gone down. Infant mortality and extreme poverty have gone down. The world is getting better. Or is it? Perhaps we are a very fragile civilization.

Related posts

Taleb’s speech also connects with other previous posts of mine. In this one I talk about stability vs rapid progress. In this one I talk about sameness.

Other Speech Highlights

Peter Thiel on predicting the future

Bret Weinstein explains cycles of civilizations

Robin Hanson explains human signaling

Jaan Tallin explains the singularity

Richard Dawkins explains the adaptive valley

Leave a Reply

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

WordPress.com Logo

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

Google photo

You are commenting using your Google 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 )

Connecting to %s