A blog is worth more than a series of posts linked to on social media sites like reddit and Hacker News in a desperate attempt to funnel in new readers. A blog, for writers, is a portfolio to hand out to propsective employers, for web developers, proof of familiarity with WordPress or similar software, and, for growth hackers, an example of one’s ability to attract visitors.
This blog, for me, has been all three of those things, but what I’m most proud of is the community it creates. A few months ago, I reviewed a paper by Vincenzo De Florio. I asked him to proof-read a draft before it went live. Delighted by my concise introduction to his otherwise complex work, he invited me to the Programme Committee of ANTIFRAGILITY 2014, the first workshop of its kind. The submission deadline, by the way, is January 31st, hint hint.
Antifragile is a relatively new term, coined by Nassim Taleb and made popular by his book of the same name. It’s receiving growing attention, and apparently even NASA has jumped on board—our keynote speaker is their Dr. Kennie H. Jones.
So what’s “antifragile” mean?
Antifragility is directly a Complex Systems topic, so it, like many properties studied by us Complex Systemsists, applies to all sorts of objects, systems, networks, and so on.
Antifragile is but one, highly desirable, side of the fragile/resilient/antifragile spectrum. Like that vase that seems to get broken in every Hollywood Romantic Comedy, fragility refers to unstability following perturbation—I bump the vase, the vase shatters. Resilience, on the other hand, refers to a resistence to perturbations—I drop a rock, the rock is fine. On the other other hand, antifragile refers to an improvement following perturbation—I jog, my muscles tear slightly, I become a better runner.
On first thought, antifragility may appear a bit too ideal, that it is a property we may strive towards but never truly introduce into future system development. In many ways it is, being a novel paradigm that asks us to view decision making contrary to existing convention.
That is, it asks us to embrace randomness, time, stress, change, and so on in order to best gain from them. This is hard. People like order. Furthermore, antifragility can be difficult to perceive because it’s possible for a system to be fragile at one scale and antifragile at another.
And this is what matters.
Cells die all the time because, on their own, they are extremely fragile. But a whole human, at a larger scale compared to individual cells, benefits from weaker cells dying off to make room for stronger ones. This continual selectivity makes the body stronger when confronted with limited stressors.
Airlines are designed in a way that mistakes are negatively correlated with future mistakes. That is, once a crash has occured, a subsequent one is less likely. Taleb also provides cities, bones, muscles, and economies as further examples of antifragile systems. Although one business might close following a market crash, that city district may improve, in time, because of the turmoil. Even social communities exhibit antifragility, growing more resilient following a disaster.
Watching television as I write this post on my phone, I stumbled across an example of my own. In the popular American series King of the Hill, Strickland Propane occasionally suffers from what the show’s characters call “goldielocks season,” during which it is neither too hot nor too cold for customers to purchase or refill tanks of propane. These sales benefit nonlinearly as extreme weather conditions threathen more people and require more fuel for survival.
Antifragile systems are more than selective, more than adaptive. They exhibit increasing rates of gains as randomness, time, and so on increase. But at what price?
In economies, a sustained period without financial trouble results in more extreme damages once a crash occurs. This is the “Turkey Problem.” The longer the butcher feeds the turkey, the friendlier the turkey will greet him the morning of Thanksgiving.
Is “Antifragile Engineering” possible? What of ethical issues? Are we expected, in some sense, to limit ourselves in the short term to better long term payouts? There are many such questions for this new and promising research area.
And this is what matters. ∎
Follow me on Twitter. Let’s chat sometime.