I started reading the book Antifragility by Nassim Nicholas Taleb recently and had some thoughts on the nature of fragility, antifragility and robustness that I wanted to capture here

Taleb remarks that some critics of antifragility want to conflate robustness and antifragility while he would prefer to maintain them separately. I’m agreeing with Taleb and I’ll try to explain why here.

Let’s start with the concept of fragility. Suppose we have a road bridge and an earthquake occurs that the bridge would reasonably have been expected to survive and it does not; we can probably conclude that the bridge was fragile in some sense. This aligns with Taleb’s assertion that fragility as a property of a system is usually fairly easy to identify and importantly, generally easy enough to predict and avoid in the design of a system.

Now suppose we have a bridge and after multiple earthquakes it continues to survive and maintain it’s utility as a transport mechanism. We might conclude that the bridge is robust, or at least robust enough, that is, this robustness is essentially not fragile in that the bridge has not demonstrated any fragility to date such as a collapse. The problem for our not so hypothetical bridge is that there could (will?) be some unusually large earthquake that exceeds the designed capability of the bridge and hence we talk about the “one thousand year event”, or black swan event in reference to Talebs previous work The Black Swan; that is, extremely rare events that have extremely high impact. This, I think, is the key to why the concept of antifragility is useful as a separate concept from robustness.

What if instead of just being not fragile (not collapsing) in response to a new earthquake, our bridge actually got stronger and was able to withstand over time larger and larger earthquakes? In our experience of bridges, this seems absurd, but in our experience of other systems - particularly organic ones - not so much. Physical exercise is an example of challenges to an organic system (our bodies) that cause tiny damage that as our bodies repair that damage result in our bodies becoming stronger. Physical trainers reference “micro tears” in the muscles during exercise that stimulate this “repair and better than before” cycle in our bodies. Conversely in the absence of exercise our bodies become weaker and less capable of withstanding sudden large shocks and as our bodies become weaker the threshold of “sudden” and “large” that our bodies can withstand is also lower.

This “getting better over time in response to challenges” is the fundamental concept of antifragilty and I hope through my examples above you can see why it is usefully distinct from fragility and robustness.

Because we’ve made the distinction between merely robust and actually antifragile, we can now start to consider the implications of how to design systems in our civilization that could actually get better over time, rather than simply not failing. Useful, no?