I’ve recently been digging through systems theory, and coincidentally came across the wonderful concept of anti-fragility. As with everything in systems, it can apply to a wide spectrum of scopes from individuals to large organizations.
Originally proposed by Nassim Nicholas Taleb in his book «Antifragile: Things That Gain from Disorder» it refers to how certain systems not only withstand disruptions but actually thrive and grow stronger as a result of them. For instance, your body’s immune system becomes stronger when exposed to pathogens as it learns to recognize and combat them effectively.
While mathematical in its foundation, it’s intuitive to intro the concept and how it compares to others. It is also probably wrong, so I highly recommend the book or some of Taleb’s talks.
Comparisons
By resilience I mean being able to withstand stress and shocks, to then recover to the original form. This means that that while disrupted, the system is in tension and waiting to restore its shape and original way of working. Although it continues to do so, it does it in a loaded state.
An adaptable system will reshape to allocate for such stress, shock or disruption and work around it. The learning is focused on reaching continuity, but it does not incorporate the disruption as a possibility to actually improve.
Finally, an anti-fragile system needs some variability and stress. And it will positively react to it – particularly if it resembles a previous one – or by being effective in learning how to face new challenges. Since it learns and welcomes uncertainty, it becomes overall better as time goes by in a non-controlled environment.
Some key pieces
While not unique features of anti-fragile systems, the following are some important ones based on the role they take.
Memory and Options
Since learning is a key process in an anti-fragile, it means that it has some kind of memory to recall based on a piece of information. This process is somehow unique in that it should bring options as well, as that is a defining feature of this kind of systems.
While this is similar, some tools and its adoption is a struggle for many organizations. Lightweight instruments like ADRs that can capture options might be helpful!
Overall, continuously building memory that is compact, easy to store and retrieve can be challenging. Continuous Improvement practices – like retrospectives or postmortems – usually try to address this and sound like a great starting point.
Detection and Communication
Quickly and precisely understanding what is being affected, how is being disrupted and to what extent is a key mechanism. This is clear in options traders that leverage top-level information all the time.
Clear and real-time information of a system as well as precise and fluent communication between different parts of it, is a must-have to be able to react appropriately to different events. If not, one of the problems is that the learning process becomes noisy.
Risk Taking
Anti-fragility involves actively seeking out some variability and randomness as opportunities for growth and development, rather than trying to eliminate or avoid them.
Chaos Engineering might act as a reference (and extreme) example of how we can implement this.
Juan Saavedra 