I’m reading Taleb’s Antifragile (finally). Well, I’m not really reading so much as listening to it as an audio book which really gives me the opportunity to completely ignore those parts of the work that irritate me while revelling in some confirmatory bias.
Antifragile certainly isn’t an easy book but it does provide a variety of interesting frameworks for consideration. The root of Taleb’s argument is that fragility is bad and that robust (or resilient) is sub-optimal. Instead, we need to exploit the ability of certain things to get stronger due to stress. Nature is a good example as evidenced by the human body. We apply stress to the body (e.g., with some heavy deadlifts), we are broken down, and the body develops additional capacity.
In an attempt to fully grok his ideas I’m attempting to break them on the wheel that I know best: IT. So what, exactly, is antifragile IT? It seems that many of our IT practices are effective for removing fragility and building robustness. We increase resiliency and exploit models that give us more (capacity, redundancy, etc.) at a reduced cost. But is that really antifragile?
I’ve been working on a project that considers the nuances of IT cost cutting. It’s not difficult to cut costs from IT but those costs will often result in infrastructure and capabilities that are more fragile. For example, we lose the ability to respond to business change and — seemingly inevitably — we have to increase our reliance on manual business processes. The black-box optimization algorithms of our ERP systems wither in the face of business change and concede to the trinity of Excel, file shares, and email. This solution is likely less efficient and seemingly becomes more fragile because it lacks standardization or repeatability.
But is this transition a bad thing? In many ways Microsoft Excel is a wonder of antifragility. I’m amazed at how it persists and permeates all enterprises and gets used for all manner of odd things. Furthermore, its shortcomings actually lead to informal continuous improvement processes. Business units develop standards and training programs. All users become power users through the mystery of tacit knowledge exchange. Where our IT-driven IT processes become rarified and fragile, Excel persists and generally gets better. Excel is perhaps more than just a technology; it is a community and a practice.
Of course, this is just an idea for how the notion of antifragility applies to IT. I could be completely off base.
I would love to hear your ideas.