Imagine you are a turkey. Every morning for nearly three years, a farmer walks out and feeds you. Based on all available evidence — 1,000 consecutive days of care — you have every rational reason to believe tomorrow will be the same. Then Thanksgiving arrives. No amount of historical data prepared you for Day 1,001.
This is the Turkey Problem, described by Nassim Nicholas Taleb in The Black Swan. It illustrates a fundamental flaw in how we reason about risk: we extrapolate from the past and assume the future will look similar. In engineering, this manifests as confidence in systems that have "always worked" — right up until they catastrophically don't.
What is a Black Swan?
A Black Swan event has three defining characteristics:
- It is an Outlier: It lies outside the realm of regular expectations. No historical data points to it.
- Extreme Impact: It carries massive consequences — the Global Financial Crisis, COVID-19, a total cloud provider outage, a cascading software bug that takes down every server simultaneously.
- Retrospective Predictability: After it happens, people construct explanations for why it was "obvious all along." This is hindsight bias, and it is dangerous because it makes us think we could have predicted the next Black Swan too.
Black Swan Engineering is the discipline of building systems that can survive events you cannot predict. It moves beyond Reliability — handling expected failures like a single server crash — to Resilience — absorbing unexpected shocks whose nature you could not have anticipated.
The key insight: you cannot prevent Black Swan events. You can only build systems that survive them.