# Black Box Thinking ## The Idea in Brief Black box thinking is the discipline of treating every failure as data. Aviation investigates every crash rigorously and shares findings openly. Most industries treat failures as ambiguous events to justify away. The difference in mindset creates radically different outcomes. Systems that learn from failure improve; systems that hide failure repeat it. --- ## Key Concepts ### Red Flags vs Ambiguous Setbacks When a plane crashes, it's difficult to pretend the system worked. The failure is stark—a "red flag." Most organisational failures are ambiguous enough to be explained away: "it was a one-off," "we did everything we could," "the market moved against us." Ambiguity is the enemy of learning. Without red flags—failures too stark to ignore—cognitive dissonance wins and nothing changes. ### Cognitive Dissonance When confronted with evidence that challenges deeply held beliefs, we're more likely to reframe the evidence than alter our beliefs. We invent new reasons, new justifications. Sometimes we ignore evidence altogether. The problem isn't just external incentives; it's the internal difficulty of admitting mistakes even when incentivised to do so. ### Systems + Culture An enlightened system alone isn't enough. Even the most beautifully constructed learning system won't work if professionals don't share the information that enables it to flourish. You need both: systems that capture failure data, and cultures that reward honesty over ego protection. ### Failure as Necessity "Sometimes, committing errors is not just the fastest way to the correct answer; it's the only way." We're hardwired to think the world is simpler than it is, but progress requires testing things that violate our beliefs. If you're not making mistakes, you're not learning. --- ## Implications **In organisations:** Build systems where failures are impossible to ignore. Track predictions against outcomes. Make the gap between expected and actual visible to everyone. **In learning:** You can't improve what you don't measure. Track your predictions, express them precisely, and review them honestly. Vague language ("probably," "might happen") makes learning impossible. **In culture:** Celebrate learning over looking good. The question "can we afford time to investigate failure?" is backwards. The real question is "can we afford not to?" **In creativity:** Creativity isn't something that happens to geniuses through contemplation. It's a response to failure—a flaw, a frustration. Without problems to solve, innovation has nothing to latch onto. --- ## Sources - [[Black Box Thinking]] — Syed's core argument: aviation's learning culture vs healthcare's defensive culture - [[Superforecasting]] — Learning to forecast requires tracking predictions and learning from errors - [[Algorithms to Live By]] — Even optimal algorithms produce bad outcomes sometimes; judge process, not single results - [[High Performance Habits]] — Elite performers obsess over process and treat failure as feedback --- ## See in Field Notes - [Decision Architecture](https://www.anishpatel.co/decision-architecture/) — Strategy as hypothesis: track what you expected versus what happened, then update - [Applied Scientific Thinking](https://www.anishpatel.co/applied-scientific-thinking/) — The posture of stating beliefs, testing them, and updating when wrong