Karl Weick, the organizational psychologist, spent decades studying how people construct meaning under uncertainty — and he found something quietly devastating: experts in high-stakes environments don't fail because they lack information. They fail because they commit too early to a single story about what's happening, then defend it against incoming contradictions rather than revising it. He called this 'cosmological episode' — the moment a mental model stops being a working hypothesis and becomes an identity. The problem isn't building models. It's that models, once useful, generate what Weick called 'enacted environments' — we literally stop perceiving inputs that don't fit the frame we've already chosen. The practical discipline he argued for is less about acquiring better models and more about maintaining what he called 'respectful interaction with your own ignorance': holding your current mental frame with just enough looseness that a disconfirming signal can still land. Today, pick the model you've used most confidently this week — about a person, a market, a product, a process — and ask yourself what evidence would have to exist for it to be wrong. Not whether that evidence exists. Just what it would look like. That prior question is harder than it sounds.

What would falsifying evidence for your most confident current belief actually look like — and have you recently seen anything that resembles it?

Drawing from Organizational Psychology / Sensemaking Theory — Karl Weick