There are two kinds of mistakes.
One is the sort where failure is not noticeable because failure means that you didn’t engage with an audience. If you do an art show and no one comes, no one realizes that your art show failed.
The other is harder to walk away from. You can’t put the toothpaste back in the tube. When you engineer credit default swaps that merge together mortgage streams, inextricably linking the healthy ones with the others, it’s a mistake with real consequences.
Confusing the two types of errors is a recipe for tragedy. If we can figure out how to organize and plan for the first type of resilient failure, it’s far easier to experiment.