We’ve pretty well mastered that part — to say the least — but what we need to work on is a) the means of informing people exactly when and how they’ve failed so they can learn from it and move on; and b) as you pointed out, ensuring these individual failures are effectively isolated.

For the first idea there are plenty of obvious metaphors; evolution and business most prominently. Evolution generates a constant churn of new mutations, business generates a constant churn of new business models. Most mutations, and most business models, fail; but, from the perspective of life as a whole, or the economy as a whole, in the long run the successes more than make up for the failures.

But the fact that failures lead to good isn’t enough to encourage people to take risks. Evolution and business are merciless, after all, and punish failure with extinction and poverty. What you’re after is a way to cushion the dangers of failure without discouraging risktaking. This is a hard problem. The question of where to calibrate a free market that makes innovation easy but dangerous versus a social safety net that makes innovation safe but hard, has no clear answer. Lacking certainty, different countries set that calibration on the basis of values rather than analysis.

Now that I think of it, though, evolution may be the right metaphor. One could argue that evolution found a solution to the problem of risk versus safety: intelligence, which lets one species try all potentially useful behaviors, without the failure of any of them endangering the species as a whole.

Perhaps what you’re after is a comparable layer of abstraction, one that lets individuals and organizations try out all potentially useful behaviors in a way that isolates the failure of any one from the failure of the whole. I submit to you that the necessary platform for this abstraction exists in the Internet, but that the necessary rules have yet to be invented. Relevance may be a good place to start.