Trial and error. The velocity of information exchange that happens on the ground from trial and error - compounded by network effects - is far higher than the e

Trial and error. The velocity of information exchange that happens on the ground from trial and error - compounded by network effects - is far higher than the exchange that is happening in substantially smaller (CDC) groups at the top with slower speed. They don't iterate as fast and have a bias to move slow. More importantly. Average knowledge is not important. What matters is that exchange happens so that the optimal decision is arrived at and communicated much faster by a large group , even if each individual knows very little information themselves. This is the same principle that rules over new tech adoption. You don't need to know how to replicate the new tech, or even how to operate it. You just need to know it works . That's what real (applied?) knowledge is about, at least in the context above.