Something hinted at in the essay is what I consider a funnel effect. If you have N people starting on some project, N/2 will get distracted after a while, some proportion of those who remain will just suck, some smaller proportion will be OK, and so on until you manage to find a few real geniuses. If you shrink the size of the funnel, you get less geniuses.
An important part of having a large funnel is giving people a way to really spend their time doing the thing. For example, writing short stories for magazines was once a reasonable way to support yourself for a few years as a young writer, and led to a very large funnel. Take away that infrastructure for young writers, and you get a smaller funnel, and an attrition in quality of the best work.
(Now, consider what happens 10-20 years after we stop hiring new grads for programming jobs...)
An important part of having a large funnel is giving people a way to really spend their time doing the thing. For example, writing short stories for magazines was once a reasonable way to support yourself for a few years as a young writer, and led to a very large funnel. Take away that infrastructure for young writers, and you get a smaller funnel, and an attrition in quality of the best work.
(Now, consider what happens 10-20 years after we stop hiring new grads for programming jobs...)