[jonfw]

Humans can certainly be self improving, both on an individual basis and in aggregate.

In humans, it seems that improvement in a new domain seems to follow a logarithmic scale.

Why wouldn’t this be the same for an AI?