Lets say you have an AI that is both self improving and unsupervised. Do we think that it is likely that as the AI gets increasingly efficient, it becomes increasingly less able to learn? Surely as the AI refines itself, it will become faster, but mor…