I could see it going both ways. You're not going to know all YC founders by default, but you will know more founders than if the batches were small.

The individual network is larger, but comprises a smaller % of the total network as the program scales.