simplest way to explain peerDAS is that sampling is basically blob sharding, which people might remember from the "eth2" context a few years ago =) we can meaningfully scale while giving a lot of breathing room bandwidth expectations (and even more networking optimizations to come)