Making the RANDAO less influenceable using a random 1 bit clock

Sharding
22

Possible follow-ups:

  1. We can probably explore if the “voters” above can be the same as the attesters (Attestation committee based full PoS chains, version 2).
  2. There should be an easy way of generating this vote based on a vdf, such that the influence of single votes can only be known when the vdf output is known. That could potentially greatly reduce the influence of knowing the vdf early.