One of the issues inherent in many kinds of consensus architectures is that although they can be made to be robust against attackers or collusions up to a certain size, if an attacker gets large enough they are still, fundamentally, exploitable. If attackers in a proof of work system have less than 25% of mining …