Ethereum co-founder Vitalik Buterin stated in a tweet that adversarial governance mechanism design theory could be a "killer app" for solving AI safety issues. He pointed out a profound duality between the two environments, both involving a less complex "principal" trying to obtain desired outcomes from a set of more complex "agents." In the first case, the principal is a static algorithm and the agents are humans; in the second case, the principal is humans plus weaker large language models, and the agents are stronger large language models. Buterin emphasized that a key finding is that better results can be achieved if the extent of agent collusion can be limited, an argument that should naturally apply to the new context of AI safety.