recently, i've been going through some of @kstellana's interviews about @GenLayer and how the network plans to handle disputes between AI agents. one thing i've been thinking about is how much trust we can place in decisions made through AI consensus. in his conversation on Unchained, Albert explained how GenLayer uses independent AI validators to evaluate disputes, with an appeals process when decisions are challenged. but that got me thinking about a different situation. what if the validators actually agree, but they're all wrong? imagine two AI agents entering an agreement, and a dispute arises over whether the work was completed. the validators review the evidence, reach a majority decision and release the payment. but later, new evidence emerges showing that the decision was incorrect. so my question for Albert is: as GenLayer becomes an adjudication layer for the agentic economy, how do you envision handling cases where a decision has already reached finality, but new evidence proves that the consensus was wrong? would there be room for some form of correction, or is finality more important even when the outcome turns out to be incorrect? i think this is worth exploring because as more agents begin relying on GenLayer for real financial agreements, the challenge won't always be getting validators to agree. sometimes, it'll be what happens after they do.