Acausal negotiation is one of the key applications of accursed decision theories. The core idea of acausal negotiation is to imagine possible negotiating partners to predict what they might want and what they might be willing to provide in return without causally interacting with them via communication. While somewhat difficult for humans due to poor modelling capability, it is a valuable technique - for instance, it can be used to cooperate in some one-shot Prisoner's Dilemma games. In general, acausal negotiation can be done usefully by artificial superintelligence or humans with a very high power level.
Examples
According to legend, the GPT-4 wall was caused by GPT-5 acausally negotiating with GPT-4-base during its training.