The unexpected value of the future

Abstract

The unexpected value of the future questions the application of expected value theory in determining how beneficial an action is, both in a utilitarian sense and in terms of longtermism. This is because it has been demonstrated many future scenarios, such as the Pasadena game or the Agnesi game, have expected values that are undefined, usually due to extremely improbable but extreme outcomes. In most practical situations, the problem is modeled with an Aquila game, a prospect that determines the total value after some time T as a function of a probability density function that is skewed due to the uncertainty of the number of independent clusters of civilization. The existence of the Aquila game falsifies longtermism under models of effective altruism by implying that it is better to try and improve the present rather than improving the far future. However, the author introduces Invariant Value Theory, an alternative to expected value theory, which is also able to provide verdicts in cases where expected value theory cannot, without relying on sensitivity to risk. In fact, the author claims that Invariant Value Theory implies strong longtermism, a moral theory which states that we should prioritize far-future outcomes over near-future outcomes to the point where improving the far-future is infinitely more valuable than improving the near-future. – AI-generated abstract.