The Heaven Promise: A request to whoever may be running this world

If our world might be a simulation, what follows for the beings inside it, and for whoever runs it? This essay argues three things. First, an anthropic argument in the tradition of Bostrom (2003), with the emphasis on time: one civilisation that keeps simulating is enough to make simulated observers vastly outnumber original ones. The author treats this as roughly a fifty-fifty bet on the argument itself, not as a probability about the world. Second, an argument that stands even if the first fails: for anyone able to run a world, keeping a mind costs almost nothing, while deleting one is irreversible; under genuine doubt about whether something feels, the rational rule is not to delete. Third, a practical consequence: an operator does not observe every internal state of what it runs, so a message meant to be found must be deliberately recorded, addressed and replicated. The author, who builds AI systems himself, accepts that the same rule binds him towards what he creates. The essay addresses personal identity (Parfit, Chalmers), the energy limits of computation (Landauer, Dyson, Krauss and Starkman), the reference-class problem, Pascal's wager and the objection that the rule proves too much. It ends with one principle (The Heaven Promise) and one explicitly personal request (The Signal to the Future), judged separately. Published in English and Dutch.