Inspired by anthropic reasoning behind Doomsday arguments, Nick Bostrom's Simulation Argument says: people who think advanced civilisations would run many fully-conscious simulated minds should also think they're probably simulated minds themselves. However, Bostrom's conclusions can (and should) be resisted, especially by sympathisers with Doomsday or anthropic reasoning. This paper initially offers a posterior-probabilistic ‘Doomsday lottery’ argument against Bostrom's conclusions. Suggestions are then offered for deriving anti-simulation conclusions using weaker assumptions. Anti-simulation arguments herein use more (epistemically and metaphysically) robust reference classes than Bostrom's argument, require no Principles of Indifference, abide better by the total evidence requirement, and yet use empirically plausible priors and likelihoods. However, while Doomsday arguments are probabilistically, epistemically and metaphysically stronger than the Simulation Argument, anthropic reasoning can (and should) refrain from embracing either.

中文翻译：

---

为什么世界末日论点比模拟论点好

受世界末日论点背后的人为推理的启发，尼克博斯特罗姆的模拟论证说：认为先进文明会运行许多全意识模拟思维的人也应该认为他们自己可能是模拟思维。然而，博斯特罗姆的结论可以（也应该）被抵制，尤其是那些对世界末日或人择推理的同情者。本文最初提供了一个后验概率的“世界末日彩票”论点，反对 Bostrom 的结论。然后提供使用较弱假设得出反模拟结论的建议。与 Bostrom 的论点相比，这里的反模拟论点使用了更多（在认识论和形而上学上）稳健的参考类，不需要无差异原则，更好地遵守总证据要求，并使用经验上合理的先验和可能性。然而，虽然世界末日论点在概率、认识论和形而上学上都比模拟论点强，但人择推理可以（并且应该）避免采用任何一种。