I think you misunderstood the comment. Or maybe I did.
My understanding: the rubberhose crypto-analysis, even if unsuccessful, will result in some major damage done. Determined attacker might try to apply it regardless of any online statements on the off chance that the statements are wrong.
You understand correctly. I suspect that in the experience of such attackers, it's not even an "off chance". They're probably up against exaggerated claims of security more often than truly well-founded ones.
My understanding: the rubberhose crypto-analysis, even if unsuccessful, will result in some major damage done. Determined attacker might try to apply it regardless of any online statements on the off chance that the statements are wrong.