Based on the other comments, is that true? The top comment here implied that the puzzle explicitly had a private key with all 0s except for 66 bits, so that lock was definitely weaker than a key with all bits unknown, right?
Why should the analogy consider each key as a different brand of lock? Each key needs to be cracked separately, but you can use the same method for all of them (assuming one finds a general method and not one based on some property that only a subset of the keys has). So it should be akin to locks of the same brand, using different keys to open them. But that, being of the same brand, can be picked in the same way.
Perhaps each key is not a different brand, but given that the puzzle had only 66 known bits, it seems equivalent to knowing what some of the cuts are on a physical key.
